BS EE S - a Consensus TR — oR RTS Management of ii i Es ia aA Hog? For sale ONLY by the Superintendent of Documents, U.S. Government Printing Office, Washington, D.C. 20402. Payment is required in advance, and check or money order should be made payable to the Superintendent of Docu- ments. NCI Monographs National Cancer Institute, Vincent T. DeVita, Jr., Director International Cancer Information Center, Susan Molloy Hubbard, Director Editorial Board Robert E. Wittes, Editor-in-Chief William J. Blot Peter Greenwald John R. Ortaldo Peter M. Blumberg Robert N. Hoover David G. Poplack John D. Boice, Jr. Susan M. Hubbard Alan S. Rabson Michael R. Boyd Steven M. Larson Paul C. Rambaut Bruce A. Chabner Brian R. Leyland-Jones Jeffrey Schlom Richard L. Cysyk Lance A. Liotta Richard M. Simon Charles H. Evans Marc E. Lippman Michael B. Sporn Eli J. Glatstein Dan L. Longo Snorri S. Thorgeirsson Michael M. Gottesman Douglas R. Lowy J. Paul Van Nevel Joost J. Oppenheim Publications Branch, Robin A. Atkiss, Chief Monograph Editors Florence |. Gregoric Anne McCarthy Editorial Policy Manuscripts from key conferences dealing with cancer and closely related research fields, or a related group of papers on specific subjects of importance to cancer research, are considered for publication, with the understanding that they have not been published previously and are submitted exclusively to NC! Monographs. All material submitted for consideration will be subject to review, when appropriate, by at least two outside reviewers and one member of the Editorial Board of the Journal of the National Cancer Institute. Opinions expressed by the authors are not necessarily those of the publisher or the editors. Symposia or related papers should be submitted to the Editor-in-Chief, Journal of the National Cancer Institute, R. A. Bloch International Cancer Information Center, Building 82, Room 211, National Cancer Institute, Bethesda, MD 20892. James O. Armitage Bruce C. Baguley Charles M. Balch June L. Biedler Clara D. Bloomfield Benjamin Bonavida Ernest C. Borden George J. Bosl Murray F. Brennan Edward Bresnick Ting-Chao Chou Harvey J. Cohen C. Norman Coleman O. Michael Colvin Thomas H. Corbett Rik Derynck Leila Diamond Johanna T. Dwyer John D. Earle Merrill J. Egorin Virginia L. Ernster Nelson Fausto James S. Felton EDITORIAL ADVISORY BOARD Isaiah J. Fidler Thomas R. Fleming @stein Fodstad Emil Frei III Richard D. Gelber David W. Golde Harvey M. Golomb E. Robert Greenberg Philip D. Greenberg Kenneth R. Harrap Ingegerd Hellstrom Maureen M. Henderson Gloria H. Heppner E. Carmack Holmes William J. Hoskins Robert C. Jackson Michael E. Johns Randall K. Johnson William W. Johnston Charles M. King Stanley J. Korsmeyer John S. Kovach Margaret L. Kripke Donald W. Kufe John S. Lazo Brigid G. Leventhal Bernard Levin Arnold J. Levine Allen S. Lichter David M. Livingston Virginia A. LiVolsi Jean-Pierre Mach Louis Malspeis Donald B. McCormick Richard S. Metzgar George K. Michalopoulos Anthony B. Miller Malcolm S. Mitchell Alan S. Morrison Lee M. Nadler C. Kent Osborne David F. Paulson Sidney Pestka Lester Peters Theodore L. Phillips Henry C. Pilot Janet S. Rasey Ralph A. Reisfeld Igor B. Roninson Ruth Sager Shigeako Sato David Schottenfeld William R. Shapiro Roy E. Shore Kendall A. Smith Paul M. Sondel Osias Stutman Herman D. Suit Raymond Taetle Ian Tannock Joel E. Tepper Cox Terhorst J. Tate Thigpen Giorgio Trinchieri Peter R. Twentyman I. Bernard Weinstein Sharon W. Weiss Raymond M. Welsh Richard E. Wilson Gerald N. Wogen National Institutes of Health Consensus Development Conference on the Management of Clinically Localized Prostate Cancer National Cancer Institute Vincent T. DeVita, Jr., M.D., Director Sponsoring Organizations Consensus Development Panel Robert B. Livingston, M.D., Chairman Alfred R. Bartolucci, Ph.D. Joshua Becker, M.D. William C. DeWolf, M.D. John T. Ellis, M.D. H. Anthony Engelbrecht, M.D. Marc B. Garnick, M.D. John T. Grayhack, M.D. Lance K. Heilbrun, Ph.D. James G. Jones, M.D. Shannon B. McGowan C. Kent Osborne, M.D. Joel E. Tepper, M.D. John R. Thornbury, M.D. National Institutes of Health Bethesda, Maryland June 15-17, 1987 Planning Committee John E. Antoine, M.D., Chairman Michael Bernstein Andrew Chiardo, Ph.D. Joyce Doherty Jerry M. Elliott Michael A. Friedman, M.D. Eli J. Glatstein, M.D. Robert B. Livingston, M.D. Elliott Stonehill, Ph.D. Patrick C. Walsh, M.D. Willet F. Whitmore, Jr., M.D. Richard D. Williams, M.D. TABLE OF CONTENTS Introduction John E. Antoine Consensus Statement: The Management of Clinically Localized Prostate Cancer Consensus Development Panel Overview: Historical and Contemporary Willet F. Whitmore, Jr. I. Staging Histologic Grade, Clinical Stage, and Patient Age in Prostate Cancer Donald F. Gleason Fine-needle Aspiration of the Prostate Mitchell C. Benson Application of Flow Cytometry and Automated Image Analysis to the Study of Prostate Cancer Mitchell C. Benson Noninvasive Imaging for Staging of Prostate Cancer: Magnetic Resonance Imaging, Computed Tomography, and Ultrasound Hedvig Hricak Lymphography in Clinically Localized Prostate Cancer Ronald A. Castellino Value of and Indications for Pelvic Lymph Node Dissection in the Staging of Prostate Cancer Richard G. Middleton II. Radiation Therapy Status of Radiation Treatment of Prostate Cancer at Stanford University Malcolm A. Bagshaw, Richard S. Cox, Gordon R. Ray Radiation Therapy Oncology Group Studies in Carcinoma of the Prostate Miljenko V. Pilepich Radiation Therapy for Localized Prostate Carcinoma: Experience at the Massachusetts General Hospital (1973-1981) W. U. Shipley, G. R. Prout, Jr, N. M. Coachman, P. L. McManus, E. A. Healey, A. F. Althausen, N. M. Heney, E. C. Parkhurst, H. H. Young II, J. W. Shipley, S. D. Kaufman Page 15 25 31 37 41 47 61 67 External-beam Radiation Therapy for Clinically Localized Prostate Cancer: Patterns of Care Studies in the United States Gerald E. Hanks Definitive Radiation Therapy in Carcinoma of the Prostate Localized to the Pelvis: Experience at the Mallinckrodt Institute of Radiology Carlos A. Perez, Miljenko V. Pilepich, Delia Garcia, Joseph R. Simpson, Fred Zivnuska, Mary Ann Hederman Local Control of Prostate Cancer With Radiotherapy: Frequency and Prognostic Significance of Positive Results of Postirradiation Prostate Biopsy Peter T. Scardino, Thomas M. Wheeler III. Surgery Selection Criteria for Radical Prostatectomy Based on Morphometric Studies in Prostate Carcinoma Fuad S. Freiha, John E. McNeal, Thomas A. Stamey Bilateral Pelvic Lymphadenectomy and Radical Retropubic Prostatectomy for Stage C or D1 Adenocarcinoma of the Prostate: Possible Beneficial Effect of Adjuvant Treatment Horst Zincke Long-term Results of Radical Prostatectomy in Clinically Localized Prostate Cancer: Experience at The Johns Hopkins Hospital Herbert Lepor, Patrick C. Walsh Total Prostatectomy for Clinically Localized Prostate Cancer: Long-term Surgical Results and Current Morbidity Robert P. Gibbons Randomized Series of Treatment With Surgery Versus Radiation for Prostate Adenocarcinoma David F. Paulson Radical Retropubic Prostatectomy With Reduced Morbidity: An Anatomic Approach Patrick C. Walsh IV. Adjuvant Therapy Radiation Therapy as Adjuvant Treatment After Radical Prostatectomy Paul H. Lange, Pratap K. Reddy, Eitan Medini, Seymour Levitt, Elwin E. Fraley Chemotherapy for Prostate Carcinoma Mario A. Eisenberger Hormone Therapy for Prostate Cancer: Results of the Veterans Administration Cooperative Urological Research Group Studies David P. Byar, Donald K. Corle Hormonal Therapy for Locally Advanced Prostate Cancer Albert B. Einstein 75 85 95 107 109 117 123 127 133 141 151 165 171 Introduction John E. Antoine’ Prostate cancer, the second most common form of malig- nant disease in American men, is one of the most frequently diagnosed cancers in men over the age of 50, with the inci- dence increasing each decade after the age of 50. Approxi- mately 96,000 new cases of prostate cancer will be diagnosed in 1988. According to the American College of Surgeons, more than one-half of these are clinically localized. More- over, approximately 26,000 deaths each year are due to prostate cancer, the etiology of which is unknown. It is a common finding at autopsy in men older than 50 years. The most frequently observed histologic tumor type is adeno- carcinoma. Localized prostate cancer is usually discovered during the rectal part of a physical examination or detected incidentally in the histologic material from a transurethral prostatectomy performed for enlargement of the prostate. A needle biopsy is often diagnostic. Patients with localized disease may be asymptomatic at the time of the diagnosis of prostate cancer or may have symptoms of urinary retention; hematuria is uncommon. The extent of tumor involvement (stage) is usually de- scribed by a classification system that has undergone evolution and modification over the years. Physicians have encountered difficulties in describing the extent of disease because of the use of more than one staging system and of staging based on clinical rather than surgical findings. Because of this lack of consistency in the staging systems, the analysis of data remains complex and difficult. Two accepted clinical staging systems are in use at this time. The American Urological System consists of A, B, C, and D stages, and the American Joint Committee uses the T, N, and M system. For proper comparison of alternative treatments for the management of localized prostate cancer, a uniform, tumor-extent, staging system is necessary. The degree of histologic anaplasia (malignancy) of the cells also appears to correlate with the patient's prognosis. The Gleason histologic scoring system is commonly but not uniformly used; the higher the Gleason score, the worse the prognosis. Standardization of a histologic grading system would also facilitate scientifically valid comparison of treatment modalities. ! Radiation Research Program, Division of Cancer Treatment, National Cancer Institute, National Institutes of Health, Department of Health and Human Services, Bethesda, MD 20892. Radical prostatectomy has a long history as the definitive management for localized prostate cancer. The retropubic approach, which readily permits staging lymphadenectomy, has become more common than the perineal approach. Recently, the surgical technique has been refined to preserve the nerves essential for sexual potency. Radiation therapy is another form of definitive treatment and is usually given by means of high-energy, external- beam photons from linear accelerators. An alternative form of radiation treatment is the implanting of radioisotopes in the cancer tissue. Hormone therapy is an effective means of palliation for advanced prostate cancers. It is under renewed investigation as an adjuvant to definitive therapy of the primary tumor in early disease. Although chemotherapy generally has been used in the palliation of patients with advanced hormone refractory disease, its role in the management of localized prostate cancer remains to be defined. Following treatment of prostate cancer, the follow-up of the patient includes physical examination, laboratory studies, and imaging examinations. Laboratory studies might include alkaline phosphatase and prostate-specific antigen. Imaging studies commonly include x-rays, bone scans, magnetic resonance imaging, computed axial tomography, and ultra- sound. The optimal combination of laboratory and imaging studies has not been determined, but this area is evolving at this time. Competing (noncancer) causes of death are common in the age group affected by prostate cancer. The natural history of this disease is variable, and some untreated patients may die without symptoms of prostate cancer. These factors make the significance of therapeutic intervention more difficult for the physician to assess for prostate than for other cancers. The quality of life and sexual functioning can be affected by treatment for prostate cancer because impotence and incontinence can occur following treatment. Prospective studies that would determine how often these side effects are associated with each treatment have not been done. These aspects should be considered when the physician helps the patient choose the best course of treatment. To evaluate new information and resolve issues regarding optimal treatment, the National Cancer Institute and the Office of Medical Applications of Research of the National Institutes of Health convened a Consensus Development Conference on the Management of Clinically Localized Prostate Cancer on June 15-17, 1987. The scientific ma- terial from the manuscripts published in this edition was presented at that consensus conference. Consensus Statement: The Management of Clinically Localized Prostate Cancer’ National Institutes of Health Consensus Development Panel Prostate cancer is the second most common form of ma- lignant disease in American men. It is one of the most common cancers in men over the age of 50, with the inci- dence increasing each decade after the age of 50. Approxi- mately 96,000 new cases of prostate cancer were expected to have been diagnosed in 1987, of which 50,000 may have been clinically localized, according to the American Col- lege of Surgeons. There are approximately 26,000 deaths from prostate cancer annually; it is a common finding at autopsy in men over the age of 50 years. The most com- mon histologic tumor type is adenocarcinoma. The etiology remains unknown at present. Localized prostate cancer is usually discovered by rec- tal examination performed during a physical examination or detected incidentally in the histologic material obtained from a transurethral prostatectomy being performed for enlargement of the prostate. Patients with localized dis- ease may be asymptomatic at the time of the diagnosis of prostate cancer or may have symptoms of urinary retention. The extent of tumor involvement (stage) customarily is described by a classification system that has undergone evo- lution and modification over the years. Difficulties in de- scribing disease extent have arisen from the use of different staging systems and from staging based on clinical as op- posed to surgical findings. Because of this lack of uniformity in the staging systems, the analysis of data remains com- plex and difficult. Two accepted clinical staging systems are in use. The system used by members of the American Urologic Association consists of A, B, C, and D stages, and the American Joint Committee on Cancer uses the T, N, and M system. Radical prostatectomy has a long history as the definitive management for localized prostate cancer. The retropubic approach, which readily permits staging lymphadenectomy, has become more common than the perineal approach. Re- cently, the surgical technique has been refined to preserve the nerves essential for potency. Radiation therapy is an- other form of definitive treatment and is usually given by means of high-energy, external-beam photons (x-rays) with linear accelerators. Hormone therapy is an effective means of palliation; its role in the management of early disease is under renewed investigation as an adjuvant to defini- tive therapy of the primary tumor. Chemotherapy generally ! Adapted from The Management of Clinically Localized Prostate Can- cer, National Institutes of Health Consensus Development Conference Statement, vol 6, No. 10, 1987. has been used in the palliation of patients with advanced hormone-refractory disease. After the completion of treatment of prostate cancer, the follow-up of the patient includes physical examination, lab- oratory studies, and imaging examinations. The exact com- bination of laboratory studies and imaging examinations has not been agreed upon, and this situation is undergo- ing evolution in the United States. The natural history of this disease is variable, and some untreated patients may die without symptoms of prostate cancer. Competing (non- cancer) causes of death are common in the age group af- fected by prostate cancer. These factors make the signif- icance of therapeutic intervention more difficult to assess than in other cancers. The general quality of life and sexual functioning (im- potence) can be affected by treatment for prostate can- cer. These aspects should be considered when the physician helps the patient to choose the best course of treatment. To evaluate new information and resolve issues regarding optimal treatment, the National Cancer Institute and the Office of Medical Applications of Research of the National Institutes of Health convened a Consensus Development Conference on the Management of Clinically Localized Prostate Cancer on June 15-17, 1987. After 1'2 days of presentations by experts and discussion by the audience, a consensus panel drawn from specialists and generalists from the medical profession and related scientific disciplines, clinical investigators, and public representatives considered the evidence. The panel agreed on answers to the following key questions: 1) What is the utility of pathologic assessment and imaging techniques in the staging of prostate can- cer? When is pelvic node dissection necessary? 2) Who is the optimal candidate for radical prosta- tectomy? What is the morbidity of the procedure, and how can it be minimized with preservation of curative potential? 3) Who are the candidates for definitive radiation therapy? What methods are optimal, and what are the long-term results in terms of local control and survival? What is the morbidity of the procedures, and how can it be minimized with preservation of curative potential? 4) Should definitive radiation therapy, hormone ther- apy, and/or chemotherapy be used as adjuvant treatment in high-risk patients? 5) What directions for future research are indicated? 3 QUESTIONS AND ANSWERS What is the utility of pathologic assessment and imaging techniques in the staging of prostate cancer? When is pelvic node dissection necessary? An accurate histologic or cytologic diagnosis of cancer of the prostate is essential prior to the planning of treat- ment. The histologic diagnosis usually is established from a specimen obtained by core biopsy or transurethral resec- tion. It may be obtained cytologically from a fine-needle aspiration. Histologic and/or cytologic grading and staging (extent of disease) are necessary in plans for appropriate treatment and provision of a prognosis. There are several systems for grading prostate cancer that yield similar information, and all are based on the degree of tumor differentiation and growth patterns. The Gleason histologic grading system is based on five tumor patterns. The values of the predominant and the lesser patterns are added together to generate a numerical histologic score. Histologic grading correlates well with the biologic degree of malignancy, i.e., tumor invasion, metastatic spread, and mortality rate. The nature of the specimen obtained by fine-needle aspi- ration usually precludes histologic but not cytologic grad- ing based on nuclear anaplasia. Experienced observers find that cytologic grading yields results and information simi- lar to those of histologic grading, but this requires further prospective correlation. The simplicity of fine-needle aspi- ration should increase its use. All tissue sampling techniques have a recognized risk of error. Flow cytometry can be used to measure DNA content. Abnormal DNA content appears to be predictive of more aggressive biologic behavior. Additional study is required for the determination of its prognostic utility in routine practice. Pelvic lymphadenectomy is a surgical staging procedure used when clinical decisions depend on accurate knowl- edge of the presence or absence of metastatic tumor in the pelvic lymph nodes. It is performed most commonly prior to a planned radical prostatectomy. Pelvic lymph node dis- section continues to provide staging information that can be obtained by no other method. Biochemical evaluation is critical to the initial staging assessment. Serum alkaline phosphatase and acid phos- phatase determinations are helpful in identifying patients with metastatic disease. The serum prostate-specific anti- gen is elevated more frequently in men with prostate can- cer than is the acid phosphatase and can be used to monitor response to both local and systemic therapies. Prostate-specific antigen is specific for prostate tissue but not for prostate cancer, which precludes its use in screening. Anatomic imaging contributes to the staging assess- ment of patients presenting with prostate carcinoma. The initial evaluation requires isotopic bone scan, chest roentgenogram, and an imaging evaluation of the upper urinary tract. Although lymphography has been used in the past, its current applications are limited. More locally directed imaging techniques for clinical staging include transrectal ultrasound, computed tomography, and mag- netic resonance imaging. It is essential that results of all imaging examinations be carefully correlated with results of digital rectal examination, cystoendoscopy, and pertinent laboratory tests in the determination of the clinical stage of the cancer. Imaging equipment should be state of the art. There is no single best procedure, and all procedures have significant limitations in their accuracy. The value of transrectal ultrasound is operator dependent; a dual-modality intrarectal probe should be used. Intra- prostatic anatomy, capsular integrity, and the seminal vesi- cles can be imaged. Pelvic adenopathy cannot be demon- strated. Computed tomography can help one assess the peripros- tatic area and lymph node size. Intraprostatic detail is poor, and masses rarely are identified. The detection of enlarged lymph nodes is highly sensitive, but normal-sized nodes may contain microscopic tumor. Magnetic resonance imaging, like transrectal ultrasound, can demonstrate intraprostatic anatomy. The technique ap- pears to be as accurate as computed tomography and trans- rectal ultrasound in the detection of periprostatic exten- sion and seminal vesicle involvement. It is competitive with computed tomography in demonstrating pelvic adenopathy. At present, scientific controlled studies are lacking in adequate patient numbers to permit other than a gross comparison of imaging methods. Clinical comparisons are biased by patient selection and nonblinded observer factors. Who is the optimal candidate for radical prostatectomy? What is the morbidity of the procedure, and how can it be minimized with preservation of curative potential? A radical prostatectomy includes the removal of the en- tire prostate and seminal vesicles with adequate resection margins by either a retropubic or perineal approach. The adequacy of surgical margins should be confirmed by a thorough pathologic evaluation. An appropriate candidate for definitive primary radical prostatectomy has a tumor that is localized to the prostate (stage A2, Bl, or B2). The patient should be an acceptable surgical candidate and have no significant “comorbid” disease. Patients should have completed a negative staging evaluation for distant metastases. Ideally, pelvic lymphadenectomy should be per- formed with a retropubic prostatectomy and for high-risk patients having a perineal prostatectomy. Adequate staging information will be obtained with a “limited node dissec- tion” that includes nodes within the boundaries of the exter- nal and internal iliac vessels and obturator fossa. This will minimize complications of lymphedema and lymphocele. Radical prostatectomy is associated with both peri- operative morbidity and late side effects. These complica- tions can include urinary incontinence, urethral stricture, impotence, and morbidity associated with anesthesia and a major surgical procedure. Significant incontinence and stricture are uncommon. The incidence of impotence can be reduced with newer surgical techniques with an anatom- ical dissection that preserves nerves necessary for erection. The preservation of potency with this technique is achieved in a majority of patients, although it is related to tumor size and patient age. Attempts to preserve potency must not compromise adequate removal of the tumor. Clinically evident local recurrence following radical NCI MONOGRAPHS, NUMBER 7, 1988 prostatectomy is not common. Cancer-free survival rates at 15 years for patients with cancer limited to one lobe of the prostate approach the expected survival of men in the general population in a comparable age group. Disease-free survival rates are less with larger tumors. Who are the candidates for definitive radiation therapy? What methods are optimal, and what are the long-term results in local control and survival? What is the morbidity of the procedures, and how can it be minimized with preservation of curative potential? Candidates for definitive radiation therapy must have a confirmed pathologic diagnosis of cancer that is clinically confined to the prostate and/or surrounding tissues (stage A2, B, or C). Patients should have completed a negative staging evaluation for distant metastases. An effective radiation therapy regimen should deliver a homogeneous dose to the entire tumor volume. Such a dose is delivered most commonly by external-beam radia- tion therapy. In selected instances, interstitial irradiation can be delivered with or without external-beam radiation ther- apy when the disease is confined to the prostate. Although lower local control rates have been reported with intersti- tial therapy, efforts are being made to improve these results by careful attention to technique and dosimetry. Newer ra- diation modalities such as particle therapy are under active investigation. Local tumor in the prostatic and periprostatic tissue can be controlled effectively by irradiation. However, the impact on overall survival of treating the regional lymph nodes is unclear. Definitive radiation therapy is associated with both acute and chronic effects on normal tissue, including proctitis, enteritis, and cystitis. These effects are generally accept- able and reversible but may be chronic; they are rarely sufficiently severe to require corrective surgical interven- tion. Other complications include occasional urethral stric- ture formation in patients who have undergone a previous transurethral resection of the prostate. Potency is preserved with definitive radiation therapy in the majority of patients but may diminish over time. Morbidity can be minimized by the use of sophisticated irradiation techniques, including the use of linear accelerators producing high-energy x-ray beams, careful treatment planning including simulation, and individualized shielding. The assessment of local control is critical in evaluation of the results of radiation therapy. Reported rates of local tumor control are a function of the diagnostic method used to establish them. By clinical criteria alone, the local con- trol rates are higher than if postirradiation positive biopsies are used as the end point. The likelihood of finding histo- logically positive biopsies following radiation therapy is of concern to physicians and is related to the clinical disease stage. The significance of this currently is being studied, but it may be predictive of subsequent distant relapse. Survival is related to the initial clinical stage and his- tologic grade. The 10-year crude survival rates of patients with low-stage disease (A2 or B) treated with radiation ther- apy are equivalent to the expected survival of men in the comparable age groups. Survival rates in more locally ad- vanced stages of disease clearly reflect the increased inci- dence of cancer death. MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER Should definitive radiation therapy, hormone therapy, and/or chemotherapy be used as adjuvant treatment in high-risk patients? No data are available to support the routine use of ad- juvant therapy after definitive surgery or irradiation. Pre- liminary data suggest that adjuvant hormone manipulation and irradiation deserve further study in locally advanced prostate cancer patients. A definition of the appropriate pa- tient population for such studies remains to be developed. What directions for future research are indicated? The Consensus Development Conference on the Man- agement of Clinically Localized Prostate Cancer provided a large body of information to optimize the diagnosis, stag- ing, and management of this prevalent disease. Whereas many controversies were addressed, numerous questions were identified that await answers and thus serve as the fo- cus for future research directions. These issues will require the collaborative input of investigators from both clinical and basic disciplines. As outlined in the consensus state- ment, many questions are being answered currently; others will be the focus for future research. Directions for Clinical Research 1) Agree to a uniform classification and schema for histologic and cytologic grading, disease staging, and response criteria that are acceptable to health care professionals caring for patients with prostate cancer. 2) Define the appropriate use of diagnostic imag- ing in staging prostate cancer patients through well-designed, controlled comparison studies. The development of imaging methods to measure tu- mor volume could provide a noninvasive predictor of the aggressiveness of cancer. 3) Encourage educational programs for pathologists and cytologists in the diagnosis of prostate can- cer with the purpose of increasing accuracy and uniformity. 4) Assess the availability and quality of surgical treat- ment and initiate surgical educational programs as needed. 5) Accept a uniform method for data reporting and statistical analyses that will allow meaningful comparisons of treatment results reported by var- ious disciplines. 6) Identify clinical and pathologic prognostic vari- ables. The identification of low- and high-risk fea- tures may allow more appropriate selection of treatments for patients with clinically localized disease. Parameters to study may include morpho- logic predictors, the correlation of DNA flow cy- tometry, prostate-specific antigen determination, and tumor cytogenetics with disease outcome. 7) Assess the clinical significance of positive postir- radiation biopsies and identify ways to reduce the incidence of these positive biopsies. 8) Assess in controlled trials the role of localized postoperative irradiation in patients with positive margins after radical prostatectomy. 9) Assess in controlled trials the role of adjuvant hor- monal therapy in patients with locally advanced disease after radical prostatectomy and/or defini- tive radiation therapy. 10) Clarify the clinical significance and therapeutic implications of the extent of nodal involvement. 11) Address the influence of treatment programs on the quality of life of patients and their loved ones. Identify appropriate psychosocial and psychosex- ual instruments and end points to assess quantita- tively the effect of treatment in patients with both localized and metastatic disease. Study and imple- ment innovative interventions to improve the psy- chological outcome. 12) Agree on a uniform clinical and pathologic def- inition of stage Al prostate cancer. Initiate stud- ies to define the natural history of untreated stage Al patients to help determine which patients may benefit from treatment. Directions for Basic Research 1) Encourage basic research to elucidate fundamental processes regulating normal prostate and prostate cancer growth and their impact on the natural his- tory of disease. 2) Assess the diagnostic and therapeutic role of prostate cancer-specific monoclonal antibodies. CONCLUSIONS Radical prostatectomy and radiation therapy are clearly effective forms of treatment in physicians’ attempts to cure tumors limited to the prostate for appropriately selected patients. Comparisons across studies suggest comparable 10-year survival rates with either form of management. What remains unclear is the relative merit of each in pro- ducing lifelong freedom from cancer recurrence. It is known that traditional radical prostatectomy can provide 15-year cancer-free survival in appropriately selected patients that is equivalent to that of a comparably aged control popu- lation. On the other hand, sufficient long-term follow-up does not exist to permit a conclusion about the ability of radiation therapy to eradicate such cancer in an equivalent proportion of patients. After appropriate primary irradiation, the long-term complication rate is now well defined and appears accept- able. The new approach to prostatectomy is clearly associ- ated with a reduction in postoperative impotence. The true comparative incidence of impotence over time, however, awaits prospective evaluation. Although impotence may re- sult from the alteration of normal anatomy, the psycholog- ical considerations should not be overlooked. Sexual reha- bilitation should address both medical and psychological needs. Information that a patient should have available when considering with his physician the choice of treatment in- cludes: 1) probability of cure, mortality, complications, and other side effects of radical prostatectomy and ra- diation therapy; 2) risk of impotence and incontinence for either treat- ment; 3) psychosocial consequences of either choice; 4) extent and risk of pretreatment staging assessment tests; and 5) economic consequences of each form of treatment. As competing, non-cancer-related causes of death (e.g., cardiovascular disease) may be expected to decrease for men over the age of 50, the issue of cure will become more important in low-stage disease. Properly designed and completed randomized trials that evaluate both disease control and quality of life after modern radiation therapy compared with radical prostatectomy are essential. Overview: Historical and Contemporary Willet F. Whitmore, Jr.! ABSTRACT—Recognition of the clinical importance of prostate cancer undoubtedly was delayed by the failure of clini- cians or pathologists to distinguish consistently between benign and malignant prostatic growths until well into the 19th cen- tury. White used castration for prostatic enlargements in 1895, but Huggins and Hodges first placed endocrine therapy on a ra- tional basis in 1941. Although a number of surgeons had at- tempted excision of prostate cancer, Young is credited with plan- ning and performing the first radical perineal prostatectomy in 1904. Orthovoltage irradiation and various techniques of intersti- tial and intracavitary radium therapy were used in the treatment of prostate cancer early in the 20th century, but it was the devel- opment of megavoltage irradiation that reopened the door to the exploration of irradiation for localized prostate cancer following World War II. Endocrine manipulation, surgery, and irradiation remain the keystones of treatment. The management of prostate cancer is controversial for several reasons: 1) The disease oc- curs in an age range in which competing causes of mortality are high. 2) The natural evolution of the disease is varied, often long, and not consistently predictable. 3) Long-term survival has been reported for each of the principal modes of therapy, but random- ized controlled studies have been limited. Uniformity in histologic grading, clinical staging, and evaluation of response to treatment would improve the quality of the data. Predictions of host life expectancy, tumor growth rate, metastatic potential, and tumor responsiveness to irradiation and endocrine therapy would en- hance the rationale of treatment.—NCI Monogr 7:7-11, 1988. This overview of prostate cancer is intended to provide a brief historical background, the major bases for uncertain- ties regarding management, some generalizations relative to treatment, resultant pertinent questions, some probable needs for clarifying current issues, and some tentative con- clusions. HISTORICAL BACKGROUND For a neoplasm as common and as important as it has proved to be, prostate cancer has been slow to achieve the attention it deserves. Some historical landmarks help to place the contemporary picture in perspective. Although the Ebers papyrus from 1500 BC (/) suggests that prostatism was recognized as a problem in the aging male, distinctions between benign and malignant prostatic enlargements were rarely made, either by clinicians or by pathologists, until the 19th century. Langstaff [cited in (2)] in 1817 may have been the first to report an authentic case of prostate car- cinoma. However, the infrequency with which the disease was recognized is suggested by Tanchon’s [cited in (2)] re- view of 8,289 cancer deaths in Paris, France, between 1830 and 1840; only 5 cases involved the prostate. In 1898, Al- ! Urology Service, Department of Surgery, Memorial Sloan-Kettering Cancer Center, 1275 York Ave., New York, NY 10021. barran and Halle [cited in (2)] demonstrated the frequency of the disease in reporting carcinoma in 14 of 100 prostatic enlargements. The independent and simultaneous publica- tions of Rich (3) and Moore (4) in 1935 called the atten- tion of clinicians and pathologists to the high prevalence of prostate cancer demonstrated by autopsy reports. Relative to surgical treatment, Kuchler (5) in 1866 and Billroth [cited in (2)] in 1867 attempted perineal excisions of prostate cancers, and Kuster [cited in (2)] in 1891 used cystoprostatectomy and ureterointestinal diversion for a lo- cally extensive prostate neoplasm. Despite these and other pioneering surgical ventures, it remained for Young [cited in (2)] to develop systematically and then perform radical perineal prostatectomy. Dr. Hugh Young first performed the operation on April 7, 1904, assisted by Dr. William Hal- sted, who may be credited with having inspired Young's efforts. The development of surgical excision for localized disease was a milestone. Radical perineal prostatectomy, however, was never widely performed. Patients meeting the appropriate selection criteria were rare, and few urologists became sufficiently adept to be comfortable with the pro- cedure. Impotence was a usual sequel, urinary incontinence was not uncommon, and rectal injury was a formidable pos- sible complication. Furthermore, irradiation and endocrine therapy later developed as alternative forms of treatment. With the description of the retropubic approach to the prostate by Millin (6) in 1945, interest in total prosta- tectomy was renewed. The problems of impotence, urinary incontinence, and possible rectal injury remained, but the familiarity of urologists with suprapubic exposures assured a wider use of the procedure than had been afforded the perineal operation. When the anatomic dissections of Walsh and Donker (7) in 1982 appeared to unravel the mystery of the impotence associated with total prostatectomy, Walsh and his associates (8) in 1983 introduced modifications in the operative technique that led not only to a high probability of preservation of sexual potency but also to a more systematic and controlled operation that restimulated the interests of clinicians in the surgical control of localized prostate cancer. Lymph node metastasis warrants special comment. Kocher [cited in (2)] operated for lumbar lymph node metastasis in a patient with prostate cancer in 1882, and Pasteau [cited in (2)] gave a remarkably complete and accurate description of the lymph node drainage of the prostate in 1897. In 1925, on the basis of the review of a large Mayo Clinic experience with prostate cancer, Bumpus (9) concluded: “. . . the lymphatic system is without doubt the earliest and most frequent site of metastatic lesions. . ..” Despite these and other evidences of the familiarity of clini- cians and pathologists with the occurrence of lymph node metastasis from prostate cancer, few surgical efforts have been directed toward the control of such metastasis. Hal- sted’s perception of female breast cancer was of a primary tumor possibly involving the regional lymph nodes of the supraclavicular fossa and axilla but still a potentially loco- regional process. He designed radical mastectomy to en- compass and possibly cure such lesions. Young’s perception of prostate cancer was of a local lesion potentially involv- ing the prostate and adjacent seminal vesicles. Although he would have been aware of the possibility of lymphatic dis- semination, the regional lymph nodes were clinically occult and clinical staging procedures for their evaluation were nonexistent. The radical prostatectomy he designed was aimed at primary tumor control and possible cure. The poli- cies regarding the regional lymph nodes that were formu- lated by surgeons treating breast cancer and by urologists treating prostate cancer may have been imprinted by these respective pioneers. It has been difficult to orient breast cancer surgeons toward the concept that regional lymph node metastasis implies systemic dissemination or urologic surgeons toward the concept that regional lymph node in- volvement may be consistent with still curable locoregional disease. One may speculate that the initial perceptions of Halsted and Young were instrumental in orienting these di- ametrically opposed treatment philosophies. The use of x-rays and radium in the treatment of cancers followed shortly after the respective discoveries of Roent- gen in 1895 and of the Curies in 1903. Loumeau in 1907 cited the use of x-rays, and Minet in 1908 and Pasteau, Wickham, and Degrais in 1910 [all as cited in (2)] used intraurethral radium to treat prostate neoplasms. The de- velopment of megavoltage irradiation during the 1940s made feasible the accurate delivery of sufficiently high ra- diation doses to stimulate the systematic investigation of external-beam irradiation in the treatment of prostate can- cer, which is epitomized by the studies of Bagshaw and Kaplan (/0) in 1962 and Budhraja and Anderson (//) in 1964. In 1952, Flocks and his associates (/2) combined surgery and the injection of colloidal "Au in the treatment of selected patients. This was the forerunner of explorations of brachytherapy with such radionuclides as '98Au, 123], and 92Ir. Although White [cited in (2)] used castration in the treat- ment of prostatic enlargements in 1893, this approach was abandoned by surgeons in the 1920s because of unconvinc- ing benefit from the procedure, failure to distinguish be- nign from malignant prostatic enlargements, and progres- sive evolution of therapeutic alternatives. It remained for Huggins and Hodges (/3) in 1941 to establish the rationale and demonstrate the effectiveness of endocrine therapy in the management of advanced prostate cancer. Surgery, irradiation, and endocrine therapy, either alone or in various combinations, remain the principal current methods of prostate cancer therapy. BASES FOR UNCERTAINTIES Current uncertainties in the management of localized prostate cancer may be justly credited to several consid- erations: i. A—B cD ~~ time FIGURE 1.—Possibilities in stage progression. 1) One factor is the rapidly increasing clinical incidence of the disease after the age of 50 years, a time when there is a progressively rising risk of death from other causes. This makes “quality survival” a legitimate alternative to “cure” in selected patients with prostate cancer. Growing older is invariably fatal; prostate cancer is only sometimes so! 2) An important consideration is the varied and unpre- dictable natural evolution of the disease (fig. 1), which ac- counts for various possible rates and patterns of local and metastatic disease progression (/4). The Patterns of Care Study of the American College of Surgeons (15) has yielded specific estimates of the contemporary relative frequency of occurrence of the various stages of prostate cancer. Experi- ence with regional lymphadenectomy in patients with stages A, B, and C prostate cancer has provided what may be con- sidered minimal estimates of the metastatic state associated with each of the respective local tumor categories (16). Nevertheless, for an individual patient with clinical stage A, B, or C prostate cancer, the rate and pattern of local and/or distant disease progression are imprecisely predict- able. The earlier the stage of the neoplasm, the greater the number of possibilities in stage progression and the greater the number of treatment options, and in the absence of con- trols, the more difficult the assessment of the contribution of therapy to the end result. 3) Another uncertainty relates to inaccuracies in clini- cal staging. Clinical methods for defining the pathologic extent of the neoplasm have progressed enormously but re- main short of ideal. In Young's early experience with radi- cal prostatectomy, suitability for the operation was largely determined by history, a careful digital rectal palpation of the prostate, perhaps cystoendoscopy, and possibly a poor-quality radiograph. Continual refinements in clinical staging have occurred since that time (fig. 2). The resul- tant impact on case selection and on the potential for stage migration require no elaboration. 4) There is also a lack of uniformity in histologic grading, in clinical staging procedures and classification, and in criteria for assessment of response to treatment. 1900 ———— — TIME —————- 1985 HISTORY PHYSICAL EXAMINATION SERUM ACID PHOSPHATASE SERUM ALKALINE PHOSPHATASE INTRAVENOUS UROGRAM BONE SCAN PELVIC COMPUTED TOMOGRAPHY SCAN ULTRASOUND HISTORY PHYSICAL EXAMINATION ? ENDOSCOPY ? RADIOGRAPH FIGURE 2.—Clinical staging for localized prostate cancer. NCI MONOGRAPHS, NUMBER 7, 1988 5) A final factor is the multiplicity of treatments, all inadequately controlled and applied differently to variously selected patients, often with incomplete assessment of the impact on the quality of life. TREATMENT GENERALIZATIONS Appropriate treatment implies that therapy is not applied if it is unnecessary or if it will be ineffective and will de- crease qualitative enjoyment of predicted life expectancy. For patients with stage A prostate cancer, the usually pro- tracted tumor evolution and the competing mortality from other causes provide uncertain bases for categorical treat- ment. Although all stages B, C, and D cancers must origi- nate from stage A lesions, the vast majority of stage A neo- plasms remain so throughout the lifetime of the host. Ep- stein et al. (/7) and Blute and associates (/8) have recently confirmed the lethal potential of some clinically identified stage A lesions but have simultaneously illustrated the far greater probability of death from other causes in these re- spective experiences. Categorical treatment of all clinically identified stage A lesions would undoubtedly constitute therapeutic “overkill.” Predictability of both the biologic potential of such a tumor and the life expectancy of its host is an ultimate requirement for rational treatment recommendations and constitutes a more urgent objective than does earlier diagnosis of such lesions (19). Furthermore, the specific effectiveness of the current treatment armamentarium in patients with stage A prostate cancers has yet to be established by clinical experience. For selected patients with stage B prostate cancers, there are gross similarities in clinical local control and overall 10- and 15-year survival rates following either external-beam irradiation (20-23) or surgical excision (24,25). The possi- ble persistence of locally positive biopsy specimens and the variable use of endocrine therapy in such radiation ther- apy experiences suggest that a great proportion of patients treated by surgical excision may actually be free of neo- plasm at 15 years but simultaneously suggest that the over- all survival advantage of cure at 15 years is, at best, small. Current information does not exclude a possibly sig- nificant survival advantage of a particular treatment after 15 years, but the age range of the prostate cancer popula- tion and competing causes of mortality suggest that the rel- evance of cure of this neoplasm to survival per se will pro- gressively diminish as follow-up intervals become longer. Apparent cures, disregarding variation in definitions, have followed external-beam irradiation, surgical excision, interstitial irradiation (26), and even endocrine therapy (27-30). Whether the subsets of patients cured following different treatments are identical is unknown, but the sal- vage by irradiation following local failure after surgery (31) and the salvage by surgery following local failure after ir- radiation (32) suggest that the subsets may be at least par- tially different. If one considers the frequency with which a watchful waiting course has probably been applied in the manage- ment of stage B prostate cancer, there is a dearth of in- formation on the subject. Endocrine therapy usually has been administered for disease progression in such circum- MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER stances. Limited evidence (27,28,33,34) suggests that, in such patients, overall 15-year survival rates (not based on no evidence of disease) are grossly similar to those achieved in apparently comparable patients managed by radical ex- cision or irradiation. Data concerning rates of local and distant disease progression in patients with stage B cancer managed either by watchful waiting or by endocrine ther- apy are extremely limited (34), and quality-of-life issues with patients so managed remain to be addressed. For stage C prostate cancer, external-beam irradiation or endocrine therapy has been the most common treatment, but a convincing survival or local control advantage to ei- ther approach is lacking (2/,22,35,36). Irradiation, surgery, or endocrine therapy has been useful for local control in some patients who have failed alternative treatment, but the relative effectiveness of the different management options in various circumstances is unclear. Variously selected pa- tients have received the following forms of treatment with some apparent clinical success: radical prostatectomy (37), cystoprostatectomy with lymph node dissection (38-41), endocrine therapy followed by radical prostatectomy (42), radical prostatectomy with lymph node dissection and en- docrine therapy (43,44), interstitial irradiation with lymph node dissection (26), and radical excision with interstitial irradiation, lymph node dissection, and endocrine therapy (45,46). Lack of control data, varied selection criteria, fre- quent use of endocrine therapy, and inconsistent response criteria confound interpretations. QUESTIONS The foregoing considerations justify the following perti- nent questions: 1) Is cure necessary? Although many more die with prostate cancer than of it, the need for cure is amply demon- strated by the significant mortality from the disease. 2) Is cure possible? On the basis of 15-year survival with- out evidence of neoplasm after treatment of selected stage B prostate cancers by surgical excision or irradiation, the answer is a qualified yes. There is evidence that such sur- vival per se must be considered to be a qualified indication of cure. This evidence includes: local or distant recurrence first recognized 15 or more years after local treatment (47), grossly similar 15-year survival rates following watchful waiting with endocrine therapy if necessary, and grossly similar 10- to 15-year survival rates after either surgical ex- cision or irradiation with variable supplemental endocrine therapy. 3) Is cure necessary in those for whom it is possible? This question is derived primarily from the grossly similar 10- to 15-year survival rates in selected patients with stage B cancer following management by watchful waiting vari- ably supplemented with endocrine therapy, by radical ex- cision, or by radiation therapy variably supplemented with endocrine therapy. No definitive answer to this question is currently possible. 4) Is cure possible in those for whom it is necessary? This question derives from the weakness of the evidence that therapy has diminished mortality from prostate cancer and from the persistently poor prognosis associated with tumors of high grade and those with regional lymph node 10 metastasis. No definitive answer to this question is currently possible. POSSIBILITIES FOR CLARIFICATION Efforts to answer unresolved questions encourage two simply stated, although not simply accomplished, consider- ations. These are: 1) standardization and adoption of uniform criteria for the pathologic characterization, clinical staging, and for assessment of responses to treatment; and 2) appropriate randomized clinical trials and/or iden- tification of tumor-host features that constitute reli- able predictors of growth rate, metastatic potential, and responsiveness to a specific form of treatment. CONCLUSIONS The end result in the patient with prostate cancer may be considered a combined consequence of host natural history, tumor behavior, and treatment effectiveness. In the absence of controlled studies or reliable predictors of either host or tumor behavior, the relative impacts of specific therapies on tumor control and survival remain ambiguous. The pos- sibility that the long survival after treatment is more a con- sequence of the natural behavior of the disease than of the nature and consequences of the treatment is one with which surgical and radiation oncologists currently must live. Quite apart from the uncertain contribution of various treatments to cancer cure, local control is a realistic and useful treat- ment objective. The relative efficiencies of different meth- ods of treatment, including endocrine therapy, in attaining this limited objective are in need of additional study, which should incorporate further attention to treatment impact on the quality of life. There is logically no best method of treatment for all patients, but rather a variety of methods, one of which is best for a particular patient and his particular tumor. Fur- thermore, the best treatment for the cancer may not nec- essarily be the best treatment for the patient. The chal- lenge for physicians is to identify features in the individual host-tumor setting that indicate the need for and specify the optimal method of management. REFERENCES (1) EDWARDS L. History of nonsurgical treatment. In Be- nign Prostatic Hypertrophy (Hinman F, ed). New York: Springer-Verlag, 1983, pp 30-34. (2) HerssST RH, POLKEY HJ. Prostatic malignancy. In History of Urology (Ballenger EG, Frontz WA, Hamer HG, et al, eds), vol. II. Baltimore: Williams & Wilkins, 1933, pp 187-208. (3) RicH AR. On the frequency of occurrence of occult carci- noma of the prostate. J Urol 1935;33:215-223. (4) MOORE RA. The morphology of small prostatic carcinoma. J Urol 1935;33:224-234. (5) KucHLER H. Uber Prostatavergroserungen. Deutsch Klin 1866;18:458-465. (6) MILLIN T. Retropubic prostatectomy, new extravesical tech- nique: Report on 20 cases. Lancet 1945;2:693-696. (7) WALSH PC, DONKER PJ. Impotence following radical prosta- tectomy: Insight into etiology and prevention. J Urol 1982;128:492-497. (8) WALSH PC, LEPOR H, EGGLESTON JC. Radical prostatectomy with preservation of sexual function: Anatomical and pathological considerations. Prostate 1983;4:473-485. (9) Bumpus HC. Carcinoma of the prostate. Surg Gynecol Ob- stet 1926;43:150-155. (10) BAGSHAW MA, KAPLAN HS. Radical external radiation ther- apy of localized prostatic carcinoma. Presented at the 10th International Congress of Radiology, Montreal, PQ, Canada, 1962. (11) BUDHRAJA SN, ANDERSON JD. An assessment of the value of radiotherapy in the management of carcinoma of the prostate. Br J Urol 1964;36:535-540. (12) Frocks RH, KErRrR HD, ELKINS HB, et al. Treatment of carcinoma of the prostate by interstitial radiation with radio-active gold ("8 Au): A preliminary report. J Urol 1952;68:510-522. (13) HuGGINs C, HopGEs CV. Studies on prostatic cancer. I. The effect of castration, of estrogen, and of androgen injection on serum phosphatase in metastatic carcinoma of the prostate. Cancer Res 1941;1:293-297. (14) WHITMORE WF JR. The natural history of prostatic cancer. Cancer 1973;32:1104-1112. (15) MURPHY GP, NATARAJAN N, PONTES JE, ET AL. The National Survey of Prostate Cancer in the United States by the American College of Surgeons. J Urol 1982;127:928-934. (16) DONOHUE RE, MANI JH, WHITESEL JA, ET AL. Pelvic lymph node dissection. Guide to patient management in clini- cally locally confined adenocarcinoma of prostate. Urol- ogy 1982;20:559-565. (17) EPSTEIN JI, PAULL G, EGGLESTON JC, ET AL. Prognosis of untreated stage Al prostatic carcinoma: A study of 94 cases with extended follow-up. J Urol 1986;136:837-839. (18) BLUTE ML, ZINCKE H, FARROW GM. Long-term follow-up of young patients with stage A adenocarcinoma of the prostate. J Urol 1986;136:840-843. (19) WHITMORE WF JR. Background for screening: Natural his- tory and treatment. In Progress and Controversies in On- cological Urology (Schroeder F, Klijn J, Kurth K, et al, eds), vol II. New York: Liss, pp 123-130. (20) vAN DER WERF-MESSING B. Prostatic cancer at the Rotter- dam Radiotherapy Institute. Strahlentherapie 1978;154: 537-541. (21) BAGSHAW MA. Radiotherapy of prostatic caricer: Stanford University experience. In Progress and Controversies in Oncological Urology (Kurth K, Debruyne F, Schroeder F, et al, eds), vol I. New York: Liss, 1984, pp 493-512. (22) BAGSHAW MA. Potential for radiotherapy alone in prostatic cancer. Cancer 1985;55:2079-2085. (23) PEREZ CA, PILEPICH MV, ZIVNUSKA F. Tumor control in definitive irradiation of localized carcinoma of the prostate. Int J Radiat Oncol Biol Phys 1986;12:523-531. (24) WaLsH PC, JEWETT HJ. Radical surgery for prostatic cancer. Cancer 1980;45:1906-1911. (25) GiBBONS RP, CORREA RJ JR, BRANNEN GE, ET AL. To- tal prostatectomy for localized prostatic cancer. J Urol 1984;131:73-76. (26) ScARDINO PT, CARLTON CE. Combined interstitial and ex- ternal irradiation for prostatic cancer. In Principles & Management of Urologic Cancer (Javadpour N, ed). Bal- timore: Williams & Wilkins, 1983, pp 392-408. (27) BARNES RW, BERGMAN RT, HADLEY HL, ET AL. Early pros- tatic cancer: Long-term results with conservative treat- ment. J Urol 1969;102:88-90. (28) BARNES R, HADLEY H, AXFORD P, ET AL. Conserva- NCI MONOGRAPHS, NUMBER 7, 1988 tive treatment of early carcinoma of prostate. Urology 1979;14:359-362. (29) FaLkowskl WS, O’CoNOR VJ JR. Long-term survivor of prostatic carcinoma with lung metastases. J Urol 1981;125:260-262. (30) JOHANSSON S, LIUNGGREN E. Prostatic carcinoma cured with hormonal treatment: A report of two cases. Scand J Urol Nephrol 1981;15:331-332. (31) RAY GR, BAGSHAW MA, FREIHA F. External-beam radiation salvage for residual or recurrent local tumor following radical prostatectomy. J Urol 1984;132:926-930. (32) KAUFMAN JJ, SMITH RB, RAZ S. Radiation therapy in car- cinoma of the prostate: A contributing cause of urinary incontinence. J Urol 1984;132:998-999. (33) GRAVERSON PH, GASser TC, MADSEN PO, ET AL. Early prostatic cancer: Radical prostatectomy versus placebo. J Urol 1987;137:364A. (34) WHITMORE WF JR, ROSENBERG S, CHOPP R. Wait and see: Experience with B lesions. In Prostate Cancer. Part B. Imaging Techniques, Radiotherapy, Chemotherapy, and Management Issues (Murphy GP, Khoury S, Kuss R, et al, eds). New York: Liss, 1987, pp 387-394. (35) BYAR DP. The Veterans Administration Cooperative Uro- logical Research Group’s studies of cancer of the prostate. Cancer 1973;32:1126-1130. (36) PauLsoN DF. Treatment of locally confined prostatic can- cer with radiotherapy versus surgery limits of curability. In Progress and Controversies in Oncological Urology (Kurth F, Debruyne F, Schroeder F, et al, eds), vol I. New York: Liss, 1984, pp 483-492. (37) ScHROEDER FH, BELT E. Carcinoma of the prostate: A study MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 11 of 213 patients with stage C tumors treated by total perineal prostatectomy. J Urol 1975;114:257-260. (38) WHITMORE WF JR, MACKENZIE AR. Experiences with vari- ous operative procedures for the total excision of prostatic cancer. Cancer 1959;12:396-405. (39) WHITMORE WF JR. The rationale and results of ablative surgery for prostatic cancer. Cancer 1963;16:1119-1132. (40) SpAULDING JT, WHITMORE WF JR. Extended total excision of prostatic adenocarcinoma. J Urol 1978;120:188-190. (41) McCuLLoUGH DG, LEADBETTER WF. Radical pelvic surgery for locally extensive carcinoma of the prostate. J Urol 1972;108:939-943. (42) Scott WW, Boyp HL. Combined hormone control ther- apy and radical prostatectomy in the treatment of selected cases of advanced carcinoma of the prostate: A retro- spective study based upon 25 years of experience. J Urol 1969;101:86-92. (43) MYERS RP, ZINCKE H, FLEMING TR, ET AL. Hormonal treat- ment at time of radical retropubic prostatectomy for stage D1 prostatic cancer. J Urol 1983;130:99-101. (44) BENSON RC Jr, TOMERA KV, ZINCKE H, ET AL. Bilateral pelvic lymphadenectomy and radical retropubic prosta- tectomy for adenocarcinoma confined to the prostate. J Urol 1984;131:1103-1106. (45) Frocks RH. The treatment of stage C prostatic cancer with special reference to combined surgical and radiation ther- apy. J Urol 1973;109:461-463. (46) FLocks RH, O'DONOGHUE PN, MILLEMAN LA, ET AL. Surgery of prostatic carcinoma. Cancer 1975;36:705-717. (47) WHITMORE WF JR. Natural history and staging of prostate cancer. Urol Clin North Am 1984;11:205-220. I. Staging mo re ee ame ES 5 R— Histologic Grade, Clinical Stage, and Patient Age in Prostate Cancer Donald F. Gleason!* ABSTRACT—The highly variable and often prolonged clini- cal course of prostate cancer poses difficult problems. Some pa- tients appear to be at such low risk that overtreatment should be avoided. Many patients must be studied for many years before 2 treatments can be compared. If the patients could be sorted into groups with predictably different survival rates, such stud- ies could be completed in less time and/or with fewer patients. Accumulated experience indicates that the survival rates for pa- tients with a diagnosis of prostate cancer are determined largely by three factors: the clinical stage, histologic grade of the tumor, and the patient’s age. Treatment is a fourth variable factor that requires further study. In this paper, the relationships and inter- actions among grade, stage, and age are analyzed and discussed, and ways are suggested in which they can be combined to enhance stratification and discrimination in clinical trials of treatment. The information can also be applied broadly to the management of individual patients, but it is painfully obvious that we need a much larger body of accumulated treatment data that must in- clude more uniform clinical staging, uniform histologic grading, and detailed patient-age reporting. These data would help ad- just for the nonuniform mixture of patients in different studies. The problem of variable patient selection processes before admis- sion to a study affects the results of many reported studies and remains a difficult problem.—NCI Monogr 7:15-18, 1988. THE PROBLEM The variable and often prolonged clinical course of prostate cancer poses a severe challenge: Many patients must be studied for many years so that 2 treatments can be compared. Some patients with stages A and B cancers, di- agnosed early, appear to live out their normal life span and die of other causes or conditions. Recent reports remind us that some patients with stage A tumors do progress to death from cancer, but these are almost always those with less differentiated tumors. The fact remains that only 2%-10% of patients with stage A tumors die of cancer, as proved in various series of studies (/). Stage B tumors are, on the average, larger and more aggressive, but there is also sub- stantial overlap with stage A in this regard. Some simple autopsy statistics emphasize the problem. Careful examination of the entire prostate at autopsy re- veals that a surprising incidence of unsuspected carcinomas begins when patients are about 40 years old and increases steadily with age. If the incidence of these cancers is about 10% at age 55, 20% at age 65, and 30% at age 75, then ABBREVIATIONS: VA = Veterans Administration; VACURG = Veterans Administration Cooperative Urological Research Group. ! Department of Laboratory Medicine and Pathology, University of Minnesota School of Medicine, Minneapolis, MN. * Reprint requests to: Donald F. Gleason, M.D., Ph.D., 6211 Logan Ave., S., Minneapolis, MN 55423. two-thirds of the tumors found in the 75-year-old men must have been present for more than 10 years and one-third must have been present for more than 20 years! The actual incidence does not affect these conclusions, which depend only on the increasing incidence with age (fig. 1). Most of these tumors found at autopsy are small and well differentiated. They obviously caused no difficulty and would presumably have remained “silent” for many more years. They apparently were not serious threats to the health of the patients, but some of these tumors will inevitably be found in prostate tissue removed during life for presumed benign obstruction and then they do become management problems. The question is: Can we distinguish those tumors which grow very slowly from those which grow more rapidly and thereby avoid unnecessary and potentially harmful treat- ment? Histologic grading offers a partial solution to that question. HISTOLOGIC GRADING Prostate cancer also displays a wide range of histo- logic appearances, and many have reported strong cor- relations between histologic and biologic malignancy, as with many other cancers. Well-differentiated tumors grow slowly; poorly differentiated tumors grow rapidly. The VA (“Gleason”) grading system is standardized by a drawing (2,3) that facilitates adoption by other workers. The drawing identifies 5 grades of tumor, but 2 grades are often present in 1 patient. The 2 grade numbers are added together to create a histologic score, which can range from 2 to 10. (For pure single-grade tumors, the grade number is doubled.) In the VACURG studies, the histologic score in the orig- inal diagnostic biopsy was the most powerful single cor- relate with subsequently observed measures of biologic malignancy (2-4). It was strongly correlated with the sub- sequently observed cancer death rates (fig. 2) and was also correlated with the initial clinical findings, such as the ex- tent of tumor, the presence of hydronephrosis, elevation of serum acid phosphatase, etc. In addition, the histologic score correlated with the rate of progression from the initial clinical stage to higher stages (4) and with the incidence and extent of metastases at autopsy years later. Other workers confirmed the reproducibility of the VA grades (5) and confirmed and extended the histologic and biologic correlations (6-9). Some (7-9) even suggested that the correlation with lymph node metastases at exploratory laparotomy was strong enough to consider foregoing the risks of staging lymphadenectomy in patients with the high- est and lowest histologic scores. Thomas et al. (10) docu- mented a sharp increase in the incidence of “upstaging” 15 16 5 8383 =n © Oo Percent With Prostate Cancer °o 8 35 45 55 65 75 85 95 Age at Death, Autopsy FIGURE 1.—Approximate incidence of prostate cancers found at autopsy at various ages. Note that one-third of the tumors in the 75-yr-old men had been present more than 20 yr! (the finding of unexpected extension of tumor outside the prostate at laparotomy for prostatectomy) between histo- logic scores 6 and 7 in the initial biopsies of 130 clinical stages A and B patients (fig. 3). The separation of patients into groups with predictable degrees of malignancy should facilitate retrospective com- parison of treatments and optimize the prospective random- ization of treatments. Use of histologic grade can assist in the management and treatment decisions for individual pa- tients. CLINICAL STAGE Other initially observed parameters also correlate well with the subsequent course of prostate cancer, including tumor size, seminal vesicle invasion, extension through the capsule, lymph node metastases, distant metastases, ele- vated serum acid phosphatase, etc. Various combinations of these findings are used by physicians to define the clini- cal stages, which also correlate strongly with the subsequent clinical course. It is worth emphasizing that all of these parameters, including the histologic score, are obviously correlated broadly with the degree of biologic malignancy of the tu- mors. The clinical parameters have the quality of “mile- posts” which indicate how far the tumor has progressed along its course. The histologic grade appears to be more directly related to the specific abnormalities at the molecular-genetic level that control the histologic structure and function of the tumor cells and is strongly correlated with the rate at which the tumors grow and progress. The clinical stage appears to be the product of the innate degree of malignancy of the tumor and the duration of the tumor (unknown) before diagnosis. Four clinical stages are usually defined at the time of the initial diagnosis, commonly labeled A, B, C, and D or I, II, III, and IV in Death Rates per Year Histologic Score ® Cancer Deaths OAIll Deaths FIGURE 2.—Histologic scores (primary plus secondary grades) vs. cancer- specific and total death rates. Percent of Cases Upstaged 2 3 4 5 6 7 8 9 10 Histologic Score FIGURE 3.—Percent of 130 clinical stages A and B patients upstaged to stage C or D at attempted prostatectomy. Note sharp increase between histologic scores 6 and 7. Graph drawn from data of Thomas et al. (10). different studies. The VA studies used the simple, numbered clinical stages of that time: Stage I: No palpable tumor, no evidence of metas- tases. Stage II: Palpable, localized nodule, no known metas- tases. Stage III: Palpably extended tumor, no known metas- tases. Stage IV: Metastases present and/or elevated prostatic serum acid phosphatase. Tumors first diagnosed in stage III (C) progress more rapidly than stage 1 (A) and II (B) tumors because, on the average, not only are they a higher grade but also time elapsed while they were (unrecognized) in stage I (A) and/or II (B), i.e., lead time bias. Stage IV (D) tumors progress more rapidly than do stage III (C) tumors for similar reasons. The histologic score and the clinical stage showed partial correlation with each other in the VA data, as would be ex- pected, with more low-grade tumors in the lower stages and more high-grade tumors in the higher stages. However, al- most all the histologic grades can be found in all the clinical stages, and both stage and grade also had strong indepen- dent predictive correlations with the biologic malignancy of the tumors. When these correlations were combined, sim- ple addition of the score number to the stage number (with a weighted value of 5 for stage IV) gave a powerful new stage-grade category number that could range from 3 to 15 and was strongly correlated with the subsequent cancer death rates (fig. 4). This category combines the clinical stage and histologic grade into a single numerical expression of the risk of progression and cancer death for patients with that same category score. It consolidates the overlapping similar clin- ical courses of the patients with lower stages and higher 0.50 0.40 0.30 0.20 Death Rates per Year 0.10 0 3 4 5 6 7 8 9 10 11 12 13 14 15 Combined Grade-Stage Category ®m Cancer Deaths O All Deaths FIGURE 4.—Combined clinical stage-histologic grade category vs. cancer- specific and total death rates. NCI MONOGRAPHS, NUMBER 7, 1988 0.40 0.30 0.20 0.10 Cancer Death Rates per Year 0 3 4 5 6 7 8 Combined Stage/Grade Category 9 10 11 12 13 14 15 ® No DES Therapy @ DES Therapy FIGURE 5.—Reduction of cancer death by treatment with diethylstilbestrol (DES) above stage-grade category 8 but no detectable effect below category 9. grades and those with higher stages and lower grades. Also, it provided a useful categorization of the patients in the VACURG studies who did or did not benefit from treat- ment with diethylstilbestrol (fig. 5). This combined stage-grade category was largely ignored in the literature on the prostate because of the enthusiasm for the subdivided clinical stages introduced by Jewett (Al, A2,B1, B2, etc). In that system, the histologic grade appears only in substage A2 (defined as more than a certain amount of tumor or poorly differentiated tumor). The combined stage-grade category always incorporates the histologic differentiation of the tumor. Its value was recently tested and reaffirmed by the group at Memorial Sloan-Kettering Cancer Center (1/1). PATIENT AGE AT DIAGNOSIS Finally, it is easy for physicians to overlook one of the most powerful predictors of any survival curve, the age of the patient. Other factors being equal, a group of older men will die sooner than a group of younger men. After age 50, the changes per decade in the normal death rate are as great as or greater than the changes related to tumor stage and grade (fig. 6). Survival after diagnosis appears to depend largely on three factors: the histologic grade, clinical stage, and age of the patient. The effect of treatment is a fourth variable factor. 17 [= S ° @® S 2 I a S °o nN oS Fraction Surviving (1969-71) = 0 10 Years #60 Years ® 80 Years ® 50 Years O 70 Years FIGURE 6.—Normal survival rates for men of specific ages (DHEW mor- tality data of 1969-71 for white United States males). Published actuarial survival data like those in figure 6 can be combined mathematically with the cancer-specific death rate for a certain tumor grade and/or stage to produce a theoretically predicted survival curve for a group of such patients. The decrease in life expectancy can be calculated. Examples (combining only age and grade for simplicity) are shown in figure 7. They could be used in patient- doctor discussions of prognosis and therapy. Series of such theoretical curves provide a mathematical model for some of our intuitive ideas about balancing stage, grade, and age for patient management considerations. For example, it is apparent from figure 7 that grade 2+3 tumor in a 75-year- old man does not increase his mortality rate drastically, but the same tumor in a 45-year-old doubles his mortality rate. Grade 3+4 tumor will be associated with an ominous prognosis at any age. Calculation of the effects on the mortality rate could be used to help select the most appropriate treatment for groups of patients defined by age, stage, and grade. How- ever, the decision points are not defined by the numbers themselves. Additional carefully accumulated data and ex- perience would be required to help the physician decide if a 10% or 20% or 30% change in the death rate is enough to select treatment A over treatment B, for example. Finally, the general state of health and the opinions of the patient and his physician on the cost/benefit ratio of the risks and side effects of each treatment to the possible FIGURE 7.—Survival curves showing effect of three grades of prostate cancer on survival for 45-yr and 75-yr-old men. Upper curve in each graph is the normal survival curve for the age group (DHEW data for United States white males, 1969-71). Lower curve is the same normal sur- vival curve reduced by the cancer-specific death rate associated with tumors of VA histologic scores 5, 6, and 7. Rate =.015 Rate =.05 Rate =-.15 1.09 1.0 1.09 .84 = .8 . 84 Fraction -61 -6 - 61 Surviving ; ] i 4] Age 45 b 1 «24 “2 24 0 0 012345678910 012345678910 1. 1.0 1.0 8- .8 Fraction .6 Jb Surviving 4 4 Age 75 : 2+ 2 0 0 1 012345678910 012345678910 012345678910 VA Grade 2-3 VA Grade 3-3 VA Grade 3-4 MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 18 increases in survival will always enter into those choices. Those difficult value judgments will always remain, but mathematical analysis of the effects of histologic grade, clinical stage, and patient’s age may narrow some of the difficult areas. SUMMARY Uniform histologic grading, clinical staging, and patient- age reporting appear essential for further progress in the management and treatment of prostate cancer to facilitate impartial comparison of the results of various clinical stud- ies with the use of the correlations described herein. ADDENDUM It is imperative, in applying the VA grading system, that one record both histologic grades and histologic score: “The tumor is grade 3 and 4, score 7,” or “The tumor is pure grade 3, histologic score 6.” The unqualified numbers 2, 3, 4, or 5 might be the his- tologic scores of several well-differentiated prostate tu- mors, but they might also be the histologic grades of 1 well-differentiated, 1 moderately differentiated, and 2 poorly differentiated tumors, and thereby present possibly disastrous misunderstandings. REFERENCES (1) BYAR D, VETERANS ADMINISTRATION COOPERATIVE URO- LOGICAL RESEARCH GROUP. Survival of patients with in- cidentally found microscopic cancer of the prostate: Re- sults of a clinical trial of conservative treatment. J Urol 1972;108:908-913. (2) GLEASON D, MELLINGER G, VETERANS ADMINISTRATION COOPERATIVE UROLOGICAL RESEARCH GROUP. Predic- tion of prognosis for prostatic carcinoma by com- bined histological grading and clinical staging. J Urol 1974;111:58-64. (3) GLEASON D. Histologic grading and clinical staging of carci- noma of the prostate. In Urologic Pathology: The Prostate (Tannenbaum M, ed). Philadelphia: Lea & Febiger, 1977, pp 171-197. (4) BYAR D, CORLE D, VETERANS ADMINISTRATION COOPERA- TIVE UROLOGICAL RESEARCH GROUP. VACURG random- ized trial of radical prostatectomy for stages I and II pros- tatic cancer. Urology 1981;17(suppl, Part II):7-11. (5) BAIN G, KocH M, HANSON J. Feasibility of grading prostate carcinomas. Arch Pathol Lab Med 1982;106:265-267. (6) CORRIERE J, CORNOG J, MURPHY J. Prognosis in patients with carcinoma of the prostate. Cancer 1970;25:911-918. (7) HARZELL W, BEAN M, HILARIS B, ET AL. Prostatic adenocar- cinoma: Relationship of grade and local extension to the pattern of metastases. J Urol 1977;118:278-287. (8) PAULSON D, PISERCHIA P, GARDNER W. Predictors of lymphatic spread in prostatic adenocarcinoma. J Urol 1980;123:697-699. (9) KRAMER S, SPAHR J, BRENDLER C, ET AL. Experience with Gleason’s histopathologic grading in prostatic cancer. J Urol 1980;124:223-225. (10) THOMAS R, LEWIS R, SARMA D, ET AL. Aid to accurate staging—histopathologic grading in prostatic cancer. J Urol 1980;128:726-728. (11) SOGANI P, ISRAEL A, LIEBERMAN P, ET AL. Gleason grading of prostatic carcinoma: A predictor of survival. Urology 1985;25:223-227. Fine-needle Aspiration of the Prostate Mitchell C. Benson!:2 ABSTRACT —Fine-needle aspiration biopsy of the prostate has been used for over 20 years as the prostate biopsy technique of choice throughout much of Europe. However, this technique has only recently gained acceptance in the United States. There is now an enlarging body of data confirming that fine-needle aspiration can be viewed as a safe, accurate, and reliable means for detec- tion and diagnosis of prostate cancer. This review summarizes the European and United States experience with fine-needle aspira- tion and supports the utility of this biopsy technique as an impor- tant addition to the urologist’s diagnostic armamentarium.—NCI Monogr 7:19-24, 1988. Despite widespread use of transrectal aspiration biopsy of the prostate throughout Europe, this technique is only slowly gaining acceptance by urologists as a safe, reli- able, and accurate means of detecting prostate cancer in the United States (/-13). It appears that lack of experience with the technique on the part of the urologist and pathol- ogist is at the root of this inertia. Thus the purposes of our presentation are to 1) discuss the technique of aspiration biopsy, 2) review the European and American experience, and 3) project the role this technique will play in the diag- nosis and treatment of prostate cancer. Biopsy can be defined as the process of removing tissue from living patients for diagnostic examination. Prostate biopsy in the United States has traditionally been performed with instruments that remove a core or piece of tissue for histologic examination. This means that a portion of a tu- mor in its natural environment (epithelia, stroma, and vas- cular network) is removed, processed, and microscopically examined for the physician to establish a diagnosis. In ad- dition, the tumor cells can be graded, and this results in our ability to prognosticate survival, though not necessar- ily in an individually predictive manner. Broders (/4) was the first to point out that grade correlated with prognosis when he stated: “The time is not far distant when not only physicians but also the patients and their relatives as well as life insurance companies will be interested in the grade of malignancy of cancer. If a surgeon knows the grade of malignancy of a cancer in addition to its size and situation, he is certainly in a better position to render more efficient treatment, and a more accurate prognosis than one with- out such knowledge.” This statement has certainly become true, and it points out that biopsy not only can establish a diagnosis but sometimes can be useful in directing therapy. ! Department of Urology, Columbia-Presbyterian Medical Center, 622 West 168th St, New York, NY 10032. 2 I thank E. Darracott Vaughan (Society of the New York Hospital, New York, NY); Perinchery Narayan (University of California, San Francisco, San Francisco, CA), and Gerald Chodak (University of Chicago, Chicago, IL) for providing data; and Ms. Mary Petway for assistance in preparation of the manuscript. Finally, physicians can use biopsy results to monitor pa- tient response to therapy. Changes caused by disease regres- sion can be seen following radiation therapy and androgen withdrawal and in many cases are thought to be predictive of patient prognosis and duration of response (15-19). Thus histologic biopsy is used by physicians to establish a diagnosis, direct therapy via predicting the aggressive na- ture of a tumor, and to monitor responses by assessing tu- mor persistence or recurrence after therapy. The following information should serve to prove that these same tasks can be accomplished with fine-needle aspiration. TECHNIQUE Although Franzen (/) was not the first to perform trans- rectal aspiration biopsy of the prostate, he was certainly the individual responsible for developing the apparatus cur- rently in use. He described the use of a 22-gauge needle in- troduced into the prostate via the rectum. Although certain advances in instrumentation have made the technique eas- ier to perform, the methods remain essentially unchanged. The following materials are used: a 22-gauge disposable aspiration needle (Cook Urological, Inc., Spencer, IN); a 20-ml disposable Luer-Lok syringe (Becton Dickinson and Co., Rutherford, NJ); a needle guide (Cook Urological, Inc.), a pistol-grip syringe holder (Precision Dynamics Corp., San Francisco, CA, or Cook Urological, Inc.); sterile gloves; and microscope slides. Inasmuch as the rectum contains no pain sensory nerves, transrectal aspiration of the prostate can be performed with- out the use of anesthesia. In general, the patient feels no more discomfort than is experienced during a rectal exam- ination. Thus this procedure can be as safely and comfort- ably performed in the physician’s office as in the operating or cystoscopy suite. No preaspiration antibiotics are nec- essary, but as with all urologic procedures, a urine culture should be checked before the procedure, and it must be sterile. No rectal preparation is required. The patient is placed in the lithotomy position, and the prostate is palpated with the index finger. The examining glove is then removed and replaced with a sterile glove. The needle guide is placed on the index finger, and a sterile finger cot or glove is placed over the tip of the needle guide. Next the finger is reinserted into the rectum, and the tip of the needle guide is placed flush against the rectal mucosa below the prostatic induration. The disposable needle with the stylet in place is passed through the guide and into the prostate nodule. The stylet is then removed, and the 22-gauge needle is connected to the 20-ml syringe already in the pistol grip. The pistol grip allows for one-handed aspiration and manipulation of the needle. The apparatus is moved back and forth approximately 20 times while negative pressure is maintained on the needle. The negative 19 20 pressure is released, and the needle is withdrawn from the prostate. The aspirated material is then deposited directly on a slide by injection of 10 ml air through the needle. If the aspiration must be repeated because there is insufficient material, a new needle need not be used if gloves were sterile because the Franzen needle is never in contact with the rectal contents. When air pressure alone is not sufficient to flush the syringe, the stylet can be replaced to clear the lumen and deposit the specimen on a slide. The slide is then smeared in a manner similar to that used in the preparation for a blood count. Some pathologists prefer the specimen to be air-dried, but others prefer an ethanol fixation. If ethanol fixation is used, the slides should be placed in the ethanol before drying takes place. Whether the specimen is air-dried or fixed, certain artifacts are introduced, so it is important that the pathologist’s preference be ascertained. If the aspirated specimen is being used for other studies, it can be placed directly into Hanks’ balanced salt solution or growth media. INDICATIONS The indications for performing a fine-needle aspiration of the prostate are the same as those for any biopsy technique. The only absolute indication is a palpable abnormality on rectal examination. Whether lesions perceived on ultrasono- graphic examination of the prostate should be biopsied is beyond the scope of this review. As with all urologic procedures, an untreated urinary tract infection or acute prostatitis is a contraindication to biopsy. The procedure should also be deferred in patients with acute infectious rectal diseases. RESULTS OF STUDIES The most important issue regarding transrectal aspiration of the prostate is its degree of accuracy in diagnosing or excluding prostate cancer, i.e., sensitivity and specificity. Specificity is the probability of a negative test in a man who is free of the disease. Sensitivity is the likelihood of positive results in a patient with prostate cancer. However, one must review data carefully because no published series uses autopsy analysis as the standard. Therefore, when the results of perineal biopsy and aspiration biopsy differ, one must decide which is the true result. The result of the aspiration biopsy can be considered to be false positive, or the result of the perineal biopsy can be considered to be false negative, depending on the standard used. The largest published series of cytologic diagnoses of prostate tumors presents reports on 1,430 biopsies in 1,110 patients reviewed by Esposti at the Karolinska Institute in Stockholm, Sweden (2). He indicates that about 10% of prostate cancers are not detected at the first aspiration biopsy, and this is in keeping with other reports. Table 1 demonstrates that in large series of aspirations (100 per- formed), a sensitivity of 89%-91% can be achieved. Prostate cancer is found to be responsible for approxi- mately 50% of palpable abnormalities within the prostate as determined by standard biopsy techniques (histologic core biopsy). The presence of cancer was demonstrated in only 37% of the patients who had fine-needle aspiration biop- sies. This lower rate of positivity should not be perceived as TABLE 1.—Sensitivity of fine-needle aspiration biopsy of the prostate in a large series? Positive Suspicious Reference ~~ Aspirations Carcinomas aspirations aspirations (2) 162 60 54 3 3) 100 45 34 5 (7) 433 198 157 1 (20) 469 203 197 10 21) 182 50 37 NR (22) 753 132 112 0 (23) 454 159 153 4 (24) 350 214 205 & Total 2,903 1,061 949 19 4 NR = not reported. No. of positive aspirations/No. of carcinomas = 89%. No. of positive + suspicious aspirations/No. of carcinomas = 91%. No. of carcinomas/No. of biopsies = 37%. b Numbers are included in positive results. lower sensitivity. It is probable that the clinicians are sam- pling extremely subtle differences in consistency because needle aspiration can be performed in the physicians of- fice without anesthesia. This interpretation is supported by Esposti and Franzen (5), who report positive biopsies in only 1,410 of 4,630 aspirations or 30%. Their lower posi- tivity rate is not secondary to missed lesions; rather, there are a larger number of true negatives. The incidence of false-positive and false-negative biop- sies needs to be addressed before fine-needle aspiration can be accepted as a reliable procedure. False-positive rates range between 0% and 2%. Clearly, false-positive results are of great concern, but as noted in the beginning, whether a result is a false or true positive depends on the reference standard. In 469 cases reported by Esposti (24), incidence was 0% when a positive aspiration was not confirmed at ei- ther autopsy or radical prostatectomy. Thus, with adequate experience, false-positive results should not occur. Even though false-negative biopsies do not result in un- necessary procedures, they may result in harmful delays in the institution of therapy and thus are as serious as false-positive results. However, the Karolinska experience (5) demonstrates that most cancers are detected on repeat biopsy and that the sensitivity of aspiration biopsy in detect- ing cancer is equivalent to standard histologic core biopsy. Thus, with repeat aspiration when clinically indicated, the false-negative rate is less than 4% of the cases (5). The recommendation for patients with a palpable nodule, sus- pected to be prostate cancer but with a negative aspiration, is a repeat aspiration and/or a histologic biopsy. Recently, reports on the American experience with trans- rectal aspiration have appeared in the literature (8-13). The experience indicates a learning curve during which aspiration samples may be inadequate for accurate interpre- tation. This learning curve is secondary to both the pathol- ogist’s need to gain experience at diagnosis and the urol- ogist’s need to learn how to obtain and prepare adequate material. In general, approximately 100 aspirations are nec- essary before the urologist consistently obtains high-quality samples. Therefore, it seems appropriate that initially as- pirations be done in conjunction with whatever technique NCI MONOGRAPHS, NUMBER 7, 1988 the urologist usually performs. Each physician could then assess at what point performance of a standard biopsy is no longer warranted. When the aspiration biopsy is posi- tive but the histologic biopsy is negative, both procedures should be repeated. If the discrepancy in results persists, the slides from the aspiration and standard biopsies should be sent to an independent pathologist with documented expe- rience in the cytologic diagnosis of prostate cancer. In excellent hands, 10% of prostate cancers are missed at first aspiration; this result is not different from that with standard histologic biopsies, which have similar false-negative results (//). When suspicious nodules con- tinue to yield negative biopsies, ultrasound-guided biopsies could be performed if the lesion can be ultrasonographically identified. This would serve to document that the aspiration needle or biopsy needle is indeed in the nodule in question. PATHOLOGIC ASSESSMENT Experience is necessary for an accurate diagnosis with any biopsy sample. With aspiration biopsy, the pathologist often examines individual cells or clumps of cells out of their natural environment; thus a special expertise is neces- sary. When false-positive reports are generated, they are most often secondary to the occurrence of atypical hyperpla- sia, squamous metaplasia, granulomatous prostatitis, or as- piration of the seminal vesicle (25). The reason for a false-negative interpretation is often secondary to the pres- ence of well-differentiated cancer. To assist in the diagnosis of well-differentiated cancer, Kline and Kannan (26) have developed major and minor criteria for the cytologic in- terpretation of well-differentiated prostate cancer (table 2). They define five major and five minor criteria that are rel- atively unique to well-differentiated cancer as opposed to benign disease. They recommend systematic use of these 10 parameters to enhance diagnostic sensitivity and allow detection of well-differentiated cancer. CYTOLOGIC VERSUS HISTOLOGIC GRADE The importance of grading of prostate cancer is self- evident because grade certainly correlates with patient TABLE 2.—Criteria for cytologic diagnosis of prostate cancer Histology of cancer, %“ Cytologic characteristic Benign, % WD MD to PD Major criteria Cellularity 15 92 75 Dyshesion 0 48 95 Nuclear irregularity 0 64 90 Anisonucleosis 7 84 100 Macronucleoli 7 72 95 Minor criteria Polarity loss 4 68 95 Crowding 7 72 75 Piling 4 68 60 Microacini 0 60 35 Cell enlargement 4 80 75 4 WD = well-differentiated; MD = moderately differentiated; PD = poorly differentiated. MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 21 TABLE 3.—Correlation of cytologic and histologic grade in 100 patients with prostate cancer? . Histology Cytologic grade WD MD PD 1 15 4 0 II 3 33 9 II 0 5 31 4 See table 2 for definitions. Results demonstrate a 79% exact correlation. prognosis (27). Therefore, before aspiration cytology can be accepted as a diagnostic tool for prostate cancer that is equivalent to histology, it must be proved that the potential for prognostication with cytologic grade is similar to that with histologic grade. In many instances during the inception of fine-needle as- piration, core biopsy and aspiration were performed simul- taneously; therefore, data correlating histologic and cyto- logic grades are available (table 3). The results demonstrate a 79% exact correlation; in no instance were cytologic grade I cells deemed to be poorly differentiated histologically or grade III cells seen as well differentiated (3,28,29). Cytologic grade can also be compared with Gleason score (table 4). There is a 75% correlation between the three cytologic grades and the Gleason sum when the Glea- son sum is divided into three categories: 2-4, 5-7, and 8-10 [(12,30-32); unpublished data]. As with histologic grade, there were no instances when cytologic grade I cells demon- strated a Gleason sum of 8-10 or cytologic grade III cells yielded a Gleason sum of 2-4. Cytologic grade not only correlates with histologic grade and Gleason score but also has a direct correlation with survival, as shown in tables 5 and 6 (28,33). In a series of 469 patients with minimum 5-year follow-up seen at the Karolinska Institute (table 5), survival was directly related to cytologic grade (28). As would be expected, the better differentiated tumors tended to be a lower stage; 47% of grade I tumors were stage B and only 10% were stage D. In distinction, only 10% of grade III tumors were stage B and 33% were stage D. No tumors in any category were stage A. A similar study from the University of Munich, Federal Republic of Germany (33), reported on survival in 488 patients (table 6). Survival was related to cytologic grade, and more importantly, survival was determined relative to cancer-related death. These authors divided cytologic grade into four categories and grade was not correlated with stage. Table 6 demonstrates that 3- and 5-year survival rates correlated with cytologic grade and that the percentage TABLE 4.— Correlation of cytologic grade and Gleason score in 149 patients with prostate cancer? Gleason score Cytologic grade 2-4 5-7 8-10 I 15 11 0 II 3 58 17 II 0 6 39 4 Results demonstrate a 75% correlation. 22 TABLE 5.—Correlation of cytologic grade and percent survival of 469 patients with prostate cancer in a study at the Karolinska Institute? TABLE 6.— Correlation of cytologic grade and percent survival of 488 patients with prostate cancer in a study at the University of Munich“ Percent survival at: Cytologic No. of grade patients lyr 2 yr 3yr 4 yr Syr I 131 89 84 73 70 68 II 265 87 75 61 58 55 111 73 73 « 45 29 15 11 ¢ Of the patients studied, 29% had stage B disease, 58% stage C, and 14% stage D (28). of patients dying of prostate cancer also increased with increasing cytologic grade. Thus, just as survival can be shown to correlate with his- tologic grade or Gleason sum, survival can also be shown to correlate directly with increasing cytologic grade. It appears that aspiration cytology has a prognostic value equivalent to that of histologic methods. CYTOLOGIC MONITORING OF RESPONSE TO THERAPY Hormonal Therapy Core biopsies performed following castration or estro- gen administration have previously been shown to demon- strate pyknotic nuclei and glycogen or squamoid histologic changes (1/9). These same findings have been seen in sam- ples obtained by transrectal aspiration (4,34). In a study performed at the University of Umea, Sweden (34), as- piration biopsies have been performed at 12, 24, and 36 months to predict patient response to estrogen therapy (ta- ble 7). Cytologic findings following therapy were classified as showing disease regression or no change and were cor- related with patient response at each biopsy interval. The results demonstrate that survival in patients with tumors showing disease regression is prolonged over survival in pa- tients whose tumors do not show regressive changes. The results also suggest that disease progression can be seen on biopsy before it becomes clinically manifest. Radiation Therapy The necessity and significance of prostate biopsy after ra- diation therapy are discussed elsewhere and in this volume (15-18), but equally important is the fact that adequate postirradiation samples can be obtained by fine-needle aspiration (35). In a series of 66 patients treated with external-beam radiation therapy, 53 consented to a postir- Percent survival at: Percent deaths Cytologic from prostate grade 3yr Syr cancer I 74 60 33 II 70 66 54 II 51 36 60 Iv 37 37 94 4 See (33). radiation aspiration biopsy every 6 months (35). There were no complications, and adequate samples were obtained in all instances. IMMUNOHISTOCHEMISTRY AND LABORATORY ANALYSIS Fine-needle aspiration does not preclude the use of im- munohistochemical stains. Both prostate-specific antigen and prostatic acid phosphatase have been used on spec- imens aspirated from the prostate and lymph nodes and other metastatic sites (36-38). It is important for one to be aware that, although fine-needle aspiration biopsy is significantly less traumatic to the prostate than is standard histologic core biopsy, it still affects laboratory tests directed toward assessment of prostate cancer. In a study of 14 patients, 8 with adeno- carcinoma and 6 with benign prostatic hypertrophy, signif- icant increases in both prostate-specific antigen and pros- tatic acid phosphatase occurred (39). Thus samples for staging blood tests should be obtained prior to aspiration biopsy. COMPLICATIONS A discussion of the relative merits of any procedure should include the incidence of complications. The com- plication rate of transrectal aspiration is significantly lower than that seen with either perineal or transrectal core biopsy. The complication rate for transrectal core biopsy of the prostate ranges from 17% to 50% with respect to in- fection, bleeding, and urinary retention (40-43). In com- parison to this very high incidence of minor and major complications via the transrectal route, transperineal core biopsy can be expected to result in complications in only 1.1%-6% of the patients (44-46). However, transrectal fine-needle aspiration carries a lower risk of complica- tions than even the perineal biopsy route. In over 3,000 TABLE 7.—Response to estrogen therapy in 44 patients with prostate cancer in a study at University of Umea“ Percent of patients with: Disease Stable Disease Cytdiogle regression disease progression finding 12 mo 24 mo 36 mo 12 mo 24 mo 36 mo 12 mo 24 mo 36 mo Regressive 50 82 73 45 0 0 5 18 27 changes No changes 32 32 32 36 27 27 22 41 50 4 See (34). NCI MONOGRAPHS, NUMBER 7, 1988 aspirations, only 12 complications occurred (47,48). These complications included: epididymitis (two), transient hema- turia (two), hemospermia (three), and infections (five). This yields an overall complication rate of 0.4%. CONCLUSIONS Fine-needle aspiration biopsy of the prostate has been shown to have sensitivity and specificity equal to those of standard core biopsy. No system in use today is individu- ally predictive, but with respect to patient prognosis, cyto- logic grading of aspiration biopsy samples appears to have a prognostic value similar to that of the histologic grading systems currently in use. The pathologic procedures fol- lowed for histologic slides for determination of the origin of a tumor can be performed on aspiration specimens, so accurate interpretation of slides becomes a matter of expe- rience. Fine-needle aspiration can be performed with less patient discomfort than can core biopsy, and this factor may lead to greater patient acceptance of serial biopsies. Also, the ease of biopsy may lower the clinician’s threshold and therefore lead to the detection of a greater percentage of localized lesions. Finally, fine-needle aspiration is associated with a significantly lower complication rate than either transrectal or transperineal core biopsy. RECOMMENDATIONS TO CONSENSUS PANEL The following recommendations were made to the Na- tional Institutes of Health Consensus Development Panel on the Management of Clinically Localized Prostate Can- cer: 1) Fine-needle aspiration biopsy should be encouraged as a standard part of the diagnostic armamentar- ium. 2) Fine-needle aspiration biopsy should continue to be performed by urologists because the “learning curve” mandates that core biopsy should be per- formed concurrently until proficiency is achieved. 3) The skills necessary to perform fine-needle aspira- tion should be acquired by urology residents during their training. 4) Pathology residents should develop an expertise in interpretation of prostate aspiration cytology slides. REFERENCES (I) FRANZEN S, GIERTZ G, ZAJNCEK J. Cytological diagnosis of prostatic tumours by transrectal aspiration biopsy: A preliminary report. Br J Urol 1960;32:193-196. (2) Esposti PL. Cytologic diagnosis of prostatic tumours with the aid of transrectal aspiration biopsy. Acta Cytol 1966;10:182-186. (3) EKMAN H, HEDBERG K, PERSSON PS. Cytological versus histological examination of needle biopsy specimens in diagnosis of prostatic cancer. Br J Urol 1967;39:544-548. (4) WILLIAMS JP, CHIR M, STILL BM, ET AL. The diagnosis of prostatic cancer: Cytologic and biochemical studies using the Franzen biopsy needle. Br J Urol 1967;39:549-554. (5) EsposTi PL, FRANZEN S. Transrectal aspiration biopsy of the prostate. Scand J Urol Nephrol [Suppl] 1980;55:49-52. (6) WILLEMS J-S, LOWHAGEN T. Transrectal fine-needle aspira- MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 23 tion biopsy for cytologic diagnosis and grading of prostatic carcinoma. Prostate 1981;2:381-395. (7) MAIER U, CZERWENKA K, NEUHOLD N. The accuracy of transrectal aspiration biopsy of the prostate: An analysis of 452 cases. Prostate 1984;5:147-151. (8) RHEINFRANK RF, NULF TH. Fine needle aspiration biopsy of the prostate. Endoscopy 1969:1:27-32. (9) KLINE TS, KOHLER P, KELSE DM. Aspiration biopsy cytology (ABC). Arch Pathol Lab Med 1982;106:136-139. (10) LiuNG BM. Fine needle aspiration biopsy of the prostate gland: Techniques and review of the literature. Semin Urol 1985;3:18-26. (11) CHopAak GW, BiBBO M, STRAUS FH, ET AL. Transrectal as- piration versus transperineal core biopsy for the diagnosis of carcinoma of the prostate. J Urol 1984;132:480-482. (12) CARTER HB, RIEHLE RA JR, Koizumi JH, ET AL. Fine needle aspiration of the prostate: A cytohistological correlation. J Urol 1986;135:294-299. (13) CHODAK GW, STEINBERG GD, BIBBO M, ET AL. The role of transrectal aspiration biopsy in the diagnosis of prostatic cancer. J Urol 1986;135:299-301. (14) BRODERS AC. Carcinoma: Grading and practical application. Arch Pathol 1926;3:376-381. (15) PHILLIPS NB, LATTIMER JK. Complications and regression of primary tumor size with radiotherapy in prostatic cancer. J Urol 1972;108:921-923. (16) Cox JD, StorreL TJ. The significance of needle biopsy after irradiation for stage C adenocarcinoma of the prostate. Cancer 1966;40:156-160. (17) FreiHA FS. Carcinoma of the prostate: Results of post- irradiation biopsy. Presented at the 78th Annual Meeting of the American Urological Association, Las Vegas, NV, April 17-21, 1983, abstr 297. (18) ScarDINO PT, WHEELER TM. Local control of prostate cancer with radiotherapy: Frequency and prognostic sig- nificance of postirradiation biopsy results. NCI Monogr 1988;7:95-103. (19) KirRCHEIM D, BRANDES D, BACON RL. Fine structure and cy- tochemistry of human prostatic carcinoma. In Male Ac- cessory Sex Organs, Structure and Function in Mammals (Brandes D, ed). New York: Academic Press, 1974, pp 397-423. (20) LiN BPC, DAViEs WEL, HARMATA PA. Prostatic aspiration cytology. Pathology 1979;11:607-614. (21) BisHOP D, OLIVER JA. A study of transrectal aspiration biop- sies of the prostate, with particular regard to prognostic evaluation. J Urol 1977;117:313-315. (22) ANANDAN N, MACKENZIE E, GINGELL SC, ET AL. Role of Franzen needle aspiration biopsy in carcinoma of the prostate. J R Soc Med 1983;76:828-830. (23) KAULEN H, DAvIDTS HH. Diagnostische Mioglichkeiten and Nachleile der transrectalen Aspirationsbiopsie mit zytolo- gischer Beurteilung beim Prostata-Carcinom. Verh Dtsch Ges Inn Med 1972;78:268-271. (24) EsposTi PL. Aspiration biopsy cytology in the diagnosis and management of prostatic carcinoma. PhD thesis. Stock- holm: Stahl Accidens Trychk 1974;33-34. (25) TANNENBAUM M. Aspiration biopsy of the prostate: The pathologist’s viewpoint. Semin Urol 1983;1:172-175. (26) KLINE TS, KANNAN V. Prostatic aspirates. A cytomorpho- metric analysis with emphasis on well-differentiated car- cinoma. Diagn Cytopathol 1985;1:13-17. (27) Mostorl FK. Problems of grading carcinoma of prostate. Semin Oncol 1976;3:161-169. (28) Esposti PL. Cytologic malignancy grading of prostatic car- cinoma by transrectal aspiration biopsy. Scand J Urol Nephrol 1971;5:199-209. 24 (29) YATANI R, ScHIRAISHI T, SOGA T, ET AL. Reliability of cy- tological grading of prostatic carcinoma compared with histological grading. Pathol Res Pract 1985;180:68-73. (30) NARAYAN P, STEIN R, LIUNG BM. Fine needle aspiration in detection of prostatic cancer: Comparison with surgical biopsy. J Urol 1987;137:abstr 361. (31) LAYFIELD LJ, MUKAMEL E, HILBORNE LH, ET AL. Cytologic grading of prostatic aspiration biopsy: A comparison with the Gleason grading system. J Urol 1987;137:abstr 362. (32) BopNER DR, HAMPEL N, MAKSEM JA, ET AL. Aspiration biopsy of the prostate. World J Urol 1987;5:62-64. (33) FauL P, SCHMIEDT E, KERN R. Prognostic significance of cytologic differentiation grading in estrogen-treated pros- tatic carcinoma diagnosed by fine-needle aspiration. Int Urol Nephrol 1980;12:347-354. (34) ToMic R, BERGMAN B, HIETALA SO, ET AL. Prognostic sig- nificance of transrectal fine-needle aspiration biopsy find- ings after orchiectomy for carcinoma of the prostate. Eur Urol 1985;11:378-381. (35) KurtH KH, ALTWEIN JE, SKOLUDA D, ET AL. Follow-up of irradiated prostatic carcinoma by aspiration. J Urol 1977;117:615-617. (36) REIFLER DM, KINI SR, Lui D, ET AL. Orbital metastases from prostatic carcinoma: Identification by immunocytology. Arch Ophthalmol 1984;102:292-295. (37) KESHGEGIAN AA, KLINE TS. Immunoperoxidase demonstra- tion of prostatic acid phosphatase in aspiration biopsy cy- tology. Am J Clin Pathol 1984;82:586-589. (38) KATZ RL, RAVAL P, BROOKS TE, ET AL. Role of immunocy- tochemistry of prostatic neoplasia by fine needle aspira- tion biopsy. Diagn Cytopathol 1985;1:28-32. (39) CHARRIE A, FLEURY-GOYON MC, DUTEY P, ET AL. The ef- fect of prostate aspiration biopsy on serum levels of pros- tatic acid phosphatase and prostate-specific antigen. JAMA 1986;256:474. (40) CRAWFORD ED, HAYNES AL, STORY MM, ET AL. Prevention of urinary tract infection and sepsis following transrectal prostatic biopsy. J Urol 1982;127:449-451. (41) EATON AC. The safety of transrectal biopsy of the prostate as an out-patient investigation. Br J Urol 1981;53:144-146. (42) SHARPE JR, SADLOWSKI RW, FINNEY RP, ET AL. Urinary tract infection after transrectal needle biopsy of the prostate. J Urol 1982;127:255-256. (43) BissapA NK. Accuracy of transurethral resection of the prostate versus transrectal needle biopsy in the diagno- sis of prostatic carcinoma. J Urol 1977;118:61-63. (44) KAUFMAN JJ, ScHULTZ JI. Needle biopsy of the prostate: A re-evaluation. J Urol 1962;87:164-168. (45) WENDEL RG, EVANs AT. Complications of punch biopsy of the prostate gland. J Urol 1967;97:122-126. (46) PARRY WL, FINELLI JF. Biopsy of the prostate. J Urol 1960;84:643-648. (47) EsposTi PL, FRANZEN S, ZAJICEK J. The aspiration biopsy smear. In Diagnostic Cytology and Its Histopathologic Basis (Koss LC, ed), 2nd ed. Philadelphia: Lippincott, 1968, pp 565-596. (48) EsposTi PL, ELMAN A, NORLEN H. Complication of trans- rectal aspiration biopsy of the prostate. Scand J Urol Nephrol 1975;9:208-213. Application of Flow Cytometry and Automated Image Analysis to the Study of Prostate Cancer Mitchell C. Benson!-2 ABSTRACT —Flow cytometry and image analysis are comple- mentary quantitative cytologic techniques that have demonstrated utility in the assessment and analysis of prostate cancer and other urologic and nonurologic tumors. This review is intended to assess the current state of the art and to project future directions and applications of these modalities in the pathologic assessment of prostate cancer. Special attention is directed toward the topics of prostate cancer detection and diagnosis, determination of patient prognosis, and the monitoring of patient response to therapy. We recommend that 1) flow cytometry and image analysis be used to determine pathologic parameters that will help in predicting poor patient prognosis and 2) quantitative cytologic determinations of DNA content be included in clinical trials so that their ultimate role in monitoring patient response to therapy can be determined. This knowledge will allow the development of protocols designed to test the value of earlier institution of multimodal therapy in high-risk populations.—NCI Monogr 7:25-29, 1988. The application of FCM and automated IA to the assess- ment and separation of heterogeneous cell populations has added new dimensions to our ability to analyze the nature of tumors. These technologies complement each other and are ideally suited to the analysis of tumors. Conventional microscope-based cytophotometry requires several hours to measure 100 cells, but with FCM and IA we can analyze cells at rates exceeding 100-200/second. In FCM, advances in instrumentation, such as the addition of second and third lasers, plus improved staining techniques now allow for the measurement of many different cell properties on an indi- vidual cell. This results in the multiparameter analysis of tumor cell populations. The speed of automated IA has in- creased because of computerization of the process. In this presentation, we focus primarily on FCM but also review the application of IA to prostate cancer by discussing the: 1) principles of FCM and cell analysis, 2) cytochem- ical stains that have been useful in FCM systems, 3) role of quantitative cytology in the diagnosis of prostate cancer, 4) ability of quantitative cytologic systems to portend prog- nosis, and 5) applicability of these systems to the monitoring of patient response. FLOW CYTOMETRY There are currently numerous FCM instruments com- mercially available. Regardless of which instrument is ABBREVIATIONS: FCM = flow cytometry; IA = image analysis; FLS = forward light scatter; PLS = perpendicular light scatter. ! Department of Urology, Columbia-Presbyterian Medical Center, 622 West 168th St., New York, NY 10032. 2 | thank Ralph deVere White and John H. Lynch for sharing the results of their studies and Ms. Mary Petway for typing assistance. used, the basic tenets are the same. These automated cell-analyzing and cell-sorting instruments combine opti- cal and electrical sensing techniques that, with the aid of computer data storage, allow several measurements to be made simultaneously on the same cell. Typical measure- ments are cell volume, multicolor fluorescence from stains bound to cellular constituents, and light scatter, with which one can assess cell size and intracellular structure. Proper sample preparation is crucial if meaningful con- clusions are to be drawn from the analysis of a specimen. An in-depth discussion of sample preparation can be found in the text by Pretlow and Pretlow (1). For analysis of DNA content only, a technique described by Deitch et al. (2) has proven reproducible and easy to perform. Multiparamet- ric analysis often requires intact, living cells, and we have found (unpublished observation) that adequate numbers of prostatic cells can be obtained from both enzymatic dissoci- ation and mechanical mincing (3). Whenever possible, it is advantageous that one avoid enzymatic digestion because the enzymes can introduce artifacts and alter the affinity of various stains for DNA and cell-surface proteins. Once a single-cell suspension has been prepared, cell viability should be assessed by trypan blue exclusion. This is necessary for all studies because the use of suspensions with poor viability (<70%) may lead to spurious results due to leakage of cellular components or DNA into the media and analysis of incomplete cells. The FCM analysis begins by introduction of the cell sam- ple into the flow chamber. Smooth, nonturbulent flow is assured by the introduction of the sample into the center of a faster moving stream of sheath fluid, as described by the Reynolds formulations of fluid flow (4). The proxim- ity of the cells to one another can be controlled by the rate at which sample fluid enters the flow chamber. The speed at which fluid exits the flow chamber is held con- stant, which allows measurement of multiple parameters at distinct sensing stations along the course of the fluid stream. The distance between the lasers is governed by the physi- cal construction of the flow system, and, because the rate at which a cell is traveling is constant, data can be corre- lated as having originated from a single source by the time elapsed between readings. One can accomplish cell sorting by carrying these con- cepts one step further. A piezoelectric transducer can be activated, and this will cause the fluid stream to vibrate. These vibrations will eventually disrupt the stream and, at a definable point, cause it to break into droplets. Some of these droplets will contain the desired cells. Because the rate at which the stream is traveling is held constant, the distance between the sensors and the deflection plates de- fines the time at which a droplet containing a desired cell 25 26 will be at the deflection plates. Then by knowing the dis- tance between the sensors and the point at which droplets are forming, one can determine the times at which to ac- tivate the deflectors and to place a charge on the desired droplets. For example, a prostate cancer cell has its vol- ume determined at time 0. At times 0 + W and 0 + W + X, various determinations are made, and it is found that this cell type is the one sought. The cell continues to move down the intact sample stream and at time 0 + W + X + Y, just before it is beginning to be disrupted, the stream is given a positive charge. The stream is then disrupted and the droplet continues to carry this positive charge. The in- tact stream is then neutralized so that future droplets are not inadvertently deflected. At time 0 + W + X + Y + Z, the charged droplet containing the desired cell passes through the charged deflector plates and reacts according to the charge placed on the droplet. This reaction allows the cell and others like it to be captured and further analyzed. The process is called cell sorting. Cells can be sorted directly onto microscope slides for visual identification or sorted into wells containing growth media and subsequently placed in cell cultures. It is our opinion that cell sorting should be a part of any experiment to ensure that one is dealing with true populations and not artifacts. Most flow cytometers are equipped with laser light sources, the most common of which is the 5-W argon laser because it has substantial power output in both the UV and visual ranges. The main drawback to laser light sources is the expense of the laser. Mercury-arc lamps offer a less ex- pensive alternative and are available on some instruments. For additional information on the techniques of FCM, see (5-8). CYTOCHEMICAL STAINS Quantitative cytochemical stains provide a rapid, accu- rate, and reproducible means of assaying cytologic, bio- chemical, and functional properties of tumor cells. Mea- surement of fluorescent dyes that bind to particular cellular constituents has received the most attention in FCM. Addi- tionally, one can combine fluorescent dyes with antibodies and lectins to study cell-surface moieties (9,10). The measurement of fluorescence has three major advan- tages in FCM: 1) Fluorescence intensity is directly propor- tional to the quantity of the dye present and, by inference, to the amount of material being bound. 2) Low concen- trations of dye per cell can be detected. 3) Nonfluorescent substrates can be activated to a fluorescent state by cellular enzymes and used to quantitate cellular function (11). In table 1 is a list of various cellular components and the fluorescent dyes that can help measure each component. This list of the most commonly used dyes is by no means inclusive of all agents in use. The review by Shapiro (12) presents a detailed discussion of the multiple probes in FCM systems. Steinkamp (6) also states some of the uses of multiple dyes in cell assessment. 3 Benson MC, Kaplan SA, Karp F: Submitted for publication. TABLE 1.—Cellular components and fluorescent dyes used to measure components Cellular component Dye DNA Hoechst 33342 Hoechst 33258 DAPI Propidium iodide Ethidium bromide Acridine orange Chromomycin Mithramycin RNA Pyronine Y Acridine orange Oxazine 1 Total protein Fluorescein isothiocyanate Sulforhodamine 101 Rhodamine 123 Rhodamine B Rhodamine 3G DiOC Mitochondria DNA CONTENT The FCM analysis of prostate cancer has focused pre- dominantly on the quantitation of the DNA content of tu- mors (/3-22).3 The rationale for investigation of the DNA content of tumors of the prostate is based on the results of studies in which researchers used the Feulgen technique of spectrophotometric analysis (23-25). These studies were limited by the small numbers of patients evaluated because of the tedious, time-consuming task of spectrophotometric analysis. DETECTION AND DIAGNOSIS OF PROSTATE CANCER To be useful in detection and diagnosis of cancer, a sys- tem must have high sensitivity and specificity. The tech- nique must be capable of detecting cancer whenever it is present in the specimen and excluding the diagnosis when cancer is absent. In experienced hands, the sensitivity of histologic and cytologic diagnosis approaches 100%. How- ever, no system can be better than the sample provided for analysis, and in this regard, pathologic examination of both fine-needle aspiration and core biopsy is about 90% ac- curate (26,27). Approximately 10% of the cancers are not detected on first biopsy, usually because the existing can- cer is not present in the biopsy specimen. Thus to be useful in the diagnosis of prostate cancer, the results with FCM or TABLE 2.— Correlation of incidence of aneuploidy and cytologic grade in 300 patients with prostate cancer No. of samples Cytologic Percent grade Diploid Aneuploid aneuploid I 79 50 39 11 41 150 78 III 2 69 97 4 See (17). NCI MONOGRAPHS, NUMBER 7, 1988 TABLE 3.— Correlation of incidence of aneuploidy and Gleason score in 190 patients with prostate cancer? 27 TABLE 4.—Correlation of DNA ploidy, Gleason score, and tumor stage with percent survival at 5 yr in 59 patients with prostate cancer No. of samples Percent survival Gleason Percent score Diploid Aneuploid aneuploid Pathology Aneuploid Diploid 2-4 17 8 32 Gleason score 5-7 28 26 48 2-4 67 93 8-10 8 A 72 5-7 33 71 8-10 29 50 See (18,22) and footnote 3. Stage IA must be equivalent to or better than those with conven- 5 3 B ° x tional histology and cytology. Use of FCM in the diagnosis of prostate cancer relies “ See (22). on the presence of an abnormal DNA histogram. Unfor- tunately, abnormalities (frank aneuploidy or >10%-15% growth fraction) occur in normal samples, and not all prostate cancers demonstrate these changes. A new tech- nique with the DNA index of nonmalignant prostatic tissue as the reference standard has shown promise as a means to enhance sensitivity.? In most studies, lymphocytes are used as a diploid reference standard. The DNA index is then de- fined as the quotient of sample DNA/lymphocyte DNA. An index of 1.00 has traditionally been considered normal. Due to variation in either available staining sites on the chro- mosomes or fluorochrome uptake into the nucleus, other normal tissues may have a slightly higher or lower index. The new technique uses the DNA index of normal prostate (1.04) as the reference standard, and a patient-specific in- dex is determined by the patient’s prostate DNA/patient’s WBC DNA. In a series of 12 patients, this new index sys- tem has shown great sensitivity in detection (100%), but specificity has not been addressed. Tables 2 and 3 present the incidence of aneuploidy as a function of cytologic grade and Gleason score. The tables demonstrate that with increasing tumor grade or Gleason score, the incidence of aneuploidy increases. The lack of sensitivity of standard DNA histogram analysis in detect- ing prostate cancer is evidenced by the fact that 2 of 71 specimens with grade III cytology and 8 of 29 specimens with Gleason scores 8-10 demonstrated seemingly normal diploid DNA patterns (/7,18,22). Systems that rely solely on DNA determination are not specific and sensitive enough to be used in the detection and diagnosis of prostate cancer. Sensitivity may have been improved by the incorporation of tissue-specific controls, but further study is necessary. QUANTITATIVE CYTOLOGY IN DETERMINATION OF PROGNOSIS Histologic aggressiveness not discernible by standard histopathology has been predicted successfully by FCM and IA (18,20-22,24,25,27,28). In a study on fine-needle aspi- ration biopsy specimens, DNA determinations were made according to a cytophotometric technique. Because this was a manually performed analysis, the number of cells assessed per patient was less than 100. Survival was correlated with percent hyperdiploid cells (25). The authors selected 43 patients with similar grade (I-III) and stage (B and C) characteristics. Twenty-one of the patients survived more than 5 years; 22 died within 3 years. Fourteen of the 21 MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER patients demonstrated predominantly diploid tumors, com- pared with only 4 of the 22. Three studies have been conducted in which patient sur- vival was correlated with DNA content as determined from analysis of paraffin-embedded tissue (21,22,28). The study by Lundberg et al. (21) correlated histopathologic grade and DNA ploidy with survival in 50 patients. No correla- tion between ploidy and grade was found, but a statisti- cally significant relationship between survival and ploidy was demonstrated. Fordham et al. (28) analyzed 72 pa- tients and found that when ploidy was correlated with Glea- son score, significant prognostication could be made. They found that for Gleason scores of 7-10, all the 19 patients with aneuploid tumors died within 3 years, compared with only 6 of 10 with diploid tumors. For Gleason scores of 2-6, they found that 2 of 6 patients with aneuploid tumors died within 3 years, but only 3 of 17 with diploid tumors died. Lastly, in a study of 54 patients by Dejter et al. (22), DNA ploidy, Gleason score, and tumor stage were corre- lated with percent survival at 5 years (table 4). These data demonstrate that within stage and grade categories the in- cidence of aneuploidy portends prognosis. In comparison to these studies, which demonstrate that the incidence of aneuploidy correlates inversely with sur- vival, deVere White and Deitch (unpublished observations), using fresh and paraffin-embedded tissue, determined that patients with aneuploidy responded better to hormonal ther- apy and, in those with stage D disease, it had a direct cor- relation with prolonged survival (table 5). It is not at first apparent why the results in this series differed from those in other series. As suggested by the authors, diploid tumors may be less hormone responsive because of lower growth fractions, as determined by pretreatment ploidy. Also pos- sible is that this represents a sampling phenomenon. How- ever, these data underscore the necessity of our stratifying patients in research protocols by DNA ploidy. TaBLE 5.— Correlation of DNA ploidy and Gleason score with percent survival at 30 mo in 77 patients with prostate cancer Percent survival Gleason score Total Aneuploid Diploid 5-7 73 100 50 8-10 50 75 20 Total 64 82 36 4deVere White RA, Deitch AP: Unpublished observations. 28 There is a 61% reported incidence of aneuploidy when any DNA pattern other than diploid is considered to be an- euploid. This value is obtained from the pooling of data from seven published series comprising 331 patients (16,18,29). The studies cited here correlated prognosis and disease extent by DNA content. Similar studies have been per- formed with light scatter measurements. LIGHT SCATTER As cells pass through the laser beams of a flow cytometer, incident light reflects, diffracts, and refracts from the sur- face and internal structures of the cell. The combination of these phenomena is known as light scatter. Although light scatter could be measured at any angle from the incident light source, the commercially available flow cytometers are designed to measure light scatter in the forward and perpendicular directions. In FCM systems, FLS has been shown to correspond directly to the size of the object be- ing studied (30). Thus the larger the cell, the more intense are the FLS signals. The PLS intensity is also related to cell size, but more im- portantly, it measures internal cellular structure (30-32). Because the largest internal structure in prostate cancer cells is the nucleus, PLS predominantly measures nuclear surface area. Nuclear surface area is a function of size and shape, and by comparing PLS to nuclear IA by manual digitization, one can see that PLS intensity is directly pro- portional to nuclear size and shape (32). Although PLS and FLS have not been widely used in the study of human prostate cancer, evidence in animal systems suggests that the combination of these parameters can be used in quanti- tative grading of tumors on the basis of cell size and nuclear shape (33). Preliminary studies in humans undergoing radical retro- pubic prostatectomy have demonstrated that the combina- tion of FLS and PLS can help predict pathologic stage (3). Some have used this technique to assess not only mean light scatter values but also intratumor variation. It is believed that the variation in signals is an estimate of tumor cell het- erogeneity and that tumor heterogeneity is related to tumor aggressiveness. The authors reported on only 7 patients, but a grading index based on FLS and PLS values correlated well with pathologic stage in the incidence of seminal vesi- cle invasion and capsular penetration. This series will have to be expanded and confirmed by others before its true use- fulness will be known. QUANTITATIVE CYTOLOGY TO MONITOR PATIENT RESPONSE Because aneuploidy is usually accepted as being associ- ated with more aggressive tumors, it is logical for one to assume that a patient responding to therapy should demon- strate a diminution in the hyperdiploid population of the tumor (34). Such a pilot FCM study was performed by Kjaer et al. (35) in patients treated with estrogenic com- pounds. Six of the 8 patients demonstrated aneuploidy and 2 had diploid tumors. Four of the 6 patients with aneuploid tumors responded to therapy, and at a biopsy 2-3 months after the initiation of therapy, this was evidenced by reduc- tion or disappearance of the aneuploid population. The 2 patients with diploid tumors also demonstrated good hor- monal response and their histograms remained unchanged. Using cytophotometric analysis, Leistenschneider and Nagel (36) examined 15 patients with cytologic grade III tumors and found that response to estramustine could be monitored by the disappearance of aneuploidy and the de- velopment of a diploid sample. Additionally, they suggested that relapse following radiation therapy could be similarly monitored. A similar conclusion was reached by Bouffioux (37), who also used a cytophotometric technique. He concluded that the persistence or reappearance of aneuploidy is evidence of therapeutic escape and, even in the absence of clinical disease progression, should be accompanied with a change in treatment. CONCLUSIONS Quantitative cytologic analysis by FCM and automated IA have been applied successfully to the analysis of prostate cancer. To date, it appears that use of these techniques does not result in accurate diagnosis, and physicians should not rely on them as a means to detect prostate cancer. Although newer techniques may have adequate sensitivity in the screening of known cancer populations, specificity remains to be determined. Quantitative cytology does appear to be useful in por- tending patient prognosis. These techniques have predictive value independent of and in some cases beyond that of cyto- logic analysis or Gleason score and tumor stage. The DNA analysis has also been shown to be a sensitive monitor of patient response. The persistence or re-emergence of aneu- ploidy following endocrine or radiation therapy appears to be an early harbinger of disease progression. RECOMMENDATIONS TO CONSENSUS PANEL The following recommendations were made to the Na- tional Institutes of Health Consensus Development Panel on the Management of Clinically Localized Prostate Can- cer: 1) Quantitative cytologic determination of DNA con- tent should be included in clinical protocols to mon- itor patient response. 2) Because these techniques seem to have the poten- tial to identify populations with a particularly poor prognosis, protocols should be developed that test the use of early multimodal therapy in high-risk populations. REFERENCES (1) PrReTLOW TG, PRETLOW TP. Cell Separation: Methods and Selected Applications, vol 1. New York: Academic Press, 1982. (2) DEITCH AP, LAW H, DEVERE WHITE RA. A stable propidium iodide staining procedure for flow cytometry. J Histochem Cytochem 1982;30:967-970. (3) BENSON MC, WALSH PC. The application of flow cytometry to the assessment of tumor cell heterogeneity and the grading of human prostate cancer: Preliminary results. J Urol 1986;135:1194-1198. (4) CROSSLAND-TAYLOR PJ. A device for counting small par- NCI MONOGRAPHS, NUMBER 7, 1988 ticles suspended in a fluid through a tube. Nature 1953;171:36-41. (5) MELAMED MR, MULLANEY PF, MENDELSOHN ML. Flow Cy- tometry and Sorting. New York: Wiley, 1978. (6) STEINKAMP JA. Flow cytometry. Rev Sci Instrumentation 1984;55:1375-1471. (7) SHAPIRO HM. Practical Flow Cytometry. New York: Liss, 1985. (8) BENSON MC. The application of flow cytometry to the study of prostate cancer. In Current Approaches to the Study of Prostate Cancer (Coffey DS, Bruchovsky N, Gardner WA Jr, et al, eds). New York: Liss, 1987, pp 765-778. (9) JuLius MH, MusupA T, HERZENBERG LA. Demonstration that antigen-binding cells are precursors of antibody- producing cells after purification with a fluorescence- activated cell sorter. Proc Natl Acad Sci USA 1972;69:134-138. (10) KRAEMER PM, TOBEY RA, VAN DILLA MA. Flow microflu- orometric studies of lectin binding to mammalian cells. J Cell Physiol 1973;81:305-311. (11) KERKER M, VAN DILLA MA, BRUNSTING A, ET AL. Is the central dogma of flow cytometry true: That fluorescence intensity is proportional to cell dye content? Cytometry 1982;3:71-78. (12) SHAPIRO HM. Multistation multiparametric flow cytometry: A critical review and rationale. Cytometry 1983;4:227-239. (13) BICHEL P, FREDERICKSEN P, KJAER T, ET AL. Flow mi- crofluorometry and transrectal fine needle biopsy in the classification of human prostatic carcinoma. Cancer 1977;40:1206-1211. (14) GOERTTLER K, EHEMANN V, TSCHAHARGANE C, ET AL. Monodispersal and deoxyribonucleic acid analysis of pros- tatic cell nuclei. J Histochem Cytochem 1977;25:560- 564. (15) TRIBUKAIT B, EsposTi PL, RONSTROM L. Tumor ploidy for characterization of prostatic carcinoma: Flow cyto- fluorometric DNA studies using aspiration biopsy mate- rial. Scand J Urol Nephrol [Suppl] 1980,55:59-64. (16) RONSTROM L, TRIBUKAIT B, EsposTi PL. DNA pattern and cytological findings in fine-needle aspirates of untreated prostatic tumors. A flow-cytofluorometric study. Prostate 1981;2:79-88. (17) TRIBUKAIT B, RONSTROM L, EsposTi PL. Quantitative and qualitative aspects of flow DNA measurements related to the cytologic grade in prostatic carcinoma. Anal Quant Cytol Histol 1983;5:107-111. (18) FRANKFURT OS, CHIN JL, ENGLANDER LS, ET AL. Rela- tionship between DNA ploidy, glandular differentiation and tumor spread in human prostate cancer. Cancer Res 1985;45:1418-1423. (19) ScHuLTZ RE, VARELLO MA, Tsou KC, ET AL. Simultane- ous flow cytometric deoxyribonucleic acid and acid phos- phatase analysis of benign and malignant lesions of the prostate. J Urol 1985;134:1133-1136. (20) DISILVERIO F, SCIARRA F. Therapeutic approach in prostatic cancer. J Steroid Biochem 1986;25:773-779. MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 29 (21) LUNDBERG S, CARSTENSEN J, RUNDQUIST I. DNA flow cytometry and histopathological grading of paraffin- embedded prostate biopsy specimens in a survival study. Cancer Res 47:1987;1973-1977. (22) DEITER SW JR, AHMED SW, CUNNINGHAM RE, ET AL. Flow cytometric measurement of cellular DNA content on paraffin-embedded human prostatic cancer tissue: Re- lationship to Gleason grade, stage and survival. J Urol 1987;137(Part 2):192A. (23) TAVARES AS, COSTA J, DECARCALHO A, ET AL. Tumour ploidy in carcinomas of the bladder and prostate. Br J Cancer 1966;20:438-441. (24) TAVARES AS, CosTA J, CosTA-MAIA J. Correlation be- tween ploidy and prognosis in prostatic carcinoma. J Urol 1973;109:676-679. (25) ZETTERBERG A, EsposTI PL. Prognostic significance of nu- clear DNA levels in prostatic carcinoma. Scand J Urol Nephrol [Suppl] 1980;55:53-56. (26) DEKERNION JB. Aspiration biopsy of the prostate: The urol- ogist’s viewpoint. Semin Urol 1983;1:166-171. (27) FORTUNOFF S. Needle biopsy of the prostate: A review of 346 biopsies. J Urol 1967,87:159-163. (28) FORDHAM MVP, BURDGE AH, MATTHEWS J, ET AL. Pro- static carcinoma cell DNA content measured by flow cy- tometry and its relation to clinical outcome. Br J Surg 1986;73:400-403. (29) BARLOGIE B, RABER MN, SCHUMANN J, ET AL. Flow cytome- try in clinical cancer research. Cancer Res 1983;43:3982- 3997. (30) KERKER M. Elastic and inelastic light scattering in flow cytometry. Cytometry 1983;4:1-10. (31) BRUNSTING A, MULLANEY PF. Differential light scattering from spherical mammalian cells. Biophys J 1974;14:439- 453. (32) BENSON MC, McDouGAL D, COrrey DS. The application of perpendicular and forward light scatter to assess nuclear and cellular morphology. Cytometry 1984;5:515-522. (33) BensoN MC, McDouGAL D, Corry DS. The use of multi- parameter flow cytometry to assess tumor cell heterogene- ity and grade prostate cancer. Prostate 1984;5:27-45. (34) MUNTZING J. DNA determinations for prognosis and ther- apy control in prostatic cancer. Semin Oncol 1983;10 (suppl):16-20. (35) KIAER TB, THOMMESEN P, FREDERICKSEN P, ET AL. DNA content in cells aspirated from carcinoma of the prostate treated with estrogenic compounds. Urol Res 1979;7:249- 251. (36) LEISTENSCHNEIDER W, NAGEL R. Cytological and DNA cy- tophotometric monitoring of the effect of therapy on con- servatively treated prostatic carcinomas. Scand J Urol Nephrol [Suppl]1983;55:197-204. (37) Bourrioux C. La mesure de I’ADN dans le produit de cyto- ponction des cancers de la prostate. Intérét diagnos- tique, prognostique et thérapeutique. J Urol (Paris) 1982;88:111-115. Noninvasive Imaging for Staging of Prostate Cancer: Magnetic Resonance Imaging, Computed Tomography, and Ultrasound Hedvig Hricak! ABSTRACT —The diagnosis of prostate carcinoma by imaging is still fraught with problems, even with the advent of highly so- phisticated techniques. Despite enthusiastic preliminary reports, no one imaging method reliably screens for this condition. The staging of prostate carcinoma is feasible, but the best imaging method remains a subject of debate. The transabdominal sono- graphic approach lacks the resolution required for detailed in- traglandular anatomic delineation. The transrectal sonographic approach excels in guiding needle biopsy and in evaluating tran- scapsular and seminal vesicle extension of known tumors. Com- puted tomography lags behind other tomographic imaging modal- ities in its accuracy for local tumor staging, but it is excellent, although nonspecific, in the detection of lymph node enlargement. Magnetic resonance detects abnormalities in the prostate in a high percentage of cases but is nonspecific. However, it can stage prostate carcinoma and detect lymphadenopathy reliably. —NCI Monogr 7:31-35, 1988. The diagnostic armamentarium for the detection and staging of prostate carcinoma is extensive, yet numerous problems persist. Despite many enthusiastic reports, no di- agnostic imaging modality is specific for the detection of prostate cancer. The decision on which is the best, most accurate, staging modality remains controversial. ULTRASONOGRAPHY The prostate gland can be examined sonographically through the transabdominal, transrectal, and transurethral approaches. The suprapubic abdominal approach is simple and does not require a dedicated, specialized transducer (1). Due to the depth of the prostate gland, the transabdominal approach is hampered by the need for a low-frequency (3.5 MHz) transducer (fig. 1), which limits the spatial resolu- tion and thus the detection and evaluation of intraprostatic disease (2). The transrectal and transurethral approaches provide a much better assessment of the prostatic parenchyma. As the distance between the rectal transducer and the prostate is short, high-frequency transducers (5-10 MHz) can be used. This markedly improves spatial resolution and allows better assessment of the anatomy and pathology of the prostate. Transrectal sonography in sagittal projection is optimal for evaluation of the craniocaudal margins of the gland. In the transrectal transverse scan, the relationships among the prostate, prostatic capsule, and levator ani are well displayed. Two planes of imaging are essential for the complete examination of the prostate by ultrasound (3). ! Department of Radiology, University of California, San Francisco, Box 0628, San Francisco, CA 94143-0628. Many reports have enthusiastically addressed the use of transrectal ultrasound for screening (4). However, in actu- ality, its use for screening lacks sensitivity (defined as the number of true-positive cases divided by the sum of the true-positive and false-negative results). When compared with findings at pathologic examination (fig. 2), the re- ported sensitivity has ranged from 62% (2) to 65% (5). Fur- thermore, sonographic tissue characterization lacks speci- ficity (defined as the number of true-negative cases divided by the sum of the true-negative and false-positive results). Biopsy remains essential (/-6). However, if an abnormal- ity within the prostate is demonstrated on the sonogram, ultrasound is an excellent modality to guide needle biopsy. In the staging of prostate carcinoma, i.e., in evaluation of transcapsular and seminal vesicle extension, transrec- tal ultrasound is advocated as having the greatest accuracy (defined as the sum of the true-positive and true-negative cases divided by the total number of patients). In the eval- uation of transcapsular tumor extension, the accuracy of this modality ranges from 64% to 90%, the specificity from 78% to 94%, and the sensitivity from 59% to 86% (6,7). The demonstration of seminal vesicle involvement is more difficult, resulting in an accuracy range of 77% to 85% and a low sensitivity ranging from 29% to 33% (6,7). The ac- curacy of ultrasound for detection of transcapsular tumor extension is significantly higher than that of computed to- mography (7). When tumor extension to the bladder neck, rectum, or pelvic lymph nodes needs to be evaluated, ultra- sound is limited. The use of transrectal ultrasound for the local staging of the prostate carcinoma (assessment of disease confined to the prostate) needs further evaluation with correlation between histopathology and ultrasound findings. However, it should be emphasized that the small field of view limits complete evaluation of the pelvis, and ultrasound is at present not acceptable for lymph node evaluation. COMPUTED TOMOGRAPHY Computed tomography can display abnormalities in con- tour of the gland, but it does not have sufficient soft-tissue contrast to provide information about disease within the gland. Staging by computed tomography has been shown to have significant limitations (8-70). Carcinoma gener- ally has the same attenuation value as the remainder of the prostatic parenchyma and will not be identifiable unless it extends outside the gland. The diagnosis of stages A and B (carcinoma confined to the gland) is derived by exclusion (11), when the lesion fails to produce changes in the gland contour. The role of computed tomography in the evalua- tion of patients with prostate neoplasms is not to confirm 3 32 FIGURE 1.—Stage B2 prostate carcinoma. Left: Transabdominal ultrasound 3.5-MHz transducer. Right: Transrectal ultrasound 7.5 MHz. Calcifications (long arrow) are seen within the prostate. On a transrectal scan, calcifications are at the junction of the central gland and peripheral zone (P). Note bulging of the left lateral contour of the gland (arrows), and, although the margin appears indistinct, no periprostatic tumor extension was seen at surgery. Spatial resolution of the transrectal scan is superior to the transabdominal. the presence of disease but to delineate the extent of tumor spread and identify local pelvic and abdominal adenopa- thy. Its sensitivity in the detection of transcapsular tumor invasion is 25%, specificity 89%, and accuracy 60% (7). FIGURE 2.—Stage B2 prostate carcinoma. Transrectal ultrasound. Gland is enlarged, but the echogenicity throughout the peripheral zone is homogeneous. At surgery, a B2 lesion was present in the right side (arrow) of the gland. These findings are significantly lower than those of either transrectal ultrasound (7) or magnetic resonance imaging (12). The sensitivity of computed tomography in detecting seminal vesicle involvement was 36%, specificity 96%, and accuracy 76% (7), similar to those of transrectal ultrasound (7). As with ultrasound, computed tomography cannot ac- curately identify extension into the bladder base, as this is a transverse interface parallel to the transverse imaging plane of the tomography. The potential for understaging and overstaging is thus significant. Obliteration of the sem- inal vesicle angle with enlargement of the seminal vesicle is suggestive of tumor extension, but the specificity of this finding is questionable. The evaluation of lymph node in- volvement by computed tomography is also limited. The reported sensitivity ranges from 50% to 75%, the speci- ficity from 86% to 100%, and the accuracy from 83% to 92% (8-10). Morphologic criteria alone (size) determine lymph node involvement; a decrease in the size threshold will give fewer false-negative results, but the false-positive rate will increase (fig. 3). Normal-sized nodes often are not detected in the pelvis. Patients with false-positive results usually have fibrotic, hyaline, or histiocytic changes in en- larged nodes. Lymphadenectomy or fine-needle aspiration is required for accurate diagnosis, with the risk of increased morbidity. MAGNETIC RESONANCE IMAGING Initial results of the evaluation of prostatic disease by magnetic resonance imaging were promising, but as more experience has been gained, reports have conflicted (12-16). Excellent tissue contrast by magnetic resonance imaging allows depiction of intraprostatic anatomy, differ- NCI MONOGRAPHS, NUMBER 7, 1988 33 La = 120 ! A IRSONICS MRI — UCSF DIRSONICS MRI — UCSF FIGURE 4.—Stage B2 prostate carcinoma. Left: T1-weighted image (spin echo) SE 500/30. Right: T2-weighted image SE 2,000/60. Prostate (P) is enlarged. On T1-weighted image, contrast is excellent between prostate parenchyma and surrounding periprostatic tissue. No tumor spread outside the prostate is identified. On T2-weighted image, central gland of the prostate (C) is of low signal intensity and can be differentiated from higher signal intensity peripheral zone. Levator ani muscles (arrows) are clearly visualized and are of normal signal intensity. MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 34 entiating the peripheral, central, and transition zones (17). Disease confined to the gland can be detected by magnetic resonance imaging; however, there is overlap between the appearance of benign and malignant conditions when this modality is used (/4). Thus findings are not specific, and magnetic resonance is not suitable as a screening modal- ity (12). Once carcinoma is histologically documented, it is valuable for staging the disease, with reported accuracy be- tween 83% and 89%, sensitivity between 75% and 87%, and specificity from 88% to 90% (12-16). Results from mag- netic resonance imaging are significantly higher than those of computed tomography or clinical examination. How- ever, an extensive imaging protocol is needed for accurate imaging results (fig. 4). This must include combinations of various imaging parameters and images obtained in two orthogonal planes. The transverse plane is valuable for as- sessment of the relationships among the prostate, prostatic capsule, and levator ani as well as for the detection of lym- phadenopathy. The sagittal plane is preferred in patients who have undergone transurethral resection of the prostate or in whom bladder or rectal involvement is suspected. Coronal images offer an additional view of the periprostatic fossa, levator ani, and possible lymph node metastasis (fig. 5). In detecting lymph node metastases, we find magnetic resonance imaging results are similar to those of computed tomography, and lymph node biopsy for histologic diagno- sis is still essential. SUMMARY At present, no imaging modality can be relied upon as a diagnostic method for prostate carcinoma. Perhaps the combination of physical examination, transrectal ultra- sound, and ultrasound-guided biopsy will be the best ap- proach for prostate cancer detection. For staging of prostate FIGURE 5.—Stage D1 prostate carcinoma. Bilat- erally enlarged lymph nodes (curved arrow) are identified. However, it is only the size of the node (1 to 1.5 cm in length) that can be detected on magnetic resonance imaging, which cannot distinguish malignant from hy- perplastic nodes; biopsy is essential. carcinoma, a uniform staging system is crucial for patient management. The use of computed tomography for local staging of the disease has not been widely accepted because the results are similar to those of clinical examination. How- ever, the accuracy rates of both transrectal ultrasound and magnetic resonance imaging for local staging of the dis- ease are significantly higher than those of either computed tomography or clinical studies. A combination of transrec- tal ultrasound and magnetic resonance imaging appears to yield the best results, but further studies are necessary for confirmation. The sensitivity of imaging needs to be ex- plored further, and a larger series should be studied for an analysis of combined staging methods by ultrasound and magnetic resonance imaging that would address not only the accuracy but also the impact of staging on the thera- peutic approach. REFERENCES (1) ABU-YOUSEF MM, NARAYAN AS. Prostatic carcinoma: De- tection and staging using suprapubic US. Radiology 1985;156:175-180. (2) HAMPER JM, DAHNERT WF, EGGLESTON JC, ET AL. Ultra- sonography of prostatic carcinoma employing amplitude- enveloped (AM) and frequency-demodulated (FM) imag- ing. J Ultrasound Med 1986;5:557-562. (3) HricAk H, JEFFREY RB, Dooms GC, ET AL. Evaluation of prostatic size: A comparison of ultrasound and magnetic resonance imaging. Urol Radiol 1987;9:1-8. (4) LEE F, GRAY JM, MCLEARY RD, ET AL. Transrectal ul- trasound in the diagnosis of prostate cancer: Loca- tion, echogenicity, histopathology and staging. Radiology 1986;158:91-95. (5) DAHNERT WF, HAMPER UM, EGGLESTON JC, ET AL. Pros- tatic evaluation by transrectal sonography with histo- pathologic correlation: The echopenic appearance of early carcinoma. Radiology 1986;158:97-102. NCI MONOGRAPHS, NUMBER 7, 1988 (6) PONTES JE, EISENKRAFT S, WATANABE H, ET AL. Preopera- tive evaluation of localized prostatic carcinoma by trans- rectal ultrasonography. J Urol 1985;134:289-291. (7) SALO JO, KivisaARI L, RANNIKKO S, ET AL. Computerized tomography and transrectal ultrasound in the assessment of local extension of prostatic cancer before radical retro- pubic prostatectomy. J Urol 1987;137:435-438. (8) SHANKAR GIRI PG, WALSH JW, HAZRA TA, ET AL. Role of computed tomography in the evaluation and management of carcinoma of the prostate. Int J] Radiat Oncol Biol Phys 1982;8:283-287. (9) GoLIMBU M, MORALES P, AL-ASKARI S, ET AL. CAT scan- ning in staging of prostatic cancer. Urology 1981;18:305- 308. (10) MORGAN CL, CALKINS RF, CavaLcanTI EJ. Computed tomography in the evaluation, staging and therapy of carcinoma of the bladder and prostate. Radiology 1981;140:751-761. (11) WALsH PC, LErorR H. The role of radical prostatec- tomy in the management of prostatic cancer. Cancer 1987;60:526-537. MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 35 (12) HricAK H, DooMs GC, JEFFREY RB, ET AL. Prostatic carci- noma: Staging by clinical assessment, CT, and MR imag- ing. Radiology 1987;162:331-336. (13) HrIiCAK H, WILLIAMS RD, SPRING DB, ET AL. Anatomy and pathology of the male pelvis by magnetic resonance imag- ing. Radiology 1983;141:1101-1110. (14) LING D, LEE JKT, HEIKEN JP, ET AL. Prostatic carcinoma and benign prostatic hyperplasia: Inability of MR imag- ing to distinguish between the two diseases. Radiology 1986;158:103-107. (15) BionDETTI PR, LEE JKT, LING D, ET AL. Clinical stage B prostatic carcinoma: Staging with MR imaging. Radiology 1987;162:325-329. (16) PHILLIPS ME, KRESSEL HY, SPrITZER CE, ET AL. Prostatic disorders: MR imaging at 1.5T. Radiology 1987;164:386- 392. (17) Hricak H, DooMs GC, MCNEAL JE, ET AL. Magnetic res- onance imaging of the prostate gland: Normal anatomy. AJR 1987;148:51-58. Lymphography in Clinically Localized Prostate Cancer Ronald A. Castellino! ABSTRACT—Lymphography demonstrates the size, position, and internal architecture of the external iliac, common iliac, para-aortic, and paracaval lymph nodes. Importantly, the “surgi- cal obturator” nodes are also routinely opacified because they are part of the external iliac chain. Analysis of the internal architec- ture permits detection of metastases in nodes of normal size, an advantage over cross-sectional imaging techniques. In a prospec- tive study of 89 unselected, previously untreated patients with carcinoma limited to the prostate or periprostatic bed, lymphog- raphy was compared with histology of lymph nodes removed at surgical staging. The sensitivity was 53% (17 of 32), specificity 93% (53 of 57), accuracy 79% (70 of 89), and positive and neg- ative predictive values were 81% (17 of 21) and 78% (53 of 68), respectively. —NCI Monogr 7:37-39, 1988. Patients with newly diagnosed carcinoma of the prostate undergo a rigorous evaluation by physicians to determine the anatomic extent of disease because this has signifi- cant implications in patient management and prognosis. In certain patients, assessment of the regional (pelvic) lymph nodes is important as most urologic surgeons will not perform radical prostatectomy in this patient population; rather, the patient will be recommended for alternative therapy, such as external-beam radiotherapy. In this sub- group of patients, clearly the best assessment of regional lymph nodes is by surgical removal and histologic evalua- tion. However, this commits the patient to a surgical pro- cedure that could be obviated, if an acceptable noninvasive evaluation of the lymph nodes could be performed. Lymphography permits accurate radiographic display of the size, position, and internal architecture of the exter- nal iliac, common iliac, para-aortic, and paracaval lymph nodes to the level of the renal vascular pedicle. Although the internal iliac (hypogastric) lymph nodes are not rou- tinely or completely opacified, importantly the so-called “surgical obturator” nodes are routinely opacified, inas- much as they are part of the medial external iliac lymph node chain (1,2). Careful analysis of the internal architec- ture of each lymph node permits detection of lymph node abnormalities compatible with metastases within nodes that are still normal in size. This is an advantage over current cross-sectional imaging techniques (computed to- mography, ultrasound, and magnetic resonance imaging), ! Department of Diagnostic Radiology and Nuclear Medicine, Room S078, Stanford University Medical Center, Stanford, CA 94305. 2 Sensitivity is the percentage of patients with histologically positive nodes whose lymphograms were correctly interpreted as positive; specificity is the percentage of patients with histologically normal nodes whose lymphograms were correctly interpreted as negative; overall accuracy is the percentage of patients whose lymphograms were correctly interpreted as positive or negative. which rely upon nodal enlargement to detect adenopa- thy. False-positive lymphographic diagnoses are relatively uncommon and are mainly the result of defects caused by nonspecific changes, i.e., fat, fibrosis, and hyperplasia. False-negative lymphographic interpretations are relatively common, due primarily to small metastatic deposits that produce minimal structural alterations within the lymph nodes or to metastases to the deeper pelvic nodes that are not opacified. Lymphography is a minimally invasive procedure that is performed on an outpatient basis. Small (1-2 cm) super- ficial incisions over the dorsum of both feet are made to isolate the small lymphatic channels in the subcutaneous tissues, followed by their cannulation. The procedure re- quires some technical skill and experience to perform, fac- tors which have discouraged many radiologists from de- veloping expertise in either the performance or subsequent interpretation of these studies. An accurate assessment of sensitivity, specificity, over- all accuracy,’ and positive and negative predictive values of lymphography, and particularly of other imaging tech- niques, is hampered by the paucity of carefully performed prospective clinical studies. Reports based upon selected patients, small numbers of patients, patients with varying stages of disease or at varying times during the course of their illnesses, and lack of pathologic correlation with the imaging study interpretation provide information that can be misleading and provide little confidence for sub- sequent decision-making. However, there are several rel- atively large series of patients with clinically localized carcinoma of the prostate, who underwent lymphography followed by lymph node biopsies, which provide some in- sights into accuracy for this technique in this patient popu- lation. The Stanford experience [(3); Castellino RA: Unpub- lished observation] encompasses 89 unselected, previously untreated patients with disease limited to the prostate (clin- ical stage A or B) or periprostatic bed (clinical stage C), without evidence of metastases based upon clinical assess- ment, laboratory studies, and skeletal radioisotopic and/or radiographic surveys. The lymphograms were interpreted as positive or negative for metastatic disease; no equivo- cal diagnostic category was utilized. The cannulation was successful bilaterally in all but 2 patients. The 2 unilateral studies are included in the analysis even though the lack of contralateral opacification compromised diagnostic effec- tiveness. Following lymphographic interpretation, the nodes of greatest concern were indicated on the radiographs. The patients underwent multiple lymph node biopsies at various anatomic sites to permit subsequent radiographic/histologic correlations. Assessment of postoperative abdominal films confirmed if the nodes of concern were adequately sam- 37 38 TABLE 1.—Lymph node histology of surgical specimens from unselected, previously untreated patients with prostate cancer: Stanford experience Histologically positive nodes Pelvic Lumbar Clinical No. of No./total No. stage patients ~~ No. Percent of patients Percent A 2 0 B 47 9 19 2/9 22 C 40 23 58 11/23 48 Total 89 32 36 13/32 41 pled. (In 2 additional patients, the lymphographically pos- itive nodes were not removed, and their results are not in- cluded in this analysis.) One or more excised pelvic lymph nodes were histologi- cally positive in 19% (9 of 47) of clinical stage B and 58% (23 of 40) of clinical stage C patients. Histologically pos- itive para-aortic or paracaval lymph nodes were noted in 22% (2 of 9) and 48% (11 of 23) of this group of patients, respectively (table 1). In no instance was there retroperi- toneal, without concomitant pelvic, lymph node involve- ment. Lymphographic-histologic correlation (table 2) for the whole group of patients revealed a sensitivity of 53%, speci- ficity of 93%, and accuracy of 79%, and predictive values of positive and negative tests in 81% and 78%, respectively. A breakdown of these accuracy data for the patients with clinical stages B and C disease is provided in table 3. A similar study was performed by the Uro-Oncology Re- search Group (4). This was a multi-institutional study and all lymphograms were interpreted by 1 referee radiologist. Their data, accumulated from 149 patients undergoing lym- phography followed by surgical lymph node sampling, re- vealed a sensitivity of 56%, specificity of 95%, and accu- racy of 81%, with predictive values of positive and negative tests in 86% and 79%, respectively (table 4). These results are quite similar to the Stanford experience. Currently, lymphography is the most accurate diagnostic imaging study in the assessment of the external and com- mon iliac nodes and para-aortic and paracaval nodes to the level of L1-2. Although increasingly accurate in patients with advanced disease (in which both the incidence and the bulk of lymph node involvement are greater), this tech- nique has significant false-negative diagnoses when patients with apparently localized prostate carcinoma are evalu- ated. A negative lymphographic interpretation is therefore unreliable, as are other negative clinical, laboratory, and diagnostic imaging evaluations, all of which are insensi- tive to relatively minimal amounts of disease. On the other hand, a positive lymphographic interpretation does possess a relatively high predictive value, which, not surprisingly, is greater in clinical stage C, as compared with clinical stage B, patients. Thus a positive lymphographic diagnosis merits strong consideration for lymph node metastases. In those in- stances in which the lymphographic findings are suspicious but not compelling, or in which the clinical presentation is discordant with the lymphographic findings, further evalua- tion can be provided with fluoroscopically guided percuta- neous needle aspiration. This technique is also performed on an outpatient basis and is associated with minimal morbid- ity and rare complications. A positive lymph node aspiration provides meaningful information; however, a negative node aspiration does not exclude the presence of metastases, and this procedure can be repeated, as indicated. Although experience with magnetic resonance imaging is still being accumulated, this technique should possess simi- lar diagnostic accuracy as does computed tomography, and perhaps may be better due to the confidence with which vessels can be distinguished from lymph nodes. Although both techniques require enlarged lymph nodes for a diag- nosis of adenopathy, they are noninvasive and require little technical skill for performance of the procedures. Further- more, many radiologists are very skilled in the performance and interpretation of cross-sectional imaging examinations. Evaluation of the prostate itself, determination of the local extension of tumor, and assessment of other organ systems are additional benefits of computed tomography and mag- netic resonance imaging studies. The following is proposed as a possible strategy in as- sessment of the regional lymph nodes in patients with ap- parently clinically localized prostate carcinoma. 1) First, and most important, only those patients need to be evaluated in whom the status of the lymph nodes will have impact upon their management. 2) Clinical evaluation and histologic assessment of the biopsy material will frequently determine a sub- group of patients in whom the likelihood of lymph node metastases is very small, or in whom lymph node metastases are likely to be relatively small in size, e.g., those with early stage disease and/or low Gleason scores. In these patients, lymph node imag- TABLE 2.—Lymphography and pathology results: Stanford experience No. of patients with: Negative Positive Positive lymphography and Negative lymphography and Clinical No. of lymphography and positive lymphography and negative stage patients lymph node biopsy lymph node biopsy lymph node biopsy lymph node biopsy A 2 2 B 47 2 7 35 3 Cc 40 15 8 16 1 Total 89 17 15 53 4 NCI MONOGRAPHS, NUMBER 7, 1988 39 TABLE 3.—Lymphographic-pathologic correlations: Stanford experience Clinical stage . Total B Cc Interpretation of results Percent Percent Percent Sensitivity 53 17/32 n 2/9 62 15/23 Specificity 93 53/57 92 35/38 94 16/17 Accuracy 79 70/89 79 37/47 78 31/40 Positive predictive value 81 17/21 40 2/5 94 15/16 Negative predictive value 78 53/68 83 35/42 67 16/24 ing studies will have a low yield indeed and could well be dismissed. 3) In patients with a likelihood of lymph node metastases, arguments could be supported that a cross-sectional imaging study be performed first that may demonstrate obvious lymph node or dis- tant metastases. This would obviate the need to per- form lymphography. 4) Lymphography would then be reserved for those patients for whom information about the lymph nodes is important for management, in whom en- larged lymphadenopathy has been excluded by computed tomographic and magnetic resonance scanning, and in whom metastases within normal size or only modestly enlarged lymph nodes may be detected by lymphographic examination. TABLE 4.—Results of 149 patients undergoing lymphography followed by surgical lymph node sampling? Interpretation of results Percent Sensitivity 56 30/54 Specificity 95 90/95 Accuracy 81 120/149 Positive predictive value 86 30/35 Negative predictive value 79 90/114 “¢ Patients were in the Uro-Oncology Research Group studies. See (4). MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER a) A positive or suspicious study can be con- firmed by percutaneous needle aspiration, if such reassurance is needed. This will obviate patients being denied a potentially curable pro- cedure on the basis of a false-positive diagnos- tic imaging study. b) A negative study does not exclude metastases. Further evaluation, such as by surgical lymph node sampling, will need to be considered, de- pending upon the importance of such informa- tion in patient management. 5) Should local expertise in performing and interpret- ing lymphograms not be available, this evaluation could be performed in regional centers by special- ists with experience. REFERENCES (1) MERRIN C, WAISMAN ZM, BAUMGARTNER G, ET AL. The clini- cal values of lymphangiography: Are the nodes surrounding the obturator nerve visualized? J Urol 1977;117:762-764. (2) ZoretiC SN, WAISMAN Z, BECKLEY SA, ET AL. Filling of the obturator lymph nodes in pedal lymphangiography: Fact or fiction? J Urol 1983;129:533-535. (3) SPELLMAN MC, CASTELLINO RA, RAY GR, ET AL. An eval- uation of lymphography in localized carcinoma of the prostate. Radiology 1977;125:637-644. (4) LiEBNER EJ, STEFANI S, URO-ONCOLOGY RESEARCH GROUP. An evaluation of lymphography with nodal biopsy in local- ized carcinoma of the prostate. Cancer 1980;45:728-734. -n — — et = BE a Value of and Indications for Pelvic Lymph Node Dissection in the Staging of Prostate Cancer Richard G. Middleton! ABSTRACT —Pelvic lymphadenectomy is valuable as a staging procedure prior to radical prostatectomy in patients with clinical stages A2, B1 (except low-grade lesions), and B2 prostate can- cer who seem to be good candidates for an attempt at curative surgery. Survival rates are promising in patients with negative pelvic lymph nodes and local tumors who undergo radical prosta- tectomy. In the presence of positive nodes, there is little reason to proceed with radical prostatectomy. Noninvasive alternatives to pelvic node dissection are appealing, but lymphangiography, ultrasound, computed tomography scanning, and magnetic reso- nance imaging are all less reliable than pelvic lymphadenectomy. Some morbidity is associated with surgical staging, and it is im- portant that this be minimized. Pelvic lymph node dissection can play a role in treatment planning for patients who will be given external-beam radiation therapy. However, the role depends on the physician’s treatment philosophy. In a recently reported se- ries of patients receiving radiation therapy for localized prostate carcinoma, prior surgical staging by pelvic lymphadenectomy is uncommonly performed.—NCI Monogr 7:41-43, 1988. There is little evidence that pelvic lymphadenectomy is beneficial therapeutically in prostate cancer. The value of pelvic lymph node dissection is as an accurate stag- ing procedure when the need to know precisely the state of the pelvic lymph nodes is vital and when significant clinical decisions depend on accurate information. Pelvic lymphadenectomy for staging is generally indicated when prostate cancer seems to be localized clinically and when the patient seems to be a good candidate for therapy with curative intent, i.e., radical prostatectomy or radiation ther- apy. When metastasis to the pelvic lymph nodes has oc- curred, the patient already has systemic disease and is un- likely to benefit from aggressive attempts to eradicate the primary prostate cancer. Although some urologic surgeons proceed with radical prostatectomy in the presence of pos- itive pelvic lymph nodes, most, including myself, recognize that the tumor at this stage is almost inevitably progressive, and we think that radical prostatectomy is not appropriate in the presence of pelvic nodal metastasis. Other palliative measures, either early or delayed, are more reasonable. The prognosis for a patient with pelvic lymph node metastasis is ominous. About 75% of the patients with pos- itive pelvic nodes develop bony metastasis within 5 years, and about 40% die of prostate cancer within 5 years of de- tection of nodal metastasis (/). Once nodal metastasis has occurred, the rate of tumor progression is related to the volume of nodal disease and to the tumor grade; local or regional therapy has little effect on the progression (7,2). ! Division of Urology, Department of Surgery, University of Utah School of Medicine, 50 North Medical Dr., Salt Lake City, UT 84132. Radical prostatectomy may decrease the incidence of local problems in the pelvis, but the benefits are slight. On the other hand, when the tumor is clinically localized to the prostate and lymph nodes are negative, as determined by lymphadenectomy, the results of radical prostatectomy are encouraging. From the University of Utah, we have re- ported (3) on 136 patients with clinical stages A2, Bl, and B2 prostate cancer and proven negative pelvic nodes who have had radical prostatectomy and follow-up for 5 years after surgery (table 1). At 5 years following radical prosta- tectomy, 128 patients (94%) were alive and 118 (87%) were alive and well without evidence of tumor recurrence. The outcome for those with stages A2 and Bl disease and neg- ative pelvic nodes was particularly good, i.e., 91% in each category were alive and free of tumor recurrence at 5 years. Patients with B2 tumors and negative pelvic nodes fared less well, but 81% were alive and free of tumor at 5 years. In our experience, patients with prostatic capsular invasion (surgical stage C disease) did as well at 5 years as those with tumor histologically confined to the substance of the prostate. Other authors (4) have confirmed that capsular invasion alone in the radical prostatectomy specimen is not particularly adverse prognostically. Microscopic tumor in- vasion into the seminal vesicles of the radical prostatectomy specimen is an ominous finding in our experience and in the reported experience of many other authors. One-half of our patients with seminal vesicle invasion had tumor recurrence within 5 years of radical prostatectomy. Fortunately, when pelvic lymph nodes are negative, seminal vesicle invasion by prostatic tumor is uncommon; in our series (3), it was 6.5%. Thus patients with positive nodes benefit little from radi- cal prostatectomy, but those with localized tumor and neg- ative nodes have a promising outlook. Further follow-up is needed in our radical prostatectomy series, but clearly, pelvic lymphadenectomy has been vital in the selection of appropriate candidates for radical surgery. Nodal assess- ment by lymphangiography, ultrasound, computed tomog- raphy scanning, and magnetic resonance imaging is not sufficiently accurate, even with fine-needle aspiration of suspicious lesions, to replace pelvic lymphadenectomy for reliable staging (5). INFLUENCE OF STAGE AND GRADE ON PELVIC NODE INVOLVEMENT The incidence of positive pelvic lymph nodes relates to clinical stage and tumor grade (table 2). We have reported 452 staging pelvic lymphadenectomies for clinically local- ized prostate cancer (6). None of 41 patients with stage Al prostate cancer or minimal focal microscopic tumor had 41 42 TABLE 1.—Radical prostatectomy in patients with negative pelvic lymph nodes and 5-yr survival No. of patients alive at 5 yr No evidence of a Clinical Total No. Total Fecunshes stage of patients No. Percent No. Percent A2 34 33 97 31 91 Bl1 44 43 98 40 91 B2 58 52 90 47 81 Total 136 128 94 118 87 ¢ Percentages represent No. of patients free of disease/total No. of patients. positive pelvic nodes. For stage A2 tumor (diffuse micro- scopic prostate cancer), no low-grade tumors were asso- ciated with positive nodes, but 26% of the patients with moderately differentiated lesions and 43% with high-grade tumors had positive pelvic lymph nodes. Few of the patients, only 4%, with low-grade B1 tumor nodules had positive pelvic nodes, but an increasing incidence of nodal metas- tasis occurred with moderate- and high-grade Bl tumors. Not surprisingly, the incidence of positive nodes is higher in B2 tumors than in Bl nodules, and tumor grade is an important factor. More than one-half of the patients with clinical stage C prostate cancer had positive pelvic lymph nodes, and virtually all of those with poorly differentiated stage C lesions had nodal metastasis. When radical prostatectomy for cure is considered, pa- tients with clinical stages A2 and B2 tumors require a stag- ing pelvic lymphadenectomy. Patients with low-grade Bl tumors have a low incidence of positive nodes, and it ap- pears that lymphadenectomy is optional and nonessential. No need is indicated for pelvic lymphadenectomy in stage Al prostate cancer, and I believe there is no need for treat- ment of any kind. Patients with clinical stage C prostate cancer are poor candidates for radical prostatectomy be- cause it is almost certain that the tumor will be incom- pletely removed. Pelvic lymph node dissection may have value for staging if radiation treatment is contemplated in clinical stage C disease. NODE DISSECTION WITH RADIATION THERAPY The emphasis of this report is on pelvic lymph node dissection and radical prostatectomy. The place of node dissection in association with radiation therapy depends considerably on the philosophy of those who manage the treatment. If a patient with localized cancer (stages A2, B, and C lesions) has a negative pelvic lymphadenectomy, radiation therapy should be directed to the prostate and periprostatic area only. There seems to be no need for the radiologist to direct radiation therapy to the site of negative pelvic lymph nodes. Furthermore, node dissection followed by full pelvic irradiation produces a high incidence of chronic lymphedema of the legs and genitalia. Many, probably most, radiation therapists treat localized prostate cancer by full pelvic irradiation with an extra boost of radiation energy to the prostate. If the full pelvis is to be irradiated, whether the lymph nodes in the pelvis are pos- itive or negative, the lymphadenectomy is not crucial and should be avoided. However, we have a continuing problem in comparing the results of radical prostatectomy with those of radiation therapy for localized prostate cancer when only the patients who have radical prostatectomy have staging pelvic lymphadenectomy and a radical prostatectomy spec- imen to be examined histologically. MORBIDITY OF PELVIC LYMPHADENECTOMY Most often, pelvic lymphadenectomy is performed prior to radical prostatectomy, so separation of the morbidity associated with the node dissection from that relating to the removal of the prostate would be difficult. However, table 3 shows the morbidity in 176 patients who had pelvic lymphadenectomies performed as an independent staging procedure. Complications occurred in 9% of the patients. However, there were no fatalities, and only 1 patient with a ureteral injury required a reoperation. CONCLUSIONS Extensive experience with pelvic lymphadenectomy has led to the following conclusions: 1) Pelvic lymphadenectomy for staging is important prior to radical prostatectomy in patients with clin- ical stages A2 and B tumors, except for low-grade B1 lesions. TABLE 2.—Incidence of pelvic lymph node metastasis by clinical stage and histologic grade Grade Well Moderately Poorly differentiated differentiated differentiated Total Stage No./total Percent No./total Percent No./total Percent No. Percent Al 0/28 0/12 0/1 0/41 A2 0/7 5/19 26 3/7 43 8/33 24 Bl 2/53 4 13/94 14 3/9 33 18/156 12 B2 5/27 18 29/106 27 9/21 43 43/154 28 C 5/10 50 18/44 41 13/14 93 36/68 53 Total 12/125 10 65/275 24 28/52 54 105/452 23 “ Values = No. of patients who had metastases/total No. who had lymphadenectomies performed. NCI MONOGRAPHS, NUMBER 7, 1988 TABLE 3.—Complications of staging pelvic lymphadenectomy in 176 patients Complication No. of patients Mortality 0 Morbidity 16 (9%) Thrombophlebitis 4 Lymphocele 4 Retroperitoneal hemorrhage 1 Pulmonary embolus 3 Ureteral injury 1 Wound hematoma 3 2) If nodal metastasis is detected, radical prostatec- tomy is unlikely to be beneficial to the patient. 3) For staging, noninvasive techniques are not as ac- curate as pelvic lymph node dissection. 4) Some morbidity associated with pelvic lymph- adenectomy is acceptable when accuracy of node staging is vital. 5) The role of pelvic lymphadenectomy prior to MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 43 external-beam radiation therapy depends on the physician’s treatment philosophy. REFERENCES (1) SMITH JA JR, HAYNES TH, MIDDLETON RG. Impact of ex- ternal irradiation on local symptoms and survival free of disease in patients with pelvic lymph node metastasis from adenocarcinoma of the prostate. J Urol 1984;131:705-707. (2) SMITH JA JR, MIDDLETON RG. Implications of volume of nodal metastasis in patients with adenocarcinoma of the prostate. J Urol 1985;133:617-619. (3) MIDDLETON RG, SMITH JA JR, MELZER RB, ET AL. Patient survival and local recurrence rate following radical prosta- tectomy for prostatic carcinoma. J Urol 1986;136:422-424. (4) JEWETT HJ. The case for radical perineal prostatectomy. J Urol 1970;103:195-199. (5) CorrREA RJ Jr, KIDD ER, BURNETT L, ET AL. Percutaneous pelvic lymph node aspiration in carcinoma of the prostate. J Urol 1981;126:190-191. (6) SMITH JA JR, SEAMAN JP, GLEIDMAN JB, ET AL. Pelvic lymph node metastasis from prostate cancer: Influence of tu- mor grade and stage in 452 consecutive patients. J Urol 1983;130:290-292. II. Radiation Therapy Status of Radiation Treatment of Prostate Cancer at Stanford University Malcolm A. Bagshaw,!* Richard S. Cox,! Gordon R. Ray? ABSTRACT—Over 900 patients have been treated with radia- tion therapy in the 30-year Stanford prostate study. Overall sur- vival, i.e., scoring death due to all causes, was 45%, 35%, 33%, 20%, and 10% for Stanford stages TO, T1, T2, T3, and T4 (nom- inal stages A, Bl, B2, C) at 15 years; lymph node status was unknown. Disease-specific survival at 15 years was 85%, 64%, 45%, 33%, and 15%, respectively, for the same patients. In 141 patients with restricted nodular disease (lymph node status un- known) equal to or less than one-half of one lobe involved (stage B1), the 15-year overall survival was 50% and identical to the ex- pected survival of an age-matched cohort of males. Potency was preserved in 86% of the patients at 15 months posttreatment, and 50% of the patients maintained erectile potency for 7 years posttherapy. Other sequelae and complications are analyzed. The incidence of second neoplasms did not exceed expectations for an age-matched population.—NCI Monogr 7:47-60, 1988. External-beam irradiation for the treatment of prostate cancer was developed in the post-World War II era, and since then, case selection, technique, the risks, and expec- tations of this treatment method have been studied exten- sively (I-10). The details of this report are based upon the Stanford experience, which spans 3 decades and appears to be a fair representation of this overall experience. PATIENTS AND METHODS Between 1956 and 1985, 1,877 patients with adenocar- cinoma of the prostate were referred for treatment (table 1). All patients were completely evaluated at the time of ac- cession to the series, and physicians used a variety of tech- niques contemporary with the times to establish the extent of disease. Initially, the metastatic process was evaluated by roentgenographic surveys; later, x-ray films were replaced by state-of-the-art scintiscans of the skeleton. Frequently lymphangiography, computed tomography, and, more recently, magnetic resonance imaging have been used as well. Of all the patients referred, 914 patients were selected for definitive treatment. Treatment.— Although the patients described here were treated with 4- or 6-MeV x-rays, 6- and 15-MeV x-rays generated by linear accelerators are currently used. Un- ABBREVIATIONS: MeV = million electron volt(s); TNM = tumor, node, metastases; TURP = transurethral resection of the prostate; TOf = cancer staging, focal subcategory; TOd = diffuse subcate- gory. ! Department of Radiation Oncology, Stanford University School of Medicine, Stanford, CA. 2 Palo Alto Medical Foundation, Palo Alto, CA. * Reprint requests to: Malcolm A. Bagshaw, M.D., Department of Ra- diation Oncology, Stanford University School of Medicine, Stanford, CA 94305. favorable rectal and bone sequelae are substantially re- duced by a 4-field technique (fig. 1) incorporating parallel- opposed anterior and posterior and left and right lateral fields (fig. 2). This arrangement provides significant pro- tection to the posterior wall of the rectum and avoids hot spots in bone. After 2,600 rad have been administered by the 4-field box technique, the radiation target is reduced to a coned down volume (fig. 3) for delivery of a booster dose of 2,000 rad limited to the prostate and seminal vesicles. Then the treatment program is expanded to the original 4-field box, which includes the entire first-echelon lymphatic sys- tem of the prostate. The composite distribution of radiation dose for the anterior, posterior, right and left lateral, and the left and right booster fields is depicted in figure 4. The ra- diation dose is delivered at the rate of 180 to 200 rad/day. This results in an absorbed dose of 7,000 rad in 7 weeks to the prostate, and 5,000 rad in 7 weeks to the first-echelon lymphatics. This program has been extremely well toler- ated with only a rare instance of significantly symptomatic prolonged urinary tract or gastrointestinal morbidity. Figure 5 demonstrates the age distribution that ranged from 35 to 86 years. The mean age was 63.4 and the median was 64 years. The series is neither prospective nor retrospective, but rather it is a continuing and contemporary study of the radiotherapy of prostate cancer as seen in the Department of Radiation Oncology at Stanford University. Appropriate demographic, diagnostic, therapeutic, and follow-up items are systematically recorded and stored in a comprehensive data bank. Follow-up evaluation of about three-quarters of the patients is conducted in-house by the author, and the remainder of the patients and/or their re- ferring physicians are interrogated regularly by a standard letter of inquiry or by telephone. Staging system.—The Stanford TNM staging system was developed of necessity in the early 1970s before other TNM systems were generally available (//-13). Even though this is a parochial system, it is necessary that we review it briefly in order to clarify data to be presented later. This system defined the primary tumor with respect to local extent and identified several categories of apparent nodular disease in terms that might serve to define criteria for either resection or irradiation. It was based upon interpretations of Jewett’s concepts, as published in his classic papers on the palpa- ble nodule of prostate cancer (/4). The original Stanford TNM system is presented in table 2. At the extremes, the T categories included TX, or status postsurgical resection, and T4, massive local tumor extending to the bony pelvis or invading adjacent organs. The intermediate T categories included cancer discovered coincident to a TURP (inciden- tal), and this was further subdivided into TOf or TOd. The T1 included nodular neoplasm clearly confined to within the capsule. This stage was divided into 4 subgroups: Tla, 47 48 TABLE 1.—Referrals for treatment of prostate carcinoma, October 1956 to December 1985¢ Total referrals 1,877 Exclusions Consultation only 198 Metastatic disease 533 Subtotal 731 Other Second primary tumor 64 Prostatectomy 38 Unusual primary tumor 5 Previous irradiation 10 Incomplete radiotherapy 14 Questionable histology 4 Implants 97 Subtotal 232 Total exclusions 963 Treated by external-beam irradiation Disease limited to prostate 506 Extracapsular extension 408 Total No. of patients 914 (in study) treated ¢ Patients were treated in the Division of Radiation Therapy, Stanford University Hospital. a nodule 1 cm or a little more in diameter; T1b, a nodule occupying at least one-half of one lobe; T1c, a nodule occu- pying more than one-half but still limited to one lobe; and T1d, involvement of both lobes. The T2 category was pur- posely defined as borderline between intracapsular and ex- tracapsular disease. This was an equivocal category, which included nodular lesions that distorted anatomic boundaries but did not definitely obliterate either the lateral sulci or the seminal vesicles. Stage T3 included tumors that clearly extended beyond the capsule, although they might involve less than 50% of a lobe or involve both lobes. Definitions of the various lymph node categories and metastatic sites are given in the table. Throughout the study, a number of parameters were an- alyzed and the status of each patient was evaluated repeat- edly. These observations included age, stage, histopatho- logic grade, treatment sequelae or complications, survival, disease-specific survival, freedom from relapse, and local control. Other items, such as lymphangiographic status or lymph node biopsy results, were also available for selected cohorts. Statistical data.— Actuarial curves for the various patient groups were calculated by the method of Kaplan and Meier (15). For these calculations, patients were considered to be at risk from the initial date of radiotherapy. Various actuarial functions, as defined herein, were determined and these are presented. 1) Survival: Patients who die of any cause are regarded as failing at the time of death. Living patients are censored at the time of last follow-up and continue in the study. 2) Disease-specific survival: Patients who die of prostate cancer are regarded as failing at the time of death. Living patients are censored at the time of last follow-up; also, patients who die of intercurrent causes are censored at the time of death. 3) Freedom from relapse: Patients who relapse either at the primary site or at a metastatic site are re- garded as failing at the time of first relapse. Pa- tients who never relapse are censored at the time of last follow-up or intercurrent death. 4) Freedom from local relapse (local control): Patients who relapse at the primary site, as determined by clinical examination, are regarded as failing at the time of local relapse. Patients who never relapse locally are censored at the time of last follow-up or death. 5) Freedom from distant relapse: Patients who relapse at a metastatic site are regarded as failing at the time of metastatic relapse. Patients whose cancers never metastasize are censored at the time of last follow-up or death. No curves in this category are presented here, but the data exist and are accessible. The significance of differences between actuarial curves was assessed by the generalized Wilcoxon test of Gehan (16). The patients were first segregated according to the 12 groupings of the staging system. Actuarial survival curves were generated and Gehan P values were determined for successive pairs. Curves not separated by significant P val- ues were coalesced, which resulted in a family of curves for each end point that remarkably, and with only one or two exceptions, grouped the patients into patterns that were originally predicted by the T stages. The method of group- ing is shown in figure 6. RESULTS Patients with TOf (Al) and TOd (A2) incidental carci- noma are presented. Although the survival curves for TOf and TOd appear divergent, the P value fails to demonstrate a significant difference. Note also that the apparent differ- ence between TOf and TOd virtually disappears when the end point is disease-specific survival rather than overall sur- vival. This is because 17 or 35% of the patients with TOd (A2) died of intercurrent disease apparently without tumor, whereas only 4 or 8% of the TOd patients died of tumor. Two of 20, or 10% of the TOf (Al), patients died of tumor, and only 5 of 20 or 10.2% of them died of intercurrent dis- ease. Therefore, the actuarial data for survival for TOf and TOd (Al and A2) were pooled because the difference be- tween like curves was not significant. Note also that the pressure of intercurrent death confounded the interpreta- tion of mortality data in this group of elderly patients who, albeit elderly, still had an opportunity for long survival. Fig- ure 7 shows the survival and disease-free survival patterns for these patients after the curves for focal and diffuse dis- ease were coalesced. Figure 8 demonstrates the overall actuarial survival of all patients as a function of the Stanford T-staging system. Note that, although adenopathy was known for a few pa- tients, it was not known for all and, therefore, in this and later figures, the status of adenopathy was disregarded as a staging parameter. However, in a previous study, surgi- cal assessment of lymphadenopathy in a cohort of these patients demonstrated histopathologically positive lymph nodes in 19% of the T1 and T2 (stage B), and 56% of NCI MONOGRAPHS, NUMBER 7, 1988 POT GAG FIGURE 1.—Parallel-opposed anterior and posterior fields in the 4-field treatment technique. MANAGEMENT OF CLINICALLY LOCALIZED PROS E CANCER 50 TUMOT me? Anal Canal = FIGURE 2.—Left and right lateral fields the T3 (stage C) patients (//). Of the T stages, Tl, in- tracapsular, and T2, equivocal (curves 1 and 2), behaved identically in overall survival. These two curves might have been coalesced; however, we purposely displayed them in- dividually to provide an easy comparison with the same stages to be displayed next as a function of disease-specific survival. From the standpoint of overall survival, little dif- ference appeared between patients with small nodules and those with extensive intracapsular involvement. Later data, however, will show that within the finer subgrouping of T1, there were significant differences in favor of lesser disease. Again, the pressure of intercurrent death tended to obfus- cate the true number of deaths caused by prostate cancer when overall survival was viewed. However, the disease-specific survival displayed in fig- ure 9 demonstrated a nearly significant difference between those with T1 or definitive intracapsular disease and those with T2 or equivocal capsular disease (P= .07). This differ- ence shows that when the capsule was distorted (T2), the disease-specific survival pattern was somewhere between T1 and T3, in keeping with the ambiguity built into the definition of the T2 category. Figure 10 depicts the pattern of freedom from relapse as a function of T stage. In this situation, the patient was scored as demonstrating no clinical evidence of disease if 1) upon physical examination, no evidence of tumor re- growth was observed on rectal examination following a prior regression after radiation therapy, and 2) routine an- cillary examinations, such as acid phosphatase determina- tions, bone scans, or roentgenograms, failed to reveal a neo- plasm. Recently, prostate-specific antibody determinations in the 4-field treatment technique. have been added as a surveillance marker. This evaluation did not include routine biopsy of the primary site. Reference to the table of P values in figure 10 segregates patients at least according to T stage into 5 highly significant groups. Therefore, freedom from relapse was the end point that most sharply defined the long-term outcome of the radia- tion treatment of prostate cancer as a function of T stage. Clearly, as with most types of cancer, prostate cancer be- comes progressively more difficult to control as the T stage escalates. Also, because tumor volume appears to be the dominant factor in the definition of T stage, tumor mass and presumably the cell number appear to play the most significant role in the determination of radiocurability. Figure 11 presents an actuarial analysis of control of tu- mor at the primary site (local control) as a function of T stage. Local control was determined by physical examina- tion only and, therefore, to a certain extent lacked authority. Some of the patients were known to have had positive biop- sies at the primary site, but these were disregarded in this analysis because only a minority of the patients were sub- mitted to biopsy, and in some of them, the clonogenicity of the apparently positive biopsies was in doubt. Clinical lo- cal control appears to be a considerably less reliable end point than either disease-specific or relapse-free survival. Local control does not clearly segregate the patients into reasonable stages. One concept of the initial Stanford staging system was to isolate solitary nodular disease as had been suggested by Jewett et al. (/4). The T1 category was comprised of nodu- lar carcinomas which were, as near as could be determined by physical examination, unequivocally confined within the NCI MONOGRAPHS, NUMBER 7, 1988 9 cm a ; prostate. The Tla nodule included those up to 1 cm or a little more in diameter, the T1b those that were larger than Tla but equal to or smaller than one-half of one lobe (2 to 2.4 cm in diameter), the Tlc those that occupied either an entire lobe or were multiple but confined to one lobe, and the T1d those that occupied a portion of both lobes or the entire gland. Inspection of figure 12 tells one that there was no significant difference in the probability of survival between stages Tla and T1b and, similarly, no significant difference in survival between T1c and T1d. Therefore, on combining the like substages in T1, the survival pattern is as depicted in figure 13. The survival pattern for intra- capsular nodular carcinoma of the prostate up to nodules at least equal in size to one-half of one lobe following ir- radiation appears identical to the expected survival of an age-matched population of males and is highly significantly better than for the larger lesions, even though the larger ones still might be presumed to be confined within the cap- sule. As stated above, disease-specific survival is perhaps a more informative way for one to view outcome in prostate cancer, inasmuch as the confounding influence of death due MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 51 FIGURE 3.—Coned down volume for de- livery of radiation booster dose to the prostate and seminal vesicles. ZIPSER AMF to intercurrent disease is removed. Disease-specific survival is presented in figure 14 for the T1 lesions. At 15 years, it is essentially 80% for Tla and T1b lesions and 50% for Tlc and T1d tumors. Although some might consider this T1b carcinoma of the prostate to be a small nodule, this is only relative. For example, if one considers a normal prostate to weigh about 30 g, then a nodule occupying about one-half of one lobe would be approximately 7.5 g or it would be a nodule of about 2.4 cm in diameter. This is a sizable nod- ule when one considers the work of McNeal et al. (7), in which 7 of 9 prostate cancer specimens of 5.4 ml or larger had at least 1 cm of complete (level III) capsular penetra- tion, and 6 of 7 had seminal vesicle invasion. Postradiotherapy Biopsy Table 3 demonstrates that 64 of 146 patients who were surgically staged for evaluation of adenopathy at Stanford had posttherapeutic biopsies on an ad hoc basis 2 years or longer following irradiation. The incidence of positive biopsies increased with increasing clinical stage. Table 4 52 00 15x 8 cm COUCH FIGURE 4.—Composite isodose distribution of the radiation dose, taken at the level of the prostate. shows that among the 39 patients who had positive biop- sies, 8 were living with metastatic disease, and 18 subse- quently died of progressive disease. Eleven others had posi- tive biopsies, but 10 have manifested no further progression. Therefore, it is apparent that conventional megavoltage ir- 100 I 90 80 | 701 60 I 50 40 30 Number of Patients 20 10 radiation in doses as high as 7,000 rad in 7 weeks, as was used in these patients, was frequently incapable of steril- izing primary prostate carcinoma, particularly in the more bulky stage T2 (B2) and T3 (C) lesions. Conversely, however, the neoplasm was sterilized in many patients, and ways to improve the ability to achieve local control at the primary site are available, one of which is to boost external-beam irradiation with transperineal iridium-192 interstitial implants, as has been advocated by Syed et al. (I8) and also by the Stanford group (fig. 15). Both use a perineal template designed to place intersti- tial iridium implants directly into the primary in stage C tumors.? Other booster methods being tested include the use of high-energy protons, neutron irradiation, negative pi meson irradiation, and heavy-ion irradiation. The addition of hyperthermia to either external-beam or interstitial irra- diation also may improve local control. Future studies may 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 Age FIGURE 5.—Age distribution of patients with prostate cancer at the time of referral to Stanford. Mean age = 63 yr. 3 As described above, 5,000 rad are delivered in 5-6 weeks to the prostatic region with the 4-field technique. Then the patient is allowed 4 weeks of rest, followed by an interstitial iridium-192 implant that adds a 3,000-rad boost to the primary site. NCI MONOGRAPHS, NUMBER 7, 1988 53 TABLE 2.—Clinical staging system for prostate cancer T staging: Primary tumor TX: Characteristic anatomic relationships distorted and/or absent secondary to major surgical intervention, i.e., radical prostatectomy. TO: Occult carcinoma: incidental finding of carcinoma in the operative specimen following TURP. f: focal = less than 5% of specimen. d: diffuse = more than 5% of specimen. T1: Palpable tumor limited by the prostatic capsule without distortion of the superior or lateral anatomic boundaries. a: Solitary nodule =< 1 cm in diameter with normal, compressible prostatic tissue on three sides (lesion amenable to radical prostatectomy). b: Palpable tumor > 1 cm occupying less than 50% of a lobe. c¢: Palpable tumor occupying > 50% of a lobe, multiple nodules limited to one lobe. d: Involvement of both lobes. T2: Palpable tumor of any size primarily limited by the prostatic capsule with minimal distortion of the lateral or superior anatomic boundaries without definite obliteration of a lateral sulcus and/or a seminal vesicle region. a: Palpable tumor occupying up to 50% of one lobe. b: Palpable tumor occupying > 50% of a lobe, multiple nodules limited to one lobe, or involvement of both lobes. T3: Palpable tumor extending beyond the prostatic capsule with definite obliteration of any extent of a lateral sulcus and/or a seminal vesicle region. a: Palpable tumor occupying up to 50% of a lobe. b: Palpable tumor occupying > 50% of a lobe, multiple nodules limited to one lobe, or involvement of both lobes. T4: Palpable tumor extending beyond the prostatic capsule with attachment to both pelvic sidewalls, rectal wall invasion, or bladder invasion. N categories: Nodes Nodal designations will involve two notations: N ____. The first blank designates the clinical impression; the second blank designates the biopsy status. Clinical: X: Lymphangiogram not performed or technically inadequate. 0: Lymphangiogram not consistent with carcinomatous involvement of either the regional or para-aortic nodes. : Lymphangiogram consistent with carcinomatous involvement of the regional nodes; para-aortic nodes uninvolved. : Lymphangiogram consistent with carcinomatous involvement of the para-aortic nodes; regional nodes uninvolved. : Lymphangiogram consistent with carcinomatous involvement of the regional and para-aortic nodes. : Pathologic adenopathy of other than regional or para-aortic nodes, as demonstrated by lymphangiography, chest x-ray, or physical examination; e.g., supraclavicular, mediastinal. (If biopsy is not performed and is negative, disregard this category.) BW —- Biopsy: : Biopsy not performed or technically inadequate. : No evidence of tumor in regional and para-aortic nodes. : Regional nodes positive, para-aortic nodes negative for tumor. : Regional nodes negative, para-aortic nodes positive for tumor. : Regional and para-aortic nodes positive for tumor. : Any of the aforementioned plus positive biopsy of nodal site outside regional or para-aortic nodes. (If not biopsied, place x.) oan gs x M sites: Distant metastases (excluding lymph nodes) MO: No evidence of metastases. MI: Isolated area of abnormal epithelial cells in bone marrow unassociated with multiple bony, soft tissue, or visceral metastases. M2: Bony, visceral, or soft tissue metastases with or without bone marrow involvement. Example: The stage of a patient who has 25% of one lobe involved with tumor and invasion of the right seminal vesicle, who had a lymphangiogram that shows normal regional and para-aortic nodes, and a biopsy that reveals tumor in the regional nodes and negative para-aortic nodes but evidence of metastases would be 3aNObMO. “ Patients with intravenous pyelogram findings suggestive of bladder involvement will undergo cystoscopy and biopsy if indicated. show that nearly all of these methods can be augmented further by the use of radiosensitizers. Radiation Sequelae Mild symptoms of urinary frequency or urgency, nocturia, or diarrhea occur in about 50% of the patients during the course of irradiation. Symptoms that persisted for more than 2 years or severe transient symptoms that required surgical intervention are tabulated in table 5. The analysis of sequelae was complicated by two is- sues. Between 1956 and 1985, of the 893 patients who were treated and analyzed for complications, 91 were par- MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER ticipants in a protocol study that required transperitoneal biopsy of the pelvic and para-aortic lymph nodes as a stag- ing procedure. This yielded valuable information on the in- cidence of para-aortic lymph node metastases; however, it became apparent that the transperitoneal approach to the para-aortic lymph nodes, combined with para-aortic radi- ation therapy (which in some patients amounted to 5,500 rad), caused excessive morbidity. For example, a substan- tially higher incidence of small bowel damage requiring surgical correction occurred in this selected cohort (19). As a result, the transperitoneal approach was abandoned in favor of retroperitoneal limited node sampling in appropri- ate cases, and the para-aortic radiation dose was reduced 54 (1) TO focal - SURVIVAL (20) (2 Toditfuse - SURVIVAL (49) (3) TO focal - DISEASE-SPECIFIC SURVIVAL (20) (@ To diffuse - DISEASE-SPECIFIC SURVIVAL (49) reer eee Gehan P values 1vs.2 .1809 3vs.4 7621 PROBABILITY (%) heehee 2 4 6 : 10 12 14 % 3 2 = 24 26 28 TIME (YEARS) FIGURE 6.—Survival and disease-specific survival, i.e., death due to prostate cancer only (node status unknown), for incidental TOf and TOd (Al and A2) carcinoma of the prostate. (1) TO - SURVIVAL (69) (2) TO - DISEASE-SPECIFIC SURVIVAL (69) rrr eee 40 += -r PROBABILITY (%) or ® ) t t 2 a 6 8 10 2 14 16 18 20 22 24 26 28 TIME (YEARS) FIGURE 7.—Survival and disease-specific survival for TO, incidental, car- cinoma of the prostate (node status unknown) after coalescing the sta- tistically similar curves for focal and diffuse disease. ttf ibiiftitit tid ded tdi de Ca i fe pi fl + + = + + + + t t + t + © Stage TO (69) (1) Stage T1 (282) reer ® Stage T2 (183) @ Stage T4 (32) (3) Stage T3 (348) (5) Expected (915) tr rrr Gehan P values Ovs.1 0112 [ Ovs.2 SURVIVAL (%) 2 4 6 8 10 12 14 16 18 20 2 2 % TIME (YEARS) FIGURES 8-14.—Prostate cancer, node status unknown. FIGURE 8.—Survival. A downward step represents death due to any cause. An upward tick represents a patient surviving at last follow-up and censored at the time indicated by the abscissa. Patient may or may not have residual or metastatic cancer. (2) stage T2 (183) (@) Stage T4 (32) (3) Stage T3 (348) (0) Stage TO (69) (1) stage T1 (282) rrr eee eee = Gehan P values » Ovs.1 0106 | = 100 4 Ovs.2 0004 | Ovs.3 | 2 » o i= > Ae wnwawace ow aa 2 3 “0000 z wT 3 0001 | 2 .4 .0000 ow .4 0000 Oo 60 | i Ln Oo lu t a0 | , » \ @ ] & ey, ® 5 wo] » L [=} aad boc be boas ben bea baa beac bona bee bean boas faa faa 20 22 24 26 28 2 4 6 8 10 12 14 16 18 TIME (YEARS) FIGURE 9.—Disease-specific survival. A downward step indicates death due to prostate cancer. An upward tick represents a patient who 1) died due to intercurrent disease, or 2) was alive at the time of last follow-up; patient may or may not have cancer. to 5,000 rad in 5 weeks. In table 5, the radiation seque- lae are presented in 2 groups: 802 patients who did not have transperitoneal lymph node biopsy, and 91 patients with excessive morbidity on whom a celiotomy had been performed. Note the striking difference in the complication rate between the 2 groups, particularly with respect to in- testinal damage and lower extremity or genital edema. The other factor that confounded the true incidence of radia- tion sequelae in these patients was the frequent inability of investigators to determine whether a given severe, persis- tent abnormality was due to persistent tumor or to radiation effect. For instance, rectal obstruction in several instances has appeared to be caused by extensive local tumor growth. However, we made no attempt to distinguish between the symptoms caused by tumor and those caused by the treat- (© Stage TO (69) (1) Stage T1 (282) (2) stage T2 (183) (3) Stage T3 (348) (@) stage Ta (32) 4 + figs 4 4 4 4 4 + 4 rrr errr errr brererrerebrerrrrfeer ery Gehan P values Ovs. 1 Ovs.2 Ovs.3 Ovs.4 Tvs. 2 Egegasaess FREEDOM FROM RELAPSE (%) feed chi ii bin sade bea bea baa baa 2 4 6 8 10 12 14 16 18 20 22 24 26 28 TIME (YEARS) FIGURE 10.—Freedom from relapse. A downward step indicates the first evidence of recurrence, either at the primary site or at a metastatic site as detected by either clinical observation or by a positive biopsy. An upward tick represents a patient who was either observed disease free or died disease free at the time of last observation. NCI MONOGRAPHS, NUMBER 7, 1988 (2) Stage T2 (183) (@ stage T4 (32) (3) Stage T3 (348) tree (© stage TO (69) (@) Stage T1 (282) REESE REESE EE rr reer Gehan P values © Ovs.1 .0016 Ovs.2 .0000 Ovs.3 .0000 | Ovs.4 .0000 1 1vs.2 0081 + 1vs.3 .0002 J @® 1vs.a 0028 | 2vs.3 3957 2vs. 4 0649 40 + = CLINICAL LOCAL CONTROL (%) sada aaa aaa aaa baa aba bens baa aaa aaa aaa 2 4 6 8 10 12 14 16 18 20 22 24 26 28 TIME (YEARS) FIGURE 11.—Local control. A downward step indicates clinical evidence of local regrowth after initial regression of tumor or after initially showing no evidence of local neoplasm. An upward tick represents a patient who demonstrated 1) no clinical evidence of local tumor at last follow-up, or 2) died without clinical evidence of local neoplasm. Patient could have had evidence of metastatic tumor either while living or at death. @ Tia (a1) (® Tb (100) © Tic (56) @ Tid (85) rere t Gehan P values avs.b .0822 avs.c .0019 avs.d .0000 SURVIVAL (%) alae bese boca toad va brea daa beau bay a Lh t t + + + + + + + + + 2 a 6 8 10 12 14 16 18 20 22 24 26 > TIME (YEARS) FIGURE 12.—A family of survival curves for T1 (intracapsular) carcinoma of the prostate is demonstrated. (See table 2 for precise definitions of the substages.) @ TiaorTib (141) @ TicorTid (141) (3) Expected Survival: Stage 1 (282) rrr rere eee Gehan P value 1vs.2 .0000 SURVIVAL (%) abana bead bera dena be a bene be en beepers bea 2 4 6 8 10 12 14 16 18 20 22 24 26 28 TIME (YEARS) FIGURE 13.—Statistically similar subgroups of stage T1 (intracapsular) carcinoma of the prostate are coalesced. MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 55 @ TiaorTib (181) @ TicorTid (141) reer errr r Gehan P value 1vs.2 .0003 fine DISEASE-SPECIFIC SURVIVAL (%) cea oa bean dee beac loca dual aaie La aa bua fois bai + + + + t 1 + t + + t 2 4 6 8 10 12 14 16 18 20 22 24 26 28 TIME (YEARS) FIGURE 14.—Disease-specific survival for the T1 lesions. ment, and all sequelae are enumerated whether due to ra- diation exposure or persistent neoplasm. In table 6 are the results of the nonprotocol patients who were examined with reference to the period during which the radiotherapy was administered. It has been the impres- sion that the incidence and severity of radiation sequelae have declined in recent years. This reduction correlates well with, among other things, the introduction in 1974 of the midcourse prostatic radiation boost, first suggested by S. Weller (personal communication). The table demonstrates that from January 1965 through December 1974, 431 pa- tients were treated, and the incidence and severity of radi- ation sequelae during that period were compared with the same factors observed for the 289 patients treated between January 1975 and December 1984. Inspection of table 6 TABLE 3.—Results of biopsy according to stage of primary tumor? Positive biopsy Stage No. No. Percent A2 1 0 0 Bl 2 0 0 Small B2 8 3 38 Large B2 22 13 59 C 31 23 74 Total 64 39 61 4 Chi-square = 4.732; P = .030. From 146 staged patients on whom laparotomies were performed, we selected 64 ad hoc for transperineal prostatic biopsy 2 yr or longer following external-beam irradiation. TABLE 4.— Current status of 64 patients Positive Negative Status biopsy biopsy Alive without metastases or progression 114 19 Alive with metastases 8 3b Dead of progressive disease 18 3b Dead of other causes, with metastases 2 0 4 Of the 11 patients, 7 are on diethylstilbestrol for local control, and 2 have received interstitial irradiation. b Four of 6 patients had positive nodes at initial staging. 56 shows a statistically significant reduction in intestinal se- quelae from 6.5% to 2.4% of the patients and a reduction in urologic sequelae from 10.9% to 7.2% of the patients during the most recent decade. Status of Erectile Potency Following External-beam Irradiation During this study, of the 434 patients who claimed erec- tile potency prior to irradiation, 86.4% remained potent at TABLE 5.—Radiation sequelae in patients with prostate cancer, September 1956 to December 1984¢ FIGURE 15.—Frontal radiograph showing in- traprostatic alignment of trocars contain- ing iridium-192 that have been introduced through a perineal template. 15 months after the start of radiation therapy. Furthermore, although erectile potency diminished with advancing age, 50% of the patients remained potent at the seventh year fol- lowing irradiation, and more than 30% maintained sexual performance for the duration of their survival (fig. 16). Table 7 demonstrates that radiation therapy for prostate cancer does not induce an increased incidence of second neoplasms. Although 10.8% of the patients developed a TABLE 6.—Radiation sequelae in nonprotocol patients with prostate cancer? Sequelae Sequelae Other Other Intestinal Urologi d Intestinal Urologi d Tope andor cause ntestina rologic (edema) or — ntestina rologic (edema) of symptoms No. Percent No. Percent No. Percent of symptoms No. Percent No. Percent No. Percent Nonprotocol patients Patient group? (802) Persistent 22 5.1 33 7.7 7 1.6 Persistent > 2 yr 30 3.7 51 6.4 12 1.5 Transient/severe 6 1.4 14 3.2 1 0.2 Transient/severe 6 0.7 25 3.1 2 0.2 Patient group (surgical) Persistent 7 24 14 4.8 4 1.4 Protocol patients Transient/severe 0 7 24 1 0.3 (91; celiotomy) Persistent > 2 yr 13 14.3 9 9.9 13 14.3 Transient/severe 5 5.5 6 6.6 0 (surgical) 4 Numbers in parentheses are numbers of patients in the group. Chi- square = 37.677; two-tailed P = .000000. 4 Chi-square = 9.08; two-tailed P= .0026. Persistent refers to symptoms of >2-yr duration. Transient/severe refers to symptoms requiring hos- pitalization and surgical correction, e.g., TUR for obstructive symptoms. b From January 1965 through December 1974, 431 patients were treated. “ Between January 1975 and December 1984, 289 patients were treated. NCI MONOGRAPHS, NUMBER 7, 1988 434 Patients Potent at Start of Treatment 86.4% Potent 15 mo after Start of Treatment POTENCY (%) 20 - 1 sda bea bu bocce been bn born bea ben pea fees pen faa 2 4 6 8 10 12 14 16 18 20 22 24 26 28 TIME (YEARS) FIGURE 16.—Status of erectile potency after external-beam irradiation for prostate cancer (node status unknown). second neoplasm, neither the incidence nor the spectrum deviated from expectations. DISCUSSION A comparison of long-term survival of the irradiated Stanford patients with the results following surgery in 2 well-documented series of patients treated by radical prostatectomy is illustrated in figure 17. Overall survival in the patients treated surgically at The Johns Hopkins Hos- pital (20) and the Virginia Mason Medical Center (21), and the patients in stages T1a and T1b treated at Stanford University Hospital did not deviate significantly from each other through 15 years of follow-up (fig. 17). Of course, there were differences in case selection as is inevitable when independent series are developed at different institutions during different times. Time-to-first-evident metastasis in surgically staged pa- tients, randomly allocated to either radical prostatectomy or definitive x-irradiation, was tested in the mid-1970s and reported by the Uro-Oncology Research Group in 1982 (22). Coincidentally, a group of patients at Stanford was TABLE 7.—Incidence of second cancers following prostate cancer? Type of second cancer Observed Expected P Lung 28 29.3 81 Colon 15 15.2 96 Bladder 11 10.2 .80 Stomach 6 54 .80 Rectum 4 7.4 21 Leukemia 4 3.8 92 Lymphoma 3 4.1 65 Pancreas 3 4.6 27 Kidney 1 29 26 Other 24 _— — Total? 99 106.2 48 ¢ Observed second cancers vs. expected incidence was calculated by the method of Monson from the Surveillance, Epidemiology, and End Results Program data for Bay Area white males. There were 914 patients at risk for more than 7,000 person-yr. b Observed second cancers occurred in 10.8% of the prostate cancer patients. MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 57 100 75 50 25 — — OVERALL SURVIVAL (percent) | | | 0 5 10 15 YEARS AFTER TREATMENT FIGURE 17.—Comparison of long-term survival of patients in stage T2a (B1) treated by radiation therapy in the Stanford University Hospital series (134 patients; open triangles) and those treated by radical prosta- tectomy in the series at The Johns Hopkins Hospital (57 patients; open circles) and Virginia Mason Medical Center (195 patients; solid circles) according to a personal communication (Shipley WU). evaluated according to a nearly identical protocol, and all 51 patients clinically staged as A2 or B who had no lym- phadenopathy as proved by surgical staging were treated with external-beam irradiation exactly as prescribed for the Veterans Administration radiotherapy group. The data from the Uro-Oncology Research Group (22) of the Veterans Administration and those from the Stanford study are plot- ted in figure 18. The pattern of treatment failure in the irradiated Stanford patients was nearly identical to that of the surgically resected patients and significantly better than that of those treated by irradiation in the Veterans Adminis- tration study. The 51 patients in the Stanford group are un- der continuing follow-up, and their recent status is demon- strated in figure 19, which shows a 60% survival at 13 0 P =.037110 Qo Z = g ® E sot & @® RADIATION THERAPY: STANFORD SERIES Q (n= 51) ® RADICAL SURGERY: VA SERIES o (n = 41) ® RADIATION THERAPY: VA SERIES (n = 56) 100 r T + r , - , ) pr—— 0 60 MONTHS FIGURE 18.—Time-to-first evidence of treatment failure. Curve 1 is gen- erated from the Stanford data and is superimposed [with the author’s (22) permission] on the original figure from the Uro-Oncology Research Group study. The P value refers to curves 2 and 3 only. 58 5 ~T T T 1 1 100 |— - ig DISEASE - bre Lua) SPECIFIC SURVIVAL > 80} LTP - E Ls 2 | @ i = 1 3 60 | aaa - 2 SURVIVAL a = 5 4 a0 | - oc w a 20 | _ 0 1 1 1 1 1 1 0 2 a 8 10 12 14 YEARS 16 FIGURE 19.—Long-term, disease-specific, and overall survival of 51 pa- tients in stages A2 and B prostate cancer who had no histopatholog- ically positive lymph nodes at surgical staging (T0d, T1, T2, NO, and MO). These patients were staged between 1970 and 1978 and are under constant follow-up. years and a disease-specific survival of 75% for the same interval. From the above, expectation of a long-term difference in survival statistics between stages A and Bl patients treated by either radical resection or external-beam irradiation ap- pears to have little precedent. What about more advanced carcinoma? The inaccuracy of clinical staging before sur- gical resection has long been a concern of thoughtful sur- geons, but the impact of the true pathologic stage not being recognized prior to surgical resection has not been equally well appreciated. The rather complex data detailed in table 8 summarize reports that contain information on clinical understaging and, in some cases, its impact on postoper- ative survival (/9,23-34). In each instance in which the impact on survival was stated, the long-term survival was reduced by more than 50% in patients who had a patho- logically proven incomplete surgical resection. Several authors (34-36) have demonstrated that residual tumor at the surgical margins appears amenable to steril- ization by postprostatectomy irradiation. For example, in the Stanford series, a 57% long-term survival was achieved TABLE 8.—Prostate cancer preoperative understaging® Pathologic Impact on survival at: capsular or Study extracapsular Seminal Decrease Principal Clinical stage disease asic No. of in percent Year investigator Reference Stage No. No. Percent invasion, % years From To 1953 Colby (23) 25 71 36 5 52 27 1957 Turner (24) 45 — — 1963 Vickery (25) 47 — 1968 Culp (26) Bl 44 8 — (exceeded expectation) 1972 Byar (27) 11 12 5 58 33 1977 Dahl (28) 22 _— — 1977 deVere White (29) 29 on we 1977 Boxer (30) 29 — 10 62 29 1980 Walsh (20) Bl 16 B2 49 15 51 5 1982 Elder (31) B2 66° 13 (minimum) 15 50 13 1982 Catalona (32) A2¢ 11° — Bl 17° —- B2 39% - d Overall 24 _— No data 1982 Middleton (33) B2¢ 12 4 — 1983 Lange (34) Af 6 3 3 BI 32 2s 68 5 8 B2 25 15 9 1984 Gibbons (2D) B 3s — -_ ¢ Carcinoma was considered confined to the prostate prior to radical prostatectomy. b Patients had extracapsular disease. ¢ Five patients had an early failure. 4 Follow-up was short. ¢ Clinical stage was node negative. Of the 6 patients with this clinical stage, 3 had extracapsular disease, and the impact on survival was not stated. & Patients received x-ray therapy. NCI MONOGRAPHS, NUMBER 7, 1988 in 13 patients who were irradiated immediately follow- ing inadequate resection, whereas in 19 patients, a 20% long-term survival was achieved following definitive irra- diation in the face of frank recurrence after prior prostatec- tomy. The suggestion by Stamey and co-workers (37) that the prostate-specific antigen may be a specific and sensi- tive marker for the determination of residual neoplasm or recurrence after prostatectomy is worthy of careful study because it might clearly identify a group of patients who would benefit from postoperative irradiation. SUMMARY AND CONCLUSIONS The interpretation of end results following various treat- ments for prostate cancer remains a confusing issue. Part of this is due to the lack of wide acceptance of a universal staging system, and part is due to a lack of definition and clarity on the selection of end-point parameters and their statistical presentation. In this paper, we have attempted to call attention to these deficiencies and to present clear alter- natives. We hope that clarification of staging and end-point reporting will become the subject of continuing effort. An actuarial survival (scoring death due to all causes) of 50% at 15 years has been achieved following external-beam irradiation in patients with early prostate cancer, i.e., TO, Tla, and T1b (stages A, A2, and Bl). This is equal to the expected survival of an age-matched population and also equal to the survival rates reported in earlier surgical series of equal longevity. Thus surgical resection and radiation therapy as related to survival appear to be equal alterna- tives in early disease. Perhaps reasonable physicians should routinely disclose this fact to the beleaguered patients. As stage and, consequently, tumor mass escalate, long- term survival following irradiation decreases; however, it still remains at 20% for T3 (stage C), and even at 10% for patients with massive T4 (stage D) local disease. Improve- ment in radiation therapy will require continuing develop- ment of adjunctive techniques for improving the efficacy of treatment. Methods, such as boosting the tumor dose with radioactive implants or high-energy particles or the addi- tion of localized hyperthermia, hold promise for improving the sterilization of more advanced neoplasms. All of these methods could be enhanced by the development of an effi- cacious radiation sensitizer. A more accurate pretherapeutic assessment of the precise extent of the primary neoplasm by an advanced imaging modality, such as ultrasound, magnetic resonance imaging, or spectroscopy, would be extremely valuable in matching patients to the appropriate treatment strategy. Surgical transection of tumor that leaves residual neo- plasm appears to be identifiable by an elevation of prostate-specific antigen titer. This could trigger the appro- priate use of postoperative radiation therapy, which appears to be a safe and effective adjunct to surgery in this situa- tion. A clinical trial that would establish this diagnostic and therapeutic combination seems warranted. REFERENCES (1) PEREZ CA, WALZ BJ, ZIVNUSKA FR, ET AL. Irradiation of carcinoma of the prostate localized to the pelvis: Analysis MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 59 of tumor response and programs. Int J Radiat Oncol Biol Phys 1980;6:555-563. (2) PEREZ CA, PILEPICH MV, ZIVNUSKA F. Tumor control in definitive irradiation of localized carcinoma of the prostate. Int J Radiat Oncol Biol Phys 1986;12:523-531. (3) McGowaN DG. Radiation therapy in the management of localized carcinoma of the prostate. Cancer 1977;39:98- 103. (4) TAYLOR WF, RICHARDSON RG, HAFERMANN MD. Ra- diation therapy for localized prostate cancer. Cancer 1979;43:1123-1127. (5) LEBEL SA, HANKS GE, KRAMER S. Patterns of care outcome studies: Results of the national practice in adenocarcinoma of the prostate. Int J Radiat Oncol Biol Phys 1984;10:401- 400. (6) HANKS GE. Optimizing the radiation treatment and out- come of prostate cancer. Int J Radiat Oncol Biol Phys 1985;11:1235-1245. (7) PILEPICH MV, KRALL JM, SAUSE WT, ET AL. Prognostic factors in carcinoma of the prostate—analysis of RTOG study 75-06. Int J Radiat Oncol Biol Phys 1987;13:339- 349. (8) RAY GR, CassaDY JR, BAGSHAW MA. Definitive radiation therapy of carcinoma of the prostate. A report on 15 years of experience. Radiology 1973;106:407-418. (9) BAGSHAW MA. External radiation therapy of carcinoma of the prostate. Cancer 1980;45:1912-1921. (10) BAGSHAW MA. Current conflicts in the management of prostatic cancer. Int J Radiat Oncol Biol Phys 1986;12: 1721-1727. (11) PISTENMAA DA, BAGSHAW MA, FRrREIHA FS. Extended-field radiation therapy for prostatic adenocarcinoma: Status report of a limited prospective trial. In Cancer of the Genitourinary Tract (Johnson DE, Samuels ML, eds). New York: Raven Press, 1979, pp 229-247. (12) INTERNATIONAL UNION AGAINST CANCER. TNM classifica- tion of malignant tumors. 2nd ed. Geneva: UICC, 1974. (13) MurPHY GP, GAETA JF, PICKREN J, ET AL. Current status of classification and staging of prostate cancer. Cancer 1980;45(suppl): 1889-1895. (14) JEWETT HJ, BRIDGE RW, GRAY GR JR, ET AL. The palpable nodule of prostatic cancer: Results 15 years after radical excision. JAMA 1968;203:403-406. (15) KAPLAN EL, MEIER P. Nonparametric estimation from in- complete observations. J Am Stat Assoc 1958;53:457- 480. (16) GEHAN EA. A generalized Wilcoxon test for comparing arbi- trarily singly-censored samples. Biometrika 1965;52:203- 223. (17) McNEAL JE, KINDRACHUK RA, FREIHA FS, ET AL. Patterns of progression in prostate cancer. Lancet 1986;1:60-63. (18) SYED AMN, PUTHAWALA A, TANSEY LA, ET AL. Manage- ment of prostate carcinoma. Radiology 1983;149:929- 033. (19) FrEIHA FS, SALZMAN J. Surgical staging of prostatic cancer: Transperitoneal versus extraperitoneal lymphadenectomy. J Urol 1977;118:616-617. (20) WALSH PC, JEWETT HJ. Radical surgery for prostatic cancer. Cancer 1980;45:1906-1911. (21) GiBBONS RP, CORREA RJ Jr, BRANNEN GE, ET AL. To- tal prostatectomy for localized prostatic cancer. J Urol 1984;131:73-76. (22) PauLsON DF, LIN GH, HINSHAW W, ET AL. Radical surgery versus radiotherapy for adenocarcinoma of the prostate. J Urol 1982;128:502-504. (23) CoLBy FJ. Carcinoma of the prostate: Results of total prosta- tectomy. J Urol 1953;69:797-806. 60 (24) TurNER RD, BELT E. A study of 229 consecutive cases of total perineal prostatectomy for cancer of the prostate. J Urol 1957;77:62-77. (25) VICKERY AL JR, KERR WS JR. Carcinoma of the prostate treated by radical prostatectomy. A clinicopathological survey of 187 cases followed for 5S years and 148 cases followed for 10 years. Cancer 1963;16:1598-1608. (26) CuLp OS. Radical perineal prostatectomy: Its past, present and possible future. J Urol 1968;98:618-626. (27) BYAR DP, MosToFI FK. Carcinoma of the prostate: Prognos- tic evaluation of certain pathologic features in 208 radical prostatectomies. Cancer 1972;30:5-13. (28) DAHL DS, WILSON CS, MIDDLETON RG, ET AL. Pelvic lym- phadenectomy for staging localized prostatic cancer. J Urol 1974;112:245-246. (29) DEVERE WHITE R, PAULSON DF, GLENN JF. The clinical spectrum of prostate cancer. J Urol 1977;117:323-327. (30) Boxer RJ, KAUFMAN JJ, GooDWIN WE. Radical prostatec- tomy for carcinoma of the prostate: 1951-1976. A review of 329 patients. J Urol 1977;117:208-213. (31) ELDER JS, JEWETT HJ, WALSH PC. Radical perineal prosta- tectomy for clinical stage B2 carcinoma of the prostate. J Urol 1982;127:704-706. (32) CATALONA WIJ, FLEISCHMANN J, MENON M. Pelvic lymph node status as predictor of extracapsular tumor extension in clinical stage B prostatic cancer. J Urol 1983;129:327- 329. (33) MIDDLETON RG, SMITH JA JR. Radical prostatectomy for stage B2 prostatic cancer. J Urol 1982;127:702-703. (34) LANGE PH, NARAYAN P. Understanding and undergrading of prostate cancer. Urology 1983;21:113-118. (35) RAY GR, BAGSHAW MA, FRrREIHA F. External beam radiation salvage for residual or recurrent local tumor following radical prostatectomy. J Urol 1984;132:926-930. (36) HANKS GE, DAwsON AK. The role of external beam ra- diation therapy after prostatectomy for prostate cancer. Cancer 1986;58:2406-2410. (37) STAMEY TA, YANG N, HAY AR, ET AL. Prostate-specific antigen as a serum marker for adenocarcinoma of the prostate. N Engl J Med 1987;317:909-916. Radiation Therapy Oncology Group Studies in Carcinoma of the Prostate Miljenko V. Pilepich!* ABSTRACT —From 1976 to 1983, the Radiation Therapy On- cology Group (RTOG) conducted 2 large-scale phase III trials of extended field irradiation in patients with carcinoma of the prostate. The first, RTOG 75-06, was designed to test the value of elective periaortic irradiation in patients in whom the tumor extended beyond the gland, but remained limited to the pelvis, and the second, RTOG 77-06, was designed to test the value of elective pelvic irradiation in patients without evidence of spread beyond the prostate. The results indicated no apparent benefit from elective periaortic irradiation in patients with detectable dis- ease confined to the pelvis and no apparent benefit from elective pelvic irradiation in patients with detectable disease confined to the prostate. Patients with extracapsular extension of the primary tumor and evidence of pelvic lymph node involvement demon- strated an outcome comparable to that in patients without evi- dence of lymphatic involvement. This observation may reflect a beneficial effect of pelvic irradiation in patients with nodal in- volvement. In contradistinction to elective irradiation of regional lymphatics, therapeutic irradiation (of the involved lymphatics) may prove strongly indicated. A prospective study testing this contention needs to be conducted. No significant correlation of treatment-related morbidity and treatment volume could be iden- tified. Analysis of the various types of treatment-related morbid- ity as to the time of onset and clinical course indicated that these behave as different disease entities characterized by a specific pattern of appearance, clinical course, and prognosis. Of par- ticular interest is the observation that most appear reversible. Doses in excess of 7,000 cGy to the prostate were associated with a significantly increased incidence of bowel morbidity. Certain treatment techniques were also associated with a significantly in- creased incidence of treatment-related problems. Although pa- tients who received concomitant hormonal management had a significantly higher proportion of high-grade lesions, their clini- cal course fared favorably in comparison with the population not receiving hormonal management. This apparent beneficial effect of adjuvant hormonal treatment needs to be tested in a prospec- tive study.—NCI Monogr 7:61-65, 1988. Since 1976, the RTOG has conducted a series of stud- ies on carcinoma of the prostate. Between 1976 and 1983, emphasis was placed on 2 large-scale phase III trials aimed at the definition of an optimal treatment volume in pa- tients with locally extensive disease (RTOG 75-06) and disease limited to the prostate (RTOG 77-06). Starting in 1981, a series of studies aimed at evaluation of adjuvant modalities was initiated. These included a study of adju- vant chemotherapy (RTOG 81-12) and a series of studies ABBREVIATION: RTOG = Radiation Therapy Oncology Group. ! Radiation Oncology Center, Washington University School of Medicine, St. Louis, MO. * Reprint requests to: Miljenko V. Pilepich, M.D., Department of Radi- ation Oncology, Catherine McAuley Health Center, P.O. Box 995, Ann Arbor, MI 48106. aimed at evaluation of adjuvant endocrine management. These include phase II studies of hormonal cytoreduction before and during a course of radiotherapy (RTOG 83-07 and 85-19) and a phase III study of hormonal cytoreduction (RTOG 86-10). The data accumulated on the first 2 studies (75-06 and 77-06) have matured sufficiently to provide a base for a series of analyses that will be summarized in this commu- nication. STUDY DESIGN AND POPULATION The primary aim of those who designed the RTOG 75-06 was to test the value of periaortic irradiation in patients in whom there was evidence of tumor extension beyond the prostate, but limited to the pelvis. Patients with clinical stage C tumor were eligible whether or not pelvic lymph nodes were involved. For these patients, assessment of nodal status was optional. Patients with stage A2 and B were also eligible but only if they had either lymphangiographically or histologically confirmed pelvic lymph node involvement. The stratification criteria included histologic grade, clin- ical stage, absence or presence of hormonal manipulation, and method of lymph node evaluation (lymphangiogram vs. laparotomy vs. no nodal evaluation). The patients were ran- domized to receive either pelvic irradiation followed by a boost to the prostate, or pelvic and periaortic irradiation fol- lowed by a boost to the prostate. The prescribed daily dose was 180-200 cGy to a total midplane dose to the regional lymphatics of 4,000-4,500 cGy. The prostatic boost tar- get volume was to receive an additional 2,000-2,500 cGy, bringing the total dose to the area to a minimum of 6,500 cGy. The end points of the study included an assessment of the effect of treatment arm assignment on disease-free survival, survival, and treatment-related morbidity. The primary aim of those who designed RTOG 77-06 was to test the value of elective pelvic irradiation in pa- tients without evidence of spread beyond the prostate. Eli- gible patients were those with clinical stages B and A2 without evidence of pelvic lymph node involvement by ei- ther lymphangiogram or laparotomy. Evaluation of the re- gional lymphatics was mandatory. Stratification criteria in- cluded histologic grade, absence or presence of hormonal manipulation, and method of nodal evaluation (laparotomy vs. lymphangiogram). The treatment entailed either irradi- ation of the pelvis followed by a prostate boost or prostatic irradiation only. The prostate was to receive a minimum of 6,500 cGy in 6% to 7 weeks. In patients randomized to re- ceive pelvic irradiation, the pelvis was to receive 4,500 to 5,000 cGy in 4% to 5 weeks. The daily tumor dose was 180 to 200 cGy. 61 62 The end points of the study included an assessment of the effect of pelvic irradiation on disease-free survival, survival, and treatment-related morbidity. RTOG-75-06.—From September 1976 to July 1983, 607 patients were accessioned to the study, and a total of 566 were analyzable. Of these, 500 had stage C, 63 stage B, and 3 stage A2 tumors. For the purposes of this and other RTOG studies, clinical stage is defined by the American Urological Staging System. Assignment of stage was based purely on clinical (physical) examination. Patients with evidence of nodal involvement were (for the purposes of this and other RTOG studies) designated as stages A, B, and C with lymph node involvement rather than stage D1. RTOG 77-06.—From March 1978 to August 1983, 484 patients were entered on this study. Of the 444 who were considered analyzable, 360 had stage B and 84 had stage A2 disease. REVIEW OF DATA ANALYSES A series of analyses of RTOG 75-06 and 77-06 data has been undertaken since 1982 and will be reviewed here in some detail. Early analyses included only part of the study population. Treatment-related Morbidity All of the reported cases of treatment-related morbid- ity were recorded and classified with the use of the clini- cal severity grading schema shown in table 1. Commonly observed reactions to irradiation, such as diarrhea, were not registered as treatment-related morbidity (i.e., compli- cations) unless they lasted beyond the first postirradiation month or unless they were grade 3 or higher. An attempt was made to separate various forms of treatment-related morbidity into specific disease entities and to define their incidence, the time of occurrence, and relation to treatment and their prognoses (1,2). For the purposes of the study, a number of clinical syn- dromes, thought to represent specific types of normal tis- sue injury have been defined. For example, cystitis is de- fined as irritative bladder symptoms (dysuria, frequency) with or without hematuria. Proctitis is defined as irritative rectal symptoms characterized by tenesmus and/or mucus discharge with or without melena and with or without di- arrhea. Diarrhea is characterized by loose, frequent bowel movements without irritative rectal symptoms. This syn- TABLE 1.—Grading of severity of complications Grade Complications 1 Minor symptoms requiring no treatment 2 Response to simple outpatient management, life-style (performance status) not affected 3 Distressing symptoms, altering of patient's life-style (performance status), possible hospitalization for diagnosis or minor surgical intervention (such as urethral dilatation) required 4 Major surgical intervention (such as laparotomy, colostomy, cystectomy) or prolonged hospitalization required 5 Fatal TABLE 2.—Summary of treatment-related morbidity? Grades Total Treatment-related morbidity 1 2 3 4 5 No. Percent Cystitis 30 28 8 66 12.5 Diarrhea 20 30 1 51 9.7 Proctitis 14 20 6 1 41 7.8 Melena 14 6 2 1 23 44 Hematuria 4 4 8 16 3.0 Urethral stricture 12 12 23 Rectal-anal stricture 4 2 1 7 1.3 Vesical neck contracture 2 1 3 0.6 Rectal ulcer 1 2 3 0.6 Bowel obstruction 2 2 0.4 ¢ In studies RTOG 75-06 and 77-06, 526 patients received treatment. drome is thought to result from the effect of irradiation on the segments of bowel proximal to the rectum. Melena is defined as any bleeding without associated proctitis symp- toms. A summary of treatment-related morbidity in 526 pa- tients accessioned prior to 1981 is shown in table 2. It is apparent that an overwhelming majority of cases of treatment-related morbidity belong to grade category 1 and 2 and did not, by definition, affect the patient’s performance status. Only a small proportion of cases (less than 1%) be- long to grade category 4. The pattern of occurrence varied considerably. Certain types of treatment-related morbidity occurred exclusively during the first year, whereas others showed an unpre- dictable pattern of occurrence (2). Various types of complications also differed considerably as to their prognosis. In most cases, the symptomatology proved reversible. Figure 1 shows the pattern of resolution of proctitis. Over 70% of the patients who developed proc- titis became asymptomatic by the end of the second year. A similar pattern of resolution was observed for diarrhea (fig. 2). The pattern of resolution of cystitis symptoms is shown in figure 3. Virtually all patients became asymptomatic by the end of the second year after occurrence. 23 ~N ® Oo O % PROBABILITY OF PERSISTENCE 05388388 1 1 i 1 1 1 Renmll—— 0 3 6 9 1215 18 21 24 27 30 33 36 MONTHS (FROM ONSET) FIGURE 1.—Pattern of resolution of proctitis symptoms. All patients who developed proctitis at any time are plotted. Time of resolution is mea- sured from onset of symptoms. Majority of patients with proctitis be- came asymptomatic within 2 yr after onset. NCI MONOGRAPHS, NUMBER 7, 1988 w € 100 Ww hr 9% ¢ 80 Ww 70 w 60 > 50 5 a0 2 3 8 20 © a 10 Fd 0 bleedin 0 3 6 9 1215 18 21 24 27 30 3B 36 MONTHS (FROM ONSET) FIGURE 2.—Pattern of resolution of diarrhea. All patients who developed the symptoms at any time are plotted. Time of resolution is measured from onset of symptoms. Majority of patients with diarrhea became asymptomatic within 2 yr after onset. The incidence of treatment-related morbidity was corre- lated extensively with a number of treatment parameters such as volumes, doses, and techniques (3,4). Extended fields were not associated with a significantly increased risk of complications. Periaortic irradiation compared with pelvic irradiation only was not associated with a signifi- cantly increased incidence of bowel injuries. Similarly, irra- diation to the pelvis did not result in an increased risk of in- testinal problems compared with irradiation to the prostate only. A detailed correlation of the incidence and severity of treatment-related morbidity and total doses did not indicate a significantly increased incidence of bowel or bladder injuries within the range of doses delivered to the regional lymphatics. The total doses to the prostate in excess of 7,000 cGy were not associated with a significantly increased incidence of bladder injuries but with an increased incidence of bowel injuries manifested by diarrhea and bleeding (fig. 4,5). Cor- relation with the treatment techniques revealed an increased incidence of bowel injuries in patients in whom pelvic ir- radiation consisted of parallel-opposed fields and also in those in whom prostatic boost consisted of opposed lateral fields or perineal fields (3,4). % PROBABILITY OF PERSISTENCE os B8838838883 1 1 1 1 1 1 A J 1 1 O03 6 9 12 15 1821 24 27 30 33 36 MONTHS (FROM ONSET) FIGURE 3.—Pattern of resolution of cystitis symptoms. All patients who developed symptoms of cystitis are plotted. Time of resolution is mea- sured from onset. Virtually all patients became asymptomatic within 2 yr after onset. MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 63 34/178 > 7000 16/141 3/65 11/132 6250-6500" 6501-6750 6751-7000 ~~ FIGURE 4.—Correlation of prostate (boost) dose and the incidence of diarrhea in RTOG 75-06. Figures above bars denote the number of cases of diarrhea per number of patients receiving the dose (6,250->7,000 cGy; P < .01 by Mantel-Haenszel test). The analysis has shown that within the range of doses and treatment volumes used in the 2 studies, neither the volumes nor doses correlated well with the normal tissues effects. A dose-response for the bowel was evident only at doses above 7,000 cGy. It could be concluded that the choice of treatment volumes and doses (up to 7,000 cGy to the prostate) should be based primarily on the anticipated therapeutic effectiveness rather than the normal tissue con- siderations. Treatment Arm Comparison Analyses of the effect of the treatment arm assignment on the course of the disease in patients accessioned to study 75-06 showed no statistically significant difference between the treatment arms (5). Moreover, no significant difference between treatment arms could be documented within a number of subpopulations characterized by grade, hormonal status, stage, age, or acid phosphatase status. The results indicated no apparent benefit from elective periaortic irradiation in patients with detectable disease confined to the pelvis. Analyses of treatment arm assignment in study 77-06 showed no statistically significant difference between treat- ment arms for the study population as a whole or within a number of subpopulations (Asbell SO, Krall JM, Pilepich MV, et al: Submitted for publication). In summary, the treatment comparison analyses in both studies indicated that elective irradiation of regional lym- phatics in carcinoma of the prostate may not be beneficial. Analysis of Prognostic Factors Extensive univariate and multivariate analyses of the factors of potential prognostic significance were also per- formed (6). These included tumor size, clinical stage, degree of histologic differentiation, nodal status, serum acid phosphatase status, hormonal management status, transurethral resection, and race. These factors were as- 6250-6500 6501-6750 6751-7000 FIGURE 5.—Correlation of prostate (boost) dose and the incidence of rectal bleeding in RTOG 77-06. Figures above bar denote the number of cases of rectal bleeding per number of patients receiving the dose (6,250->7,000 cGy; P < .01 by Mantel-Haenszel test). 64 sessed as to their interdependence and correlation with the clinical course (study end points). For the purposes of the study, tumor size is defined as a product of two perpendicular tumor dimensions in cen- timeters, as determined by physical examination on presen- tation. Hormonal management status (either administration of estrogens and/or orchiectomy) was broken down into two categories dependent on the date of the initiation in refer- ence to the first day of radiotherapy. If started within 60 days of the base date (first day of radiotherapy), the hor- monal management was labeled as concomitant; otherwise, it was labeled as antecedent. Univariate analysis identified the degree of histologic differentiation, age, and tumor size as factors of prognostic significance for local recurrence. Patients 60 years of age or younger were at a significantly higher risk of developing local recurrence. Patients with a tumor size characterized by a product of tumor dimensions of over 25 cm were also at a significantly higher risk of local recurrence (fig. 6). The factors predictive of distant metastases included tu- mor size, histologic differentiation, serum acid phosphatase, and hormonal management status. Patients who received hormonal management prior to consideration of radio- therapy (antecedent hormonal management) exhibited a higher rate of distal metastases than those who had no hormonal manipulation and those who received hormonal manipulation in conjunction with radiotherapy (concomi- tant hormonal management). Populations receiving either antecedent or concomitant hormonal manipulation had a significantly higher proportion of cases with high-grade le- sions, so that a higher rate of tumor progression could be anticipated. The observation that the incidence of distant metastases was the same in patients who received concomi- tant hormonal management as in those who did not may indicate a beneficial effect of endocrine manipulation when used as an adjuvant. This observation needs to be tested in a prospective study. Elevation of serum acid phosphatase and larger tumor size were also associated with a high incidence of distal metastases. Survival correlated well with the degree of histologic differentiation but did not correlate well with the serum acid phosphatase status and tumor size. ~ 0 PERCENT WITHOUT RECURRENCE on Oo <1.00 25 eeeenenen 1.01-9.00 ----= 901-2500 —— 2500+ 0 1 1 1 1 1 1 J oO 1 2 3 4 5 6 7 YEARS FIGURE 6.—Incidence of local-regional recurrence as a function of tumor size expressed as a product of tumor dimensions. In the multivariate analyses of prognostic factors, use of a Cox regression model singled out the degree of histologic differentiation (expressed in the form of Gleason score) as the most important prognostic factor for all assessed end points. Tumor size and patient’s age were also found predictive of local control. Serum acid phosphatase was found predictive of the incidence of distal metastases, but the degree of histologic differentiation proved to be the only factor predictive of survival. A surprising observation was the lack of prognostic sig- nificance of nodal status in patients with clinical stage C disease. Although patients with positive nodes, proved by either lymphangiogram or lymphadenectomy, had a signif- icantly higher proportion with high Gleason scores and el- evation of serum acid phosphatase, their outcome proved comparable to those of patients with no evidence of nodal involvement and to those in whom the regional lymphatics were not evaluated (7). These findings may indicate that irradiation of the pelvis exerted a beneficial effect, possi- bly by curing a proportion of patients with limited nodal involvement. This hypothesis may need to be tested in a prospective study assessing the effect of lymphatic irradia- tion in patients with known involvement of regional lym- phatics. In contradistinction to the above observation, assessment of regional lymphatics in patients with primary tumor lim- ited to the prostate (77-06 study population) proved valu- able. It allowed identification of a subpopulation of patients who would ordinarily be considered candidates for radical prostatectomy and have an exceedingly good prognosis. There were 104 such patients who had negative stag- ing lymphadenectomy, detectable disease limited to the prostate, and normal serum acid phosphatase. With the me- dian follow-up of 3 years and 9 months, only 5 patients de- veloped evidence of distal metastasis (8). The disease-free survival and survival in this population compares favorably with the best of the surgical series. The observation indi- cates that the observed outcome in patients treated with definitive radiotherapy is at least equivalent, if not supe- rior, to that in patients treated with radical prostatectomy. CONCLUSIONS Data accumulated in the 2 large-scale randomized phase III trials indicate that elective (prophylactic) irradiation of regional lymphatics in carcinoma of the prostate may not be beneficial. The term “elective” or “prophylactic irradiation” refers to irradiation of the periaortic area in patients with locally advanced disease confined to the pelvis and to pelvic irradiation in patients with known disease confined to the prostate. The studies were not designed to test the value of therapeutic irradiation, i.e., irradiation of the involved lymphatics. Each patient with known involvement of the pelvic lymphatics received pelvic irradiation. The outcome in patients with clinical stage C disease and involvement of pelvic regional lymphatics proved compara- ble to that in node-negative patients. This observation may indicate a beneficial effect of therapeutic lymphatic irradi- ation. This contention must be tested in a prospective trial. Irradiation of the regional lymphatics does not result in a significantly increased incidence of treatment-related morbidity compared with prostate irradiation only. NCI MONOGRAPHS, NUMBER 7, 1988 The treatment-related morbidity has been qualified and quantified in great detail. The incidence of serious treatment-related complications proved to be low. Most of the treatment-related symptomatology proved reversible. Within the range of doses used in the RTOG studies, no dose-response could be identified for doses ranging from 6,500 to 7,000 cGy. Doses in excess of 7,000 cGy seem to be associated with an increased risk of bowel morbidity. Local-regional control is satisfactory in the limited size primaries. Novel approaches are necessary in patients with bulky primaries. In populations characterized with disease clinically lim- ited to the capsule and negative lymphadenectomy, the out- come appears at least comparable, if not superior, to that in surgically treated patients. REFERENCES (1) PILEPICH MV, PAJAK T, GEORGE FW, ET AL. Preliminary re- port on phase III RTOG studies of extended field irradia- tion of carcinoma of the prostate. Am J Clin Oncol 1983;6: 485-491. (2) PILEPICH MV, KRALL JM, GEORGE FW, ET AL. Treatment MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 65 related morbidity in phase III RTOG studies of extended- field irradiation for carcinoma of the prostate. Int J Radiat Oncol Biol Phys 1984;10:1861-1867. (3) PiLEPICH MV, KRALL JM, SAUSE WT, ET AL. Correlation of radiotherapeutic parameters and treatment related morbid- ity in carcinoma of the prostate—analysis of RTOG study 75-06. Int J Radiat Oncol Biol Phys 1987;13:351-357. (4) PiLEPICH MV, ASBELL SO, KRALL TM, ET AL. Correlation of radiotherapeutic parameters and the treatment related morbidity—analysis of RTOG study 77-06. Int J Radiat Oncol Biol Phys 1987;13:1007-1012. (5) PiLEPICH MV, KRALL JM, JOHNSON RJ, ET AL. Extended field (periaortic) irradiation in carcinoma of the prostate— analysis of RTOG 75-06. Int J Radiat Oncol Biol Phys 1986;12:345-351. (6) PILEPICH MV, KRALL JM, SAUSE WT, ET AL. Prognostic fac- tors in carcinoma of the prostate—analysis of RTOG study 75-06. Int J Radiat Oncol Biol Phys 1987;13:330-349. (7) PiLEPICH MV, KRALL JM, SAUSE WT, ET AL. Prognostic sig- nificance of nodal involvement in locally advanced (stage C) carcinoma of the prostate—RTOG experience. In press. (8) PILEPICH MV, BAGSHAW MA, ASBELL SO, ET AL. Definitive radiotherapy in resectable (stage A2 and B) carcinoma of the prostate—results of a nationwide overview. Int J Radiat Oncol Biol Phys 1987;13:659-663. Radiation Therapy for Localized Prostate Carcinoma: Experience at the Massachusetts General Hospital (1973-1981) W. U. Shipley,” G. R. Prout, Jr., N. M. Coachman, P. L. McManus, E. A. Healey, A. F. Althausen, N. M. Heney, E. C. Parkhurst, H. H. Young II, J. W. Shipley, S. D. Kaufman! ABSTRACT —The success following irradiation in 370 patients with clinically localized prostate carcinoma was measured by overall patient survival as well as the cumulative incidence with time of treated patients who developed either local tumor re- growth or progression with distant metastases. With a mini- mum follow-up of 5 years in living patients, we evaluated the cumulative frequency curves using both univariate and multi- variate (Cox) analyses. Overall patient survival and probability of progression with distant metastases were significantly influ- enced by initial tumor stage and the degree of histologic dif- ferentiation. The results at 8 years are significantly better for patients with T2 (B) tumors (local regrowth in 8%, distant metas- tases in 18%) than for patients with T3-T4 (C) tumors (local re- growth in 28%, distant metastases in 60%). Patient tolerance of external-beam radiation therapy was carefully analyzed in 121 consecutively treated patients in 1980 and 1981 for subsequent radiation-related sequelae. Minor transient intestinal and uro- logic sequelae were observed in 21% and 23% of the patients, respectively. These mild to moderate symptoms resolved in all but 7% of the patients who are continuing with mild symptoms. One patient had a major complication, i.e., a cystectomy required for persistent bleeding. Erectile potency has been maintained in 63% of potent patients. No specific benefit or detriment in out- come was seen in the minority of 51 patients who were irradi- ated by iodine-125 implantation. We conclude from these results that, for patients with stage T2 tumors, the results with radia- tion therapy and surgery are good and are similar for at least 8-10 years of follow-up; thus patients should decide which treat- ment they would prefer after being fully informed. For patients with T3 and T4 tumors, the results are significantly poorer than for those with T2 tumors. Rigorous clinical research will be nec- essary in patients with T3 and T4 tumors to document possible benefit from either more aggressive local treatment or from adju- vant systemic therapy (such as androgen deprivation) against the undetected distant metastases, which are the major clinical prob- lem for men with tumors of this stage. —NCI Monogr 7:67-73, 1988. The success following local treatment (by either irra- diation or surgery) for patients with clinically localized prostate carcinoma may be evaluated by several methods. ! Radiation Medicine Service (W. U. Shipley, N. M. Coachman, P. L. McManus, E. A. Healey, J. W. Shipley), Urologic Surgery Service (G. R Prout, Jr., A. F. Althausen, N. M. Heney, E. C. Parkhurst, H. H. Young II), and Medical Oncology Service (S. D. Kaufman), Massachusetts General Hospital Cancer Center, Boston, MA. 2 We thank Ms. Anne Macaulay for assistance in preparation of the manuscript and Ms. Erica O. Moulton for assistance in statistical analysis. * Reprint requests to: William U. Shipley, M.D., Radiation Medicine Service, Massachusetts General Hospital Cancer Center, Cox Bldg. 3, Boston, MA 02114. These include overall survival, survival without evidence of tumor progression, or by separate evaluation of the curves of the cumulative incidence with time of treated patients who developed either local regrowth or distant metastasis. We have used the latter approach to evaluate the clinical patterns of failure following radiation therapy for patients with localized prostate carcinoma. For an analysis of the incidence of either local regrowth or the development of distant metastases, the assessment depends on both the du- ration of the observation period and the physician’s ability to discriminate clinically between local regrowth and other local changes, as well as between distant metastases and other abnormalities. The prognostic factors of tumor grade and clinical stage have been reported in many surgery and radiation therapy series to influence significantly overall survival and progression to distant metastasis (/-10). How- ever, few investigators have evaluated additional prognostic factors by multivariate analysis. This present retrospective analysis is of survival and of the cumulative frequency of the development of local tumor regrowth and of distant metastases observed in 370 patients irradiated at this in- stitution for prostate cancer. The possible independent and significant prognostic influence of six tumor and patient characteristics was evaluated by multivariate analysis. Ob- servation periods range from 5 to 12 years. PATIENTS AND METHODS This retrospective analysis is on ‘all 370 patients irradi- ated at the Massachusetts General Hospital Cancer Center from 1973 to 1981 for clinically localized and palpable car- cinoma of the prostate. We have grouped our patients into the 1983 American Joint Committee TNM Staging System (11). All patients at diagnosis were stage M0. One hundred thirty-five, or 36%, of the patients in this series underwent a pretreatment staging lymphadenectomy, and thus the ma- jority (235 of 370) at diagnosis were stage NX. The radia- tion therapeutic techniques have been described (6,12,13) for both the group of 319 patients treated by external-beam radiation therapy as well as the 51 patients who were treated by low-dose, preoperative external-beam radiation therapy and retropubic iodine-125 interstitial implantation. Conventional external-beam irradiation to the true pelvic lymph nodes (5,040 cGy) by a contoured four-field box technique was followed by a boost to the primary tumor volume, from 6,400 to 7,000 cGy, in 259 patients with pho- ton beams of 10-42 megavolts. Sixty additional patients had the boost to the prostatic tumor volume by the per- ineal portal with 160-million electron volt protons to total doses up to 7,700 cobalt cGy equivalents (/2). Another 51 67 68 TABLE 1.— American Joint Committee TNM Staging System, 1983 Stage Definition Comment T1 No palpable tumor; T1a: no more than 3 high-power fields of carcinoma found on histologic sections; T1b: more than 3 high-power fields of carcinoma found on histologic sections T2 Palpable tumor limited to (within) the prostatic capsule T2a Palpable tumor occupying <50% of one lobe and < 1.5 Same as, or nearly identical to, what has been reported as stage cm in diameter Bl by the Memorial system or as stage Tla and T1b by the Stanford system T2b Palpable tumor occupying > 50% of one lobe, > 1.5 cm Same as, or nearly identical to, what has been reported as stage in diameter, and/or more than one lobe or containing B2 by the Memorial system or as stage Tlc and T2 by the more than one nodule Stanford system T3 Palpable tumor extending into and usually beyond the Same as stage C in the Memorial system and stage T3 in the prostatic capsule Stanford system T4 Palpable tumors with attachment to the pelvic side walls Same as stage D1 in the Memorial system and as stage T4 in the or with invasion of the rectal wall or bladder Stanford system 4 See (11). patients were treated by 1,050 cGy as preoperative radia- tion therapy and an iodine-125 interstitial implantation of 16,000 cGy, total decay (/3). Patients with stage T1 (A) tumors were omitted from this analysis because in patients with a palpable tumor, we wished to evaluate the possi- ble significance of symptomatic bladder outlet obstruction requiring a transurethral resection on local regrowth or dis- tant metastases. Because, by definition, all stage T1 patients would have had a transurethral resection and no palpable tumor, their inclusion for this prognostic variable was not appropriate. The follow-up has been done by the radiotherapist and the referring urologist. The response of the primary tumor is based on the serial digital rectal examination as well as on the urologic evaluation, when clinically indicated, of blad- der outlet or ureteric obstruction or of pain as a possible result of local tumor regrowth. The development of distant metastases, which occurred in 99 of 370 patients, was usu- ally diagnosed by development of a positive bone scan, of abdominal lymph nodal metastases diagnosed by computed tomography or, rarely, development of pulmonary or he- patic metastases. Routine interval bone scans were not done on asymptomatic patients. These examinations were done only for symptomatic patients or for those asymptomatic patients on whom rising serum acid phosphatase levels were detected. Hormonal and radiation therapy were used simul- taneously in 77 of the 370 patients. Hormonal therapy was routinely reserved for those patients with evidence of tumor progression. The six tumor and patient factors evaluated in this report include: primary tumor stage (table 1), degree of tumor differentiation (well and moderately well vs. poor), age at diagnosis (Zthe population mean of 67.8 yr), the method of diagnosis (needle biopsy vs. transurethral resec- tion), and stable or elevated serum acid phosphatase levels at the time of diagnosis. Whether initial hormonal therapy (diethylstilbestrol or bilateral orchiectomy) was used did not influence the evaluation. The actuarial tumor regrowth or recurrence time distri- bution curves were calculated by the Kaplan-Meier life ta- ble method, and the tests for differences in the curves were by log rank. We assessed the relative prognostic signifi- cance of the six patient and treatment variables using the Cox proportional hazard model (/4). RESULTS Overall Patient Survival The overall survival by actuarial calculation for patients grouped by clinical stage is shown in figure 1, with a sta- tistically significant difference by univariate analysis of P less than .0001 and by multivariate analysis of P = .0013 (table 2). The only other prognostic factor significantly and independently influencing overall survival was degree of tumor differentiation (P = .0059; table 2). This influence of degree of differentiation is shown for the stage T2 pa- tients in figure 2. Figure 3 illustrates that the overall sur- vival for T2 patients treated either by external-beam radia- tion therapy or by retropubic iodine-125 interstitial implant was similar. Having symptomatic bladder outlet obstruc- tion that required a transurethral resection at the time of 100 — ® N— ma 80 = STAGE T2 S_ N=164 3 © 60 3 & 3 5 ‘07 a STAGES T3 + T4 5 N=206 £ © 8 20 a STAGE AT DIAGNOSIS p<.0001 0 T T T T —T T T T 0 2 4 6 8 10 years after irradiation FIGURE 1.—Overall survival of 370 irradiated patients with tumors sub- grouped by stage (T2 vs. T3-T4). NCI MONOGRAPHS, NUMBER 7, 1988 TABLE 2.—Overall survival Regression Variable coefficient P Tumor stage 32 .0013 Histologic differentiation 275 .0059 Transurethral resection of the prostate 1.89 .06 Initial hormonal therapy 0.68 49 Elevated serum acid phosphatase —043 67 Age at diagnosis —0.28 27 diagnosis was an unfavorable prognostic sign of borderline significance (P = .06; table 2). Local Tumor Regrowth The stage (local tumor extent and/or invasiveness) was the one and only significant prognostic factor of local tumor regrowth in this series (see fig. 4 and table 3). However, its influence was highly significant (P = .0006) and showed a 28% incidence of local recurrence by 8 years in those patients with stages T3 and T4 tumors, as compared with only 8% for patients with clinical stage T2 tumors. The differences in the length of posttreatment observation time among any 2 groups being compared makes either the de- velopment of local tumor regrowth or the development of distant metastases better evaluated by the cumulative fre- quency method (i.e., actuarial recurrence time distribution curve) rather than the proportion of patients developing ei- ther local regrowth or distant metastases. For instance, in figure 4, the cumulative frequency of local tumor regrowth at 8 years is 28% in the patients with T3 and T4 tumors, whereas the proportion of T3 and T4 patients who have developed a local regrowth (34 of 206 or 17%) is much lower. The curve over time for patients with T2 tumors may be nearly flat beyond 5 years, but the curve for stage T3-T4 patients is continuing to increase up to and probably beyond 8 years of observation. The cumulative frequency of local regrowth is similar in stage T2 patients, whether the treatment is by external-beam irradiation or by iodine-125 implantation (fig. 5). In addition, the cumulative frequency ~ S E WELL + MODERATE > = a N=126 = POOR & N=38 3 = 40 2 3 STAGE T2 § 20 HISTOLOGIC DIFFERENTIATION s p= .006 0 ¥ T T > T T Y T 0 2 4 6 8 10 years after irradiation FIGURE 2.—Overall survival for stage T2 patients subgrouped by histologic differentiation (well and moderate vs. poor). MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 69 < 1-125 < N=38 5 EXTERNAL BEAM 1 N=126 wn 60 ~J 3 & > O 404 Ss = 3 3 20 — S STAGE T2 METHOD OF TREATMENT 0 r T T T T ee 1 0 2 4 6 8 10 years after irradiation FIGURE 3.—Overall survival of stage T2 patients treated either by external- beam irradiation or by interstitial iodine-125 implantation. of patients developing local tumor regrowth for those with specified initial tumor stage is not influenced by previous initial hormonal therapy or lack of it (fig. 6). Development of Distant Metastases The cumulative incidence of progression with distant metastases is also significantly influenced by initial tumor stage (P = .00007 by multivariate analysis). Within tumor stage (T2 vs. T3 or T4), histologic differentiation (well or moderate vs. poor) is a statistically significant variable pre- dicting progression to distant metastases (fig. 7). Only 16 of the 164 stage T2 patients had elevated acid phosphatase levels at diagnosis. By univariate analysis, an increase in the cumulative incidence of progressing with distant metas- tases at the P = .05 level was predicted. However, elevated acid phosphatase at diagnosis “co-varied” with poor dif- ferentiation, such that it was of no independent statistical significance by multivariate analysis. One hundred thirty-five of the patients in this series had surgical staging of their lymph nodes prior to irradiation. 100 A 80 4 STAGE AT DIAGNOSIS Zz g 8 p=.0006 © $60 oc I S 40 3 STAGES T3 + T4 > N=206 3 8 STAGE T2 S N=164 years after irradiation FIGURE 4.—Cumulative incidence (actuarial time distribution curve) of local regrowth of tumor subgrouped by clinical stage (T2 vs. T3 and T4). 70 TABLE 3.—Local regrowth? Regression Variable coefficient P Tumor stage 3.41 .0006 Elevated acid phosphatase —0.85 39 Transurethral resection of the prostate 0.77 44 Age at diagnosis —0.69 49 Histologic differentiation 0.51 61 Initial hormonal therapy ~0.28 .80 4 Data for 370 patients were evaluated by multivariate analysis. For stage T2 patients, microscopic lymph nodal status was a significant predictor of progression to distant metastases (P = .02; fig. 8) but was not significant in the 72 patients undergoing such evaluation who were stage T3 and T4 (fig. 9). Radiation Sequelae Following High-energy, Photon-beam Irradiation Both the acute and late radiation reactions in patients treated by iodine-125 implantation (/3) and by perineal proton boost irradiation (6) have recently been reported. The present analysis is on 121 patients treated on a lin- ear accelerator with a photon beam of 25-megavolt peak energy. These patients were treated from January 1980 through December 1981. We treated all with 180-cGy frac- tions, 5 sessions per week using the previously described four-contoured field box technique to include the primary tumor volume and the pelvic lymph nodes below the bifur- cation of the common iliacs to a dose of 5,040 cGy in 28 fractions. The prostatic tumor volume was then treated by a cone-down boost with contoured parallel-opposed lateral fields to a total tumor dose of 6,480 cGy to 6,840 cGy in 180-cGy fractions, 5 sessions per week. The genitourinary radiation sequelae are shown in tables 4-6. Only 1 of 121 patients (i.e., 0.8%) sustained a major complication, which was a radical cystectomy for persistent bleeding in a patient with an original stage T3 tumor who 100 1 STAGE T2 METHOD OF TREATMENT 80 + = .66 60 40 20 probability of LOCAL REGROWTH (%) EXTERNAL BEAM (126) €— 1-125 (38) X T 7 r T 7 T 0 2 4 6 8 10 years after irradiation FIGURE 5.—Cumulative incidence of local tumor regrowth for patients with stage T2 tumors treated either by external-beam irradiation or iodine-125 implantation. 100 STAGES T3 + T4 INITIAL HORMONAL TREATMENT 80 p= .49 4 60 40 YES (68) NO (138) probability of LOCAL REGROWTH (%) years after irradiation FIGURE 6.—Cumulative incidence of local tumor regrowth for patients with stage T3 and T4 tumors subgrouped by treatment with initial hormonal therapy plus radiation or no treatment, or by radiation therapy alone. had no microscopic cancer residual in his prostate, but had a preexisting thrombocytopenia due to hypersplenism. No patient is incontinent (table 4). Of the 10 urethral strictures or bladder neck contractures seen, all were observed in the first 2 years following irradiation (table 5), and 9 were read- ily corrected by transurethral surgery. One patient still has moderately severe urinary frequency but is continent. The overall incidence of stricture or bladder neck contracture was 8.3% (10 of 121). However, the incidence was 14% (9 of 63) in those patients who had undergone a transurethral resection prior to radiation therapy. By contrast, there was a significantly lower incidence of stricture (1.7% or 1 of 58) in those patients not having to undergo transurethral pros- tatic resection prior to radiation therapy. Erectile potency for 3 or more years has been maintained in 34, for an inci- dence of 63% (table 6). At diagnosis, erectile potency was 100 ® STAGE AT DIAGNOSIS » 804 p<.0001 » = 0 = Ww 60 = STAGES T3 + T4 = N=206 i 9D 40 a S > E eo oe Q a STAGE T2 N=164 0 T T T T 1 0 2 4 6 8 10 vears after irradiation FIGURE 7.—Cumulative incidence of progression with distant metastases for all irradiated patients with tumors subgrouped by stage (T2 vs. T3 and T4). NCI MONOGRAPHS, NUMBER 7, 1988 100 A STAGE T2 NODAL STATUS p= .02 80 + 60 N+ (12) 40 + 20 probability of DISTANT METASTASIS (%) No (51) 0 in ur — T T T e— 0 2 4 6 8 10 years after irradiation FIGURE 8.—Cumulative incidence of progression with distant metastases for patients with stage T2 tumors subgrouped by histologic evaluation of lymph nodal status subsequent to staging pelvic lymphadenectomy. present in 54 of the irradiated patients. The initial tumor stage did not influence the incidence with which potency was maintained. The Radiation Therapy Oncology Group grading system was used for the radiation sequelae shown in table 7. Diar- rhea persisting 3 or 4 months following radiation therapy or rectal bleeding that did not require any medication (pa- tient’s wish) occurred in 12%, although in an additional 9% of the patients, these symptoms were judged “moder- ate” because the patient wished to use a medication (ei- ther an antispasmodic agent or an anal suppository). These mild to moderate radiation sequelae have been resolved in all but 8 patients; 4 have persistent but controlled (with medication) diarrhea, and 4 have infrequent mild episodes of hemorrhoid-like rectal spotting with blood. No patient has required surgery for any intestinal complication. Within the diarrhea subcategory, 11 of 15 patients have had their symptoms resolved and, among the 22 patients with infre- quent rectal bleeding, 18 have had this symptom completely resolved. Thus radiation sequelae that have persisted are infrequent (table 8). Minor rectosigmoid irritation symptoms persist 100 7 8 1 STAGES T3+T4 wn 2 4] NODALSTATUS % p= .25 wn E S60 N+ (30) 5 = 3 No (42) @ 2 40 o 2 3 © S20 Q 0 T T — T T T - 1 0 2 4 6 8 years after irradiation FIGURE 9.—Cumulative incidence of progression with distant metastases for patients with stage T3 and T4 tumors subgrouped by histologic eval- uation of lymph nodal status subsequent to staging pelvic lymphadenec- tomy. MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 71 TABLE 4.— Genitourinary radiation sequelae Incidence in 121 patients Sequelae No. Percent Hematuria 21 17.0 Resolved 20 95.0 Persistent 1 0.8 Incontinence 0 0 in 6.6% of the patients. Minor urologic symptoms persist in 0.8% of them, and a major complication occurred and was surgically corrected by cystectomy in 0.8%. There have been no major sequelae in the small intestine. Erectile im- potence has occurred in 37% of previously potent patients. Two patients underwent an uneventful hip replacement for progressive degenerative joint disease, although it is not likely that development of this disease was influenced, for better or worse, by the radiation (4,500 cGy to the femoral heads) that all 121 patients received. DISCUSSION Many observations in this single-institution, retrospective analysis are similar to those reported (/-10). This is espe- cially true for the prognostic influence of the primary tu- mor size as well as the degree of histologic differentiation on the subsequent development of disease progression and on overall patient survival. Most of these reports, including the present one, tell of a follow-up that is very limited be- yond 10 years from treatment. However, the results from studies by researchers at Stanford University Medical Cen- ter have follow-up after radiotherapy for up to 27 years (5,15), and those at the M. D. Anderson Hospital and Tu- mor Institute have 77 patients who have been evaluated for more than 10 years after treatment (/6). A histologic sam- pling by either transurethral resection or Silverman needle biopsy may undergrade the primary tumor relative to the findings at radical prostatectomy in 25% to 33% of the pa- tients (/ 7,18). However, the consistent and statistically very significant influence that poorly differentiated tumor has on treatment outcome in virtually all the reported series sug- gests the important usefulness of the clinical grade of the tumor. Our results give additional documentation to the impor- tance of primary tumor size as influencing the subsequent probability of patients developing either a local tumor re- growth or progression with distant metastases. For the stage TABLE 5.—Genitourinary radiation sequelae Incidence in 121 patients Sequelae No. Percent Stricture 10 8.3 Prior transurethral resection 9/63 14.0 of the prostate No prior transurethral resection 1/58 1.7 of the prostate Corrected strictures 9 90 Persisting symptoms 1 0.8 72 TABLE 6.—Genitourinary radiation sequelae TABLE 8.—Persistent radiation sequelae Sequelae No. Percent Symptoms Percent Potent prior to radiation therapy 54/121 45 Urologic Potency maintained 34 63 Minor 0.8 Stages T1-T2 22/39 56 Major 0.8 Stages T3-T4 12/15 80 Rectal Minor 6.6 Major 0.0 Erectile impotence 37.0 T2 (B) tumors, the local control rate is 92%, compared with only 72% at 8 years for those with stages T3 and T4 (C) tumors. Likewise, initial tumor size was highly significant with regard to the probability of patients developing pro- gression with distant metastases; e.g., at 8 years, this was 18% for patients with stage T2 tumors and 60% for those with stage T3 or T4 tumors. In all multivariate analyses, no single factor was more significant than initial tumor stage, including when such an analysis was done on the subgroup of 135 patients who had surgical staging of the pelvic lymph nodes. The implications of these observations are that ear- lier diagnosis, improvements in the local treatment of sub- groups of patients with T3-T4 tumors, and development of some effective, tolerable, systemic, cytotoxic treatment will be of enormous benefit. Our frequency of local regrowth agrees quite well with most other series. For instance, our permanent local tumor control rate for patients with stage T2 (B) tumors is similar to that following radical prostatectomy reported by Gib- bons et al. (19). They report local tumor regrowth in 6% of 195 stage B patients with from 4 to 30 years of observation, although they do not use the actuarial technique of calcula- tion, and some of those patients received postprostatectomy irradiation. Most radiation therapy series use actuarial cal- culations and have similar results. However, this is not true for the series from the Baylor College of Medicine, in which a combination of external-beam irradiation and radioactive gold-198 interstitial brachytherapy was used (20). In 124 patients ranging in stage from A2 to C1, the actuarial lo- cal recurrence rate at 10 years was 49%. This rate con- trasts sharply with the actuarial 10-year local recurrence rates from other radiation series, i.e., at Stanford Univer- sity Medical Center, this recurrence rate was 21% in 799 patients with stages B and C tumors (/5); at the M. D. TABLE 7.—Late bowel radiation sequelae Incidence in Persisting Sequelae 121 patients symptoms Grade? Manifestation No. Percent No. Percent 1 Mild diarrhea or bleeding 14¢ 12 5 4 (no medication required) 2 Moderate diarrhea or 12¢ 10 3 2 bleeding (treated with medication) 3 Surgery for bleeding 0 0 0 0 4 Surgery for fistula or 0 0 0 0 obstruction 5 Death 0 0 0 0 4 Sequelae were resolved in 69% of patients (18 of 26). b The Radiation Therapy Oncology Group grading system was used. Anderson Hospital and Tumor Institute, this rate was 19% for 551 stage C patients (16), and here at Massachusetts General Hospital, we reported a recurrence rate of 22% for 370 stage T2-T4 (B-C) patients. This lower recurrence frequency was also seen in patients who did not receive initial hormonal therapy (fig. 6, table 3). The reason for the high regrowth rates following gold-198 brachytherapy combined with external-beam irradiation may be related to radiation dose inhomogeneity because a significant propor- tion of the prostate received no more than 5,000 cGy (21). The correlation shown in the Baylor series between a pos- itive postirradiation biopsy and the development of distant metastases most likely results because both are known to be strongly influenced by initial tumor volume (20). In small tumors, the incidence of progression with distant metastasis is low as is the incidence of positive tumor biopsy postir- radiation. In larger tumors both incidences are high. Only when a large series of patients with similarly sized tumors are analyzed for the relationship between positive tumor biopsy following irradiation and progression with distant metastasis could a causal relationship be suggested. For patients with tumors of comparable stage, e.g., T2a or the B1 nodular tumors, the 15-year overall survival fol- lowing either radical prostatectomy (/9,22,23) or radiation therapy (4) is similar and ranges from 51% to 57%. Al- though some differences in patient selection may exist in these series, the overall local efficacy of either radiation therapy or surgery is high and similar. Therefore, we judge that the patients with stage T2 (B) tumors should decide which treatment they prefer after being fully informed as to available therapy and the possible sequelae. For patients with the larger stage T3-T4 tumors, we believe rigorous clinical research is necessary in any attempt to document techniques of improved therapeutic efficacy. At present, we are running a randomized phase III trial for patients with T3-T4 tumors of conventional 25-megavolt photon ther- apy to doses in the conventional dose range of 6,800 cGy, compared with the perineal proton boost allowing the tu- mor volume to be raised safely to exposure to a total dose of 7,600 cGy. We believe that undetected distant metas- tases are the major clinical problem in patients with locally advanced prostate adenocarcinoma and await the develop- ment of effective and tolerable adjuvant systemic therapies. REFERENCES (1) VICKERY AL Jr, KERR WS Jr. Carcinoma of the prostate treated by radical prostatectomy. Cancer 1963;12:1958- 1968. (2) BYAar DP. Identification of prognostic factors. In Cancer NCI MONOGRAPHS, NUMBER 7, 1988 Clinical Trials: Methods and Practice (Buyse M, Staquet MJ, Silvester JR, eds). Oxford: Oxford Univ Press, 1984, pp 423-443. (3) GLEASON DF, MELLINGER GT, VETERANS ADMINISTRATION COOPERATIVE UROLOGICAL RESEARCH GROUP. Prediction of prognosis for prostatic adenocarcinoma by combined histologic grading and clinical staging. J Urol 1974;111: 58-64. (4) BAGsHAW MA. Potential for radiation therapy alone in can- cer of the prostate. Cancer 1985;55:2079-2085. (5) BARZELL W, HILARIS BS, WHITMORE WS JR, ET AL. Prostatic carcinoma: Relationship of grade and local extent to pat- tern of metastases. J Urol 1977;118:278-282. (6) DUTTENHAVER JR, SHIPLEY WU, PERRONE TL, ET AL. Pro- tons or megavoltage x-rays as boost therapy for pa- tients irradiated for localized prostatic carcinoma. Cancer 1983;51:1599-1604. (7) Perez CA. Carcinoma of the prostate. Int J Radiat Oncol Biol Phys 1983;9:1427-1438. (8) ROSEN EM, CAssaDY JR, CHAFFEY JT, ET AL. Radiotherapy for localized prostatic carcinoma. Int J Radiat Oncol Biol Phys 1984;10:2201-2210. (9) HANKS GE, LEBEL SA, KRALL JM, ET AL. Patterns of care studies: Dose-response observations for local control of adenocarcinoma of the prostate. Int J Radiat Oncol Biol Phys 1985;11:153-157. (10) PiLericH MV, KRALL JM, SAUSE WT, ET AL. Prognostic factors in carcinoma of the prostate—analysis of RTOG study 75-06. Int J Radiat Oncol Biol Phys 1987;13:339- 349. (11) AMERICAN JOINT COMMITTEE ON CANCER. A Manual for Staging of Cancer. Philadelphia: Lippincott, 1983. (12) SHIPLEY WU, TEPPER JE, PROUT GR JR, ET AL. Proton radi- ation as boost therapy for localized prostatic carcinoma. JAMA 1979;241:1912-1915. (13) DELANEY TF, SHIPLEY WU, O'LEARY MP, ET AL. Preoper- MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 73 ative irradiation, lymphadenectomy, and iodine-125 im- plantation for patients with localized carcinoma of the prostate. Int J Radiat Oncol Biol Phys 1986;12:1779- 1785. (14) Cox DE. Regression models and life tables. J R Stat Soc B 1972;34:187-220. (15) SHIPLEY WU, BAGSHAW MA, PROUT GR JR. The success of radiation therapy in controlling prostatic cancer within the treated field. In Proceedings of the Second International Symposium on Prostatic Cancer (Murphy GP, Khoury F, eds). New York: Liss, 1988, pp 199-212. (16) ZAGARS GK, VON ESCHENBACK AC, JOHNSON DE. The man- agement of stage C adenocarcinoma of the prostate with external beam radiation therapy. J Urol 1987;37:abstr 199A. (17) CATALONA WI, STEIN AJ, FAIR WR. Grading errors in pros- tatic needle biopsies. J Urol 1982;127:919-922. (18) MILLS SE, FOWLER JE. Gleason histologic grading of pros- tatic carcinoma: Correlations between biopsy and prosta- tectomy specimens. Cancer 1986;57:346-349. (19) GiBBONS RP, CORREA RJ, BRANNEN GE, ET AL. To- tal prostatectomy for localized prostatic cancer. J Urol 1984;131:73-76. (20) SCARDINO PT, FRANKEL JM, WHEELER TM, ET AL. The prognostic significance of post-irradiation biopsy results in patients with prostatic cancer. J Urol 1986;135:510-516. (21) CARLTON CE JR, HUDGKINS PT, GUERREIRO WG, ET AL. Radiotherapy in the management of carcinoma of the prostate. J Urol 1976;116:206-210. (22) JEweTT HJ. Radical perineal prostatectomy for palpable, clinically localized, non-obstructive cancer. Experience at The Johns Hopkins Hospital: 1909-1963. J Urol 1980;124:492-497. (23) JEwerT HJ, WALSH PC. Radical perineal prostatectomy for clinical stage T2 carcinoma of the prostate. J Urol 1982;127:704-709. External-beam Radiation Therapy for Clinically Localized Prostate Cancer: Patterns of Care Studies in the United States Gerald E. Hanks' ABSTRACT —Data are presented from the Patterns of Care Study and other sources that define the role of external-beam irradiation in the management of localized prostate cancer as prac- ticed in the United States as a whole. Patients must be treated with complex treatment techniques and high-energy linear accelerators and careful adjustment of radiation dose. Transurethral resection of the prostate should be avoided in the intermediate and poorly differentiated subgroup of stage C patients. The excellent 5- and 10-year survival for patients treated by radiation therapy is demon- strated for all stages of prostate cancer and for T1 or early stage B patients. It is noted that the national averages for survival have improved between 1973 and 1978. Stages A2 and B patients with negative lymph node dissections show freedom from recurrence that is equal to patient reports for radical surgery. Complications resulting from radiation therapy were modest, and potency was maintained in 73% of the patients. Adjuvant irradiation is necessary for pathologic stage C patients after recovery from surgery. Radia- tion therapy is equally effective though less costly than surgery for early prostate cancer. A particular need of future research is the study of the patterns of care in the United States regarding the surgical management of prostate cancer so that health profes- sionals can determine if this care is generally available throughout the United States and if good outcome and acceptable morbidity result after it is given.— NCI Monogr 7:75-84, 1988. These data were collected for this National Consensus Panel to help answer the following questions: What methods are optimal for definitive radiation therapy? What are the long-term results in control and survival? What is the morbidity of the procedures? How can it be minimized? Should definitive radiation therapy be used as an adju- vant in high-risk patients? What directions for future research are indicated? When is pelvic node dissection necessary and who are the candidates? In addition, some unique data will be presented on the relative costs of the alternative forms of management of early prostate cancer. PATIENTS AND METHODS Data are presented from the Patterns of Care Study in Radiation Therapy, the author’s former practice in Sacra- mento, California, and from the literature. ABBREVIATION: TURP = transurethral resection of the prostate. ! Department of Radiation Therapy, Fox Chase Cancer Center, Central and Shelmire Ave., Philadelphia, PA 19111. The Patterns of Care Study technology has been reported (/-3). Suffice to say that this two-level random sample of the practice in the United States represents true national averages for 2 surveys of patients who were treated either in 1973 and 1974 or in 1978 (4). A third survey was conducted on patients treated in 1973 in the five facilities in the United States where the largest numbers of patients with prostate cancer were treated. These special data are added to those of the 2 previous surveys only for the determination of dose-response relationships for infield control. A summary of patient records that were reviewed is shown in table 1. The information presented concerning the relative cost of radiation therapy and surgery was gathered by an assessment of the average cost of 1) treatment with external-beam irradi- ation for 128 patients, 2) radical prostatectomy for 12, and 3) '°] seed implantation for 8 patients undergoing lymph node dissection in Sacramento, California, during 1984 (5). The costs of each method of treatment were determined from the point of our having obtained a positive biopsy and a negative metastatic workup. They include all technical, pro- fessional, and hospital inpatient and outpatient charges and are accurate for the private practice of radiation therapy and urologic surgery at the time and site studied. Staging is by the A, B, C system common to practice in the United States (6). RESULTS Need for Complex Technology Table 2 illustrates the reduction in complications observed when patients are treated with multiple field techniques or multiple fields per day (7,8). An advantage in any recurrence and in infield recurrence when linear accelerators are used with energies of 6 million electron volts or greater as con- trasted to 4-million electron volt linear accelerators or cobalt-60 units is shown in table 3 (9). Importance of Dose Selection and Method of Diagnosis Tables 4-6 depict dose-response relationships for infield failure in 1,516 patients (/0). Looking at all the patients, one would conclude that the range of 6,500 to 7,000 cGy is appropriate. When one considers individual stages, there is no advantage to exceeding 7,000 cGy for stage B cancer, although for stage C there is a decrease in local failure when the maximal dose to the center of the prostate is 7,000 cGy or greater. Thus dose must be adjusted to tumor volume (4). These same investigators have observed a doubling of com- plications from 3.5% to 6.9% (P = .03) when dose exceeds 7,000 cGy, which emphasizes that these high doses should be restricted to patients with extensive local disease (4,7,10). In 75 76 TABLE 1.—Patterns of Care Study data base TABLE 6.—Patterns of Care Study: Relationship of dose to infield recurrence in 624 stage C patients No. of patient St Years of records es Survey sample treatment reviewed A B C National average 1973, 1974 674 66 312 296 Special purpose 1973 188 5 62 121 National average 1978 733 115 381 237 Percent actuarial free of recurrence? TABLE 2.—Patterns of Care Study No. of infield i ————— Dose, cGy failures/total Percent 3yr Syr 7 yr <6,000 17/69 25 72 63 — 6,000-6,499 31/111 28 76 64 64 6,500-6,999 46/200 23 83 72 68 7,000+ 42/244 17 86 81 76 No. of Radiation therapy patients Complication, % Technique Anterior/posterior- 301 6.0 posterior/anterior Other 328 3.0 Fields/day 1 187 6.0 2-6 426 35 TABLE 3.—Correlation of photon energy with recurrence Total Percent recurrence Treatment No. of source patients Any? Locoregional® Cobalt-60 309 33 20 6 MeV 305 31 18 6-8 MeV 108 19 10 20 MeV 138 24 10 4 MeV = million electron volts. bp= 02. tp OL TABLE 4.—Patterns of Care Study: Relationship of dose to infield recurrence, all stages, 1,516 patients Percent actuarial free of recurrence” No. of infield Dose, cGy failures/total Percent Syr 7 yr <6,000 41/178 23 71 62 6,000-6,499 51/269 19 76 73 6,500-6,999 85/548 16 83 79 7,000+ 74/521 14 84 79 ¢ By linear trend P = .0001; Mantel P = .0009. TABLE 5.—Patterns of Care Study: Relationship of dose to infield recurrence in 724 stage B patients Percent actuarial free of local recurrence? No. of infield Dose, cGy failures/total Percent 3yr Syr Tyr <6,000 22/817 25 85 71 — 6,000-6,499 16/120 13 94 82 77 6,500-6,999 37/283 13 91 88 77 7,000+ 32/234 14 93 84 78 4 For unstratified, linear trend, P = .0021 and Mantel, P = .0203. 4 For unstratified, linear trend, P = .0021 and Mantel, P = .0203. addition, we could calculate that there would have been an 8% improvement in local control and a 10% improvement in complications if optimal doses were administered to patients in each stage who were treated in 1973 and 1974 (4). Patients with intermediate and poorly differentiated stage C cancer who undergo TURP as a method of diagnosis have a poor prognosis as illustrated by figures 1-3 (4,11,12), whereas no adverse effects have been observed in those with well-differentiated cancers. Significant increases are ob- served in both metastasis and death. How the subset of patients is affected is shown in table 7. Possible mechanisms include: The cancer is disseminated by the procedure or, in some indirect way, obstruction helps to select a poor progno- sis subgroup separate from known prognostic indicators (4). Investigators (13-15) who reported no effect from TURP may not have looked at the appropriate subgroups of patients. 1001 98 78 —-EMODIM™M MC —r >» 0 T T TT T T T T T T T 8 12 24 3% 48 68 72 84 96 108 128 132 MONTHS FROM ONSET OF TREATMENT FIGURE 1.—Survival by method of diagnosis. Solid line = TUR (56 of 87 patients; dashed line = no TUR (21 of 52 patients). P < .01. NCI MONOGRAPHS, NUMBER 7, 1988 “cox — — = LEGEND: GROUP RECURRENCE TOTAL NO TUR 47 178 25 - TUR 48 [AR P< 01 mMozZzmMT Too MD yr TT ry 0 on 24 36 48 60 n 84 96 FIGURE 2.—Freedom from metastatis by method of diagnosis, RTOG 75-06. A 1 L 501 : : fo v £ LEGEND: GROUP DEAD TOTAL NO TUR 42 178 TUR 38 mn P< .01 25 1 mm T 7 T ¥ T ¥ T ¥ T ¥ T T T 0 12 2H 36 48 60 7? FIGURE 3.—Survival by method of diagnosis, RTOG 75-06. oo - © > MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 4 — 77 TABLE 7.—Patterns of Care Outcome Studies: Adverse effect of TURP Observed in T3, T4, or stage C cancers Local recurrence unchanged Metastatic recurrence doubled Death rate doubled Not explained by differences in histology or tumor extent National Averages for Survival The survival of 60 patients with mixed stages Al and A2 disease treated in 1973-1974 in the United States is shown in figure 4 to be equal to survival for their expected age group (16). One patient experienced an isolated local recurrence, and 3 patients developed metastasis as a first failure. No lymph node dissections were performed, and patients were not selected for favorable functional status as is common in surgical reports. The national averages for survival of stage B patients are shown in figure 5 (16); no patient had lymph node staging. No significant decrease in survival was observed over that expected for a similar age group of patients without cancers at 5 years, but a 15% decrement is noted at 10 years. Again, these patients are not selected for favorable prognostic indi- cators. Figure 6 shows the national average for survival of patients with stage C cancer. Increased mortality was expressed in the first 5 years with a 25% decrement below the expected survival at both 5 and 10 years (16). This is a population that is neither selected for its good general health nor assessed by lymph node dissection. Improvement in Survival Between 1973 and 1978 The survival of stage B patients treated in 1973 and in 1978 shows improvement (fig. 7), but the difference is not 58+ —ZmMmOoomMmT 49 30 me —r > — 1 T T 1 T T T TT T TT 8 1 2 3 4 S 6 7 8 9 18 1 YEARS FROM TREATMENT START LECENDs SURVIVAL —— EXPECTED — —— OBSERVED FIGURE 4.—National averages: Stage A survival compared with expected survival (60 patients). 78 98+ 78 “~—ZmMOo3xMO 58 —_ZMO DMO 49 Mm < =r > Mm —r > YEARS FROM TREATMENT START T T T T T T T T T T T LEGEND: SURVIVAL ~~ —— EXPECTED ~~ --- OBSERVED 8 12 24 3 9 60 72 84 9 18 128 FIGURE 5.—National averages: Stage B survival compared with expected survival (312 patients). MONTHS FROM ONSET OF TREATMENT FIGURE 7.—Survival for stage B patients treated in 1973 and 1974 (312) compared with those treated in 1978 (347). Solid line = 1973; dashed significant and was predicted when we compared the distri- line = 1978. bution of favorable independent variables for survival between the 2 series (4,7). Figure 8 illustrates the difference in survival of stage C patients who were treated in 1973 versus 1978 (4) that approaches statistical significance (P= P E p R £ ‘ R C N € T N 1 A Mm L A an “== 38 1 | v 38 v 38 E £ 28] 20-| 18 10 2-1 2 T T T T T T T T 1 IY T T T T T T T T ¥ Y e 1 2 3 4 5 8B 7 8 9 Ww um 9 12 kl 36 8 60 72 84 98 18 128 YEARS FROM TREATMENT START MONTHS FROM ONSET OF TREATMENT LEGEND: SURVIVAL ~~ —— EXPECTED ~~ —- — OBSERVED z : 2 FIGURE 8.—Survival for stage C patients treated in 1973 and 1974 (296) FIGURE 6.—National averages: Stage C survival compared with expected compared with those treated in 1978 (215). Solid line = 1973; dashed survival (296 patients). line = 1978. NCI MONOGRAPHS, NUMBER 7, 1988 100 75 ——— OBSERVED SURVIVAL 254 ------ EXPECTED SURVIVAL 0313 267 118 32 T T T T T T T T T T 0 1 2 3 4 5 6 7 8 9 10 YEARS FROM START OF TREATMENT FIGURE 9.—Survival of T1 NO MO patients treated with external-beam irradiation. .10). When we analyzed the distribution of independent vari- ables for survival between these 2 series, we saw no differ- ence to explain the improved survival, and it may well be due to improved treatment. Stage migration is always a concern in sequential studies but is difficult to rule in or out (1/7). Survival of a Subgroup of Early Stage B Patients The survival data of 313 early stage B (TI NO MO) patients (shown in fig. 9) represent a pool of data from 138 treated in 1973 and followed up to 10 years with 175 treated in 1978 and followed up to 5 years (/8). There is no signifi- cant difference from the expected survival at 5 years, although a 10% decrement is noted at 10 years. Indeed, this would be expected as no patient had a lymph node dissection, and 30% of 40 patients with grade III histology developed metastasis as a first failure site. Outcome for Groups of Patients Directly Comparable to Surgical Series Table 8 shows the outcome of 65 patients treated in 1978 who were given external-beam irradiation or '>’I implants after negative lymph node dissection. They did extremely well in local control, with 1 of 65 patients experiencing isolated local recurrence, and any type of recurrence was noted in only 9 of 65 patients (14%). A comparison of the survival of the subgroup of patients treated with external-beam irradiation with that of patients in the randomized study of the Uro-Oncology Research Group 79 as reported by Paulson et al. is presented in figure 10 (19). Patients in the Patterns of Care Study show the same nonpro- gression rate as those in the surgical arm of the Uro- Oncology Research Group study. Patients in this latter group who received radiation therapy remain anomalous and exhibit a disease progression rate commonly seen in stage C patients. Table 9 lists the serious problems of the investigation, analysis, and reporting of the Uro-Oncology Research Group study that inevitably lead the observer to the conclusion that the radiation arm contained stage C patients, whereas the surgical arm contained lower stage patients (20,21). It is unfortunate that this trial (19) has been reported to show that radical prostatectomy is superior to external-beam radiation therapy. Sequelae of External-beam Radiation Therapy Table 10 lists the posttreatment sequelae observed in 1,393 patients treated with external-beam radiation therapy. There was indeed only 1 death among 1,393 patients, and these complications were observed for treatment given to patients in 1973,1974,and 1978. It is clear that current modern tech- nology and technique will result in even fewer serious com- plications. To keep in perspective the comparable sequelae, we have included data from several recent surgical series in table 11 (22-25). The maintenance of potency after treatment is a most important goal, and table 12 illustrates that potency is gener- ally more frequently preserved following external-beam radiation therapy than following radical prostatectomy even with nerve-sparing modification (26,27). This is even more impressive when you consider that the radiation group is an unselected group of patients who frequently have other medi- cal illnesses and variable levels of general health, whereas the radical prostatectomy group is highly selected for both general health and desire to maintain potency. Adjuvant Irradiation After Prostatectomy for Pathologic Stage C Cancer Table 13 illustrates the clinical inaccuracy of one’s judging the extent of prostate cancer among patients in various insti- tutions as evaluated by the overall fraction of surgical patients who have pathologic stage C cancer. It also reflects a variation in the tendency of surgeons to operate on patients with stage B2 cancer (e.g., Catalona 44%, Walsh 22%), when they will indeed observe the predicted two-thirds frequency of extension out of the gland (26-28). In tables 14 and 15, we (29) have tabulated the local con- trol of prostate cancer when adjuvant irradiation was given before or after local recurrence. Control is obtained in 94% of TABLE 8.—Results of 65 patients with stages A and B after negative lymph node dissection Failure Total Stage failure Treatment ————— Isolated Infield and group A B infield metastases Metastases No. Percent External-beam irradiation 10 27 1 1 3 5 14 1251 implant 5 23 0 3 1 4 14 MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER @ o | ¢ | 1 1 1 ] PERCENT WITHOUT FAILURE 8 1 i ©—@ VA SERIES (Prostatectomy) —_—r Ss . O---0 VA series) (Radiotherapy) n Qo 0 1 | 1 | 1 YEARS FIGURE 10.—Comparison of Patterns of Care Study (PCS) series (stage A2, B with negative lymph node dissection) to Uro-Oncology Research Group trial. the patients when treated shortly after surgery versus 70% of patients when treated after clinical recurrence. It is also commonly observed that higher doses of radiation are required after clinical recurrence and more complications will result from that high-dose treatment. Table 16 is a summary of local recurrence reported in 943 patients with pathologic stage C cancer categorized by TABLE 9.—Problems of the study, analysis, and reporting of the Uro-Oncology Research Group comparison“ Many lost or dead of intercurrent disease (56%) in 4-7 yr Incontinence rate of 12%-40% Radiation therapy committee not involved in data analysis Suggestions of radiation therapy committee ignored Inappropriate end point Questionable method of randomization Assigned treatment received by 90 of 106 (85%) Radiation received by 4 of 47 assigned surgery Surgery received by 3 to 59 assigned radiation Positive surgical margins eliminated (2 patients) TABLE 11.—Complications of radical prostatectomy“ Complication Frequency, % Death 0-2 Infection, fistula, rectal injury 1-8 Bladder stricture 9-18 Incontinence 2.5-15 Impotence 90-100 4 No complications involving the lymphatic system were observed. See (22-25). treatment (adjuvant irradiation); a consistent reduction in local recurrence was noted when postoperative radiation was given. The reason no differences are reported in survival between the irradiated and nonirradiated pathologic stage C patients may be explained. With reasonable basic assumptions (30% develop metastasis in 5 yr, 15% die of intercurrent disease in 5 yr, and 5% develop local recurrences after irradiation, but 23% do so without it), it is possible for one to calculate that 150 patients would be needed in each experimental arm to show the difference in a prospective random comparison. Relative Costs of Radical Prostatectomy and External-beam Radiation Therapy Table 17 illustrates that the surgical treatment of early prostate cancer doubles the cost of treatment whether it is radical prostatectomy or lymph node dissection with >I seed implantation (5). Equivalent survival results are obtained with external-beam radiation therapy and either of the tech- niques mentioned before. This escalation of cost may not escape the notice of perceptive analysts at health main- tenance organizations, insurance companies, and the Federal Government who are approving payment of the bills for this care. Future Research Patterns of care studies are needed in the surgical treat- ment of prostate cancer so that physicians can observe what the national averages are for outcome and whether differ- ences exist in success of treatment in different types of facilities. Are the good results observed in five or six reporting institutions also obtained in the usual community hospitals? TABLE 12.—Maintenance of potency after treatment 4 See (19). TABLE 10.— Complications of external-beam radiation therapy Percent Actual complications from: free of i A A licati Radiati No. of oti iaapions adiation Surgery patients ~~ Yearof ———_" No.of Per- No.of Per- treated treatment 5 yr 10 yr patients cent patients cent 619 1973, 1974 93 86 28 45 16 25 674 1978 92 40 6.0 17 25 Treatment Potent, % Previous Standard radical prostatectomy 0-10 External-beam radiation therapy 60-70 Recent Nerve-sparing prostatectomy“ 72 Nerve-sparing prostatectomy” 52 External-beam therapy“ <3/mo, partial erections 47 >3/mo, full erections 73 4 See (26). b See (27). ¢ Banker FL: Personal communication. NCI MONOGRAPHS, NUMBER 7, 1988 81 TABLE 13.—Frequency of pathologic stage C disease observed in patients having radical prostatectomies No. of No. of patients Institution prostatectomies with pathologic reporting Reference performed stage C Frequency, % The Johns Hopkins Hospital (26) 100 41 41 Barnes Hospital (27) 77 35 45 Virginia Mason Medical Center (28) 148 45 30 Duke University Hospital 319 159 50 4 Anscher MS, Prosnitz LR: Submitted for publication. TABLE 14.—Postprostatectomy radiation therapy before local recurrence No. locally Principal controlled/ Five-yr investigator Reference No. of patients survival, % Follow-up, yr Ray (30) 10/13 57 3-15+ Pilepich 30 18/18 ~45 1-10 Rosen 32) 15/16 ~94 7 (median) Gibbons (28) 21/22 ~73 5-15 Hanks (29) 10/10 86 2-11 Total 74/79 (94%) The national averages for outcome of radiation therapy should be continuously monitored for further evidence of improvement with time and if more appropriate treatment is given with time. Studies should be designed to test the hypothesis that TURP may disseminate cancer. Studies should be developed that are directed at improving the rate of local control and suppressing the rate of metasta- sis in stage C cancer. Urologists and radiation therapists must agree on common staging, end points of outcome reporting, and methodology of data analysis. DISCUSSION Problems of Comparison Comparison of the outcome of external-beam radiation therapy and management of radical prostatectomy of clini- cally localized prostate cancer is confounded by the reporting of different end points and different statistical methods of analysis (including some that are incorrect and misleading), and a general absence of reporting of complications in the literature. One of these problems of comparison is illustrated by the frequently quoted Hopkins series (22) and the recent report of Middleton et al. (24), in which data on patients lost to follow-up are discarded and survival is calculated on those who remain. Thisis a misleading and an improper method of analysis of the data. As an example, data on 18% of the patients in the Hopkins series of 70 and 11% of patients in the Middleton series of 156 patients lost to follow-up were discarded and treated in this manner. The data of Gibbons etal. (23) are correctly analyzed and can be used in compari- son to the long-term data reported by radiation therapists. Surgical survival is expected to exceed that of equivalent treatment by irradiation as surgeons treat a generally more healthy subset of patients who will experience less intercur- rent disease mortality. Our patterns of care data also show that Karnofsky functional status is an independent variable in recurrence, which means that stage for stage, patients with low Karnofsky status (included in radiation series) will more likely die of cancer. Optimization of Technical Management With Radiation Therapy Our studies (4,7-9) have shown that prostate cancer must be treated with high-energy linear accelerators, the use of treatment simulation, and the use of complex field arrange- ments and dosimetry if optimal results are to be obtained. Excessively simple treatment programs are clearly associated with increased numbers of complications or recurrence and should not be used. As expected, we (10) believe that radiation dose is a criti- cal determinant for obtaining local control. A significant problem in the United States as a whole has been the use of an excessive dose in early stage prostate cancer that only con- tributes to an excess of morbidity without improving local control. Tumor dose must be adjusted to tumor volume, and it appears that only the patients with locally extensive stage C cancers benefit from radiation doses of 7,000 cGy or more. The optimization of dose and tumor volume has been shown to offer the opportunity for significantimprovements in local control and complications (4,7,10). TABLE 15.—Postprostatectomy radiation therapy after local recurrence Principal No. locally controlled/ investigator Reference No. of patients Survival Percent Ray (30) 11/19 Disease free at 5 yr 40 Disease free at 10 yr 26 Rosen 32) 10/13 Alive at 4 and 13 yr 15 Gibbons 28) 16/23 Alive 5-25 yr 40 (9 of 23) Hanks (29) 8/10 Alive at 5 yr 71 Total 45/65 (70%) MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 82 TABLE 16.—Effect of postoperative irradiation on local recurrence in patients with pathologic stage C prostate cancer Local recurrence No irradiation Postoperative irradiation No. of patients/ No. of patients/ Treatment Investigators at Reference total No. of patients total No. of patients Radiation and ~~ Duke University Hospital a 39/160 3/94 surgery University of Utah Medical Center 33) 39/160 3/94 Virginia Mason Medical Center (28) 39/160 3/94 Radiation only ~~ The Johns Hopkins Hospital (34) 3/88 Stanford University Medical Center (30) 3/88 Washington University Clinics an 3/88 Sacramento (author's practice) (29) 3/88 Joint Center for Radiation Therapy 32) 3/88 University of Southern California (35) 3/88 (36)° 1/45 Surgery only Mayo Clinic 37) 127/556 Memorial-Sloan Kettering Cancer Center 38) 127/556 University of Virginia Hospital 39) 127/556 University of California Hospital (Los Angeles) (40) 127/556 Bowman Gray 41) 127/556 Massachusetts General Hospital (Harvard) 42) 127/556 166/716 (23%) 7/227 (3%) 4 Anscher MS, Prosnitz LR: Submitted for publication. b Authors present an editorial comment without reference. It is important that urologists remain aware of the subset of patients for whom an increased frequency of metastasis and death is recorded when their diagnosis has been established by TUR. This subset represents the group of intermediate and poorly differentiated stage C patients; the effect is not observed in stage B patients or those with well-differentiated tumors of any stage (7,/1,12). Precisely what is responsible for this adverse outcome is not known but is subject to ongoing study. In addition to suggesting that dissemination of the cancer cells follows TURP, Hoffman et al. (43) believe that this subgroup of patients has an increased frequency of lymph node metastasis. This latter observation awaits con- firmation by researchers at a second institution. Conversely, the patient who has intermediate or poorly differentiated stage C cancer who does not present with obstruction and does not require a TUR has a favorable prognosis, almost TABLE 17.—Relative costs of treatment for cancer of the prostate, Sacramento, California, 1984 Relative costs in dollars Treatment Average Treatment Range Lymph node dissection 13,900 12,000 10,200-26,400 and '2I implant Radical prostatectomy 14,100 14,400 10,000-18,900 External-beam radiation therapy Prior to October 1984 6,650 6,750 6,200-6,750 After October 1984 5,500 5,600 5,300-5,600 (CPT-4)¢ 4 Government-mandated billing change to CPT-4 system. equivalent to the survival of stage B patients (4). As we see reports of the management of stage C patients with various modalities, it will be important to know the distribution of patients who had undergone TURP versus needle biopsy within each series as another variable that could falsely cause treatments to appear different. In prospective studies of stage C prostate cancer, method of diagnosis should be a stratification point. Long-term Outcome With Radiation Therapy The survival and disease-free survival of patients (16,18) treated in the United States is excellent and approxi- mates that of single-institution reports, even though this national average is composed of investigators who obtain results better and worse than the national average. Appar- ently, the technology and methodology of treatment of pros- tate cancer with external-beam radiation therapy have been well distributed throughout the United States, even though we have clearly identified certain types of practice and equipment that are associated with poor results. Major ques- tions remain as to whether the technology and surgeons’ expertise to perform radical prostatectomy, and particularly the nerve-sparing radical prostatectomy, have had equal dis- tribution around the United States, and whether those com- plex procedures are performed with acceptable morbidity and cure in facilities outside the half-dozen from which researchers report their results in the literature. It is possible that the improvement in survival observed in stage C patients between 1973 and 1978 is due to the improvement of therapy technique and dose selection that we (4) have shown to occur. One can never exclude the possibil- ity that stage migration has accounted for this improvement, NCI MONOGRAPHS, NUMBER 7, 1988 but when we evaluated the difference in survival against the distribution of independent variables in the 2 series, we found no explanation for the improved results. This is in contrast to what we have seen in stage B cancer, i.e., the improvement in results between 1973 and 1978 is explained by increased representation of low-grade tumors in 1978. Results of Radiation Therapy in Candidates for Radical Surgery Our studies with early stage B (T1) tumors show only a 10% decrease in survival at 10 years over that expected for this age distribution of patients, and we have also shown that, within this group of patients, there were 40 with poorly differentiated histology, of whom 30% expressed initial fail- ure as metastasis in the first 5 years. This latter group should certainly be screened by noninvasive techniques for lymph node metastasis, but, if finding small amounts of lymph node metastasis at surgery would not alter the treatment the patient receives, then it does not seem worthwhile for the physician to proceed with lymph node dissection and its associated 10% morbidity. (Survivorship at 10 yr, which is only 10% below expected, would seem comparable to that in the surgical series equaling the 10-yr expected survival after all patients with lymph node metastases are eliminated.) In our survey of patients treated in 1978, we identified 65 patients who are directly comparable to patients receiving radical surgery because their health allowed lymph node dissections; the results of those node dissections were negative. These patients show a nonprogression rate at 5 years that is the same as many radical surgery series and indeed the same as reports of the surgical arm of the Uro- Oncology Research Group prospective trial. Morbidity of Radiation Therapy Radiation sequelae are modest, and we know there has been a reduction in morbidity for patients treated more recently as opposed to 10-15 years ago. We would expect a similar reduction in surgical morbidity as a result of improvement in technology and general supportive care. The complications of surgery and radiation are quite different, and the patient is going to need consultation with specialists of both disciplines before he can evaluate what complications are acceptable or unacceptable and to select the procedure he wishes. Potency appears to be better preserved after radiation ther- apy than after the nerve-sparing prostatectomy. Our current data indicate that 73% of truly potent patients retain their potency, and Bagshaw et al. (44) have demonstrated that over 80% of their patients retained potency. The 2 earlier reports of the nerve-sparing operation show 52% and 72% of the patients maintaining potency (26,27). Radiation as an Adjuvant in Pathologic Stage C Patients We have shown that it might take as many as 150 patients in each arm of a prospective random trial to demonstrate an overall advantage of administering external-beam therapy following radical prostatectomy for pathologic stage C dis- ease. The local control rates observed in data from nearly 1,000 patients tabulated from the literature indicate a consis- tent difference in local recurrence of roughly 25% versus 5% when postoperative radiation is added. It also seems clear from MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 83 the literature that the patient with seminal vesicle involvement has a greater risk of local recurrence than does the patient with tumor at the surgical margins, and it would seem that these 2 groups of patients should receive adjuvant radiation therapy. The advantage of administering adjuvant therapy following maximal recovery of urinary function and before local recurrence has occurred is that doses of 6,000 cGy can be administered with a high local control rate and low mor- bidity. If one waits until there is gross recurrence of the cancer, then doses in the range of 7,000 cGy are needed and morbidity is markedly higher. Research Needs Clear research needs exist for improvement of our treat- ment of prostate cancer, an understanding of this cancer’s particular biology, and documentation of the availability of high quality care to everyone in the United States. For these goals to be achieved, the Federal Government and private sources must increase funding of clinical and basic research in prostate cancer to a level appropriate for this major national health problem that affects 90,000 and kills 25,000 men each year. REFERENCES (1) KRAMER S, HERRING D. The Patterns of Care Study: A nationwide evaluation of the practice of radiation therapy in cancer management. Int J Radiat Oncol Biol Phys 1976;1:1231-1236. (2) KRAMER S. The study of the patterns of cancer care in radia- tion therapy. Cancer 1977;39:780-787. (3) KRAMER S, HANKS GE, DIAMOND JJ. Summary results from the Fourth Facilities Master List Survey conducted by the Patterns of Care Study. Int J Radiat Oncol Biol Phys 1983;9:1881-1883. (4) HANKS GE. Optimizing the radiation treatment and outcome of prostate cancer. Int J Radiat Oncol Biol Phys 1985; 11:1235-1245. (5) HANKS GE, DUNLAP K. A comparison of the cost of various treatment methods for early cancer of the prostate. Int J Radiat Oncol Biol Phys 1986;12:1879-1881. FOWLER JE, WHITMORE WF JR. The incidence and extent of pelvic lymph node metastases in apparently localized pros- tate cancer. Cancer 1981;47:2941-2945. (7) LEIBEL SA, HANKS GE, KRAMER S. Patterns of care outcome studies: Results of the national practice in adenocarcinoma of the prostate. Int J Radiat Oncol Biol Phys 1984; 10:401-409. (8) HANKS GE, DIAMOND JJ, KRAMER S. The need for complex technology in radiation oncology. Cancer 1985;55:2198- 2201. (9) KRAMER S, HANKS GE, MACLEAN CJ. Patterns of failure: Results of the Patterns of Care Study in cancer of the larynx, prostate, and cervix and in Hodgkin's disease. Cancer Treat Symp 1983;2:157-168. (10) HANKS GE, LEBEL SA, KRALL JM, ET AL. Dose response observation for local control of adenocarcinoma of the prostate. Int J Radiat Oncol Biol Phys 1985;11:153-157. (11) HANKS GE, LEIBEL SA, KRAMER S. The dissemination of cancer by TUR of locally advanced prostate cancer. J Urol 1983;129:309-311. (12) PILEPICH MV, KRALL JM, HANKS GE, ET AL. Correlation of pretreatment TUR and prognosis of patients with stage C carcinoma of the prostate treated with definitive radio- (6 ~ 84 therapy—RTOG experience. Int J Radiat Oncol Biol Phys 1987;13:195-199. (13) KUBAN DA, EL-MAHDI AM, SCHELLHAMMER PF, ET AL. The effect of transurethral prostatic resection on the incidence of osseous prostatic metastasis. Cancer 1985;56:961-964. (14) PAULSON DF, Cox EB. Does transurethral resection of the prostate promote metastatic disease? J Urol 1987;138: 90-91. (15) FOWLER JE, FISHER HAG, KAISER DL, ET AL. Pelvic lymph node metastases and probability of survival without distant metastases in patients treated with '?’ implantation for localized prostate cancer: Influence of pretreatment TUR of the prostate. Presented at the American Urological Associa- tion Annual Meeting, Kansas City, MO, May 16-20, 1982, abstr 243. (16) HANKS GE, DIAMOND JJ, KRALL JM, ET AL. A ten-year follow-up of 682 patients treated for prostate cancer with radiation therapy in the United States. Int J Radiat Oncol Biol Phys 1987;13:499-505. (17) FEINSTEIN AR, SOSIN DM, WELLS CK. The Will Rogers phe- nomenon: Stage migration and new diagnostic techniques as a source of misleading statistics for survival in cancer. N Engl J Med 1985;312:1604-1608. (18) HANKS GE, KRALL JM, MARTZ KL, ET AL. The outcome of treatment of 313 patients with T-1 (UICC) prostate cancer treated with external beam irradiation. Int J Radiat Oncol Biol Phys 1988;14:243-248. (19) PAULSON DF, LIN GH, HINSHAW W, ET AL. Radical surgery versus radiotherapy for adenocarcinoma of the prostate. J Urol 1982;128:502-504. (20) SCARDINO PT, CARLTON CE JR. Combined interstitial and external irradiation for prostate cancer. In Principles & Management of Urologic Cancer (Javadpour N, ed), 2nd ed. Baltimore: Williams & Wilkins, 1983, pp 392-408. (21) BYHARDT RW, JENSEN R, ROSWIT B, ET AL. Letter. J Urol 1983;130:1205-1206. (22) WALSH PC,JEWETT HJ. Radical surgery for prostatic cancer. Cancer 1980;45:1906-1911. (23) GIBBONS RP, CORREA RJ JR, BRANNEN E, ET AL. Total prosta- tectomy for localized prostatic cancer. J Urol 1984; 131:73-76. (24) MIDDLETON RG, SMITH JA JR, MELZER RB, ET AL. Patient survival and local recurrence rate following radical prosta- tectomy for prostatic carcinoma. J Urol 1986;136:422-424. (25) ELDERIJS,JEWETT HJ, WALSH PC. Radical perineal prostatec- tomy for clinical stage B2 carcinoma of the prostate. J Urol 1982;127:704-706. (26) EGGLESTONJC, WALSH PC. Radical prostatectomy with pres- ervation of sexual function: Pathological findings in the first 100 cases. J Urol 1985;134:1146-1148. (27) CATALONA WJ, DRESNER SM. Nerve sparing radical prosta- tectomy: Extraprostatic tumor extension and preservation of erectile function. J Urol 1985;134:1149-1151. (28) GIBBONS RP, COLE BS, RICHARDSON RG, ET AL. Adjuvant radiotherapy following radical prostatectomy: Results and complications. J Urol 1986;135:65-68. (29) HANKS GE,DAWSON AK. The role of external beam radiation therapy after prostatectomy for prostate cancer. Cancer 1986,58:2406-2410. (30) RAY GR, BAGSHAW MA, FrREIHA F. External beam radiation salvage for residual or recurrent local tumor following radi- cal prostatectomy. J Urol 1984;132:926-930. (31) PILEPICH MV, WALTZ BJ, BAGLAN RJ. Postoperative irradia- tion in carcinoma of the prostate. Int J Radiat Oncol Biol Phys 1984;10:1869-1873. (32) ROSEN EM, CASSADY JR, CONNOLLY J, ET AL. Radiotherapy for localized prostate cancer. Int J Radiat Oncol Biol Phys 1984;10:2201-2210. (33) JACOBSON GM, SMITH JA JR, STEWART JR. Postoperative radiation therapy for pathologic stage C prostate cancer. Int J Radiat Oncol Biol Phys 1987;13:1021-1024. (34) FORMAN JD, WHARAM MD, LEE DJ, ET AL. Definitive radio- therapy following prostatectomy: Results and complica- tions. Int J Radiat Oncol Biol Phys 1986;12:185-189. (35) SKINNER DG, LIESKOVSKY G. Carcinoma of the prostate: An opinion on management of early stage disease with a com- mentary on the meaning of capsular penetration. J Urol 1985;134:1183-1184. (36) SKINNER L. Editorial comment. J Urol 1985;134:891. (37) MYERS RP, FLEMING TR. Course of localized adenocarci- noma of the prostate treated by radical prostatectomy. Pros- tate 1983;4:461-472. (38) WHITMORE WF JR, MACKENZIE AR. Experiences with var- ious operative procedures for the total excision of prostatic cancer. Cancer 1959;12:396-405. (39) SCHELLHAMMER PF. Radical prostatectomy for carcinoma of the prostate and 15 year analysis of survival and local control. J Urol 1986;135 (Part 2):247A. (40) BELT E, SCHROEDER FH. Total perineal prostatectomy for carcinoma of the prostate. J Urol 1972;107:91-96. (41) TOMLINSON RL, CURRIE DP, BoyCE WH. Radical prostatec- tomy: Palliation for stage C carcinoma of the prostate. J Urol 1977;117:85-87. (42) VICKERY AL JR, KERR WS JR. Carcinoma of the prostate treated by radical prostatectomy. A clinicopathologic sur- vey of 187 cases followed for 5 years and 148 cases fol- lowed for 10 years. Cancer 1963;16:1598-1608. (43) HOFFMAN GS, SCARDINO PT, CARLTON CE JR. Effect of TURP on survival and dissemination of disease in prostatic cancer. Presented at the American Urological Association Annual Meeting, Las Vegas, Apr 17-21, 1983. (44) BAGSHAW MA, Cox RS, RAY GR. Status of radiation treat- ment of prostatic cancer at Stanford University. NCI Monogr 1988;7:47-60. Definitive Radiation Therapy in Carcinoma of the Prostate Localized to the Pelvis: Experience at the Mallinckrodt Institute of Radiology Carlos A. Perez,l* Miljenko V. Pilepich,' Delia Garcia,' Joseph R. Simpson,’ Fred Zivnuska,’ Mary Ann Hederman' ABSTRACT —Definitive radiation therapy was administered to 577 patients with histologically confirmed carcinoma of the pros- tate localized to the pelvis between January 1967 and December 1983. All patients were available for a minimal 3-year follow-up, and the median period of observation is 6.5 years. The actuarial survival without tumor in stages A2 and B at 5 years was 78% and at 10 years 60%. In stage C, the corresponding survival figures were 60% at 5 and 40% at 10 years. The overall actuarial survival in stage B patients was 76% at 5 and 56% at 10 years, which is similar to the life expectancy of a comparable cohort of normal males. In stage C, the actuarial survival was 65% at 5 and 35% at 10 years. The pelvic failure rate in stage A2 was 12% (5 of 41), 17% in stage B (31 of 185), 28% (93 of 328) in stage C, and 48% (11 of 23) in stage D1. Distant metastases were noted in 12% of the patients with stage A2, 20% stage B, 42% stage C, and 65% stage D1. In stage B, patients who had control of the pelvic tumor exhibited an 85% actuarial 5-year survival and a 60% one at 10 years. This compares with an actuarial survival of 30% at 5 and 10 years when there was evidence of pelvic recurrence alone or combined with distant metas- tases. In stage C patients with pelvic tumor control, actuarial survival was 81% at 5 and 50% at 10 years, in comparison with 25% at 5 and 10% at 10 years when there was development of pelvic recurrence or distant metastases or a combination of both. There was a strong correlation between the survival and appearance of distant metastases with the histologic degree of differentiation of the tumor in all stages. However, the probability of tumor control in the pelvis was not significantly correlated with this parameter. The administration of hormones concomitantly with radiation ther- apy did not significantly influence the probability of tumor control, appearance of distant metastases, or survival. Major sequelae of therapy were noted in 2.2% of the patients, whereas minor sequelae were observed in approximately 12% of the patients. Radiation therapy has been shown to be an effective therapeutic alternative to radical prostatectomy or hormonal manipulation in patients with carcinoma of the prostate.—NCI Monogr 7:85-94, 1988. Over the past 25 years, radiation therapy has been demon- strated to be an acceptable therapeutic alternative for patients with stages A2, B, and C carcinoma of the prostate. The survival and tumor control are comparable to those observed with radical prostatectomy in patients with early stages and hormonal manipulation or orchiectomy, or both, for more advanced stages. Sequelae of treatment have been acceptable with a good quality of life in most patients (/-6). ! Radiation Oncology Center, Mallinckrodt Institute of Radiology, Washington University School of Medicine, Missouri Baptist Hospital, and St. Luke's Hospital, St. Louis, MO. 2 St. John’s Mercy Medical Center, St. Louis, MO. * Reprint requests to: Carlos A. Perez, M.D., Radiation Oncology Center, Mallinckrodt Institute of Radiology, Suite 411, 4511 Forest Park Blvd., St. Louis, MO 63108. PATIENTS AND METHODS Patients.—Between January 1967 and December 1983, a total of 577 patients with histologically proven adenocarci- noma of the prostate localized to the pelvis were treated with definitive radiation therapy at the Radiation Oncology Cen- ter of the Mallinckrodt Institute of Radiology and affiliated hospitals. All patients have been followed for a maximum of 16 years, minimum of 3 years, with a median follow-up of 6.5 years. All radiation oncology and hospital records were reviewed as required, and the information was coded on computer-compatible forms and analyzed on a VAX 8600 computer. A Biomedical Department Program statistical package was used for computation of the data (7,8). All survivals and survival functions utilize the actuarial life table as applied by Cutler and Ederer (9), and test statistics pro- vided are Generalized Wilcoxon (Breslow), Generalized Savage (Mantel-Cox), and Tarone-Ware (10,11). Trend analysis was performed by the Tarone method (1/2). Each patient had a complete physical examination, rectal examination, routine blood count, chemistry profile, urinaly- sis, and determination of serum prostatic and acid phospha- tases as well as alkaline phosphatases. Chest x-ray, intrave- nous pyelogram, and radioisotope bone scan were also obtained. Initially, the regional lymph nodes were evaluated by pedal lymphangiogram, but later this was replaced by computed tomography scans of the pelvis and abdomen. The patients were staged according to a modification of the American Urological Association classification of 1978 (table 1). Cystoscopy was routinely performed; in most patients, the diagnosis was obtained by needle biopsy, although in about 25% of them, transurethral resection of the prostate provided definitive diagnosis. The tumors were clas- sified according to the degree of histologic differentiation into well, moderately, or poorly differentiated (table 2). Patients were followed periodically during and after ther- apy or until death by the staff of the Radiation Oncology Center, the referring urologist, and occasionally the family physician. When follow-up information was not available, contact was made directly with the patient or relatives. Follow-up was obtained in 98% of the patients, and those lost to follow-up were considered to have died of the disease. Tumor control was assessed by periodic rectal examina- tion, acid phosphatase plasma determinations, and, when indicated, radiographic or radionuclide scan studies. Initially, surgeons performed postirradiation biopsy of the prostate on each patient almost routinely to ascertain the effect of radia- tion on the tumor. However, later this procedure was limited only to patients suspected of recurrence following rectal 85 86 TABLE 1.—Clinical staging of carcinoma of the prostate? Stage Description Al Well-differentiated tumor, incidental finding A2 Poorly differentiated tumor, or more than 5% positive chips found on transurethral resection; incidental finding Bl Tumor < 1.5 cm in diameter or localized to one lobe B2 Tumor > 1.5 cm in diameter or extending beyond one lobe Cl Tumor extending into periprostatic tissues; normal to high acid phosphatase levels C2 Tumor invading seminal vesicles or periprostatic extension, diameter of > 6 cm D1 Tumor extension into bladder or rectum, pelvic disease extending to pelvic wall D2 Abnormal bone scan or survey or distant metastases % This is the staging system used at the Radiation Oncology Center of the Mallinckrodt Institute of Radiology, which is modified from the American Urological Association Classification of 1978. examination (tumor persistence over 18 mo after irradiation or evidence of regrowth). Irradiation techniques.— The methods of treatment with either 22-megavolt photons from a betatron or 16- to 25- megavolt photons from linear accelerators have been de- scribed (1/3). With the betatron, the pelvic tissues received 4,500-5,000 cGy through anterior/posterior-posterior/ante- rior ports (15 X 15 cm for stage Bor 18 X 15 cm for stages C and D1 tumors). Following this, the posterior port was dis- continued and the anterior port decreased to 14 X 14 cm to deliver an additional 1,000 cGy to the external iliac and hypogastric lymph nodes. The anterior/posterior port was further reduced to 6 X 8 cm or 8 X 10 cm (depending on the size of the prostate) for an additional 1,000 cGy, for a total dose of 6,000-6,500 cGy to patients in stages A2 and B or 7,000 cGy to those in stage C. In 1974, a 25-megavolt photon beam from a linear accel- erator was used, and after 4,500 cGy was delivered through a large anterior/posterior-posterior/anterior port, an addi- tional 2,000-2,500 cGy were delivered with 270° anterior 100 D @® o Oo H o Percent Actuarial Survival, NED TABLE 2.— Distribution of patients by clinical stage and pathologic tumor differentiation Stage Degree of differentiation A2 B C Dl Well 12 78 90 3 Moderate 18 78 131 4 Poor or undifferentiated 10 27 102 14 Ungraded 1 2 5 2 Total 41 185 328 23 arc rotation (7 X9- or 8 X 10-cm ports). Starting in 1980, the rotational boost was given with bilateral 120° arcs, with 60° skiped vectors anteriorly and posteriorly. In a few patients, when there was lymphangiographic or surgical evidence of periaortic lymph node metastasis, this volume was treated with a dose of 5,000 cGy in 5 to 6 weeks through anterior/posterior-posterior/anterior ports. The usual daily dose was 180 cGy in 5 weekly fractions. In a small number of patients, irradiation was temporarily dis- continued for 1-2 weeks after the 4 or 5 weeks of treatment because of acute side effects (acute cystitis, proctitis, skin reaction, etc.). RESULTS Survival The overall, actuarial, tumor-free survival according to the various stages of the disease is illustrated in figure 1. In stage B, a 5-year 75% actuarial tumor-free survival and a 10-year 55% rate are similar to the normal life expectancy observed in a comparable cohort of individuals (fig. 2A). In stage C, the 5-year survival was 65% and at 10 years 35%, which is approximately 10%-15% below the normal life expectancy (fig. 2B). Tables 3 and 4 illustrate the direct and adjusted survival for intercurrent disease in patients available for 5 or 10 years of evaluation, respectively. The survival figures closely parallel those noted in the actuarial survival computations. There was a strong correlation between the FIGURE 1.—Tumor-free actuarial (NED) sur- vival by stage for 577 patients with carci- noma of the prostate localized to the pelvis and treated with definitive irradiation at the Mallinckrodt Institute of Radiology. PTS. (41) a - A B, (15) 20 v B, (70) 0 Ci (en ¢ C2 (67) + D0; (23) 1 1 1 1 1 1 1 | 1 J 0 1 > 3 4 5 6 7 8 9 10 Years After Initial Therapy NCI MONOGRAPHS, NUMBER 7, 1988 87 100, 80 g > 60 > 5 wn § o 40 (185 patients) [3 a &—A NED O&A ACTUARIAL 20 ®-8 EXPECTED SURVIVAL 1 1 1 1 1 1 1 1 cd 0 1 2 3 4 5 6 8 9 10 Years After Therapy FIGURE 2.—Survival for 185 stage B patients (A) and 328 stage C patients (B) with car- cinoma of the prostate localized to the pel- vis and comparison with normal life expec- 100 B tancy. NED = no evidence of disease. 80 o E 2 60 2 @ Ta E 8 40 (328 patients) $ A—A NED O—4 ACTUARIAL 8-8 EXPECTED SURVIVAL 20 1 1 ki 1 1 1 1 1 1 J 0 1 2 3 4 5 6 8 9 10 Years After Therapy probability of survival and the histologic degree of tumor differentiation (fig. 3). Seventy-five patients in this series had a staging lymph- adenectomy prior to the initiation of radiation therapy (37 stage B and 38 stage C, of whom 6 and 15 had positive nodes, respectively). The incidence of positive nodes was similar to TABLE 3.—Survival at 5 yr No. of Five-yr direct that reported in other studies (/4). The relapse-free actuarial survival in stage B is comparable to that for patients with negative nodes (fig. 4A). However in stage C, the actuarial 10-year survival was 85% in the 23 patients with negative nodes in contrast to only 17% in 15 patients with positive nodes (fig. 4B). TABLE 4.—Survival at 10 yr Ten-yr direct . . survival } Percent Co No. of survival Percent Clinical patients Death-intercurrent adjusted Clinical patients Death-intercurrent adjusted stage atrisk No. Percent disease survival stage atrisk No. Percent disease survival A2 29 23 79.3 1 82.1 A2 3 2 66.7 1 100.0 Bl1 94 75 79.8 12 91.5 Bl1 30 21 70.0 2 75.0 B2 51 42 823 4 89.4 B2 16 10 62.5 4 83.3 Cl 133 89 66.9 17 76.7 Cl 76 30 395 16 50.0 C2 126 80 63.5 11 69.6 C2 61 20 32.8 10 39.2 D1 21 2 9.5 3 11.1 Dl 12 2 16.7 2 20.0 MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 88 1008 80 g > 60 > 5 wn 7 8 40 a AO WELL DIFFERENTIATED (78 patients) ©—@ MODERATELY DIFFERENTIATED (78 patients) 20[- AA POORLY DIFFERENTIATED (27 patients) 1 1 1 1 1 1 1 1 1 ] 0 1 2 3 4 5 6 7 8 9 10 Years After Therapy FIGURE 3.—Tumor-free actuarial survival by histologic grade for patients with stage B B (A) and stage C (B) disease. 100 80 oo £ > 60 > 5 [9] i 5 40 Lo a AA WELL DIFFERENTIATED (90 patients) 20 @—@ MODERATELY DIFFERENTIATED (131 patients) A—A POORLY DIFFERENTIATED (102 patients) 1 1 ] ] 1 1 1 1 1 J 0 1 2 3 4 5 6 7 8 9 10 Years After Therapy 100A 1002 & sot Q 80 2 2 £ | > 2 60 z 60 & 5 2 a = 40 z 40 w . . 5 9—< Negative Nodes (31 patients) © O—< Negative Nodes (23 patients) & 20}— a—a Positive Nodes (6 patients) & 20[~ ®— Positive Nodes (15 patients) p=.3l p
    2 a eof (8) © 5 = 2 40} = QO 2 (5) (4) (3) (2) m m Qa 20F 1 1 1 1 1 1 1 1 1 a 0 1 2 3 4 5 6 7 8 9 10 Years After Failure FIGURE 6.— Actuarial survival after failure for patients with stage B (A) and stage C TUMOR CONTROL PTS. (B) carcinoma of the prostate. ee NED (124) AA Pelvic Only (24) 100 B #4 Concurrent Pelvic + DM (16) W¥—¥ Pelvic then DM B18 DM then Pelvic #4 DM Only ° > > 5 a ° 5 2 © < < [5] oO © a 5 Years After Recurrence 6 MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 90 probability of survival after failure in the prostate alone in 24 patients was 35% because most of them developed distant metastases later. For patients who had an initial distant fail- ure, alone or combined with pelvic recurrence, the 5-year survival probability was 25%. Patients with stage C tumors and pelvic failure, distally or in combinations, had a 5%-10% probability of survival. Of the stage B patients without pelvic recurrence, 20% developed distant metastases (30 of 154), which is compara- ble to those who had a pelvic failure with or without distant metastases (7 of 31). However, in stage C patients without a pelvic failure, 78 of 235 had significantly less distant metas- tases (33%) in comparison with 60 of 93 patients (65%) with pelvic failure alone or with concomitant distant metas- tases (P=.81 and P<.01, respectively) as shown in figure 7. Tumor Control and Survival Correlated With Concomitant Hormonal Therapy The concomitant use of hormonal manipulation, either orchiectomy, administration of exogenous estrogens (usually 3-5 mg diethylstilbestrol daily), or both, in conjunction with radiation therapy, had no significant impact on the incidence of pelvic recurrence, distant metastases, or survival (fig. 8,9). Causes of Death The cause of death was known to be related to tumor in 140 patients, to myocardial infarction or cerebral or other vascular accidents in 50, and in 53 to other intercurrent diseases or senility (table 5). Stage B | Stage C | I 70 p=.8l p<.0l 93 I ‘0 I I I $ sof | a I oO 5 I 2 | 2 of c | 60 > | 3 | 2 sof ro] I 8 [ 5 31 | & I 20} I I 7 I 10 A : | | | Al A 0 No Pelvic Pelvic DM No Pelvic Pelyict DM Failures Failures Failures Failures FIGURE 7.—Incidence of distant metastases (DM) correlated with pelvic tumor control for patients with stages B and C disease. In stage C, the difference is statistically significant (P < .01). TABLE 5.—Causes of death Disease No. of patients Cerebrovascular accident 14 Myocardial infarction 23 Other vascular accidents 13 Other or unknown 53 Tumor 140 Site of failure Pelvis 14 Pelvis and distant metastases 59 Distant metastases only 67 Correlation of Technical Parameters With Tumor Control in the Pelvis Patients at various times received slightly varying doses of irradiation because of some changes in technique and calibra- tion procedures or differing philosophy of treatment over the 20 years of analysis. Figure 10 illustrates that 4 of 13 patients with stage B (31%) receiving less than 6,000 cGy developed a pelvic failure, compared with 27 of 172 patients (16%) who developed local recurrence with higher doses (P = .24), but 20 of 50 patients (40%) with stage C tumors treated with less than 6,500 cGy developed a pelvic recurrence, in comparison to 73 of 278 (27%) treated with 7,000 cGy. These differences are almost statistically significant (P = .07). No significant correlation was seen between the doses of irradiation given to patients in stage B and survival. At 5 years, the disease-free survival in the patients in stage C receiving 7,000 cGy was 50% compared with approximately 40% of those treated with doses between 6,000 and 6,500 cGy. However at 10 years, the difference was not significant between the 2 groups (data not shown). There was no significant correlation of the tumor control and the size of the reduced fields used for the prostatic boost in either stage B or C tumors (6 X 8-cm, 7 X 9-cm, or 8 X 10-cm ports). Sequelae of Therapy Major sequelae of irradiation were noted in 13 patients (2.2%), consisting mostly of small bowel obstruction (4), vesical fistula (2), hemorrhagic cystitis (2), and ureteral stric- ture (2). One patient treated in 1968 with anterior reducing portals received 7,800 cGy to the pubic bone and developed localized bone necrosis about the pubic symphysis (table 6). TABLE 6.—Major definitive complications from radiation therapy in 577 patients Complications No. occurring Small bowel obstruction 4 Proctitis requiring a colostomy Fatal enteritis? Rectovesical fistula? Cystitis Bladder fistula Ureteral stricture Pubic bone necrosis —0 em DN DN 4 This acute syndrome developed after a total dose of 4,020 cGy. b One patient also had a recurrent pelvic tumor. NCI MONOGRAPHS, NUMBER 7, 1988 91 100A 801 oO w Zz @ 60 > 2 = an EO) PTS. oO 3 A—A XRT ALONE - NO HORMONES (160) ¥—¥ HORMONES + XRT (13) 20 A&A PREVIOUS HORMONES +XRT >90 DAYS (12) 1 1 1 1 1 1 1 1 1 1 0 1 2 3 4 5 6 7 8 9 10 Years After Therapy FIGURE 8.—Tumor-free survival for patients with stage B (A) and stage C (B) disease 8 correlated with concomitant hormonal 100 therapy. 80 oO Ww Zz g 60 > ‘S 5 a 8 40 PTS. S AA XRT ALONE - NO HORMONES (238) a ¥—¥ HORMONES + XRT (54) Nfl eeellid sok A—{ PREVIOUS HORMONES + XRT >90 DAYS (36) 1 1 1 1 1 1 1 1 1 J 0 1 2 3 4 5 6 7 8 9 10 Years After Therapy Minor sequelae of therapy (table 7) were noted in approxi- mately 12% of the patients and consisted of transient proctitis and hemorrhagic cystitis; a few patients experienced other gastrointestinal side effects and were treated conservatively. Urethral stricture was noted in 4.9% of the patients on whom Stage A; Stage B 60 Stage C 90 50 m Distant Metastases 238 Ea Pelvic + DM B Pelvic W 40 x S = 3 < “30 = Zz wl & i 20 10 24 4 10= 23 0 No Hormones No Hormones Hormones Hormones Hormones No Hormones FIGURE 9.—Patterns of failure correlated with concomitant hormonal ther- apy for stages A2, B, and C tumors. MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER a transurethral resection was performed in contrast to 3.6% of those not undergoing this procedure. Leg edema was observed in 8 of 83 (9.6%) patients on whom a staging laparotomy was performed, compared with 1 of 494 (0.2%) without this procedure. Impotence was noted in 82 of 210 patients (39%) recorded to be potent prior to initiation of radiation therapy. or% p=.24 p=.07 o o ~n o PERCENT PELVIC FAILURES S PERCENT PELVIC RECURRENCES 0 <6,500 6,501-7000 DOSE TO PROSTATE (£5%) cGy 5,500-6,000 6,001-6,750 DOSE TO PROSTATE (£5%) cGy FIGURE 10.—Correlation of pelvic recurrence and dose of irradiation for patients with stages B (A) and C (B) carcinoma of the prostate. 92 TABLE 7.—Minor definitive complications from radiation therapy in 577 patients Complications No. occurring Proctitis 29 Rectal ulcer 1 Rectal stricture 1 Enteritis 7 Anal stricture/fissure 2 Cystitis/hematuria 14 — Ureteral stricture Urinary incontinence After transurethral resection 2/164 (1.2%) No transurethral resection 4/413 (1%) Urethral stricture 23 After transurethral resection 8/164 (4.9%) No transurethral resection 15/413 (3.6%) Scrotal/penile edema 8 Leg edema After staging laparotomy No staging laparotomy aN \o 8/83 (9.6%) 1/494 (0.2%) Subcutaneous fibrosis 2 Scrotum, skin necrosis 1 Bladder ulcer 1 Impotence 82/210 DISCUSSION Multiple reports have described results with radiation ther- apy comparable to those obtained with radical prostatectomy (14,16-19) in patients with stages A2 and B, or with hor- monal manipulation in patients with stage C carcinoma of the prostate (/-4,6). Obviously, direct comparison of the vari- ous series is not possible because of patient selection, diver- sity of staging, pathologic classifications, and different eval- uation end points. Only 1 prospectively randomized study has been published (20) on 97 patients with operable carcinoma of the prostate (stage A2 or B), with better survival noted at 5S years in the patients treated with radical prostatectomy. Unfortunately, this study has limited patient accrual and no prospective stratification for degree of differentiation of the tumor or stage; for unknown reasons, the number of patients in the surgical group is only 41 with 56 in the radiotherapy arm. The follow-up was short (5 yr) and the data have not been available for independent review. Finally, 11 of the failures in the radiation therapy group were described as abnormal bone scans and 3 as progressive elevation of acid phosphatase only, both of which are unrelated to local- regional treatment. In a recent article, Pilepich et al. (21) compared the results of researchers at several institutions on patients treated with external-beam irradiation that were comparable to those included by Paulson in his study; the survival obtained with irradiation is similar to that reported with radical prostatec- tomy. Furthermore, the 5- and 10-year survival observed after definitive irradiation is higher than that reported by Thomp- son (22) and Hanash et al. (23) with transuretheral resection alone in patients with probably more advanced tumors. Hanash et al. (23) reviewed the data on 200 patients with histologically proven carcinoma of the prostate treated with transurethral resection of the prostate only between 1934 and 1942 (probably the same population studied by Thompson). Patients were not clinically staged. In those with clinically latent (occult) tumors, the 5-year survival was about 50%, and the 10-year survival for all histologic grades was 30%. This was similar to the expected survival for a comparable normal population. In contrast, patients with clinically manifest tumors had a 5-year survival of 20% for grades I, II, and III and less than 5% for grade IV. The 10-year survival was less than 10% for all grades, significantly below the expected normal survival (about 40%). The 5-year survival of 75% for patients with stage B and 58% for stage C tumors after definitive irradiation is compar- able to rates reported by the Veterans Administration Coop- erative Urological Research Group with prostatectomy in the first group and hormonal manipulation in the second (20,24). Thus it appears that radiation therapy is a reasonable therapeutic alternative for patients with localized carcinoma of the prostate. Irradiation has been noted to yield control of the prostate tumor in approximately 85%-90% of the patients with stages A2 and B tumors and in 70%-75% of those with stage C lesions (/-3). Our figures are slightly lower than those reported by others, probably because of the longer follow-up on our patients; yet, they are comparable to those reported by Bagshaw and co-workers (/) in the Stanford experience. We have noted that patients in whom control of the tumor in the pelvis is observed have a lower probability of develop- ing distant metastases and better survival. Whether this represents a better immune response of the patient or prolif- erating characteristics of the tumor cannot be elucidated from our data. Furthermore, it is possible that pelvic failures antecede distant dissemination in a small proportion of the patients. Nevertheless, in our experience, pelvic tumor con- trol is a critical parameter affecting the prognosis of these patients. The doses of irradiation given to patients in stage B did not significantly correlate with probability of tumor control in the pelvis. However in stage C, higher doses of irradiation to the prostate resulted in better pelvic tumor control. Hanks et al. (25) reported a better 4-year, recurrence-free survival in 574 patients with various stages of carcinoma of the prostate treated with doses over 6,500 cGy. In general, these dose levels are accepted as optimal because, as noted by Hanks and associates (2) in the Patterns of Care Study analyses of patients treated with irradiation for carcinoma of the pros- tate, the major complication rate was 3.5% with doses below 7,000 cGy in comparison with 7% with higher doses. These observations coincide with those reported by us (26) in carci- noma of the uterine cervix, in which doses above 7,500-8,000 cGy even to small volumes were associated with a signifi- cantly greater incidence of treatment sequelae in the bladder or the rectum. The volume of tissue to be treated, as outlined by Pilepich and associates (27), is imperative to achieve optimal tumor control. In the present series, we did not observe a significant correlation between the sequelae of therapy and doses of irradiation given. The development of sexual impotence is a disturbing and vexing sequela of therapy that greatly affects the quality of life of those patients who were potent at the time irradiation was initiated. We (28) performed serial plasma testosterone and dihydrotestosterone determinations before and after radiation therapy without significant NCI MONOGRAPHS, NUMBER 7, 1988 changes noted. Other possible causes of this undesirable sequela, such as vascular injury impairing the blood flow to the corpora cavernosa, radiation injury to the neurovascular autonomic branches governing erection, or psychologic fac- tors, need to be elucidated further in carefully conducted prospective studies (29,30). Previous reports have shown that concomitant administra- tion of hormonal therapy with irradiation did not significantly affect survival or patterns of failure in these patients (14,31,32). Our analyses confirm this observation, and we reaffirm our recommendation to treat patients with localized carcinoma of the prostate with definitive radiotherapy alone, reserving hormonal manipulation for those who subsequently develop a pelvic tumor recurrence or distant metastases. This concept does not conflict with an approach suggested by Green et al. (33) and Pilepich (personal communication), who advocate a short course of hormonal therapy (3 mo) before irradiation to decrease the volume of the tumor and improve sensitivity of the neoplastic cells to irradiation. Paul- son and co-workers (34) reported on a randomized study in which radiation to the pelvis was compared with delayed hormonal manipulation in 73 patients with stage C tumors. Although the recurrence-free survival was initially higher for the patients treated with irradiation, survival was comparable at 3 years, mostly because of the development of distant metastases. However, longer follow-up was not offered, and the patients’ quality of life was not assessed. CONCLUSIONS Obviously, the need is great for more biologic research leading to a better understanding of the natural history of prostate carcinoma and the response of cells to various therapeutic agents, methods for early detection, more ac- curate staging, and more sensitive procedures to detect early recurrences on follow-up. Furthermore, clinical urologists and oncologists have a critical need for standardized pretreatment evaluation cri- teria, staging system, pathologic classification, and statistical methodology for the analyses of therapeutic outcomes. Ideally, patients should be evaluated conjointly by urolo- gists and radiation oncologists, and they should be presented with explanations of the therapeutic alternatives (effective- ness, morbidity, costs), so that they can participate intelli- gently in the selection of their therapy. We strongly recom- mend properly designed and prospectively randomized studies to resolve existing controversies in the management of these patients, particularly the use of prostatectomy (radi- cal or nerve-sparing retropubic procedures) or irradiation of stages A2 and B lesions. The potential, albeit limited, benefit of pelvic irradiation in patients with metastatic pelvic lymph nodes deserves further evaluation. Because of the high incidence of distant metastases, partic- ularly in patients with stage C tumors, and the high death rate in patients with pelvic failure, itis important that we continue to search for effective cytotoxic therapy to control micro- metastases or overt clinical metastases. In summary, in our experience, definitive radiation therapy has been shown to be an effective therapeutic modality in localized carcinoma of the prostate, with results similar to those reported in surgical series. Biologic or technical factors MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 93 that affect tumor control and prognosis should be clearly identified and, whenever required, the techniques of treat- ment should be modified to enhance therapeutic results. A continued dialogue between the various specialists involved in the management of patients with carcinoma of the prostate is critical if optimal multidisciplinary care is to be fostered. REFERENCES (1) BAGSHAW MA, RAY GR, Cox RS. Radiotherapy of prostatic carcinoma: Long- or short-term efficacy (Stanford Univer- sity experience). Urology 1985;25(suppl 2):17-23. (2) HANKS GE. Optimizing the radiation treatment and outcome of prostate cancer. Int J Radiat Oncol Biol Phys 1985; 11:1235-1246. (3) PEREZ CA. Carcinoma of the prostate. A vexing biological and clinical enigma. Int J Radiat Oncol Biol Phys 1983;9:1427-1438. (4) EL-MAHDI AM, KUBAN DA, SCHELLHAMMER PF. The treat- ment of choice for localized poorly differentiated adenocar- cinoma of the prostate. Am J Clin Oncol 1985;8:477-480. (5) PILEPICH MV, KRALL JM, SAUSE WT, ET AL. Correlation of radiotherapeutic parameters and treatment related morbid- ity in carcinoma of the prostate—analysis of RTOG study 75-06. Int J Radiat Oncol Biol Phys 1987;13:351-357. (6) TAYLOR WIJ, RICHARDSON RG, HAFERMANN MD. Radiation therapy for localized prostate cancer. Cancer 197943: 1123-1127. (7) DIXON WIJ, ed. BMDP Statistical Software. Los Angeles: Univ Calif Press, 1982. (8) DixoN WJ, ed. Release of BMDP Statistical Software. Los Angeles: Univ Calif Press, 1986. (9) CUTLER SJ, EDERER F. Maximum utilization of the life table method in analyzing survival. J Chronic Dis 1958; 8:699-713. (10) MILLER R. Survival Analysis. New York: Wiley, 1981. (11) TARONE R, WARE J. On distribution free tests for equality of survival distributions. Biometrika 1977;64:156-160. (12) TARONE R. Tests for trend in life table analysis. Biometrika 1975;62:679-682. (13) PEREZ CA, WALZ BJ, ZIVNUSKA FR, ET AL. Irradiation of carcinoma of the prostate localized to the pelvis: Analysis of tumor response and prognosis. Int J Radiat Oncol Biol Phys 1980;6:555-565. (14) MIDDLETON RG, SMITH JA JR, MELZER RB, ET AL. Patient survival and local recurrence rate following radical prosta- tectomy for prostatic carcinoma. J Urol 1986;136:422-424. (15) PEREZ CA, PILEPICH MV, ZIVNUSKA F. Tumor control in definitive irradiation of localized carcinoma of the prostate. Int J Radiat Oncol Biol Phys 1986;12:523-531. (16) ELDER JS, GIBBONS RP, CORREA RJ JR, ET AL. Efficacy of radical prostatectomy for stage A2 carcinoma of the pros- tate. Cancer 1985;56:2151-2154. (17) CATALONA WJ, DRESNER SM. Nerve-sparing radical prosta- tectomy: Extraprostatic tumor extension and preservation of erectile function. J Urol 1985;134:1149-1151. (18) JEWETT HJ. Radical perineal prostatectomy for palpable, clin- ically localized, non-obstructive cancer: Experience at The Johns Hopkins Hospital 1909-1963. J Urol 1980;124: 492-494, (19) GIBBONS RP, MASON JT, CORREA RJ JR, ET AL. Carcinoma of the prostate: Local control with external beam radiation therapy. J Urol 1979;121:310-312. 94 (20) PAULSON DF, LIN GH, HINSHAW W, ET AL. Radical surgery versus radiotherapy for adenocarcinoma of the prostate. J Urol 1982;128:502-504. (21) PILEPICH MV, BAGSHAW MA, ASBELL SO, ET AL. Definitive radiotherapy in resectable (stage A2 and B) carcinoma of the prostate—results of a nationwide overview. Int J Radiat Oncol Biol Phys 1987;13:659-663. (22) THOMPSON GJ. Transurethral resection of malignant lesions of the prostate gland. JAMA 1942;120:1105-1109. (23) HANASH KA, Utz DC, Cook EN, ET AL. Carcinoma of the prostate: A 15-year followup. J Urol, 1972;107:450-453. (24) PAULSON DF, HODGE GB JR, HINSHAW W, ET AL. Radiation therapy versus delayed androgen deprivation for stage C carcinoma of the prostate. J Urol 1984;131:901-902. (25) HANKS GE, LEIBEL SA, KRALL JM, ET AL. Patterns of care studies: Dose-response observations for local control of adenocarcinoma of the prostate. Int J Radiat Oncol Biol Phys 1985;11:153-157. (26) PEREZ CA,BREAUX S, BEDWINEK JM, ET AL. Radiation therapy alone in the treatment of carcinoma of the uterine cervix. II. Analysis of complications. Cancer 1984;54:235-246. (27) PILEPICH MV, PRASAD SC, PEREZ C A. Computed tomography in definitive radiotherapy of prostatic carcinoma. Part I. Definition of target volume. Int J Radiat Oncol Biol Phys 1982;8:235-240. (28) GRIGSBY PW, PEREZ CA. The effects of external beam radio- therapy on endocrine function in patients with carcinoma of the prostate. J Urol 1986;135:726-727. (29) MITTAL B. A study of penile circulation before and after irradiation in patients with prostate cancer and its effect on impotence. Int J Radiat Oncol Biol Phys 1985;11: 1121-1125. (30) WALSH PC, LEPOR H, EGGLESTON JC. Radical prostatectomy with preservation of sexual function: Anatomical and pathological considerations. Prostate 1983;4:473-485. (31) VAN DER WERF-MESSING B, SOUREK-ZIKOVA V, BLONK DI. Localized advanced carcinoma of the prostate: Radiation therapy versus hormonal therapy. Int J Radiat Oncol Biol Phys 1976;1:1043-1048. (32) NEGLIA WJ, HUSSEY DH, JOHNSON DE. Megavoltage radia- tion therapy for carcinoma of the prostate. Int J Radiat Oncol Biol Phys 1977;2:873-882. (33) GREEN N, MELBYE RW, GEORGE FW III, ET AL. Radiation therapy of inoperable localized prostatic carcinoma: An assessment of tumor response and complications. J Urol 1974;111:662-664. (34) PAULSON DF, CLINE WA JR, KOEFOOT RB JR, ET AL. Extended field radiation therapy versus delayed hormonal therapy in node positive prostatic adenocarcinoma. J Urol 1982; 127:935-937. Local Control of Prostate Cancer With Radiotherapy: Frequency and Prognostic Significance of Positive Results of Postirradiation Prostate Biopsy Peter T. Scardino,"* Thomas M. Wheeler? ABSTRACT —The best available data indicate that, although it is imperfect, the postirradiation biopsy performed at a suffi- cient interval after radiotherapy can provide accurate prognos- tic information useful in the determination of the success or failure of radiotherapy in an individual patient as well as the measurement of overall efficacy of any particular radiothera- peutic regimen. Needle biopsy of the prostate was performed routinely in 510 patients with clinical stage A2, B, or C1 prostate cancer treated with a combination of radioactive gold seed implan- tation and external-beam irradiation. Of the 140 patients who had one or more needle biopsies performed 6-36 months after comple- tion of radiotherapy, who had no evidence of local recurrence or distant metastases at the time of biopsy, and who had received no hormonal therapy before documented recurrence of the tumor, 45 (32%) had one or more biopsies positive for cancer. The frequency of positive biopsy results correlated significantly with the size of the local tumor but not with the grade. The correlation between biopsy results and the eventual development of recurrence was highly significant. If any biopsy was positive, 60% of the patients eventu- ally developed local recurrence; if all biopsies were negative, only 19% developed local recurrence during the period of follow-up. The poor prognosis associated with a positive biopsy result was found within almost every subset of stage, grade, or nodal status exam- ined, although the results varied because of the small number of patients in some groups.—NCI Monogr 7:95-103, 1988. Despite numerous reports of the use of radiotherapy to treat localized prostate cancer, there remains no consensus as to how the efficacy of radiotherapy should be measured. Various end points have been used, such as the survival rate, the disease-free survival rate, and the time to development of distant metastases. Because radiotherapy, like surgery, is a form of local or regional therapy, the best result that can be expected is complete eradication of tumor within the treated field. Consequently, the most direct measure of the efficacy of radiotherapy is the rate of local treatment failure, and the only unequivocal proof of local treatment failure is progres- sive growth of the primary tumor. In any series of patients treated with radiotherapy for local- ized prostate cancer, the actual incidence of local treatment failure is difficult for one to identify and tends to be obscured ! Scott Department of Urology, Baylor College of Medicine; Urology Services of The Methodist Hospital and St. Luke’s Episcopal Hospital, Houston, TX. 2 Scott Department of Urology, Baylor College of Medicine; Department of Pathology, The Methodist Hospital, Houston, TX. * Reprint requests to: Peter T. Scardino, M.D., Scott Department of Urology, Baylor College of Medicine, Suite 1003, 6560 Fannin St. Houston, TX 77030. by the long natural history of the disease, the shortened life expectancy of the patients, and the effects of other treatments so often used, especially hormonal therapy (fig. 1). Because of these factors, investigators have sought other, more imme- diate, measures of the ability of radiotherapy to eradicate the local tumor, such as palpation (digital rectal examination), which is notoriously subjective (/), and postirradiation biopsy of the prostate, which remains controversial. In this review, we will outline the problems of accurately assessing the local treatment failure rate after radiotherapy, describe our results of postirradiation needle biopsy of the prostate after a combination of gold seed implantation and external-beam irradiation, and summarize previous reports of the prognostic value of postirradiation biopsy results, dem- onstrating that, though it is imperfect, the postirradiation biopsy performed at a sufficient interval after radiotherapy provides accurate prognostic information that can be used to determine the success or failure of radiotherapy in an individ- ual patient as well as the overall efficacy of any particular radiotherapeutic regimen. BACKGROUND Disease Progressive growth of the primary tumor (local recur- rence) after radiotherapy is best defined as a morbid clinical phenomenon manifested, e.g., by obstruction of the bladder outlet or ureter, and proved by biopsy. Local recurrence may not become apparent for many years. Even untreated local- ized prostate cancer has a long natural history [(2); Thorsen BJ, Fritjoffsen A: Personal communication]. Treatment with radiotherapy or hormonal therapy may delay the appearance of local recurrence even further (3-5). In patients treated with gold seed implantation and external-beam radiotherapy at hospitals affiliated with the Baylor College of Medicine, local treatment failure developed progressively up to 12 years after therapy (fig. 2). Consequently, the local failure rate reported in a given series depends heavily on the length of patient follow-up. Clinical staging of patients with localized prostate cancer is inaccurate. Understaging occurs far more frequently than overstaging. Many patients with apparently localized tumors have unsuspected pelvic lymph node metastases or distant metastases at the time of treatment that will shorten their lives and alter their treatment programs when hormonal therapy is introduced to control symptomatic metastases. These patients will not be eligible to manifest local recur- rence. In a calculation of the local recurrence rate, they would 95 96 Life Expectancy Technique of of the Patients Radiotherapy LOCAL TREATMENT FAILURE Natural History Effects of of the Disease Other Treatments FIGURE 1.—Four major factors influencing apparent rate of local treatment failure after radiotherapy (see text). remain part of the denominator, but they would not become part of the numerator. Patients Because prostate cancer is a disease of older men, who have a high incidence of comorbid conditions, many patients who present with clinically localized prostate cancer will die of other causes before the appearance of signs or symptoms of local recurrence. In our 510 patients with clinically localized prostate cancer treated with radiotherapy, the mean age was 64 years (6). These patients had follow-up for an average of 8.6 years; 22% died of prostate cancer and 21% died of other causes (table 1). Hormonal Therapy When hormonal therapy (androgen ablation) is used in addition to radiotherapy, the effect on the primary tumor is profound. This is a phenomenon witnessed by every urologist who administers hormone therapy to patients with metastatic prostate cancer. Often the nodular malignant prostate becomes palpably normal within several months. The degree and rate of shrinkage have been precisely documented by ultrasonography (7). Several investigators have examined the results of biopsy of the prostate after hormonal therapy alone 20 % With 2 Local Recurrence 60 80 Direct calculation: 170/510 = 33% Years FIGURE 2.— Actuarial rate of local treatment failure in Baylor series of 510 patients treated with gold seed implantation and external-beam irradia- tion. At 15 yr, the actuarial rate of local recurrence was 53%. If direct calculation had been used, eventual risk of local treatment failure would have been substantially underestimated because 170 (33%) of 510 pa- tients followed for a mean of 8.6 yr developed local recurrence. TABLE 1.—Cancer status and vital status by clinical stage Percent dead of: Clinical No. of Positive Any Prostate Other stage patients nodes, % recurrence, % cancer causes A2 +B 393 25 42 17 21 C1 117 44 74 39 18 Total 510 30 49 22 21 for patients with locally extensive or metastatic prostate cancer and have reported that the biopsy will be converted to negative in 36%-47% of the patients (3,4). The concomitant use of hormonal therapy will profoundly alter the incidence and apparent significance of positive post- irradiation biopsy results after radiotherapy as well as the rate of apparent local recurrence. POSTIRRADIATION BIOPSY RESULTS Patient Population To evaluate the frequency and prognostic significance of positive biopsy results after definitive irradiation, we re- viewed the results of routine needle biopsy of the prostate in our patients with clinical stage A2, B, or C1 prostate cancer (table 2) treated between 1966 and 1979 with a combination of radioactive gold seed implantation and external-beam irradiation (5,8-10). All 510 patients underwent staging pel- vic lymph node dissection so that we could determine the dose and fields of irradiation to be used (table 3). The calcu- lated dose of radiation to the prostate averaged 6,927 + 842 (SD) cGy by the combined technique (table 4) (8,9). The clinical stage and nodal status of all 510 patients are listed in table 5. Following treatment, prostatic tissue was obtained by nee- dle biopsy or transurethral resection from 263 patients. To determine the prognostic significance of the identification of malignant glands in a routine needle biopsy of the prostate, we selected for analysis only those patients who had had one or more needle biopsies performed 6-36 months after the completion of radiotherapy, who had no evidence of local recurrence or distant metastases at the time of biopsy, and who received no hormonal therapy before documented recur- rence of tumor. A total of 170 biopsies were performed TABLE 2.—Staging system for prostate cancer Stage Disease characteristics Al No tumor palpable; <3 microscopic foci of well-differentiated tumor A2 No tumor palpable; >3 foci and/or tumor less than well differentiated BIN Tumor = 1.5 cm; confined to one lobe Bl Intracapsular tumor > 1.5 cm; confined to one lobe B2 Tumor involving both lobes; confined to prostate Cl Tumor extending beyond capsule, with or without invasion of lateral sulci or seminal vesicles; <6 cm in maximum diameter C2 Same criteria as stage C1 disease except tumor >6 cm NCI MONOGRAPHS, NUMBER 7, 1988 97 TABLE 3.—Protocol of combined gold seed implantation and external-beam radiotherapy for patients with no evidence of metastases Gold seed implantation, cGy External-beam radiotherapy, cGy“ Total dose, cGy Pelvic lymph nodes 1966-1979 After 1979 1966-1979 After 1979 1966-1979 After 1979 Negative 2,500-3,000 3,500 4,000 4,500 7,000 8,000 Positive 2,500-3,000 3,500 5,000 5,400 8,000 8,900 4 Radiotherapy to prostate was given 18-21 days after gold seed implantation. among the 140 patients who met these criteria. The clinical stage, nodal status, and grade of tumors in these patients were not significantly different from those in the remaining 370 patients, except that fewer (25%) had positive lymph nodes. The biopsy specimens were reviewed by a single patholo- gist and were considered positive if any neoplastic glands were seen (fig. 3), regardless of the presumed viability of the cells or the extent of changes due to radiation (71). End Points For this analysis, the end point used was time to local treatment failure. Local treatment failure (local recurrence) was defined as a clinical phenomenon causing discrete signs or symptoms, such as hematuria, pelvic pain, ureteral obstruc- tion, bladder outlet obstruction, or a progressively enlarging palpable mass. In each case, these signs or symptoms were confirmed by a tissue diagnosis of persistent prostate cancer. Neither a positive biopsy nor an abnormal digital rectal examination of the prostate was considered evidence of local treatment failure. Results Of the 140 patients, 45 (32%) had one or more biopsies positive for cancer (table 6). The frequency of positive biopsy results correlated significantly with the size (stage) of the local tumor; the larger the tumor was, the greater the fre- quency of positive biopsies, but biopsy results did not corre- late with the grade of the tumor (table 7), perhaps because poorly differentiated tumors, although generally larger, may be more radiosensitive. The correlation between biopsy results and the eventual development of local recurrence (or of any recurrence, local or distant) was highly significant (table 8). If any biopsy was positive, 60% of the patients eventually developed local re- currence; if all biopsies were negative, only 19% developed local recurrence during the period of follow-up. When the results were analyzed according to the actuarial life-table TABLE 4.— Characteristics of 510 patients and type and dose of radiotherapy Characteristic Mean Range Age, yr 64 43-82 Follow-up, yr 8.6 25-171 Mean dose of radiation, cGy Gold seeds 2,602 500-6,000 External-beam 4325 3,000-5,000 Total dose, cGy 6,927 + 842 (SD) 4,500-10,500 method (fig. 4), the probability of local recurrence for patients with a positive biopsy was 52% at 5 years and 72% at 10 years; for those with a negative biopsy, it was only 12% at 5 years and 30% at 10 years (P<<.0001). The poor prognosis associated with a positive biopsy result was found in almost every subset of stage, grade, or nodal status examined, although the results varied because of the small numbers of patients in some groups. Of 105 patients with proven negative lymph nodes (table 9), 27 (26%) had one or more positive biopsy results after irradiation, and 50% of these patients have developed local treatment failure. Among patients carefully matched for the features of local- ized prostate cancer known to have prognostic importance (stage, grade, nodal status), biopsy results still retained their prognostic significance (table 10). For example, we analyzed the results among 47 patients with a small palpable tumor confined to one lobe of the prostate (stages BIN and B1) with proven negative pelvic lymph nodes. Within this subset of patients, the risk of local recurrence was significantly greater for those with a positive biopsy (fig. 5). Of these patients, 57% developed local recur- rence at 5 years, compared with only 8% if the biopsy was negative (P < .02). Biopsy results correlated significantly with the results of the digital rectal examination at the time of biopsy (table 11), but even among patients with a normal digital rectal exami- nation of the prostate, the risk of local treatment failure was significantly greater if the biopsy was positive. At 5 years, 40% of those with a positive biopsy had developed local recurrence, compared with only 7% of those with a negative biopsy (fig. 6). Finally, we tested the hypothesis that a patient with a positive biopsy result 1 year after radiotherapy that was converted to negative at 2 years has the same favorable prognosis as a patient whose initial (1 yr) biopsy result was negative. In our series, 26 patients with a positive biopsy result 1 year after radiotherapy received no subsequent treatment but underwent a second needle biopsy approxi- TABLE 5.—Clinical stage and lymph node status Clinical Pativits Percent with stage No. Percent positive nodes A2 130 25 22 BIN 25 5 8 Bl 140 28 24 B2 98 19 37 Cl 117 23 44 Total 510 100 30 MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 98 "os 4 " ihe SE - 7 Ee, mL o~ — FIGURE 3.—High-power photomicrographs (X315) of radiation atypia (A) and persisten t carcinoma (B). NCI MONOGRAPHS, NUMBER 7, 1988 TABLE 6.— Biopsy results by clinical stage? 99 TABLE 10.—Prediction of local recurrence by biopsy results within each stage in patients with negative lymph nodes Clinical No. of Percent with stage patients positive biopsy Local recurrence A2 22 36 Negative biopsy Positive biopsy EN E 7 Clinical ~~ Total No. No. of No. of B2 25 3) stage of patients patients Percent patients Percent P¢ Cl 37 51 A2 17 12 0 5 60 .003 Total 140 32 BIN 8 7 29 1 0 NS Bl 39 33 12 6 67 .002 a Pol B2 19 14 20 5 0 NS Cl 22 12 25 10 70 .08 TABLE 7.—Biopsy results by grade? Total 105 78 15 27 52 .0001 No. of Percent with 4 NS = not significant. Grade patients positive biopsy I 53 28 II 43 35 1 28 36 TABLE 11.—Correlation of patients’ biopsy results with rectal A : : p 4] = well-differentiated tumor cells; II = moderately differentiated; examination at time of biopsy III = poorly differentiated. Grade was not known for 16 patients. P was not significant. TABLE 8.—Prognostic significance of prostate biopsy in 140 patients? Patients Recurrence, % Biopsy results No. Percent Local Any Negative 95 68 19 3s Positive 45 32 60 76 4p <.0001. TABLE 9.—Biopsy results by stage for patients with negative lymph nodes? No. of Percent with Stage patients positive biopsy A2 17 29 BIN 8 13 Bl 39 15 B2 19 26 Cl 22 45 Total 105 26 p< 02. ors Negative [ (N=95) 0.2} £ oa} u, s | ~ Positive S ose} ST e—a _(N=45) a J L 2. bt 21] p<.0001 1 1.0 lt : 0 5 10 Time to Recurrence (Years) FIGURE 4.— Actuarial analysis showing that probability of clinical local recurrence was significantly greater if result of routine needle biopsy performed 6-36 mo (shaded area) after completion of therapy was posi- tive for cancer. MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER Biopsy results Patients Negative Positive Prostate examination No. Percent No. Percent No. Percent Normal 101 72 81 80 20 20 Abnormal 39 28 14 36 25 64 Total 140 100 95 68 45 32 e P< 0001. 0 -9 Negative | (N=40) 0.2} \ | \ > \ 3 0.4 [ “ $ = = =m = = = £ °5t Positive (N=7) 0.8 _ 1.0 i ———- — . 0 5 10 Time to Recurrence (Years) FIGURE 5.— Among patients carefully matched for volume of tumor (stage B1, negative nodes), positive biopsy result 6-36 mo (shaded area) after therapy was associated with significantly greater risk of eventual local recurrence. 0 Negative (N=81) 0.2 > 2 04 SN == === o © Positive a 06 (N=20) 0.8 1.0 L A — i i i 1 A A i 0 5 10 Time to Recurrence (Years) FIGURE 6.— Among patients with normal digital rectal examination of pros- tate at time of biopsy, positive biopsy result was associated with signifi- cantly greater risk of local recurrence. 100 TABLE 12.—Prognostic significance of 2nd biopsy result in patients with 1st biopsy positive Recurrence Second Patients Local Distant Any“ biopsy result No. Percent No. Percent No. Percent No. Percent Negative 8 31 4 1 13 4 50 Positive 18 69 9 5 28 12 67 ¢ Any recurrence refers to local, or distant, or both. mately 1 year later. In 8 of these patients, the biopsy was converted to negative, but in 18, the result remained positive. However, the risk of subsequent local treatment failure was identical (50%) in these 2 groups (table 12), which indicated that if any biopsy was positive in the 6- to 36-month interval, the prognosis was significantly worse, regardless of subse- quent biopsy results. DISCUSSION The value of postirradiation biopsy results as a measure of local treatment failure has been one of the most controversial topics in urologic oncology. The issues involved are impor- tant and have direct clinical relevance. If postirradiation biopsy results are reliable, they offer an objective way for physicians to determine the efficacy of a particular radio- therapeutic regimen without waiting 10-15 years and, thereby, to eliminate the confounding effects of death from other causes or from metastatic prostate cancer and the fre- quent introduction of hormonal therapy. In our experi- ence (12), the results of postirradiation biopsy performed at a sufficient interval after radiotherapy correlate significantly with the ultimate outcome of therapy. Since we published our initial report (/2), our conclusions have been criticized on the basis that the patients selected for biopsy were not representative of the entire group treated. In fact, we have not been able to find any significant differences in the stage, grade, nodal status, risk of recurrence (either local or distant), or survival rates between the 140 patients who had biopsies and the 370 who did not (fig. 7). Several unselected series in which every patient underwent biopsy have reported results similar to ours (13-15). Other investigators (/6) have questioned the quality of the particular radiotherapeutic regimen that we used, but the long-term survival rates of our patients are virtually identical to those reported after external-beam irradiation (table 13). Among patients with carefully staged disease who had proven negative pelvic lymph nodes and were treated with external-beam irradiation alone (/6) or with combined gold seed implantation and external-beam radiation therapy (our studies), there were no differences in time to appearance of distant metastases or cancer-specific survival rate (fig. 8). In fact, a comparison of postirradiation biopsy results in our patients with those reported from Stanford University Medi- cal Center (16,17) indicates that, on a stage-for-stage basis, positive biopsy results after irradiation were less frequent in our series (table 14). Certainly the meaning of a positive biopsy result after irradiation will depend on the quality of the radiotherapeutic technique used. If few patients ever developed local recur- rence, routine needle biopsies after irradiation would be irrelevant. To date, however, local treatment failure has been a problem in every series in which it has been carefully analyzed and in which the confounding effects of hormonal therapy, death from other causes, and death from prostate cancer have been eliminated. A number of concerns have been raised about the interpre- tation of postirradiation needle biopsies. A biopsy often pro- vides only a random sample of the prostate, especially in patients who have no palpable nodule to be used as a target, and may therefore miss foci of residual tumor, a problem that may be alleviated with the advent of ultrasonically guided biopsy of the prostate. A second concern is the difficulty with histologic interpre- tation of biopsy specimens; radiation-induced atypia may be confused with residual carcinoma in the irradiated prostate (fig. 3). In particular, Cox and Stoffel (/8) and Cox and Kline (19) have questioned the interpretation of postirradiation biopsy, strongly arguing the impossibility of distinguishing viable from nonviable, lethally irradiated cells on histologic grounds, but this issue has been re-examined by Bostwick et al. (11), who have clearly defined the criteria for recog- nizing carcinoma in postirradiation biopsy specimens and dis- tinguishing it from radiation-induced atypia of nonmalignant glands. They have criticized the concept of “nonviable” cancer, which indeed is not a recognized phenomenon in other human tumors, and have argued that this concept SURVIVAL CANCER-SPECIFIC SURVIVAL 100 a 80 60 40 Percent PROBABILITY OF LOCAL RECURRENCE PROBABILITY OF ANY RECURRENCE SN Probability Years FIGURE 7.— Actuarial survival rate, cancer-specific survival rate, probability of any recurrence (local or distant), and probability of local recurrence among patients treated with gold seed implantation and external-beam radiotherapy. Comparison of 140 patients who had a postirradiation needle biopsy (closed circles) of prostate with 379 patients who had no needle biopsy (closed squares) shows no differences between groups. NCI MONOGRAPHS, NUMBER 7, 1988 TABLE 13.—Comparison of actuarial survival rates after external-beam radiotherapy and combined gold seed implantation plus external-beam radiotherapy Actuarial survival rates, %“ Clinical No.of —m@8 MM stage Technique patients 5yr 10 yr 15 yr A2 +B External® 491 814 60+ S 34+ 6 Combined 393 86+t4 59+ 6 28+£13 C External? 407 615 35+ 5 17+ 5 Combined 117 74+8 34+12 17%13 4 Values are means + 2 SE; 95% confidence intervals. b See (16). ¢ Data are from our present study. should no longer be applied to postirradiation biopsy of the prostate. A third major concern is that the time after radiotherapy that a biopsy can be considered reliable has not been clearly defined. Again, Cox and Stoffel (/8), who are among the strongest proponents of the hypothesis that postirradiation biopsy results have no prognostic significance, have argued to a great extent on the basis of their finding that the incidence of positive biopsies decreased with time after radiotherapy. In their series, the incidence of positive biopsies decreased from 65% at 6 months to 19% at 24 months or more. This perplex- ing phenomenon has often been cited as proof that a postirra- diation biopsy less than 24 months after radiotherapy is un- reliable. In our series, we found a similar but more rapid decline with time after radiotherapy (fig. 9), but with a pla- teau by 12 months at a level of about 35% positive biopsies (5). There are several problems with the analysis of Cox and Stoffel; the most damaging to their conclusion is that nearly one-half of their patients received hormonal therapy. Their patients were seen and followed regularly by the urologists at the Walter Reed Army Medical Center, who administered hormonal therapy to many patients before and after radio- therapy. Consequently, we cannot attribute the decreasing A 100 pwr 1 80 — ] 4 1 ‘>. J 60 ve 40 —— Stanford, external beam (N=51) E 20 Baylor, gold seeds (N=254) Percent 100 r TH e.g 80 ‘». 60 Percent 40 —e— Stanford, external beam (N=51) 20 --m- Baylor, gold seeds (N=254) 0 dill ————————__" 0 5 10 15 Years FIGURE 8.—Comparison of time to distant metastases (A) and cancer- specific survival (B) rates among patients with carefully staged disease (stages A2 and B, negative pelvic lymph nodes) treated with external- beam radiotherapy alone (/6) or combined gold seed implantation and external-beam radiotherapy (our series) during comparable years. MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 101 TABLE 14.—Postirradiation biopsy results Gold seeds plus external-beam radiotherapy? External beam radiotherapy alone? Clinical No. of Percent with No. of Percent with stage patients positive biopsy patients positive biopsy A2 23 35 1 0 BIN 9 11 2 0 Bl 48 21 8 38 B2 27 33 22 59 C 40 55 31 74 Total 147 34 64 61 4 Data are from our present study. b See (17). incidence of positive biopsies reported by Cox and Stoffel to the effects of radiotherapy alone; we must include the com- bined effects of radiotherapy and hormonal therapy (3,4). In our series, in which hormonal therapy was not used before documented recurrence of tumor, patients with a positive biopsy at 1 year had a poor prognosis regardless of the results of a second biopsy done 1 year later (table 12). Such late biopsies may simply miss small foci of residual carcinoma, which will eventually regrow and cause local recurrence. Not all patients with a negative biopsy result after irradia- tion remain free of local treatment failure, nor do all patients with a positive biopsy develop local recurrence. There are a number of valid explanations for false-negative biopsy results. If a sampling error occurs, especially in a patient with no palpable nodule, persistent tumor not included in the specimen will eventually grow and cause local tumor recur- rence. In other instances, a negative biopsy may be accurate, but the prostate is left in situ after radiotherapy, because a second primary tumor may develop and be attributed to persistence of the original tumor. Lastly, a false-negative result may be reported if persistent carcinoma is misinter- preted as radiation-induced atypia. False-positive results can also be explained. Some patients with a positive biopsy may not develop local recurrence because of the limited time of follow-up or because of early death from metastatic prostate cancer or other causes. In other patients, metastases develop before clinical local recur- rence, as we have defined it, becomes apparent. These % Postitive 30 Months FIGURE 9.—Incidence of positive postirradiation biopsy results declined with time after radiotherapy, reaching a minimum at 12 mo in our series in which no hormonal therapy was given (closed squares), and 24 mo in the series of Cox and Stoffel (/8) in which hormonal therapy was widely used (closed diamonds). Values in parentheses = numbers of patients who had biopsies at each interval. 102 TABLE 15.—Prognosis based on postirradiation biopsy results after various radiotherapy techniques No. of patients Negative biopsy Positive biopsy Total Recurrence Total Recurrence No. of Technique References patients No. Percent No. Percent No. Percent No. Percent External-beam radiotherapy 13-15,17,20-22 273 126 46 18 14 147 54 75 51 Gold seed + external-beam radiotherapy 12 124 81 65 25 31 43 35 28 65 lodine-125 23-25 109 71 65 15 21 34 35 19 56 Total 506 278 55 58 21 224 45 122 54 patients receive hormonal therapy, which may obscure the manifestation of local recurrence during the remainder of their lives. Lastly, a false-positive result may occur if radiation-induced atypia is misinterpreted as persistent carcinoma. In a review of the literature (/2-15,17,20-25), we have identified 506 patients, including those in our study, who were treated with either external-beam irradiation alone, combined gold seed implantation plus external-beam irradia- tion, or iodine seed implantation (table 15). Of these patients, 45% had a positive biopsy result after irradiation, and 54% of these had evidence of treatment failure at the time of the report (mean, 4 yr), compared with only 21% of the patients with a negative biopsy. Recently, Schellhammer et al. (26) reported their experience with postirradiation biopsy results after external-beam irradiation alone or iodine- 125 implanta- tion in patients who received no hormonal therapy before documented recurrence. The frequency and the prognostic significance of positive biopsy results after irradiation were similar to ours. In summary, the best available data indicate that a physician can perform needle biopsy of the prostate at a sufficient interval after radiotherapy to determine the effi- cacy of a particular radiotherapeutic technique. Biopsy results offer the clinician a way to determine the success or failure of radiotherapy in the individual patient. Further efforts of physicians should be directed toward defining the earliest time after the initiation of radiotherapy that biopsy results can be considered reliable and, most importantly, identifying how patients with positive biopsy results after irradiation should be treated. REFERENCES (1) FuiNo A, SCARDINO PT. Transrectal ultrasonography for prostatic cancer. II. The response of the prostate to defini- tive radiotherapy. Cancer 1986;57:935-940. (2) WHITMORE WF JR. Overview: Historical and contemporary. NCI Monogr 1988;7:7-11. (3) COSGROVE MD, GEORGE GW III, TERRY R. The effects of treatment on the local lesion of carcinoma of the prostate. J Urol 1973;109:861-865. (4) DHOM G, DEGRO S. Therapy of prostatic cancer and histo- pathologic follow-up. Prostate 1982;3:531-542. (5) SCARDINO PT, CARLTON CE JR. Combined interstitial and external irradiation for prostatic cancer. In Principles & Management of Urologic Cancer (Javadpour N, ed), 2nd ed. Baltimore: Williams & Wilkins, 1983, pp 392-408. (6) LERNER SP,SCARDINO PT, GERVASIL,ET AL. The risk of dying of prostate cancer in patients with clinically localized dis- ease. J Urol. In press. (7) CARPENTIER PJ, SCHROEDER FH, BLOM JH. Transrectal ultra- sonography in the follow-up of prostatic carcinoma patients. J Urol 1982;128:742-746. (8) CARLTON CE, HUDGINS PT, GUERRIERO WG, ET AL. Radio- therapy in the management of stage C carcinoma of the prostate. J Urol 1976;116:206-210. (9) HUDGINS PT. Irradiation of prostate cancer combined with abdominal exploration. In Textbook of Radiotherapy (Fletcher GH, ed). Philadelphia: Lea & Febiger, 1975, pp 768-772. (10) SCARDINO PT, GUERRIERO WG, CARLTON CE JR. Surgical staging and combined therapy with radioactive gold grain implantation and external irradiation. In Genitourinary Tumors: Fundamental Principles and Surgical Techniques (Johnson DE, Boileau MA, eds). New York: Grune & Strat- ton, 1982, pp 75-80. (11) Bostwick DG, EGBERT BM, FAJIARDO LF. Radiation injury of the normal and neoplastic prostate. Am J Surg Pathol 1982;6:541-551. (12) SCARDINO PT, FRANKEL JF, WHEELER TM, ET AL. The prog- nostic significance of post-irradiation biopsy results in patients with prostatic cancer. J Urol 1986;135:510-516. (13) JacoBl GH, HOHENFELLNER R. Staging, management and post-treatment reevaluation of prostate cancer: Dogma questioned. In Prostate Cancer (Jacobi GH, Hohenfellner R, eds). Baltimore: Williams & Wilkins, 1982, pp 31-56. (14) KurTH KH, ALTWEIN JE, SKOLUDA D, ET AL. Follow-up of irradiated prostatic carcinoma by aspiration biopsy. J Urol 1977;117:615-621. (15) KIESLING VJ, MCANINCH JW, GOEBEL JL, ET AL. External beam radiotherapy for adenocarcinoma of the prostate: A clinical follow-up. J Urol 1980;124:851-854. BAGSHAW MA. Radiation therapy for cancer of the prostate. In Diagnosis and Management of Genitourinary Cancer (Skinner DG, Lieskovsky G, eds). Philadelphia: Saunders, 1988, pp 425-445. (17) FREIHA FS,BAGSHAW MA. Carcinoma of the prostate: Results of postirradiation biopsy. Prostate 1984;5:19-25. (18) Cox ID, STorreL TI. The significance of needle biopsy after irradiation for stage C adenocarcinoma of the prostate. Cancer 1977;40:156-160. (19) Cox JD, KLINE RW. Do prostatic biopsies 12 months or more after external irradiation for adenocarcinomas, stage III, predict long-term survival? Int J Radiat Oncol Biol Phys 1983;9:299-303. (20) SEWELL RA, BRAREN V, WILSON SK, ET AL. Extended biopsy follow-up after full-course radiation for resectable prostatic carcinoma. J Urol 1975;113:371-373. (16) NCI MONOGRAPHS, NUMBER 7, 1988 (21) COSGROVE MD, KAEMPF MI. Prostatic cancer revisited. J Urol 1976;115:79-81. (22) NACHTSHEIM DA JR, MCANINCH JW, STUTZMAN RE, ET AL. Latent residual tumor following external radiotherapy for prostate adenocarcinomas. J Urol 1978;120:312-314. (23) LYTTON B, COLLINS JT, WEISS RM, ET AL. Results of biopsy after early stage prostatic cancer treatment by implantation of 125] seeds. J Urol 1979;121:306-309. (24) SCHELLHAMMER PF, LAGADA LE, EL-MAHDI AM. Histologi- MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 103 cal characteristics of prostatic biopsies after iodine-125 implantation. J Urol 1980;123:700-705. (25) HERR HW, WHITMORE WF JR. Significance of prostatic biop- sies after radiation therapy for carcinoma of the prostate. Prostate 1982;3:339-350. (26) SCHELLHAMMER PF, EL-MAHDI AM, HIGGINS EM, ET AL. Pros- tate biopsy after definitive treatment by interstitial iodine- 125 implant or external beam radiation therapy. J Urol 1987;137:897-901. III. Surgery Selection Criteria for Radical Prostatectomy Based on Morphometric Studies in Prostate Carcinoma! Fuad S. Freiha,* John E. McNeal, Thomas A. Stamey” ABSTRACT —Morphometric reconstruction of 122 consecutive radical prostatectomy specimens were analyzed for cancer volume and grade, seminal vesicle (SV) invasion, lymph node (LN) metasta- sis, and complete capsular penetration. The mean cancer volume for 91 specimens without SV invasion or LN metastasis was 3.7 cm?; for 14 with only SV invasion, 9.0 cm3; for 17 with LN metastasis, 15.2 cm3; and for 12 with both SV invasion and LN metastasis, 17.8 cm. The mean cancer volume for 60 specimens without capsular penetration was 2.5 cm, and for 62 it was 9.0 cm3. Grade of cancer correlated well with tumor volume. We believe that radical prostatectomy for cure should be performed on patients with tumors less than 3.8 cm? in volume. Methods for accurate assessment of tumor volume before surgery should be given research priorities. —NCI Monogr 7:107-108, 1988. The malignant potential and capacity of prostate carci- noma to metastasize are functions of degrees of differentia- tion and tumor volume (1). Prostate cancer dedifferentiates as tumor volume increases (2). Morphometric analyses of 122 consecutive radical prostatectomy specimens confirm these findings and identify tumor volumes that influence the selection of appropriate therapy for the patient with clinically localized disease. MATERIALS AND METHODS One hundred twenty-two consecutive radical prostatec- tomy specimens were studied by the step-section technique. After fixation in 37% formalin, the gland was serially sec- tioned at 3-mm intervals perpendicular to the rectal surface and separated into right and left halves. Sections were embedded in paraffin and cut at 7 um and stained with hematoxylin and eosin. We determined cancer volume by outlining the tumor on every slide, calculating its surface area with the use of a Compaq computer, and multiplying that by the section thick- ness. For correction for tissue shrinkage caused by fixation, the sum total was multiplied by a factor of 1.5. ! Supported by a grant from the Richard M. Lucas Cancer Foundation. 2 Division of Urology, Department of Surgery, Stanford University School of Medicine, Stanford, CA. * Reprint requests to: Fuad S. Freiha, M.D., Division of Urology, Depart- ment of Surgery, Stanford University School of Medicine, 300 Pasteur Dr., Stanford, CA 94305-5118. Seminal vesicle invasion and lymph node metastasis were quantitated and recorded. Complete penetration of the capsule by cancer cells, referred to as level III capsular penetration, was quantitated and measured by linear centimeters. The grade of cancer was determined according to the Gleason classification (3), and the percent of each histologic pattern was recorded. RESULTS Seventeen of the 122 specimens had lymph node metas- tases and 26 had seminal vesicle invasion. Twelve of the 26 with seminal vesicle invasion also had lymph node metas- tases. Complete capsular penetration was present in 62 spec- imens; 16 had less than 1 cm and 46 had more than 1 cm of linear penetration along the capsule. The mean cancer volume for all 122 was 5.8 cm?3, with a range of 0.01 to 42.10 cm3. The mean percent Gleason histologic pattern 4 and/or 5 present in all 122 specimens was 29% with a range of 0% to 100%. Tables 1 and 2 show the correlation between tumor volume and grade, seminal vesicle invasion, and lymph node metastasis, all of which were closely interrelated. The incidence of lymph node metastasis and seminal vesi- cle invasion in relation to the presence or absence and extent of capsular penetration is shown in table 3. The mean cancer volume for 60 specimens without capsu- lar penetration was 2.5 cm?, with a range of 0.01 to 23.60 cm’. The large tumors that exhibited no capsular penetration had little or no Gleason histologic pattern 4 and/or 5. Tables 4 and 5 show the correlation between volume, grade, and capsular penetration. Of the 46 specimens with more than 1 cm of penetration, 17 had neither lymph node metastasis nor seminal vesicle invasion, and the mean cancer volume for these 17 was 7.3 cm?, a figure which remains significant when compared with the mean of those samples without capsular penetration. DISCUSSION Before recommending radical prostatectomy to a patient with clinically localized prostate cancer, the urologist must try to answer two questions: 1) How much cancer does the 107 108 TABLE 1.—Correlation between tumor volume, seminal vesicle (SV) invasion, and lymph node (LN) metastasis 3 Metastasis/ No. of Volume, um invasion specimens Mean Range P Negative LN Negative SV 91 3.7 0.01-26.50 Positive LN Positive SV 12 17.8 491-42.10 <.001 Positive LN 17 152 3.69-42.10 <.001 Negative LN Positive SV 14 9.0 1.33-24.40 <.001 TABLE 2.—Correlation between Gleason grade, seminal vesicle (SV) invasion, and lymph node (LN) metastasis Percent >, Metastasis/ No. of Oleason grade 24 invasion specimens Mean Range P Negative LN Negative SV 91 17 0-100 Positive LN Positive SV 12 63 20-100 <.001 Positive LN 17 62 20-100 <.001 Negative LN Positive SV 14 34 0-100 <01 patient have? 2) What is the largest cancer that may be cured by radical prostatectomy? The urologist continues to rely on the rectal examination to answer the first question. Even to the most experienced among us, rectal examination is highly inaccurate in assess- ing the extent of local disease and often underestimates its volume (4). Transrectal ultrasound and magnetic resonance imaging are promising new modalities that may prove to be more accurate than the rectal examination. The use of tumor markers such as the prostate-specific antigen may also aid in the preoperative assessment of cancer volume and extent. To answer the second question, we performed morphomet- ric reconstructions of radical prostatectomy specimens and analyzed them for cancer volume and grade, extent of capsu- TABLE 3.—Capsular penetration and incidence of lymph node metastasis and seminal vesicle invasion TABLE 4.— Correlation between tumor volume and capsular penetration Volume, cm? Capsular No. of penetration ~~ specimens ~~ Mean Range P None 60 25 0.01-23.60 <1 cm 16 43 1.46-16.32 Not significant >1 cm 46 10.6 1.47-42.10 <.001 TABLE 5.—Correlation between tumor grade and capsular penetration Percent Gl de >4 Capsular No. of Sleasor grace 22. penetration specimens Mean Range P None 60 9 0-80 1 cm 46 46 5-100 <.001 Lymph Seminal Capsular No. of node metastasis vesicle invasion penetration specimens No. Percent No. Percent None 60 0 0 1 2 1 cm 46 17 37 24 52 lar penetration, seminal vesicle invasion, and lymph node metastasis to find the largest tumor amenable to cure by radical prostatectomy. Analysis of these morphometric studies has identified ranges of tumor volumes that may influence selection of appropriate therapy for the patient with clinically localized disease. Patients with tumors that are less than 3.8 cm? in volume do not have seminal vesicle invasion or lymph node metastasis and, therefore, will have an excellent chance for cure by radical prostatectomy. Tumors with volumes between 2.6 and 3.7 cm? exhibit complete capsular penetration. The exact significance of complete capsular penetration in the absence of seminal vesi- cle invasion or lymph node metastasis is not known. We now have 17 such patients who are being followed closely with examinations, prostate-specific antigen determinations, and bone scans; we hope to learn more about the significance of this pathologic finding. Tumors with volumes up to 2.5 cm? are localized and exhibit neither capsular penetration nor lymph node metastasis. Now that we have identified ranges of tumor volumes with prognostic significance, the need for methods to measure preoperative tumor volume accurately becomes apparent. REFERENCES (1) MCNEAL JE, BOSTWICK DG, KINDRACHUK RA, ET AL. Patterns of progression in prostatic carcinoma. Lancet 1986;1:60-63. (2) MCNEAL JE: Origin and development of carcinoma in the prostate. Cancer 1969;23:24-34. (3) GLEASON DF: Histologic grading and clinical staging of pros- tatic carcinoma. In Urologic Pathology: The Prostate (Tan- nenbaum M, ed). Philadelphia: Lea & Febiger, 1977, pp 171-197. (4) SPIGELMAN SS, MCNEAL JE, FREIHA FS, ET AL. Rectal examina- tion in volume determination of carcinoma of the prostate: Clinical and anatomical correlations. J Urol 1986;136: 1228-1230. Bilateral Pelvic Lymphadenectomy and Radical Retropubic Prostatectomy for Stage C or D1 Adenocarcinoma of the Prostate: Possible Beneficial Effect of Adjuvant Treatment Horst Zincke' ABSTRACT —Limited clinical stage C (T3 NX M0) disease can be treated surgically, and morbidity can be acceptable. When appropriate adjuvant therapy (orchiectomy and/or radiation) is administered, residual cancer can be controlled locally for at least a limited period. The incidence of local progression in pathologic stage C or D1 disease may be negligible after early adjuvant orchiectomy and/or radiation treatment. The combination of immediate orchiectomy and radical prostatectomy has been shown to limit progression significantly (P= .0009) in many patients with D1 (T0-3 N1,2 M0) disease. However, some patients do not re- spond to this combination treatment, which suggests that systemic dissemination of heterogeneous tumor cells is unresponsive to adju- vant androgen ablation therapy. The DNA ploidy pattern may be a valuable predictor of disease outcome after treatment in stage D1 disease. Other pathologic variables (including acid phosphatase levels) have not been useful in predicting disease outcome or treat- ment response. Finally, patients with limited clinical stage C disease and those with pathologic C or D1 disease should be enrolled in a prospective randomized protocol so that the possible beneficial effects of adjuvant treatment programs can be evaluated. Apart from the usual pathologic variables and prostate-specific antigen testing, the DNA ploidy pattern should be included as a stratifica- tion factor.—NCI Monogr 7:109-115, 1988. Adenocarcinoma of the prostate is the second most fre- quent cause of death from cancer in the United States. For 1987, it has been projected that 96,000 new patients will have tumor of the prostate and that 27,000 will die of prostate cancer (1). Thus cancer of the prostate is not an innocuous disease. However, depending on its grade and stage at the time of presentation, the tumor can be curable. Conceptually, only a small percentage of patients with cancer of the prostate qualify for radical prostatectomy (2). The rationale for this attitude originated from reports of small series in which survival rates were favorable only for patients with small lesions (=1.5 cm, stage B1/T1) and were poor for patients with larger lesions. In most of these studies, however, accurate staging (by bone scan and pelvic lymphadenectomy) was not done. Thus many of the patients who were believed to have localized lesions had locally extended disease at surgery, and some also might have had metastatic disease. Hence the concept evolved that patients with clinical stage B2 (T2) disease (e.g.,=2 cm, clinically localized to prostate, 50% of whom have pathologic local extraprostatic disease) are unsuitable surgical candidates. Conservative treatment (hormonal or radiation, or both) was recommended for these ! Department of Urology, Mayo Clinic and Mayo Foundation, 200 First St., S.W., Rochester, MN 55905. patients. The short-term (5-10 yr) survival when irradiation is used for this stage of the disease (3-6) is similar to that achieved by radical prostatectomy without adjuvant treat- ment for localized cancer. However, radical prostatectomy for localized cancer preceded by bilateral pelvic lymphade- nectomy recently has been preferred by an increasing number of surgeons because of the observed favorable long-term local and systemic control (7) unmatched by conservative treatment modalities and its potential as a potency- preserving procedure (8). Many patients have neither localized nor distant metastatic cancer but have locally extracapsular disease with or without regional lymph node involvement. Treatment for these patients has been discussed with considerable controversy during recent years (9-12). This controversy involves the question whether additional treatment is needed for local stage C (pT3) disease which, before and during surgery, had been perceived as a localized lesion only, but which unex- pectedly was discovered on histopathologic examination of the prostate after radical prostatectomy to be extracapsular disease with or without regional lymph node involvement. The treatment approach for primary limited clinical stage C (T3 NX MO) disease and the management for stage D1 (T0-3 N1,2 M0), i.e., regional lymph node involvement, recognized only at pelvic lymphadenectomy and usually preceding an originally planned radical prostatectomy, have been of par- ticular concern to many investigators (9,11,13). Surgery as a treatment option for clinical stage C disease has usually been excluded because of the historically poor results when perineal prostatectomy only was used. However, reports of physicians using this approach lacked appropriate staging (by pelvic lymphadenectomy), and the opportunity for appropriate adjuvant treatment was missed. Because of the high incidence of residual cancer after surgery for this stage of the disease and the high incidence of pelvic lymph node involvement (50%), which many consider to be a contraindication to radical surgery, radiation had been pre- ferred to radical surgery for the treatment of clinical stage C cancer of the prostate. The problems associated with local and systemic control with this large tumor bulk and usually high-grade cancer have been recognized, and augmentation therapy has been suggested (1/2). Of particular concern has been the persistence and/or local recurrence of the disease after radiation that is usually associated with an ominous prognosis (14). Morbidity in patients who fail after local radiation has been high and has resulted in significant impairment in the quality of life (15). In this context, the goals of therapy for patients with clinical stage C disease should 109 110 include not only prolongation of survival but also control of local progression with its associated decrease in complica- tions and thus possible improvement in quality of life. Hence monotherapy for stage C disease of the prostate currently seems to be less attractive, particularly when a cancer is locally in an advanced clinical and pathologic stage. Patients with limited but unequivocal clinical stage C cancer of the prostate (without involvement of bladder base at cystoscopy) can be treated successfully with bilateral pel- vic lymphadenectomy, radical prostatectomy, and adjuvant treatment without significant morbidity and with excellent long-term results in nonprogression and survival (9,16). TREATMENT OF CLINICAL STAGE C ADENOCARCINOMA OF THE PROSTATE A previous report (9) involved 101 patients with a mean age of 64 years (range, 40-74 yr) who had limited but unequivocal clinical stage C (T3 NX M0) adenocarcinoma of the prostate on clinical examination. These patients under- went bilateral pelvic lymphadenectomy and radical prosta- tectomy. Patients had clinical stage C disease based on digi- tal rectal examinations and no evidence of metastasis, as determined by radionuclide bone scanning and roentgeno- grams as well asby laboratory tests, including measurements of total and tartrate-inhibited fractions of acid phosphatase. Follow-up ranged from 0.5 to 17 years (mean, 4.9 yr). Fifty patients were followed for a minimum of 5 years and 10 were followed for 10 years. Of these patients, 35 were under observation at 5 years and 7 were under observation at 10 years. At these respective intervals, 21 and 5 were free of disease. Regular follow-up included quarterly examinations during the first postoperative year, biannually during the second year, and usually annually thereafter if the disease had not progressed. In addition to the usual tests, a bone scan was performed at least twice annually during the first 2 years and annually thereafter. Disease progression was judged to be present if there was biopsy-proven local recurrence or if (75) 100 80 — (35) _Expected survival —=——{0 60 50 (27) Observed survival 5 0 0F (5) 50 Observed nonprogression 20 | Clinical stage C cancer of prostate (N=101) 10 1 1 1 1 1 1 1 1 1 1 Years Figures: Numbers in parentheses represent numbers of patients under observation at that time. FIGURE 1.—Survival curves of 101 patients with clinical stage C adenocarci- noma of prostate. Curves of expected survival, overall survival, and observed nonprogression are shown. Expected survival is based on 1970 United States life tables for North Central United States. Modified from Zincke et al. (9). Reproduced with permission of the publisher. 100 A 80 | 60 o 50 Cc > 40 £ 5 30 n & 20 + Residual cancer © L 1 L L 1 1 1 1 0 1 2 3 4 5 6 ? 8 oe Years 100 F 80 + TTI Eo Tn EF Ss 60 Te MR ® so} Te 2 sof TTT gor ee : No (N=85) © 30 Qa 5 Cc 20 I * Residual cancer © ; i 1 1 1 1 JX 1 1 0 1 2 3 4 5 6 ? 3 ’ ° Years FIGURE 2.—Residual cancer and survival (A) and nonprogression (B) in 101 patients with prostate cancer. Reproduced with permission of the pub- lisher (9). radionuclide bone scans or roentgenograms (or both) became positive. An abnormal acid phosphatase level alone was not considered to be consistent with metastasis. Analyses of local, systemic, and overall progression as well as overall survival and survival according to cause-specific death were always performed according to the grade and extent of the tumor (measured in cubic centimeters, seminal vesicle involvement) and whether margin-positive disease (residual cancer) was histopathologically identified at radical prostatectomy and was left behind. We attempted to remove all residual tumor. However, removal was not possible in all patients, and tumor was left behind rather than risking inten- tional incontinence. Lymph node involvement was quantified. Of the 101 patients, 52 had positive lymph nodes. One-third of these 52 did not receive adjuvant treatment, whereas almost two-thirds of the patients who had patho- logic stage C lesions received no adjuvant treatment. The most frequently administered adjuvant treatment was orchiec- tomy with or without additional treatment (30 and 12 patients with pathologic stage D1 and C lesions, respec- tively). Statistical analysis was done by the Kaplan-Meier method, and regression analysis was performed according to the Cox proportional hazards model. The observed survival of the 101 patients was similar to the expected survival of an age- and sex-matched control group of the general popula- tion of the North Central United States (fig. 1). Absence of cure is demonstrated in the continuously declining observed nonprogression curve. Nonprogression and survival curves were similar for stages C and D1 disease. Residual cancer was present in 16 patients and did not affect survival and nonprogression significantly (fig. 2). Patients who had under- gone either orchiectomy or irradiation had no evidence of NCI MONOGRAPHS, NUMBER 7, 1988 local recurrence during follow-up. Conversely, none of the 16 patients with local recurrence had evidence of residual cancer, and 13 of the 16 had not received adjuvant treatment (table 1). Of 47 patients who received no adjuvant treatment, 13 (28%) had local recurrence. None of the 48 patients who had received adjuvant radiation and/or orchiectomy with or without additional diethylstilbestrol had a local recurrence, but 3 of 6 patients treated with diethylstilbestrol did. Local recurrence seemed to be related to tumor grade but not to number of involved nodes. Residual disease was not asso- ciated with local recurrence because these patients had received adjuvant treatment. Specifically, adjuvant orchiec- tomy was not associated with local recurrence. Of the 52 patients who had positive pelvic lymph nodes, 30 underwent immediate adjuvant orchiectomy (91% were disease-free at 5 yr), which significantly delayed progression, compared with the 22 patients who did not have immediate orchiectomy (11% were disease-free at 5 yr; P < .0001; fig. 3A). Adjuvant orchiectomy resulted in a 5-year survival of 91%, compared with 58% (P= .09) when it was not used (fig. 3B). In a more recent analysis of 384 patients who had patho- logic stage C disease (pT3 NO MO), 70 originally had clinical stage C disease (18%); the mean age of this group was 64 years. The mean follow-up was 4.5 years, with a median of 3.6 years and a range of 1 to 18 years. Of the 784 patients, 326 are still alive and are being followed. Thirty patients (8%) have died of prostate cancer. As in previous experience, 5- and 10-year survival rates (87% and 67%, respectively) were similar to the expected survival (83% and 64%, respec- tively; fig. 4). Analysis for cause-specific deaths only revealed a 5- and 10-year survival of 94% and 81%, respectively (fig. 5). The continuously declining curve (the last patient TABLE 1.—Data on 16 patients with local recurrence after radical surgery for clinical stage C adenocarcinoma of the prostate Time to local Patient Pathologic Residual Adjuvant recurrence, Survival, No. stage cancer treatment? mo mo 1 Cc’ No No 26 97 2 ch No No 67 95 3 C No DES 16 47¢ 4 C No No 12 61 5 C No No 25 49 6 C No No 81 99 7 C No No 30 42 8 C No No 52 104 9 DI No DES 58 202 10 D1? No No 56 88 11 D1? No No 13 46 12 D1? No No 9 36¢ 13 D1 No No 4 664 14 D1? No DES 18 514 15 D1? No No 10 76 16 DI No No 7 48¢ “ DES = diethylstilbestrol. b Seminal vesicle involvement was observed. ¢ Patient died of cancer. 4 Patient died of unrelated cause. MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 111 100 80 Orchiectomy (N=30) 60 TT TTT» 40 = “No orchiectomy (N=22) 20 % nonprogression w o T Stage D;, (N=52) p= .0001 10 1 1 1 1 2 3 4 5 6 7 8 9 10 o 100 80 60 50 40 YT" TTrTTrju No orchiectomy (N=22) 30 % surviving 20 Stage D,, (N=52) p= .09 10 1 1 1 1 1 1 1 1 1 0 1 2 3 4 5 6 ? Years FIGURE 3.— Adjuvant treatment and nonprogression (A) and survival (B) in 52 patients with pathologic stage D1 prostate cancer. Reproduced with permission of the publisher (9). died at 13 yr after initial treatment) lends credence to the impression that cure cannot be expected in this difficult patient population. Overall nonprogression at 5 and 10 years was 71% and 52%, respectively, whereas local recurrence was absent in 85% and 71% of the patients at 5 and 10 years, respectively. Local recurrence was significantly associated with seminal vesicle involvement as well as tumor grade. In patients with grade 3 or 4 disease, local recurrence was absent at 10 years in only 67%. Similar to a previous study, residual cancer (usually treated by adjuvant therapy) was not significantly related to systemic or local progression. Furthermore, among 280 patients who did not receive adjuvant treatment (usually having low-grade disease and no seminal vesicle involvement), overall progres- sion occurred in 25% but in only 10.8% (4 of 37 patients) of (143) = Expected h ( Survival, % Years FIGURE 4.—Kaplan-Meier survival curve of 384 patients with pathologic stage C (pT3) prostate cancer who were treated with bilateral pelvic lymphadenectomy and radical retropubic prostatectomy, compared with expected survival of an age- and sex-matched control group of the North Central United States. 112 Survival, % Years FIGURE 5.—Kaplan-Meier survival curve of 384 patients with pathologic stage C (pT3) prostate cancer who were treated with bilateral pelvic lymphadenectomy and radical prostatectomy, according to cause- specific death (prostate cancer). those who underwent orchiectomy with or without diethyl- stilbestrol and/or radiation treatment (usually having high- grade disease and seminal vesicle involvement). More than 40% of the patients receiving oral hormones only had pro- gression. Radiation alone (25 patients) and orchiectomy (37 patients) effectively reduced local recurrence (0% and 2.7%, respectively) in these patients, but 41 of the 280 patients (14.6%) without adjuvant treatment had local recurrence (P= .02). Furthermore, evidence from the analysis indicated that clinical stage did not affect the decision in regard to adjuvant treatment. Rather, the pathologic finding seemed to dictate the physician’s decision whether to apply additional treatment. TREATMENT OF STAGE D1 ADENOCARCINOMA OF THE PROSTATE Reported experience of the surgical and nonsurgical man- agement of stage D1 (pT0-3 N1,2 MO) cancer of the prostate is limited. Randomized studies for this stage have not been performed. Observation only, as shown by Paulson (/1), led to progression in 50% or more of the patients at less than 2 years, and most single treatment regimens (surgical or conservative) have resulted in progression in the same per- centage of patients at 5 years (/ 7). Only recently have results of combination treatments been published by others (18,19). Of 33 patients who underwent radical prostatectomy in the series of deKernion et al. (18), 25 received adjuvant treat- ment in the form of radiation or hormones. The fact that only 4 of 10 patients are free of disease after radical prostatectomy and hormone and radiation therapy may reflect disease vari- ables rather than the treatment itself, in particular because all 8 of the patients who underwent radical prostatectomy only are still alive and free of disease at follow-up of 5 years. In a series of 27 patients with stage D1 disease who underwent radical prostatectomy, Lange and colleagues (19) administered adjuvant radiation in 2 1; the median follow-up period was 48 months. Recurrence was observed in 3 of the 21 patients, 2 of whom had a response to bilateral orchiec- tomy. These results suggest that adjuvant treatment com- bined with cytoreductive surgery may be beneficial in these patients, when comparison is made with patients treated by radical prostatectomy only (/1). The only other large series of patients who had been treated for stage D1 prostate cancer is that involving 61 patients reported by Bagshaw (12), who achieved 5- and 10-year survival rates of 50% and 20%, respectively, with external-beam irradiation alone. Disease- free survival was not discussed. The generally poor results observed (/1,12,17) with con- servative monotherapy have led to disillusionment among many investigators, and some have advised observation only until symptomatic progression occurs. Many investigators believe that surgical treatment is inappropriate for a disease that has become systemic, albeit also larger locally. This fatalistic view has not been shared by my colleagues and me because of our favorable early experience with combi- nation treatment (radical prostatectomy plus immediate orchiectomy). Our initial experience in the treatment of stage D1 disease by means of bilateral pelvic lymphadenectomy and radical prostatectomy with or without adjuvant treat- ment was reported in 1978 (20). Since then, my colleagues and I have extended the indications to include patients with limited clinical stage C lesions (9) and some with multiple positive pelvic lymph nodes (=10 nodes). We do not include patients with involved periaortic and pericaval lymph nodes, a determination that can be made at pelvic lymphadenectomy. From 1966 to 1985, my associates and I performed pelvic lymphadenectomy and radical prostatectomy, with or with- out adjuvant treatment, on 1,473 patients who had all stages of prostate cancer. Of these, 195 patients (13%) had stage D1 disease (fig. 6). The ages of the 195 patients ranged from 40 to 77 years (mean, 64 yr). Follow-up ranged from 1 to 19 years (mean, 5 yr). Patients were seen quarterly for the first year, biannually during the second year, and annually there- after if there was no evidence of metastatic disease. Thirty- five percent (68 patients) had clinical stage C disease, and 90% (175 patients) had pathologic stage C disease. Most patients had only one lymph node involved, two-thirds had a large tumor bulk 10 cm?), and one-half of the total had undifferentiated cancer. Unfavorable variables, such as residual cancer (31%) and elevated acid phosphatase level (17%), were also noted (table 2). Of those patients who underwent surgery, 5 or more years had elapsed for 93 of them, 10 or more for 28, and 15 years or more for 9. Of the various adjuvant treatment modalities (orchiec- tomy, diethylstilbestrol, irradiation), orchiectomy was the most beneficial. Although orchiectomy was eventually per- formed in almost all patients who had progression, overall immediate orchiectomy was performed in only 109 patients (56%). Overall, the probability of nonprogression at 5, 10, and 15 years was 56%, 47%, and 47%, respectively. Projected Procedures performed, no. Year © 25 50 75 100 125 150 175 200 225 T T T T T T T T 1 1966 TOTAL '68 Pelvic lymphadenectomy 70 plus radical prostatectomy ‘72 74 76 78 '80 '82 '84 [ For all stages (n= 1, 473) M8 For stage D ¢ (n= 195) FIGURE 6.—Number of radical retropubic prostatectomy procedures per- formed at the Mayo Clinic from 1966 to 1985. NCI MONOGRAPHS, NUMBER 7, 1988 TABLE 2.—Data on 195 patients with pathologic stage D1 (pT0-3 N1,2 MO) prostate cancer Patients Pathology No. Percent Grade 2 99 51 3,4 96 49 Tumor bulk (prostate only) >10 cm? 74 38 <10 cm? 121 62 No. of positive nodes? 1 85 44 2 48 23 3 23 12 4 16 8 =5 20 11 Residual cancer Yes 61 31 No 134 69 Elevated acid phosphatase Yes 33 17 No 162 83 4 Data are for 192 patients only. absence of local progression at 10 years was 80%, with 12 patients still under observation at this time (fig. 7). Projected survival (fig. 8) at 5 and 10 years was 83% and 68%, respec- tively, which is similar to the expected survival of an age- matched control group of the North Central United States. Analysis according to cause-specific death projected a 5- and 10-year survival of 92% and 83%, respectively (fig. 9). Fourteen patients (7.2%) died of cancer. Their age (mean, 61 yr) and number of positive nodes (mean, 2.14 + 2.0) were not significantly different from the age and positive nodes of those who did not die of the disease. Only 3 patients had received early adjuvant orchiectomy, 5 had received diethyl- stilbestrol, 3 had irradiation and orchiectomy in 2 of these 3, and 3 had no early adjuvant treatment. The period of progression-free survival ranged from 9 to 115 months (mean, 53 mo), and the survival ranged from 36 100 eo oS #0 (118) 2 (12) [3] o 60 Systemic e Q 3 40 Systemic/local (10) 2 R 20 0 | 1 ] | | 1 l | 1 J 0 1 2 3 4 5 6 7 8 9 10 Years after surgery FIGURE 7.—Kaplan-Meier curves of nonprogression for 195 patients with stage D1 prostate cancer who had undergone radical prostatectomy according to systemic, local, and overall progression. MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 113 o 2 © RX al 20 0 | | | | | | | | ] J 0 1 2 3 4 5 6 7 8 9 10 Years after surgery FIGURE 8.—Kaplan-Meier survival curve of 195 patients with stage D1 prostate cancer who had undergone radical prostatectomy with or without adjuvant treatment, compared with the expected survival. to 212 months (mean, 80 mo). As seen previously in the treatment of stage pT3 disease, adjuvant treatment signifi- cantly reduced local recurrence, which was noted in 22 (11%) of the 195 patients. Of these, 17 had no adjuvant treatment originally, 3 had orchiectomy, and 2 had received diethylstil- bestrol. Again, residual cancer did not correlate with progres- sion, in particular local recurrence (fig. 10), because most patients with residual cancer received adjuvant treatment. Of the patients with residual cancer, adjuvant orchiectomy or radiation resulted in a 90% and 100% local recurrence-free survival, respectively. Conversely, overall, 20% of the patients had local recurrence at 5 years; this group included patients who had received diethylstilbestrol or who had no adjuvant treatment because they originally were without residual disease. This percentage was significantly higher (P < .03) than for patients receiving adjuvant orchiectomy. A presurgical elevated acid phosphatase level (33 patients), which is assumed to be associated with an extremely poor prognosis, was not related to progression or survival. At5 and 10 years, survival rates for patients with an elevated acid phosphatase level or elevated tartrate-inhibited fraction of the acid phosphatase, or both, were not significantly different from survival rates for patients who had no elevated acid phosphatase level, with or without adjuvant orchiectomy. This applied also to the results of nonprogression. Once patients experienced progression, the probability of survival 100 (131) 80 | (67) n= 19 ES i (16) o 8 £ 2 I 2 I g 4% a I 20 | 0 L | L L L ' L s s J 0 t+ 2 3 4 5 6 7 8 9 10 Years after surgery FIGURE 9.—Kaplan-Meier survival curve of 195 patients with pathologic stage D1 prostate cancer, according to cause-specific death (prostate cancer). 114 % without local recurrence 1 1 1 1 1 0 1 2 3 4 5 Years after surgery FIGURE 10.—Kaplan-Meier curves of survival until local recurrence in 195 patients with stage D1 prostate cancer who had undergone radical prosta- tectomy, according to presence (61 patients) or absence of residual cancer. was low. Of the 57 patients who had progression, usually without initial adjuvant treatment, 20 died after a median follow-up time of only 2.5 years. Five-year survival of patients who did not have immediate orchiectomy (although they usually underwent orchiectomy later) was 64% and was only 33% for patients who underwent immediate orchiec- tomy (fig. 11). The numbers in these 2 groups are too small for any definite conclusion to be made, but for patients who failed after early adjuvant orchiectomy (14 of 109 or 13%), the probability for longer survival tended to be poorer than for those who had not received immediate adjuvant orchiectomy (43 of 86 or 50% failed). Of the pathologic variables examined, i.e., tumor grade, seminal vesicle involvement, tumor bulk, and number of nodes, only the number of nodes was associated with progres- sion but not survival (fig. 12). Patients with one positive node who did not receive adjuvant treatment had a significantly better disease-free survival (P< .02) than did those with two or more nodes involved. Because follow-up of many patients had been less than 5 years, projected overall survival rates (in particular, sur- vival rates according to cause-specific deaths) may not show significant differences because of the lag period. Therefore, data on the 93 patients who had follow-up of 5 years or more were analyzed. Immediate orchiectomy resulted in a 5- and 10-year nonprogression rate of 84% and 78%, respectively, compared with 46% and 35% (P = .0009) when no imme- diate orchiectomy was performed (fig. 13). Survival at 5 and 100 (24) 8 No (n = 43) an © 60 = © Yes (n = 14) Ra 20 — P= .1 @ a 1 | | | | | J 0 1 2 3 a 5 6 7 Years after first progression FIGURE 11.—Kaplan-Meier curves of survival after first progression in 57 patients with stage D1 prostate cancer who had undergone radical prosta- tectomy, according to whether they had immediate adjuvant orchiectomy. (19) S 80 1 node (n= 29) ‘a 2 nodes (n= 12 op 2 60 @ a x gw @ 3 23 nodes (n= 12) ® ® o 0 1 2 3 4 5 Years after surgery FIGURE 12.—Kaplan-Meier curves of nonprogression for 53 patients with stage D1 prostate cancer who had undergone radical prostatectomy with- out immediate adjuvant treatment, according to number of positive nodes. 10 years for the orchiectomy group was 87% and 76.5%, respectively, and for the nonimmediate orchiectomy group, 77.5% and 48.5%, respectively. Survival in regard to cause- specific death only (fig. 14) at 5 and 10 years in the imme- diate orchiectomy group was 92% and 80%, respectively, and was 91.5% and 84.5%, respectively, for the nonimmediate orchiectomy group (no significant difference). These results may reflect not only various treatment modalities but also various disease characteristics. Therefore, in the absence of reliable tumor/host variables (e.g., grade, bulk, and number of nodes) and in the search of a tumor variable that might be a reliable predictor, flow cytometric analysis was performed on the prostate cancer specimens of our patients in regard to DNA ploidy pattern (21). Prelimi- nary results indicated that ploidy pattern is related to disease outcome, particularly in regard to adjuvant hormonal (i.e., orchiectomy) treatment. If these initial data can be confirmed by an extensive analysis of more patients and longer follow- up, the use of DNA pattern determination may be highly significant as a prognostic indicator. This analysis is particu- larly important because all pathologic variables in stage D1 disease have not proved useful in the prediction of disease outcome. In conclusion, prostatectomy as a means for effective local control provides tumor debulking; immediate orchiectomy (as opposed to diethylstilbestrol, which can be associated with a compliance problem) is a most effective form of 100 — Immediate orchiectomy - 26) ® 80 (18) {n=26) 5 (10) 7) o - @ a 60 [— o l Delayed treatment (n = 67) o © a0 [- = Jt (1) @ 2) S P= 0009 2 20 5 1 1 L 0 2 4 6 8 10 Years after surgery FIGURE 13.—Kaplan-Meier curves of nonprogression for 93 patients with stage D1 prostate cancer (follow-up = 5 yr) who had undergone radical prostatectomy, according to immediate adjuvant orchiectomy or delayed treatment. NCI MONOGRAPHS, NUMBER 7, 1988 (20) Immediate orchiectomy (n= 26) — (47) 60 — Delayed treatment (n= 67) Not dying from prostate cancer, % T Years after surgery FIGURE 14.—Kaplan-Meier survival curves of cause-specific death (prostate cancer) of 93 patients with stage D1 prostate cancer (follow-up = 5 yr) who had undergone radical prostatectomy, according to immediate adjuvant orchiectomy or delayed adjuvant treatment. androgen ablation, which may achieve some systemic con- trol of androgen-responsive cells. The favorable results in the treatment of stage D1 disease with adjuvant orchiectomy suggest that tumors of many patients retain, to some degree, androgen-dependent homogeneity, despite pelvic nodal involvement and thus systemic disease. Accumulated evi- dence from experience at the Mayo Clinic and that at other institutions indicates that orchiectomy or prostatectomy alone is ineffective for this stage of the disease. REFERENCES (1) AMERICAN CANCER SOCIETY. Cancer clusters: Cancer statis- tics 1987. CA 1987;37:1-64. (2) CuLp OS, MEYER JJ. Proceedings: Radical prostatectomy in the treatment of prostatic cancer. Cancer 1973;32:1113- 1118. (3) BAGSHAW MA. External radiation therapy of carcinoma of the prostate. Cancer 1980;45(suppl):1912-1921. (4) GUERRIERO WG, CARLTON CE JR, HUDGINS PT. Combined interstitial and external radiotherapy in the definitive man- agement of carcinoma of the prostate. Cancer 1980; 45(suppl):1922-1923. (5) WHITMORE WF JR. Interstitial I-125 implantation in the man- agement of localized prostatic cancer. Prog Clin Biol Res 1984;153:513-527. (6) Cupps RE, Utz DC, FLEMING TR, ET AL. Definitive radiation MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 115 therapy for prostatic carcinoma: Mayo Clinic experience. J Urol 1980;124:855-859. (7) BENSON RC Jr, TOMERA KM, ZINCKE H, ET AL. Bilateral pelvic lymphadenectomy and radical retropubic prostatectomy for adenocarcinoma confined to the prostate. J Urol 1984; 131:1103-1106. (8) WALSH PC, DONKER PJ. Impotence following radical prosta- tectomy: Insight into etiology and prevention. J Urol 1982;128:492-497. (9) ZINCKE H, Utz DC, TAYLOR WF. Bilateral pelvic lymphade- nectomy and radical prostatectomy for clinical stage C prostatic cancer: Role of adjuvant treatment for residual cancer and in disease progression. J Urol 1986;135: 1199-1205. (10) ZINCKE H, Utz DC. Radical surgery for stage D1 prostate cancer. Semin Urol 1983;1:253-260. (11) PAULSON DF. The prognostic role of lymphadenectomy in adenocarcinoma of the prostate. Urol Clin North Am 1980,7:615-622. (12) BAGSHAW MA. Potential for radiotherapy alone in prostatic cancer. Cancer 1985;55:2079-2085. (13) OLssON CA. Editorial comment. J Urol 1986;135:1205. (14) SCARDINO PT, WHEELER TM. Prostatic biopsy after irradia- tion therapy for prostatic cancer. Urology 1985;25(suppl): 39-46. (15) RAINWATER L, ZINCKE H. Radical prostatectomy after radia- tion therapy for adenocarcinoma of the prostate: Complica- tions and prognosis. J Urol. In press. (16) PATTERSON DE, ZINCKE H. Perioperative complications of pelvic lymphadenectomy and radical retropubic prostatec- tomy for stages C and DI prostate cancer. Urology 1984;23:243-246. (17) ZINCKE H, UTZ DC, THULE PM, ET AL. Treatment options for patients with stage D1 (T0-3,N1-2,M0) adenocarcinoma of prostate. Urology 1987;30:307-315. (18) DEKERNION JB, HUANG MY, KAUFMAN JJ, ET AL. Result of treatment of patients with stage D1 prostatic carcinoma. Urology 1985;26:446-451. (19) LANGE PH, MOON TD, NARAYAN P, ET AL. Radiation therapy as adjuvant treatment after radical prostatectomy: Patient tolerance and preliminary results. J Urol 1986;136:45-49. (20) ZINCKE H. Updated experience of management of cancer of the prostate at the Mayo Clinic. In Proceedings of the International Symposium on the Treatment of Carcinoma of the Prostate. West Berlin: Univ Press, 1978, pp 57-67. (21) WINKLER HZ, RAINWATER LM, MYERS RP, ET AL. Stage D1 prostatic adenocarcinoma: Significance of nuclear DNA ploidy patterns studied by flow cytometry. Mayo Clin Proc 1988;63:103-112. Long-term Results of Radical Prostatectomy in Clinically Localized Prostate Cancer: Experience at The Johns Hopkins Hospital Herbert Lepor,* Patrick C. Walsh"? ABSTRACT —The objectives of our retrospective long-term anal- ysis of radical prostatectomy at The Johns Hopkins Hospital are to determine the efficacy of radical prostatectomy and the optimal statistical method for ascertaining survival following therapeutic intervention for men with clinically localized prostate cancer. The duration of survival and the cause of death were ascertained for 57 men with clinical stage B1 prostate cancer who had radical prosta- tectomies at The Johns Hopkins Hospital between 1951 and 1963. The absence of metastatic disease was determined by radiographic survey of the bones only. The survival curve determined by the direct method was virtually identical to the projected survival curve for a 62-year-old man in 1960. The cause-specific actuarial survival analysis indicated that only 14% of the men with stage B1 disease and a 15-year life expectancy will develop metastatic prostate cancer following radical prostatectomy. The cause-specific survival curve plateaued after 10 years, which indicated that the majority of men surviving 10 years free of disease are cured of the disease. Survival analysis was also determined by the direct method for 48 men with clinical stage B2 prostate cancer who had undergone radical prostatectomy between 1951 and 1963. Overall, the survival rates for these men were considerably lower than those for similarly treated men with clinical stage B1 disease. The survival curves following radical prostatectomy for men with stage B1 disease and clinical stage B2 disease pathologically confined to the prostate were similar. Radical prostatectomy for stage B1 disease was per- formed with minimal morbidity, and potency was preserved in most patients with the use of nerve-sparing modifications. A review of the clinical experience at The Johns Hopkins Hospital indicates that excellent local disease control was achieved following radical pros- tatectomy. On the basis of curability, morbidity, and control of local disease, radical prostatectomy represents a very effective treatment for clinically localized prostate cancer. Cause-specific survival analysis represents a useful statistical method for the ascertainment of survival rates for men with this disease.—NCI Monogr 7:117-122, 1988. The optimal treatment of clinically localized prostate cancer represents one of the most controversial issues in the management of genitourinary cancers. The therapeutic options that have been advocated for the management of this disease include radical prostatectomy, external-beam radio- therapy, interstitial radiotherapy, hormonal therapy, trans- urethral resection of the prostate, and no treatment (/-6). A recent survey of the patterns of care for prostate cancer ! Department of Urology, The Johns Hopkins Medical Institutions, Balti- more, MD. 2 We thank Ms. Karon Hertlein for editorial assistance and Dr. Ellen Shapiro for review of the manuscript. * Reprint requests to: Herbert Lepor, M.D, Division of Urology, Depart- ment of Surgery, The Jewish Hospital of St. Louis at Washington Univer- sity Medical Center, 216 South Kingshighway, St. Louis, MO 63110. sponsored by the American College of Surgeons revealed that there is no consensus for the treatment of clinically localized disease (7). Many factors contribute to the controversy regarding the optimal management of clinically localized prostate cancer. The natural history of the disease remains poorly defined. The metastatic potential of stage Al disease was not recognized until recently (8,9). Because the majority of studies reported in the literature are nonrandomized and the clinical series reported lack standardization of staging criteria and methods of statistical analysis, it is difficult for one to compare the efficacy of various therapeutic options and to determine the efficacy of specific therapeutic modalities due to the indis- criminate use of adjuvant hormonal therapy. The resolution of the controversy regarding the optimal treatment requires a randomized study with long-term clinical follow-up. Alter- natively, the relative efficacy of therapeutic options can be inferred from nonrandomized, nonconcurrent clinical series that have been reported with long-term follow-up. Our objectives in this retrospective long-term analysis of radical prostatectomy at The Johns Hopkins Hospital are to determine curability, morbidity, and control of local disease following radical prostatectomy for clinically localized pros- tate cancer. The lack of standardized methods for reporting survival analysis compounds the potential misinterpretation of nonrandomized, nonconcurrent studies of men with this disease. Therefore, we analyzed the survival data using various statistical methods to determine appropriate methods for ascertaining survival (curability) following therapeutic intervention. HISTORICAL PERSPECTIVE Although radical perineal prostatectomy was originally described by Kuchler in 1866 (1/0), Young in 1905 (11) modified the operative procedure and made it more anatomic and practical. Since that time, there have been a few minor but important modifications of Young's technique (12-14). The criteria for patient selection were clarified in 1968 when Jewett (15) reported The Johns Hopkins Hospital expe- rience with radical prostatectomy for the palpable nodule of prostate cancer. Between 1909 and 1951, 103 men with a palpable nodule of prostate cancer underwent radical prosta- tectomy, and all had follow-up for at least 15 years or until the time of death; 27% of the men with a palpable nodule of histologically differentiated and limited prostate cancer sur- vived 15 years free of recurrence. Walsh and Jewett (16) reported a more recent series of radical prostatectomies per- formed for men with stage Bl disease between 1951 and 1963. The survival of patients with clinical stage B1 prostate 117 118 cancer who had undergone radical prostatectomy paralleled the survival curve of a 62-year-old man in the year 1960. Elder et al. (17) recently reviewed the long-term survival of men with clinical stage B2 prostate cancer who had under- gone radical perineal prostatectomy at The Johns Hopkins Hospital between 1951 and 1963. The survival curve for men with clinical stage B2 disease confined pathologically to the prostate paralleled that for men with clinical stage B1 disease who had this radical surgery performed. The presence of extraprostatic extension was associated with a marked decrease in survival. Despite the curative potential of radical prostatectomy, many patients and their physicians selected alternative forms of treatment, inasmuch as most patients were impotent after the operation. On the basis of anatomic studies performed in the male fetus and stillborn infants, Walsh modified the technique for radical retropubic prostatectomy to preserve sexual function (14,18). Sexual function can now be pre- served in 70% of the men with clinically localized prostate cancer who undergo radical prostatectomy (/9). STAGE B1 DISEASE Survival Analysis The duration of survival and the cause of death were determined for 57 men with clinical stage B1 prostate cancer who underwent radical prostatectomy at The Johns Hopkins Hospital between 1951 and 1963. The absence of metastatic disease was determined by a radiographic survey of the bones. The majority of cases were staged with serum acid phosphatase determinations. Staging pelvic lymphadenec- tomies and bone scans were not performed. Of the 57 patients, 54 had pathologic evidence of stage B disease. This is a unique series because all patients were carefully staged clinically and pathologically, and none received adjuvant hormonal therapy. Survival analysis was calculated according to the direct, the all-cause actuarial (Kaplan-Meier), and the cause- specific actuarial methods (20). The percentage of patients alive at the end of a specific interval is calculated with the direct method. In the all-cause and cause-specific methods, the record of each patient for the duration of follow-up is used, and therefore they include patients still alive at the end of the study and those lost to follow-up. The procedure is nonparametric because it does not assume a particular func- tional form, such as a negative exponential function, for the survival curve. A distinction is made between deaths from prostate cancer and deaths from causes unrelated to cancer. The primary advantage of cause-specific analysis is that this method controls for the effect on survival of various covaria- bles, such as general medical condition and age. Therefore, cause-specific survival analysis precisely indicates the impact of a therapeutic modality on survival because only death from a specific disease entity is evaluated. The 5-, 10-, and 15-year all-cause survival rates were determined with the direct method (fig. 1). The 5-, 10-, and 15-year all-cause survival rates were 83%, 63%, and 51%, respectively. The all-cause survival curve was compared with a projected average survival curve for a 62-year-old man in the year 1960 (fig. 1). The observed survival curve for men with clinical stage BI prostate cancer following radical peri- 100%, N a tony a 0 I he 0 L sO a 75 5 oOo ~ i > Ta z N® 57 CASES 1951-63 3 sof X US WHITE MALES, 1960 w oO ES — z w 25 Oo x w a 0 : . ‘ lo} 5 10 15 YEARS AFTER OPERATION FIGURE 1.—All-cause survival curves for 57 men with stage Bl prostate cancer who had radical prostatectomies at The Johns Hopkins Hospital between 1951 and 1963. Curves were determined by the direct method. neal prostatectomy was virtually identical to the projected survival curve of age-matched men selected from the general population. The survival rates were calculated with the use of the Kaplan-Meier procedure (fig. 2). Deaths directly attributed to prostate cancer and deaths unrelated to cancer are not differ- entiated when the all-cause actuarial survival analysis is used. The 5-,10-, and 15-year all-cause actuarial survival rates are virtually identical to the values calculated with the direct method. The all-cause survival curve decreases at a constant rate throughout the entire follow-up interval. This is not unexpected due to the advanced age of the men with prostate cancer who were followed up over a long interval. The survival rates were also calculated by cause-specific actuarial survival methods (fig. 2). A comparison of the all- cause and cause-specific survival curves indicates that the primary cause of mortality in surgically treated men with stage Bl disease is not cancer related. The cause-specific actuarial survival curve indicates that only 14% of the men with stage B1 disease and 15-year life expectancy will die with metastatic prostate cancer following radical prostatec- tomy. The cause-specific survival curve plateaus after a follow-up of 10 years, which indicates that the majority of men surviving 10 years are cured of their disease. Using disease-specific survival analysis, Bagshaw (27) has recently analyzed the survival of men with clinically localized pros- tate cancer (stages A and B) undergoing radiotherapy. The disease-specific survival curve following radiotherapy con- tinues to decline at a constant rate. Freiha and Bagshaw (22) recently observed that the incidence of positive biopsies fol- lowing external-beam radiotherapy for stage B disease was 50%. A positive biopsy after irradiation is associated with a significantly greater tendency for disease progression (23). It is possible that the continuous decline of the disease-specific survival curve following radiotherapy is attributed to the failure of radiotherapy to eradicate completely the local disease in these patients. The major limitation of all-cause survival analysis is that this procedure does not allow for the inclusion of potentially significant covariables such as age, race, and general medical NCI MONOGRAPHS, NUMBER 7, 1988 SURVIVAL PROBABILITY 0.0 1 1 denen Ad A 0 5 10 15 20 25 FOLLOW-UP TIME IN YEARS FIGURE 2.— Actuarial survival rates for 57 men with stage B1 prostate cancer who had radical prostatectomies at The Johns Hopkins Hospital between 1951 and 1963. Curves were determined by all-cause (—) and cause- specific (----) actuarial survival analysis. condition. The all-cause survival rate for a group of young healthy men with localized prostate cancer undergoing any therapeutic intervention will exceed that of an elderly debili- tated group of men receiving comparable treatment. This phenomenon is illustrated by determination of the nonpros- tate cancer, prostate cancer, and all-cause mortality rates as a function of age at the time of radical prostatectomy (table 1). The likelihood of mortality from nonprostate cancer-related causes increases stepwise as a function of increasing age. The mortality rate of prostate cancer (cause-specific) is inde- pendent of age. The all-cause mortality rate according to age and that of cancer other than prostate are parallel. We used the Cox proportional hazards method (24) as modified by Breslow (25) to evaluate the effect of age on survival and to compute the relative risk for age (table 2). This analysis clearly indicates that the all-cause mortality rates are age dependent (P = .003), whereas the cause-specific (prostate cancer) mortality rates are independent of age (P = .85). When comparing nonrandomized, nonconcurrent studies, one must consider that parameters other than therapeutic measures may account for survival differences. Specifically, if there is a substantial discrepancy in the interval before patients are initiated into studies, factors such as general advances in medical treatment may account for some of the observed survival differences (26). When nonconcurrent comparative studies of men with clinically localized prostate cancer are analyzed, improvements in the criteria for select- TABLE 1.—Mortality rate by age and by death category Mortality, % Age, No. of Nonprostate Prostate All yr patients cancer cancer causes 50-54 9 11.1 11.1 222 55-59 18 389 222 61.1 60-64 15 60.0 6.7 66.7 65-69 12 58.3 16.7 75.0 70-74 3 100.0 0.0 100.0 Total 57 47.4 14.0 61.4 119 TABLE 2.—Risk of death for patients over 50 yr of age relative to 50-yr-old patients? Relative risk Age, yr All causes Prostate cancer 50 1.0 1.0 55 1.37 (1.02, 1.85) 0.94 (0.48, 1.82) 60 1.89 (1.04, 3.41) 0.88 (0.24, 3.30) 65 2.59 (1.06, 6.29) 0.83 (0.11, 5.98) 70 3.55 (1.09, 11.62) 0.78 (0.06, 10.86) 4 Risk was estimated from Cox model. Values in parentheses are 95% confidence intervals. ing candidates for treatment may also account for observed survival differences. The limitations of nonconcurrent studies are illustrated by a review of The Johns Hopkins Hospital experience with radical prostatectomy for stage Bl disease (table 3). During the 1909-1951 interval, 103 men with stage Bl disease underwent radical prostatectomy, and the 15-year disease-free survival was 27%. Fifty-seven men with stage B1 disease underwent radical prostatectomy between 1951 and 1963, and the 15-year disease-free survival was 51%. The staging criteria were similar in these series except that the majority of men in the more current study had serum acid phosphatase determinations. The significant survival differ- ence observed in these series may be explained by advances in medical care and surgical technique. When modern radio- therapy series are compared with The Johns Hopkins Hospi- tal experience with radical prostatectomy, significant ad- vances in staging criteria and overall improvements in medical care are likely to favor the survival data following radiotherapy. Therefore, when modern radiotherapy series are compared with historical radical prostatectomy series, the impact of the time interval and the improvement in patient selection must not be overlooked. Morbidity Jewett (15) reviewed the morbidity following radical prostatectomy for a series of radical perineal prostatectomies performed at The Johns Hopkins Hospital for stage BI dis- ease between the years 1909 and 1951. The operative mortal- ity in this series was only 1%. Anastomotic strictures and thrombophlebitis occurred in 7% and 2% of the patients, respectively. Although intraoperative injury to the rectum occurred in 4% of the patients, none developed a rectal fistula. Mild stress incontinence occurred in 3% of the patients, and no man was totally incontinent. Following radical perineal prostatectomy, 90% of the men were impotent. Overall, even in the preantibiotic era, radical prostatectomy was performed TABLE 3.—Disease-free survival at 15 yr following radical prostatectomy for stage Bl disease Date of No. of prostatectomy patients Survival, % 1909-1951 103 27 1951-1963 57 51 MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 120 with minimal intraoperative and perioperative complications or long-term sequelae other than impotence. Impotence following radical prostatectomy occurs from injury to the autonomic innervation of the corpora cavernosa. On the basis of anatomic dissections performed in the male fetus and stillborn infants, Walsh and Donker (/8) delineated the anatomic pathway of the cavernous nerves, and the tech- nique for radical retropubic prostatectomy was modified so that the autonomic innervation to the corpora cavernosa was preserved (/4). Walsh and Lepor (/) recently reported the morbidity in 290 men on whom radical prostatectomies were performed at The Johns Hopkins Hospital between 1982 and 1985 using the nerve-sparing modifications. There were no intraoperative deaths or rectal injuries. Pulmonary emboli and ureteral injury occurred in less than 1% of the patients. No patient who has been followed up for at least 1 year was totally incontinent. No patients wear urinary appliances or have required anti-incontinence surgery; less than 5% of the patients wear a small pad in their trousers that may be changed once or twice a day. Seventy percent of the patients are potent. The complication rate following radical prostatec- tomy with modern surgical techniques remains low. The decision for alternative therapy based on relative morbidity does not seem to be justified. Local Disease Control Unfortunately, not all men with clinically localized pros- tate cancer are cured following radical prostatectomy because of the presence of micrometastatic disease. How- ever, complete excision of the local tumor burden in these patients may provide excellent local disease control. There- fore, when the relative efficacy of radical prostatectomy for men with clinically localized prostate cancer is considered, these parameters are relevant because local disease control and prolongation of disease-free survival may enhance the quality of life. The long-term experience at The Johns Hop- kins Hospital provides some insight into the local disease control following radical prostatectomy. Local recurrence following therapeutic intervention may represent histologic, palpable, or symptomatic local recur- rence. The clinical significance of local recurrence is the development of symptoms of bladder outlet obstruction and the potential for malignant cells to metastasize. Because the prostatic fossa is empty after radical prostatectomy, any local recurrence of significant volume is readily discernible by rectal examination. On the other hand, after radiotherapy, the prostatic fossa is indurated, and the distinction between changes due to radiation effects and those due to recurrence is not readily apparent. Because most consider the digital rectal examination inadequate for assessing local disease control following radiation therapy, surgeons at some centers are routinely performing postirradiation needle biopsies. Al- though a consequence of local recurrence is the development of symptomatic bladder outlet obstruction, few long-term studies following radiotherapy or radical prostatectomy address the issue of symptomatic local recurrence. Local (palpable) disease recurrence developed in 14 of 86 men (16%) with pathologic stage B disease who underwent radical prostatectomy between 1909 and 1951 (13). Because of improvementsin clinical staging that have developed since 1951, the vast majority of men with clinical stage B1 disease have tumors pathologically confined to the prostate. Overall, we can expect long-term local disease control in at least 85% of the men with stage Bl disease who receive this surgical treatment. STAGE B2 DISEASE Radical prostatectomy has been offered for the treatment of stage B2 prostate cancer with the realization that some men with clinically undetectable micrometastatic disease will not be cured. Elder et al. (17) recently reviewed the long-term experience with radical prostatectomy for stage B2 disease at The Johns Hopkins Hospital. The duration of survival and the cause of death were determined for 48 men with clinical stage B2 prostate cancer who underwent radical prostatectomy at The Johns Hopkins Hospital between 1951 and 1963 (fig. 3). The absence of metastatic disease was ascertained by plain-film radiograms of the bones only. Sur- vival was determined by the direct method. We are currently evaluating this clinical series using cause-specific survival analysis. The 5-, 10-, and 15-year disease-free survival rates following radical prostatectomy were 75%, 50%, and 25%, respectively. The observed survival curve calculated by the direct method for men with clinical stage B2 disease is far less favorable than that for men with clinical stage B1 disease. The survival data were stratified according to the presence or absence of pathologic evidence of extraprostatic extension (fig. 4). When the tumor was pathologically confined to the prostate, the 5-, 10-, and 15-year disease-free survival rates were 63%, 63%, and 50%, respectively; when the tumor extended beyond the prostate (seminal vesicle invasion), the rates were 75%, 40%, and 13%, respectively. Thus the cure of men with clinical stage B2 disease is influenced by pathologic stage. The observed 15-year disease-free survival rate calcu- lated by the direct method for clinical stage B2 disease patho- logically confined to the prostate was equivalent to the observed survival rate for men with clinical stage B1 disease. Men with clinical stage B2 disease pathologically confined to the prostate are excellent candidates for radical prostatec- 100 75 50 Percentage Survivorship 25 0 5 10 15 Years After Operation FIGURE 3.—All-cause disease-free survival curves for 48 men with clinical stage B2 prostate cancer who had radical prostatectomies at The Johns Hopkins Hospital between 1951 and 1963. Curves were determined by the direct method. NCI MONOGRAPHS, NUMBER 7, 1988 100 BS £75 Se Oo 2 Ss 3 8 nN 1 © 50 B2 (Pathologically 2, Confined to Prostate) Oo © O oS 25 : © a Nes, (Extraprostatic Extension) 0 5 10 15 Years After Operation FIGURE 4.—Influence of pathologic stage on disease-free survival in men with clinical stage B2 prostate cancer who had radical prostatectomies at The Johns Hopkins Hospital between 1951 and 1963. Survival data are analyzed for men with stage B1 disease (—); stage B2 disease pathologi- cally confined to prostate (---); and stage B2 disease pathologically extending beyond prostate (—-—). tomy. Recent advances in clinical staging criteria have improved physicians’ ability to recognize those with clinical stage B2 disease who have no pathologic evidence of extra- prostatic extension. When clinical staging criteria were limited to plain-film radiograms and serum acid phosphatase determinations, only 31% of the men with clinical stage B2 disease were found to have pathologically confined tumors. Between 1982 and 1984, 22 men with clinical stage B2 prostate cancer underwent radical prostatectomy. The stag- ing criteria in this group of patients included staging lymphadenectomies, serum acid phosphatase determinations, and radionuclide bone scans. Only 25% of the prostatectomy specimens had pathologic evidence of seminal vesicle exten- sion (19). Despite improvements in staging criteria, radical prostatectomy is not an optimal treatment for all men with clinical stage B2 disease. The clinical dilemma is whether to withhold a potentially curative treatment for 73% of the patients or offer a potentially noncurative treatment for 27% of those with presumed micrometastasis. Hopefully, future improvements in imaging modalities, such as transrectal ultrasound or magnetic resonance imaging, in conjunction with new approaches to pathologic evaluation of biopsy specimens will aid in the identification of the ideal candidate for surgical cure. CONCLUSIONS The long-term experience with radical prostatectomy at The Johns Hopkins Hospital was reviewed for evaluation of the efficacy of this therapeutic option for the treatment of localized prostate cancer. The criteria for evaluation follow- ing radical prostatectomy included curability, morbidity, and local disease control. Direct method calculations showed that the survival curves for men with clinical stage Bl disease and for men with clinical stage B2 disease pathologically confined to the pros- MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 121 tate were identical to the projected survival curve for a 62- year-old man in 1960. Radical prostatectomy usually results in cure when the tumor is pathologically confined to the prostate. Although men with stage B1 disease are the optimal candidates for this radical surgery, it is likely that most men with clinical stage B2 disease, according to modern staging criteria, may also be cured of their disease following radical prostatectomy. The survival rate of men with clinical stage B1 disease who had undergone radical prostatectomy was ascertained by cause-specific survival analysis; the primary advantage of this analysis is that the impact of a specific therapeutic modal- ity on survival is more precisely determined when death from a specific disease entity is evaluated. The cause-specific sur- vival curve for men with clinical stage B1 disease plateaued at 10 years, which indicated that the majority of men surviv- ing 10 years were cured of their disease. On the basis of data from this series of patients, men with clinical stage B1 pros- tate cancer who undergo radical prostatectomy may be told that the chance of death from prostate cancer within 15 years of surgery averages 14% + 5%. This long-term retrospective analysis of The Johns Hop- kins Hospital experience with radical prostatectomy illus- trates several fundamental caveats for comparison of non- randomized, nonconcurrent studies of radical prostatectomy and radiotherapy. Cause-specific survival analysis represents a useful statistical method for us to ascertain survival rates for men with clinically localized prostate cancer. When nonran- domized, nonconcurrent studies are compared, one must rec- ognize that parameters other than therapeutic measures may account for survival differences. Researchers’ attempts to compare modern radiotherapy series with the historical radical prostatectomy series at The Johns Hopkins Hospital are faulted due to significant advances in overall medical care and staging criteria that were available in the modern radio- therapy series. A review of the historical data and the more recent expe- rience with radical prostatectomy at The Johns Hopkins Hospital indicate that the operative procedure can be per- formed with minimal morbidity. This therapeutic option can be offered with the virtual assurance of the patient achieving urinary continence. Historically, despite the curative poten- tial of radical prostatectomy, many patients and their physi- cians selected alternative forms of treatment because most patients were impotent after the surgery. Because of modern modifications in surgical technique, radical prostatectomy can be performed with preservation of potency in the major- ity of patients. On the basis of curability, morbidity, and control of local disease, radical prostatectomy represents an effective treat- ment for clinically localized prostate cancer. REFERENCES (1) WALSH PC, LEPOR H. The role of radical prostatectomy in the management of prostate cancer. Cancer 1987;60:526-537. (2) BAGSHAW MA. Potential for radiotherapy alone in prostatic cancer. Cancer 1985;55:2079-2985. (3) GROSSMAN HB, BATATA M, HILARIS B, ET AL. 25] implanta- tion for carcinoma of the prostate: Further follow-up of first 100 cases. Urology 1982;20:591-598. 122 (4) BARNES RW, NINAN CA. Carcinoma of the prostate—biopsy and conservative therapy. J Urol 1972;108:897-900. (5) HANASH KA, Utz DC, Cook EN, ET AL. Carcinoma of the prostate: A 15-year follow-up. J Urol 1972;107:450-453. (6) CHoprp RT, WHITMORE WF JR. Delayed treatment of prostatic cancer. Presented at the American Urological Association Meeting, 1980, abstr 234. (7) SCHMIDT JD, MEHLIN CJ, NATARAJAN N, ET AL.Trends in patterns of care for prostatic cancer, 1974-1983: Results of surveys by the American College of Surgeons. J Urol 1986;136:416-421. (8) EPSTEIN JI, EGGLESTON JC, PAULL G, ET AL. Prognosis of untreated stage A1 prostate carcinoma: A study of 94 cases with extended follow-up. J Urol 1986;136:837-839. (9) BLUTE ML, ZINCKE H, FARROW GM. Long-term follow-up of young patients with stage Al adenocarcinoma of the pros- tate. J Urol 1986;136:840-843. (10) KuckLER H. Uber prostataver-Grosserungen. Deutsch Klin 1866;18:458. (11) YOUNG HH. The early diagnosis and radical cure of carcinoma of the prostate—a study of 40 cases and presentation of a radical operation which was carried out in four cases. Johns Hopkins Hosp Bull 1905;16:315-321. (12) BELT E. Radical perineal prostatectomy in early carcinoma of the prostate. J Urol 1942;48:287-297. (13) HODGES CV. Vesicourethral anastamosis after radical prosta- tectomy: Experience with the Jewett modification. J Urol 1977;118:209-210. (14) WALSH PC, LEPOR H, EGGLESTON JC. Radical prostatectomy with preservation of sexual function: Anatomical and pathological considerations. Prostate 1983;4:472-485. (15) JEWETT HI. The results of radical perineal prostatectomy. JAMA 1969;210:324-325. (16) WALSH PC, JEWETT HI. Radical surgery for prostatic cancer. Cancer 1980;45:1906-1911. (17) ELDERIJS,JEWETT HJ, WALSH PC. Radical perineal prostatec- tomy for clinical stage B2 carcinoma of the prostate. J Urol 1982;127:704-706. (18) WALSH PC, DONKER PJ. Impotence following radical prosta- tectomy: Insight into etiology and prevention. J Urol 1982;128:492-496. (19) EGGLESTON JC, WALSH PC. Radical prostatectomy with preservation of sexual function: Pathological findings in the first 100 cases. J Urol 1985;134:1146-1148. (20) BYAR DP. The Veterans Administration Cooperative Urologi- cal Research Group's studies of cancer of the prostate. Cancer 1973;32:1126-1130. (21) BAGSHAW MA. The role of radiation therapy in the manage- ment of prostatic carcinoma. In Problems in Prostatic Cancer Control (Paulson DF, ed), vol 1. Philadelphia: Lip- pincott, 1987, pp 181-192. (22) FREIHA FS,BAGSHAW MA. Carcinoma of the prostate: Results of postirradiation biopsy. Prostate 1984;5:19-25. (23) SCARDINO PT, WHEELER TM. Prostatic biopsy after irradia- tion therapy for prostatic cancer. Urology 1985;25(suppl): 39-46. (24) Cox DR. Regression models and life tables. J R Stat Soc B 1972;34:187-220. (25) BRESLOW NE. Covariance analysis of censored survival data. Biometrics 1974;30:88-99. (26) LEPOR H, Ross A, WALSH PC. The influence of hormonal therapy on survival of men with advanced prostatic cancer. J Urol 1987;128:335-338. Total Prostatectomy for Clinically Localized Prostate Cancer: Long-term Surgical Results and Current Morbidity’ Robert P. Gibbons? ABSTRACT —The outcome for the first 57 successive patients who underwent total perineal prostatectomy for clinically local- ized prostate cancer at the Virginia Mason Clinic and who have been followed up for a minimum of 15 years is reviewed for evaluation of the long-term impact of this operation on the dis- ease. Twenty percent of the patients had pathologic stage C dis- ease. Recurrence developed in 11 of the 55 patients (20%) who could be evaluated, and death from prostate cancer occurred in 6 (11%) during this interval. The actual observed overall sur- vival at 15 years or more was 60%, the actuarial survival 67%, and the cause-specific survival 89%. The current morbidity of this operation was evaluated by review of the last 50 consecutive patients who underwent this procedure and had follow-up of at least 6 months. Operative time averaged 140 minutes, and blood loss averaged 660 ml; 22% of the patients required a transfu- sion. Average postoperative hospitalization was 5 days. Two pa- tients required a temporary colostomy for unrecognized rectal injury, and 2 developed a stricture requiring more than one dila- tion. Three patients (6%) wear pads for mild stress incontinence. One patient died of a cerebral vascular accident. —NCI Monogr 7:123-126, 1988. The goal of those who manage localized prostate cancer should be to provide long-term disease-free survival, with minimal risk of long-term morbidity from the treatment. Of all the patients with prostate cancer, the disease is clinically localized (stage A or B) in 40%-50% of them (/). For better definition and clarification of the current role of total prostatectomy in the management of patients with clinically localized prostate cancer, the long-term surgical results and current morbidity data from the Virginia Mason Clinic are reviewed. PATIENTS AND METHODS We reviewed the data on 57 successive patients who underwent total perineal prostatectomy between 1954 and 1971 at the Virginia Mason Clinic and who have been fol- lowed up a minimum of 15 years to evaluate the long-term impact of this operation on this disease. No patients who underwent total prostatectomy during this interval were ex- cluded. Two patients were lost to follow-up, leaving 55 patients who could be evaluated, 44 with clinical stage Bl and 11 with clinical stage B2 disease. Patients were judged to have a stage Bl tumor if it was only pal- pable in one lobe or a stage B2 tumor if it was palpable in both lobes. The average age was 59 years (range, 45-73). ! Supported by funding from the Virginia Mason Research Center. 2 Section of Urology and Renal Transplantation, Department of Surgery, Virginia Mason Medical Center, 1100 Ninth Ave., Seattle, WA 98111. 31 thank Gloria Bailey, Ph.D., for assistance with data collection and analysis. Four patients who had pathologic stage C disease received adjuvant external-beam radiation therapy in the immedi- ate postoperative period; 1 also received postoperative hor- monal therapy (1 mg diethylstilbestrol daily). Otherwise, radiation therapy or hormone therapy was given only if re- current prostate cancer was documented. None of these 57 patients had staging lymphadenectomy. The current morbidity of total perineal prostatectomy was evaluated by a review of the records of the last 50 consecutive patients who underwent this procedure and had follow-up of at least 6 months. During this same interval, an additional 19 patients elected to undergo the modified retropubic prostatectomy in an effort to preserve sexual function. Five patients (10%) had evaluation of the pelvic lymph nodes prior to perineal prostatectomy. Patient selection.—A patient was considered to be a candidate for total (radical) prostatectomy if he had 1) biopsy-proven tumor without palpable evidence of extension beyond the capsule or into the semi- nal vesicles in the opinion of 2 experienced urol- ogists, 2) normal level of serum acid phosphatase, 3) no evidence of metastatic disease on bone x-ray or in later years on bone scan supplemented by radiographs of any abnormal areas, 4) expected survival of at least 15 years, and 5) willingness to undergo an operation. In gen- eral, a staging pelvic lymphadenectomy is only performed in patients with clinical stage A2 or B2 or high-grade (Glea- son score, 8-10) lesions. Surgical procedure and follow-up.—All patients under- went a total prostatectomy via the perineal route, which was described initially by Young (2). Following recovery from the operation, patients were followed up at 6-month intervals, with history, acid and alkaline phosphatase deter- minations, and rectal examinations. If the history suggested bone involvement or systemic disease, repeat bone x-rays or scans were obtained. Evidence of local failure was con- firmed by biopsy. Patients were judged to be free of disease if they had 1) no history of failing health or bone pain and no evidence of local or regional disease on rectal exam- ination; 2) normal acid phosphatase and prostate-specific antigen; and 3) normal bone scan in light of questionable bone disease. RESULTS Long-term Of the 44 patients with clinical stage Bl and 11 with clinical stage B2 disease, 5 (11%) and 6 (55%), respectively, had microscopic pathologic stage C disease, for an overall incidence of 20% (table 1). Four patients with pathologic stage C disease received adjuvant external-beam radiation 123 124 TABLE 1.—Final pathologic stage of 55 patients with clinical stage B disease who were treated by total perineal prostatectomy TABLE 2.—Histologic grades of prostate cancer in 55 patients with clinical stage B disease? Pathologic stage C Clinical No. of No. of stage patients patients Percent Bl 44 5 11 B2 11 6 35 Total 55 11 20 therapy within 90 days of surgery. One patient died of metastatic prostate cancer 12 years later, 1 died of prostate cancer 5 years later, 1 died without evidence of prostate cancer 9 years later, and 1 is alive with no evidence of disease 17 years later. The histologic grades of prostate cancer in all 55 eval- uated patients are shown in table 2. The vast majority of patients had disease in the intermediate grades II and IIL Eleven patients (20%) developed recurrent disease; it was local in 2 patients, distant in 4, and local and distant in 5. Average time to recurrence was 7.3 years (range, 2-15). The influence of the pathologic stage on recurrence is seen in table 3. The recurrence rate was 16% during a mean of 7.6 years in patients with pathologic stage B disease and 36% during a mean of 6.8 years in patients with pathologic stage C disease. Two patients developed a local recurrence 6 and 7 years following surgery; both patients then received external- beam radiation therapy. They are disease free 20 and 12 years, respectively, following completion of radiation ther- apy. Another patient received radiation therapy for local recurrence 3 years following surgery and 1 year later de- veloped distant disease. He died of prostate cancer 3 years later. The overall actual observed survival rate 15 years af- ter surgery for the 55 patients with clinical stage Bl or B2 disease was 60% (table 4). During this interval, 6 pa- tients (11%) died of recurrent prostate cancer. The mean interval to death was 10.3 years (range, 7-14). The causes of death other than recurrent cancer were cardiovascular disease in 10 patients, carcinoma of the lung in 3, and other causes (leukemia, suicide, and pneumonia) in 3. The mean interval to death from a cause other than prostate cancer was 7.3 years (range, 1-14). Overall, disease-free, and cause-specific actuarial survival curves are seen in fig- ures 1-3. Current Morbidity The current morbidity observed after this operation was evaluated by review of the last 50 consecutive patients who Histologic grade Pathologic com —————— stage I II II Iv Total B 5 19 19 1 44 Cc = 4 5 2 11 Total 5 23 24 3 55 ¢ Grading system is that of Gaeta et al. (3). underwent this procedure and had follow-up of at least 6 months. The average operative time for a total perineal prostatectomy is 140 minutes, and estimated blood loss averages 660 ml; 22% of the patients require a transfu- sion. Average postoperative hospitalization is 5 days. The catheter is left indwelling for an average of 14 days fol- lowing surgery. In the current series, 24 of the 50 pa- tients (48%) had pathologic stage C disease. Eight per- cent of the patients had well-differentiated tumors (Gleason score, 2-4), 72% had moderately well-differentiated tumors (Gleason score, 5-7), and 20% had poorly differentiated tu- mors (Gleason score, 8-10). Three patients were rehospitalized following discharge for problems relating to their surgery. The most serious postoperative complication was the rectal cutaneous fistula seen in 2 patients that required colostomies to be performed. These complications undoubtedly represented unrecognized rectal injury at the time of surgery. The colostomy was tem- porary in 1 patient. The second patient had a cerebrovas- cular accident 9 days following the colostomy that resulted in death. This was the only death in more than 350 radical perineal prostatectomies performed at this institution. After recovery, 7 patients developed a stricture, but only 2 (4%) required more than one dilation. Forty-seven of the 50 patients (94%) have normal urinary control, and 3 (6%) wear pads for dampness. No patient has a degree of urine loss that requires the wearing of an appliance. None of the 26 patients asked has had an erection satisfactory for inter- course, but all 20 patients surveyed note that the sensation of orgasm is intact. To regain erectile function following surgery, 3 patients have had a penile prosthesis placed, and 4 use pharmaceutical means to achieve satisfactory erec- tions. DISCUSSION The results of a treatment for prostate cancer can be measured by its effect on local control of the primary tu- mor, recurrence rate (progression-free interval), or overall, disease-free, or cause-specific survival. To understand bet- TABLE 3.—Influence of pathologic stage B vs. stage C disease on recurrence rate, disease-free interval, and death rate? No. of yr Deaths from No. of years . Patients Recurrences to recurrence prostate cancer to death Pathologic stage No. Percent No. Percent Mean Range No. Percent Mean Range B 44 80 7 16 7.6 3-15 4 9 9 7-14 C 11 20 4 36 6.8 2-12 2 18 13 12-14 Total 55 100 11 20 7.3 6 11 10.3 ¢ Influence was determined 15 yr after total perineal prostatectomy in patients with clinical stage B disease. NCI MONOGRAPHS, NUMBER 7, 1988 TABLE 4.— Actual observed survival status of 55 patients with clinical stage B1 or B2 disease 15 yr after total perineal prostatectomy Clinical stage Patients Bl B2 Total No. Percent Alive Disease free 27 5 32 58 With disease 0 1 1 2 Dead Disease free 12 2 14 25 With disease 1 1 2 4 Of disease 4 2 6 11 Total 44 11 55 100 4 Expected 15-yr survival for men aged 59 yr during these study years was 56% (4). ter the influence of treatment for localized disease, it is necessary that one evaluate the natural history of untreated localized disease and the clinical course of patients who develop recurrences after treatment. Larson and Norlen (5) reported a 77% disease progression rate in 31 patients with clinical stage B disease who received no treatment and who had been followed for a mean of 6.5 years. Whitmore (6) reported a 54% progression rate (local and distant) in 13 untreated patients with stage B1 disease who had been fol- lowed up for 5 years. By 15 years, all the patients had disease progression. By comparison, only 20% of our surgi- cally treated patients had progression of their disease during a similar 15-year interval. The importance of disease control is emphasized fur- ther by the outcome for patients with newly diagnosed metastatic disease. Kramer et al. (7) have shown that 50% of the patients with positive regional lymph nodes are dead at 39.5 months, and Murphy and co-workers (8) re- ported that 50% of the patients with newly diagnosed bone 1.0 1 +——— Overall Survival o— — —o Expected Survival (Age 59) .80 2 S 60 Q e a s 2 40 = n .20 17 + 8% Sv 0.0 oO 5 10 15 20 25 30 35 Follow-up Time (years) FIGURE 1.—Overall actuarial survival of 57 patients with clinical stage B prostate cancer at =15 yr after total perineal prostatectomy examined with the expected survival of 59-yr-old males in a comparable popula- tion. MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 125 1.0 ~——— Disease-Free Survival 91 + 4% o— ——o Expected Survival (Age 59) 80+ 2 £60 0 ° a Ea Ss 40 3 @ 20 0.0 T T T T ¥ T 1 0 5 10 15 20 25 30 35 Follow-up Time (years) FIGURE 2.—Disease-free actuarial survival of 57 patients with clinical stage B prostate cancer at =15 yr after total perineal prostatectomy examined with the expected survival of 59-yr-old males in a comparable population. metastases are dead at 23 months. When patients develop hormone-refractory metastatic cancer, the average survival time is 10 months (9). The morbidity of hormone-refractory metastatic prostate cancer is significant. Common sequelae in the months prior to death are bone pain and anemia from bone marrow invasion; bladder dysfunction (reten- tion, incontinence, and hematuria); urinary tract infection; anorexia; and uremia from obstructed ureters. The 80% recurrence-free control rate observed 15 years or more af- ter total prostatectomy simply has not been recorded with the other available treatment modalities for prostate cancer and represents a significantly improved quality of life over that of the patient who develops a recurrence. Further evidence that treatment can influence the natural history of the disease can be seen by a comparison of the 1.01 .80 2 3 60- 0 o a © = 4 Z 40 = wn 204 0.0 T T T 1 oO 5 10 15 20 25 30 35 Follow-up Time (years) FIGURE 3.—Cause-specific actuarial survival of 57 patients with clinical stage B prostate cancer at =15 yr after total prostatectomy. 126 survival of treated patients with the life expectancy of the general male population of the same age in the same region during the same interval (fig. 1). The better overall sur- vival of these surgically treated cancer patients can be ex- plained by effective treatment and by selection criteria that exclude poor-risk patients with significant cardiovascular, respiratory, metabolic, and other diseases. Disease-free or cause-specific survival addresses better the effectiveness of the treatment alone in preventing recurrence. Disease-free survival in this surgically treated group is comparable to expected survival for up to 30 years of observation (fig. 2). Cause-specific survival did not plateau until 17 years had elapsed because these patients with clinical stage B dis- ease were treated with total perineal prostatectomy (fig. 3). These results emphasize the need for long-term observa- tion in the evaluation of any new form of treatment of this disease. Patients who undergo total prostatectomy at this time ac- cept a very small risk for long-term permanent complica- tions or mortality. Mild stress incontinence is infrequent. Although erectile function is usually lost following total perineal prostatectomy, Weldon and Tavel (/0) recently re- ported a modification to the perineal approach that spared erectile potency in 56% of their patients. Walsh et al. (11) reported an 83% potency rate after modified radical retro- pubic prostatectomy. If voluntary erectile function is lost following surgery or any of the other treatment modalities used, a penile prosthesis or pharmacologic-induced erec- tions have proved satisfactory because the sense of orgasm is intact. An additional advantage of total prostatectomy is the benefit of pathologic study of the neoplasm. Twenty per- cent of the patients in the long-term series and 48% in the current series had pathologic stage C disease. When this occurs, adjuvant radiotherapy can decrease the incidence of subsequent local recurrence, and its use should be con- sidered, particularly in patients whose tumors extend to the surgical margins or who have seminal vesicle invasion (12). The experience at the Virginia Mason Clinic demon- strates that if disease-free, long-term survival is most important to the patient with clinically localized prostate cancer, total perineal prostatectomy remains the proven treatment of choice. Minimal risk is incurred from the oper- ation, and the potential complications are well defined and manageable. REFERENCES (1) MurPHY GP, NATARAJAN N, PONTES JE, ET AL. The national survey of prostate cancer in the United States by the American College of Surgeons. J Urol 1982;127:928-934. (2) YOUNG HH. The early diagnosis and radical cure of carci- noma of the prostate: A study of 40 cases and presentation of a radical operation which was carried out in four cases. Johns Hopkins Hosp Bull 1905;16:315-321. (3) GAETA JF, ASIRWATHAM JE, MILLER G, ET AL. Histologic grading of primary prostate cancer: A new approach to an old problem. J Urol 1980;123:689-693. (4) Washington State Life Tables for Males. State Life Tables, US Decennial Life Tables for 1969-71. DHEW Publ No. (HRA)75-1151. Washington, DC: National Center for Health Statistics, 1975. (5) LARSON A, NORLEN BI. Five year follow-up of patients with localized prostatic carcinoma initially referred for expec- tant treatment. Scand J Urol Nephrol [Suppl] 1985;93:30. (6) WHITMORE WF. A multidisciplinary analysis of controversies in the management of prostate cancer. Presented at the Organ Systems Program of the National Cancer Institute, Prouts Neck, ME, Oct 17-19, 1986. (7) KRAMER SA, CLINE WA JR, FARNHAM R, ET AL. Prognosis of patients with stage D-1 prostatic adenocarcinoma. J Urol 1981;125:817-819. (8) MURPHY GP, BECKLEY S, BRADY M, ET AL. Treatment of newly diagnosed metastatic prostate cancer patients with chemotherapy agents in combination with hormones ver- sus hormones alone. Cancer 1983;51:1264-1272. (9) EISENBERGER MA, SIMON R, O'DWYER PJ, ET AL. A reevalua- tion of nonhormonal cytotoxic chemotherapy in the treat- ment of prostatic carcinoma. J Clin Oncol 1985:;3:827- 841. (10) WELDON VE, TAVEL FR. Potency sparing radical perineal prostatectomy: Anatomy, surgical technique and initial results. J Urol 1987;137:225. (11) WALSH PC, LEPOR H, EGGLESTON JC. Radical prostatectomy with preservation of sexual function: Anatomical and pathological considerations. Prostate 1983;4:473-485. (12) GiBBONS RP, COLE BS, RICHARDSON RG, ET AL. Adjuvant radiotherapy following radical prostatectomy: Results and complications. J Urol 1986;135:65-68. Randomized Series of Treatment With Surgery Versus Radiation for Prostate Adenocarcinoma David F. Paulson! ABSTRACT —In the early 1970s, a multicentered cooperative group effort was established by urologists and oncologists to ex- amine the relative disease control provided by surgery, radiation therapy, or observation for patients with localized or regional disease. The data derived from this trial were controversial be- cause they 1) did not support previous concepts regarding the relative impact of treatment and 2) raised provocative questions as to the interpretation of previous institutional reports that pro- moted a single treatment modality. The data from the random- ized trial demonstrated that: 1) bipedal lymphangiography could not demonstrate accurately the presence or absence of micro- scopic involvement of pelvic lymphatic structures; 2) treatment selection should be based on the anatomic distribution of disease; 3) a clinician’s use of first appearance of local or distant dis- ease in a patient who was supposedly disease free after receiving the chosen therapy served as an accurate way to define the im- pact of the initial treatment; 4) radical surgery was more effective than radiation therapy in controlling disease that was clinically confined to the primary organ of origin; and 5) the apparent disease control produced by radiation on large-volume, localized disease might only reflect the natural history of the disease. —NCI Monogr 7:127-131, 1988. Early in the 1970s, a large multi-institutional study in- volving investigators at 13 major medical centers and their associates at Veterans Administration Medical Centers set out to ascertain their ability to determine the accuracy of current staging studies in assessing the anatomic distri- bution of the host at risk and, upon identification of the anatomic distribution of disease, to compare accepted treat- ment modalities in the management of that disease (/,2). Patient accession began February 1975, and the last patient was entered on September 8, 1978, the day that funding was withdrawn by the National Cancer Institute. In this study, 509 men with newly diagnosed, biopsy-proven prostate adenocarcinoma were assigned a preliminary clinical stage based on rectal examination, colorimetric serum acid phos- phatase levels, a plain chest x-ray, and a metastatic bone survey. The scheduling of staging studies, designed to de- termine the anatomic distribution of disease, was sequenced to progress from those global studies that were designed to show the most widespread evidence of disease and to focus gradually on the prostate. Thus all men with an elevated acid phosphatase were believed to have systemic disease and were excluded from the randomization schema for pa- tients with apparently localized or regional disease. In ac- cordance with the sequential schema, all men who demon- strated no evidence of osteoblastic disease on skeletal films were subjected to a technetium-99-mentodextra posterior ! Division of Urologic Surgery, Department of Surgery, Duke University Medical Center, P.O. Box 2977, Durham, NC 27710. bone scan. Approximately 25% of all patients with no bony disease identified on routine skeletal x-rays had bony dis- ease on the basis of isotopic bone scanning. The frequency of bone disease increased as the volume of local disease increased [fig. 1; (3)]. Following the exclusion of disease in the axial and ap- pendicular skeleton, the next region assessed involved the nodal structures of the pelvis that drained the prostate lymphatics. Patients who demonstrated no evidence of bone-positive disease underwent bipedal lymphangiogra- phy, with the accuracy of the lymphogram determined by pelvic lymphadenectomy. The hypothesis was that prostate carcinoma, like most other human solid tumors, has the propensity to metastasize early to the regional lymphatics. The lymphatics of the prostate exit the posterior as- pect of the gland and involve initially the hypogastric (pri- mary), obturator (secondary), external iliac (tertiary), and presacral (quaternary) lymphatics (4). The anatomic lim- its of the staging lymphadenectomy were designed not only to encompass the primary and secondary areas of nodal drainage but to leave undisturbed tertiary lymphatics lat- eral to the external iliac artery and vein. Thus the mar- gins of the node dissection were limited to the triangle bor- dered by the external iliac vasculature, the pelvic floor, and the hypogastric vasculature; all node-bearing tissue in this area was removed, including the node-bearing tissue sur- rounding the obturator nerve and vessels (fig. 2). As an- ticipated, the incidence of node-positive disease was found to increase as the volume of local disease increased (fig. 3). When the lymphograms were interpreted by a radiolo- gist (single review) and the results compared with the in- cidence of node-positive disease as determined by pelvic lymphangiography, it was noted that, when the nodes were called positive, they were positive 90% of the time. How- ever, when they were called negative, there was a 12% false-negative call. The success of the review radiologist should be contrasted with that of the institutional lymphan- giographer who had a 27% false-positive rate and a 44% false-negative rate (table 1). In an attempt to determine whether histopathologic char- acteristics of the tumor tissue biopsied (to establish the diagnosis of cancer) could be used to predict the pres- ence or absence of nodal extension, the investigators an- alyzed the incidence of positive nodes as a function of the Gleason sum. The product of such an analysis demon- strated that both high- and low-grade disease (determined by the Gleason sum) functioned as relatively accurate pre- dictors of node-positive disease, equivalent to that of cur- rent imaging modalities. Eighty-seven percent of the pa- tients with a Gleason sum less than 5 had node-negative disease, whereas 100% of those with a Gleason sum of 9 127 128 E223 Negative Bone Survey EZZ2 Positive Bone Scan 100 Xl Ts} Z A I-8 II Preliminary Clinical Stage FIGURE 1.—Patients with bone disease detected by radioisotopic scanning who were considered disease free by skeletal survey. N.A. = not appli- cable. Reproduced with permission from (3). or 10 had node-positive disease (table 2). With some mi- nor variations, this general trend has been substantiated by other investigators. Patients whose disease was confined to the prostate, de- termined by digital examination, and who had no evidence of bone- or node-positive disease were assumed to have organ-confined disease and were randomized to radical prostatectomy or external-beam radiation therapy. Patients with clinical stage A2 or stage B (T1-2 NO MO) were ran- domized in balanced groups of 4 by institution to either rad- ical prostatectomy or megavoltage radiation therapy (5). Any patient with an occult focal cancer was excluded from this randomization scheme, as were any patients who had clinical stage C (T3 NO MO) disease. Prostatectomy could be accomplished by a perineal or retropubic route; however, the anatomic limits of the dissection had to in- clude the apex of the prostate and the seminal vesicles (fig. 4). Patients assigned to radiation therapy were treated with megavoltage equipment, i.e., the highest available en- ergy (cobalt-60, linear accelerator, and/or betatron x-ray beam), with a minimum surface-to-axis distance of 80 cm. The field size was to have included the prostate, peri- prostatic region, and pelvic lymph nodes as determined Obturator n Hypogastric - Super ficial inguinal NL $4 W» 7 £xt hac d A E223 Negative Pelvic Nodes [7771 Positive Pelvic Nodes 100 % 50 II NA IV-8 II-C I-A I-B I Preliminary Clinical Stage FIGURE 3.—Incidence of pelvic node extension as a function of preliminary clinical stage in patients with no bone disease as determined by isotopic bone scan. Reproduced with permission from (3). by lymphograms and localization films. The upper mar- gin of the radiation field was at the level of the iliac crest, the lateral margin at least 1 cm beyond the external iliac nodal chains, and the lower margin 1 cm below the infe- rior prostate. All fields were treated with currently accepted schemata with a total tumor dose of 4,500-5,000 rad in 40 days total elapsed time. The dose was specified from the appropriate isodose curve as a minimum dose in the vol- ume of the prostate. An additional treatment boost of 2,000 rad was delivered to a reduced volume which had to include the prostate. Patients were followed at 2-month intervals for the first year and at 3-month intervals thereafter. Serum bio- chemical profiles with acid phosphatase determinations, Karnofsky’s performance ratings, and a physical exam- ination were obtained at each follow-up; chest x-rays and isotopic bone scans were obtained at 6-month in- tervals. The impact of treatment was determined with first-evidence-of-treatment-failure as the end point. Inves- tigators chose use of first-evidence-of-recurrent disease rather than survival to avoid confounding the impact of the first therapy by the subsequent application of a second FIGURE 2.—Shaded area indicates area of limited pelvic lymph node dissection for staging of prostate carcinoma. Repro- duced with permission from (3). Lymph nodes Superfigial and deep subinguinal NCI MONOGRAPHS, NUMBER 7, 1988 TABLE 1.—Lymphangiogram vs. node dissection Node dissection Lymphangiogram Radiologist results reviewer Positive, % Negative, % Positive Referee 90 10 Positive Institutional 73 27 Negative Referee 21 78 Negative Institutional 44 56 therapy that could alter the survival experience. Survival data had not been accrued because funding was withdrawn by the National Cancer Institute, and patient follow-up and data analysis could not be completed as had originally been projected. Furthermore, lack of funding precluded analy- sis of morbidity data and follow-up of accrued morbidity data. Treatment failure was identified by acid phosphatase elevation on two consecutive follow-ups or by the appear- ance of bony or parenchymal disease with or without con- comitant acid phosphatase elevation. The appearance of increased isotope uptake in an area previously not demon- strating such increased uptake was identified as the pres- ence of bony disease. The appearance of such metastatic bony disease was either progressive on subsequent scans or was accompanied by elevation of serum acid phosphatase levels. Identification of cancer in the prostate on follow-up after radiation did not signify treatment failure in assess- ment of relative treatment efficacy. Curves representing nonparametric estimates of time-to-first-evidence of treat- ment failure were generated according to the Kaplan-Meier method (6). Censored survival values representing patients without evidence-of-treatment failure at the time of last follow-up were represented by a single vertical tick on the treatment graph. Treatment efficacy in pairs of subgroups was assessed for differences by the Cox-Mantel test (7). Fifty-six patients received external-beam radiation therapy, and 41 underwent radical prostatectomy. An analysis of the time-to-failure curves of the 2 treatment groups indi- cates that radical surgery possessed a distinct advantage over radiation therapy in controlling disease and was sig- nificant at the .037 level (fig. 5). This study was critiqued; detractors argued that patients with more aggressive dis- ease had been assigned to radiation therapy and the length of follow-up was too short. Analysis of the relative Glea- son grade between the 2 groups indicated that the average Peritoneum Denonvilliers’ fascio,2 layers Fused FIGURE 4.—Limits of dissection for a radical prostatectomy. Reproduced with permission from (3). MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 129 TABLE 2.—Node biopsy; incidence of nodal extension as a function of Gleason grade No. of Gleason sum Positive, % Negative, % patients/group 2-5 13.9 86.1 36 6 324 67.6 34 7 499 50.1 21 P<.0005 8 75.0 25.0 12 9-10 100.0 0 7 No diagnosis 33.0 66.1 12 Gleason grade in the radiation therapy arm was 5.1, with an average Gleason sum of 5.5 in the radical surgery arm. If the Gleason sum functions as an indicator of biologic ag- gressiveness, the 2 groups seem equivalent, but those under- going surgery had a slightly higher Gleason sum than those treated by radiation. An additional 20 months of follow-up continued to demonstrate a disease control advantage of radical surgery over radiation therapy (fig. 6). In the original study design, persistent evidence of dis- ease after biopsy and after definitive radiation therapy did not constitute failure. However, current information would suggest that this does indicate a patient population at in- creased risk for progression. If the biopsy-positive patients (7 of 21 who had repeat biopsies at 1 yr were positive) are included as treatment failures, the curves diverge even far- ther (fig. 7). The imbalance that resulted, with respect to the statistical distribution of patients, was not intentional; the study was not weighted in favor of radical prostatec- tomy. The study was designed for researchers to use bal- anced groups of 4 within each of the 13 participating insti- tutions. Patient accession was halted on a specific day, and the randomization scheme was not allowed to go to com- pletion. At the time the study was completed, 42 patients 01 P=.037110 —_—— ft mb eh de te © = - 4 < [V _ 50 = wi - oo x wl a d 100 T T T Y T T T T T — 0 60 MONTHS Time -to-first-evidence of treatment failure for Radiation Therapy (mm) N= 56 Time- to -first-evidence of treatment failure for Radical Surgery (===) N=4] FIGURE 5.—Time-to-treatment failure for patients randomized to radiation therapy or radical surgery whose disease is confined to the prostate. Reproduced with permission from (3). 130 P= 054489 wn o 1 Percent failing 1 100 T T T T T T T T v 1 80 Time-to-first-evidence of failure MONTHS TOTAL Radiation—56 Surgery--—-- 41 FIGURE 6.—Time-to-treatment failure for patients randomized to radical surgery or radiation therapy when the analysis is completed with an additional 20 months of follow-up. The difference between the 2 treat- ment arms remains distinct. Reproduced with permission from (3). had been randomized for radical prostatectomy and 55 for radiation therapy. Four patients refused radical prostatec- tomy after they were randomized and demanded radiation therapy. Three patients who had been randomized to ra- diation therapy demanded radical prostatectomy. Thus, as the study analysis was designed to determine the impact of treatment, the final population pool under study encom- passed 41 patients who received radical prostatectomy and 56 who received radiation therapy. An additional interesting aspect of this randomized trial focused on large-volume, node-negative, localized disease (T3); the treated population was randomized to receive ei- ther radiation therapy or delayed hormonal therapy (no im- mediate treatment). Again, with first-evidence-of-treatment failure as the end point, the failure rates for patients actively treated by radiation therapy versus those treated by obser- vation only are superimposed (fig. 8). One may interpret these data to demonstrate that the purported disease con- trol impact of radiation therapy on large volume prostate cancer may reflect nothing other than the natural history of the disease. One may ask whether these randomized trial data pur- porting the superior impact of radical surgery reflect insti- 100 = .021602 on o Percent Failing Total ~—— Radiation, N = 56 «=== Surgery, N = 41 80 Months Time-to-first-evidence of failure FIGURE 7.—Time-to-failure curves for patients randomized to either definitive radiation or to radical prostatectomy. Patients failing radi- ation in this efficacy projection include the 7 of 21 patients who had postradiation-positive biopsies. % FAILING | — Deloyed Hormonal Therapy 2 == Radiotherapy 100 Ta 40 seo MONTHS FIGURE 8.—Difference in interval to first-evidence-of-treatment failure be- tween both groups was not statistically significant (P= .71). Reproduced with permission from (3). tutional experience, wherein radical surgery is the primary form of therapy chosen for apparently localized disease. Recent review of 280 patients at Duke University Medi- cal Center who underwent radical prostatectomy demon- strates an overall disease control response (fig. 9) similar to that seen within the randomized trial. However, the pop- ulation base within this single institutional experience can be manipulated in a retrospective manner to project a most favorable outcome. When the patients were retrospectively analyzed to reflect the volume of disease in the patients being subjected to radical surgical control, patients who had small-volume, organ-confined disease demonstrated a 6% failure rate at 6 years, whereas those who had micro- scopic stage C cancers confined to the surgical specimen experienced a 22% failure rate at 6 years, in contrast to patients with margin-positive disease who experienced a 35% failure rate (fig. 10). These retrospective data projec- tions document the difficulty any investigator has in provid- ing an unbiased data analysis of nonrandomized prospec- tive studies. Any retrospective nonrandomized series may be biased if the population is weighted by patients who have small-volume disease as identified by transurethral re- section; approximately 15% to 20% of these patients will be disease free after transurethral resection alone, whereas another 15% to 20% will only have a single microscopic focus of disease in the excised surgical specimen. Inclu- sion of such patients in any nonrandomized series will weight the study for success. Similarly, if the population is weighted with either high- or low-grade disease (as scored by Gleason sums), that population which is weighted with o [=] ® Probability of Not Failing Years FIGURE 9.—Time-to-failure curves for patients undergoing radical prosta- tectomy at Duke University Medical Center (January 1, 1987; stage A vs. B) as a function of diagnosis by transrectal needle biopsy (204; heavy solid line) or transurethral resection of the prostate (76; slanted broken line). NCI MONOGRAPHS, NUMBER 7, 1988 o = ‘© uw - oO Zz - ° > E a === Unknown, N/A n @ | 7 Organ contined 154 8 mmm Specimen confined 43 2 a #225 Margin positive 72 0 — . . , i — 0 1 2 3 4 5 6 7 8 9 10 Years FIGURE 10.—Time-to-failure curve for the population at Duke University Medical Center (January 1, 1987) undergoing radical prostatectomy. Patient population has been segregated as to whether the tumor was organ confined, specimen confined, or margin positive. high-grade disease will be destined to fail at a more rapid rate than that which is weighted with low-grade disease. In conclusion, although the randomized trial as conducted at the Veterans Administration Medical Centers may be flawed, it is an unbiased, randomized, prospective trial in which physicians based treatment selection on the anatomic distribution of disease and used the most current methodol- ogy to determine the local and regional extent of malignant growth. Differences in the application of surgical and ra- diotherapeutic treatment at the various test sites may not MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 131 have been equivalent; however, the treatment application does reflect the real world situation and, as such, should be viewed as presenting data that are current with respect to the relative impact of the 2 treatments. For one to deny the study because it does not fit preconceived notions is improper. REFERENCES (1) PAULSON DF, URO-ONCOLOGY RESEARCH GROUP. The impact of current staging procedures in assessing disease extent of prostatic adenocarcinoma. J Urol 1979;121:300-305. (2) PAULSON DF, URO-ONCOLOGY RESEARCH GROUP. Predictors of lymphatic spread in prostatic adenocarcinoma. J Urol 1980;123:697-699. (3) DEKERNION JB, PAULSON DF, eds. Genitourinary Cancer Management. Philadelphia: Lea & Febiger, 1987. (4) SMITH MJV. The lymphatics of the prostate. Invest Urol 1966;3:439. (5) PauLsoN DF, LIN GH, HINSHAW W, ET AL. Radical surgery vs. radiotherapy for stage A2 and stage B (T,_,M(N,) ade- nocarcinoma of the prostate. J Urol 1982;128:502-504. (6) KapLAN EL, MEIER P. Nonparametric estimation from incom- plete observations. J Am Stat Assoc 1958;53:457-481. (7) Cox DR. Regression models and life tables. J R Stat Soc [B] 1972;34:187-220. Radical Retropubic Prostatectomy With Reduced Morbidity: An Anatomic Approach Patrick C. Walsh! ABSTRACT—The morbidity of radical retropubic prostatec- tomy for prostate cancer has been reduced through improved un- derstanding of the surgical anatomy of the prostate. Delineation of the anatomy of the dorsal vein complex has led to modifica- tions in the surgical technique that have reduced blood loss and improved surgical exposure. The addition of epidural anesthesia and presurgical donation of autologous blood has limited the need for the homologous transfusion of blood to 2% of the patients and has reduced the frequency of serious perioperative complications such as pulmonary emboli to 0.3%. Delineation of the anatomy of the pelvic plexus and identification of the neurovascular bun- dles as the macroscopic landmark of the microscopic cavernous nerves have made it possible for the surgeon to make an informed decision at the time of surgery whether the neurovascular bun- dles can be preserved safely or excised widely with the specimen. In all surgical approaches to prostate cancer, the primary goal must be excision of all tumor; preservation of sexual function should be of secondary concern. These considerations were ad- dressed in the treatment of 320 consecutive patients; 74% of the men are potent postoperatively. It was necessary to excise one neurovascular bundle widely in 49 patients; 69% are potent. In addition to improvements in postoperative sexual function, the incidence of incontinence following surgery has been reduced. The total medical expenses for patients undergoing radical pros- tatectomy range from $8,500 to $9,500 and are similar to those for external-beam radiotherapy. With a reduction in overall mor- bidity, it is hoped that more men with localized disease will accept radical prostatectomy as the form of treatment having the great- est likelihood for cure with the expectation of an excellent quality of life postoperatively.—NCI Monogr 7:133-137, 1988. Although radical prostatectomy is effective in the cure of localized prostate cancer, the procedure has never gained widespread popularity because the operation was perceived to be technically difficult and associated with major com- plications. Thus some believed that the surgery was worse than the disease itself. For this reason, patients were often advised to do nothing or were encouraged to seek treatment with potentially less effective forms of therapy because they had fewer side effects. In offering any surgical procedure to a patient, the physi- cian (and patient) must weigh the potential benefits versus the risks. During the past decade, we have attempted to re- duce the risks of radical retropubic prostatectomy through investigations of the surgical anatomy of the prostate and adjacent structures. Over the past 10 years, the morbidity of the procedure has been reduced by lowered blood loss and improved operative exposure, reduced postoperative com- plications such as pulmonary emboli, and assured urinary control in virtually every patient and potency in most. At ! Department of Urology, Marburg 134, The Johns Hopkins Medical Institutions, Baltimore, MD 21205. the same time, this anatomic information has made it pos- sible for the surgeon to obtain wider margins of resection where indicated. REDUCED BLOOD LOSS BY IMPROVED VASCULAR CONTROL Delineation of the anatomy of the dorsal vein complex was the single most important observation in the evolu- tion of this surgical technique. Previously, most surgeons dreaded this operation because of the profound blood loss encountered when the dorsal vein complex at the apex of the prostate was transected. Once this had occurred, the remainder of the procedure was performed blindly and bluntly. Dissatisfied with this approach, we (/) examined the anatomy of the dorsal vein complex and Santorini’s plexus and, in 1979, published an anatomic approach to the surgical management of this venous complex. The modifi- cations in surgical technique included incisions in the en- dopelvic fascia near the pelvic sidewall away from the lat- eral venous plexus, division of the puboprostatic ligaments close to the pubis with care to avoid laceration of the su- perficial branch of the dorsal vein complex, and ligation of the main dorsal vein complex and its surrounding fas- cia anterior to the urethra. These maneuvers reduced blood loss and improved exposure so that the remainder of the procedure could be performed anatomically. More recently, we (2) undertook further efforts to re- duce blood loss during the operation. Bulldog clamps are placed temporarily on the hypogastric arteries to reduce arterial perfusion of the prostate during the early stages of the procedure. Epidural anesthesia has been shown to be effective in reducing blood loss during the operation, and all patients are encouraged to donate 3 units of their blood for autologous transfusion during the surgery. With this ap- proach, only 2% of the patients undergoing radical retropu- bic prostatectomy require transfusion of a homologous unit of blood. In addition, as noted by Modig (3), there has been a marked reduction in the frequency of pulmonary em- boli since the introduction of these recent advances. Of the first 125 patients, 3 developed pulmonary emboli postop- eratively; one occurred 3 weeks postoperatively (when the patient was at home) and resulted in sudden death. How- ever, following institution of routine epidural anesthesia and autotransfusions, only 1 of the next 300 patients had a pul- monary embolus. This nonfatal event occurred in a patient with a familial history of pulmonary emboli in two gener- ations. Thus radical prostatectomy can be performed today in a relatively bloodless, controlled fashion and is rarely associated with serious postoperative complications. 133 134 Shortly after the anatomy of the dorsal vein complex was delineated and this technique was applied during radi- cal retropubic prostatectomy, several patients reported that they were fully potent postoperatively, and more stated that some sexual function was present. This new finding encour- aged the suggestion in 1980 that preservation of potency following radical prostatectomy is possible (4). SEXUAL FUNCTION AND URINARY CONTINENCE Anatomy of Pelvic Autonomic Plexus For years it was assumed that impotence after radical prostatectomy resulted from injury to the autonomic nerves that innervate the corpora cavernosa; however, the exact anatomic location of these nerves was uncertain. In 1982, Walsh and Donker (5) suggested that impotence was in- duced by injury to the pelvic nerve plexus that provides autonomic innervation to the corpora cavernosa. On the basis of this observation, the technique of radical retro- pubic prostatectomy was modified slightly (6-9). Although the branches of the pelvic plexus that innervate the cor- pora cavernosa are microscopic, they can be recognized intraoperatively because of their rather constant associa- tion with the capsular vessels of the prostate (6,10). These neurovascular bundles are located in the leaves of the lat- eral pelvic fascia and previously were unknowingly injured during standard radical prostatectomy. However, once the precise location of these neurovascular bundles was deter- mined, the technique of radical retropubic prostatectomy could be modified in a way that enabled the surgeon to vi- sualize these bundles intraoperatively and to decide whether it was safe to preserve the neurovascular bundles or nec- essary to excise them with a specimen. This approach has resulted in preservation of sexual function in the major- ity of patients and has also improved postoperative urinary continence. Preservation of Sexual Function The surgical technique, which is described in detail else- where (8,9), has now been used on 320 men who have been followed up for 1 year or longer. In all procedures, the sur- geon made every attempt to excise all tumor; preservation of sexual function was of secondary concern. Of these 320 men, 259 were potent preoperatively and had sexual part- ners. Postoperatively, 192 (74%) are potent. Potency, which is defined as the ability to achieve an erection that is sufficient for vaginal penetration and or- gasm, returned gradually over the 1-2 years following surgery. Potency correlated with both the age of the pa- tient and the stage of the disease (table 1). Based on clinical stage, potency returned in 93% of the patients with stage Al disease, 72% with stage A2, 92% with stage BIN, 72% with stage B1, and 56% with stage B2. The tumor was con- fined within the capsule in virtually all men with stage Al disease; therefore, it is reasonable for one to assume that both neurovascular bundles can be preserved completely in almost all men with clinical stage Al disease and that this is responsible for the excellent return of sexual function in this group of patients. Conversely, the reduced rate of potency in men with stage B2 disease indicates that attempts at re- section of more advanced tumors result in greater injury to the pelvic nerve plexus. In 49 men, we had to excise one neurovascular bundle; 69% are potent postoperatively (/1). Thus one can conclude that potency can be maintained fol- lowing radical prostatectomy even when it is necessary to excise one neurovascular bundle widely. The information in table 1 is very useful in our preoperative advice to patients, on the bases of age and the stage of the lesion, as to the likelihood for the return of sexual function postoperatively. Preservation of Urinary Continence The most disabling and feared complication following radical prostatectomy is total urinary incontinence. Fortu- nately, it occurs infrequently in the hands of the experi- enced surgeon. There are several mechanisms of urinary continence in men: the vesical neck, the passive urethral mechanism, and the external sphincter-pelvic floor mecha- nism. All passive mechanisms, such as the bladder neck and the intrinsic urethral mechanism, are most effective when elevated by the tonic activity of the pelvic floor. The incon- tinence that occurs in patients after radical prostatectomy can often be related to rigidity of the remaining posterior urethra and nonelevation of the bladder base after voluntary cessation of urination (/2). Consequently, for the patient to achieve continence after radical retropubic prostatectomy, the surgeon must avoid injury to the pelvic floor mecha- nism, reconstruct the vesical neck so that it will provide a passive mechanism of continence, and avoid stricture for- mation by coapting the bladder neck to the urethra accu- rately. To ensure accurate coaptation of the bladder mucosa TABLE 1.—Influence of age and clinical stage on postoperative potency in 320 men with 1-yr follow-up after radical retropubic prostatectomy Age, yr? . 30-39 40-49 50-59 60-69 70-79 Total Clinical stage No Percent No. Percent No. Percent No. Percent No. Percent No. Percent Al — _ 2/2 100 9/10 90 3/3 100 _ — 14/15 93 A2 — — — _— 9/10 90 4/7 57 0/1 0 13/18 72 BIN 2/2 100 4/5 80 30/31 97 11/12 92 0/1 0 47/51 92 Bl — _— 11/14 79 42/51 82 39/60 65 1/5 20 93/130 72 B2 nn —_— 2/3 67 15/22 68 8/20 40 — — 25/45 56 Total 2/2 100 19/24 79 105/124 85 65/102 64 1/7 14 192/259 74 “ Values are numbers of men who had postoperative potency/total number who had preoperative potency. NCI MONOGRAPHS, NUMBER 7, 1988 to the urethra, thereby avoiding a vesical neck contracture, we have modified our technique by exteriorizing the bladder mucosa over the raw edges of the detrusor musculature (8). Recently, the role of the external sphincter mechanism has come under greater scrutiny, and we (9) have suggested that the autonomic innervation to the intrinsic skeletal muscle is important in ensuring urinary continence postoperatively. Of the 320 patients in this series who have been followed up for 1 year or longer, 296 (93%) are completely conti- nent and wear no pads or appliances, and 21 (7%) have mild stress incontinence for which they wear a small pad in their trousers. No patient 1) is totally incontinent, 2) wears an external drainage device, or 3) has undergone placement of an artificial sphincter. The frequency of stress inconti- nence was similar in patients with stages A and B disease, which suggests that a prior transurethral resection of the prostate was not an important risk factor in patients with mild stress incontinence postoperatively. Several authors have reported an improvement in results with urinary continence following adoption of this surgi- cal technique. Fowler et al. (/3) noted a reduction in total incontinence from 17% to 0% and occurrence of mild in- continence in only 4% of the patients. Similarly, O’Donnell and Finan (/4) reported a reduction in total incontinence from 12% to 0%, with only 6% of the patients having mild stress incontinence. They also noted an increase in the func- tional urethral length and an increase in the peak urethral pressure profile with this technique, compared with their prior standard surgical approach. These data suggest that preservation of the pelvic nerves during radical retropu- bic prostatectomy plays a role in functional maintenance of urinary continence postoperatively. PATHOLOGIC FINDINGS The autonomic branches of the pelvic plexus to the cor- pora cavernosa (cavernous nerves) are located outside the capsule of the prostate and Denonvilliers’ fascia; this is the first principle of this procedure. Thus if one knew that all tumors were intracapsular, it should be possible to preserve sexual function in all such patients without injury to these nerves. However, prostate cancer frequently penetrates the prostatic capsule into adjacent soft tissue. Thus a series of important interrelated questions will demonstrate whether preservation of sexual function compromises the removal of tumor. 1) Were the neurovascular bundles resected com- pletely during standard radical prostatectomy in the past? 2) What are the pathologic findings and surgical margins compared with those of standard prostatectomy? 3) Does this technique provide local and distant control of disease? The neurovascular bundles are located in the lateral pelvic fascia with the major tributaries of the dorsal vein of the penis and Santorini’s plexus. During standard radi- cal perineal prostatectomies, the lateral pelvic fascia was reflected off the prostate so that injury to the dorsal vein of the penis and Santorini’s plexus could be avoided (6). Thus the excellent long-term control of disease that has been reported following radical perineal prostatectomy (control of local disease and distant metastases) has been achieved without routine resection of the lateral pelvic fascia and neurovascular bundles. Consequently, one must assume that the neurovascular bundles were damaged but not com- MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 135 pletely resected during radical perineal prostatectomy. Also, according to the descriptions of radical retropubic prosta- tectomy available in standard textbooks, there is no evi- dence that the neurovascular bundles were completely re- sected with the retropubic approach either (15,16). The neurovascular bundles are intricately attached to the rec- tum, and to resect them completely, one must dissect imme- diately and deliberately very close to the anterolateral and lateral surfaces of the rectum. Previously, this was not part of the surgical procedure. However, one of the great advan- tages of the anatomic approach to radical prostatectomy is the accurate control of blood loss, which provides a clear surgical field. With this technique, all structures are clearly visualized, and a deliberate decision can be made on the basis of the operative findings as to whether structures can be preserved or resected more widely with the specimen. Thus one can resect the entire neurovascular bundle (9,71) if necessary by dividing the fascia and enclosed neurovas- cular bundle lateral to the urethra and resecting the fascia lateral to the rectum (table 2). In this manner, a wider mar- gin of resection can be achieved with this technique than was previously possible by blunt dissection. Recently, we evaluated the pathologic specimens re- moved from the first 100 consecutive patients who under- went nerve-sparing radical retropubic prostatectomy (/7). Although 41% of the patients had established capsular pen- etration, only 7 had positive surgical margins (17% with stage A disease; 18% with stage B2). These 7 patients had extensive periprostatic involvement by tumor; 5 had in- volvement of the seminal vesicles, but in none were the sur- gical margins positive only at the site of the nerve-sparing modification. On the basis of these findings, our surgical pathologist saw no indication that the nerve-sparing modi- fication compromised the adequacy of the removal of can- cer, which was determined primarily by the extent of the tumor rather than the operative technique (/7). Recently, we updated our results based on pathologic findings in 414 consecutive patients who had radical prostatectomies per- formed for clinical stages A and B cancer (table 3). Overall, 10% of the patients had positive surgical margins. The fre- quency of positive surgical margins was always equal to or less than the percent of patients with involvement of the TABLE 2.—Percent of neurovascular bundles resected in 482 consecutive patients having radical retropubic prostatectomy“ Percent bundles resected? Clinical No. of ee stage patients ++ +0 00 Al 21 81 19 0 A2 54 69 22 9 BIN 85 79 16 5 Bl 220 61 35 4 B2 98 20 67 13 DO¢ 4 50 25 25 Total 482 57 36 7 “On the basis of clinical findings at surgery, neurovascular bundles were either preserved or partially or completely resected. b ++ = both neurovascular bundles preserved; +0 = 1 neurovascular bundle preserved; 00 = no neurovascular bundle preserved. “DO indicates elevated acid phosphatase with no other evidence of metastatic disease. 136 TABLE 3.—Pathologic findings in 414 consecutive patients having radical prostatectomies Patients with pathologic findings, % Seminal Positive Positive Clinical No. of Confined Capsular vesicle surgical lymph stage patients to organ penetration involvement margin nodes Al 16 94 6 0 0 0 A2 40 83 17 15 15 10 BIN 78 82 18 3 3 0 Bl 196 65 35 10 9 3 B2 84 29 71 31 19 23 Total 414 64 36 13 10 7 seminal vesicles and similar to the frequency of patients with positive pelvic lymph nodes. This confirms our im- pression that patients with positive surgical margins have extensive disease and are in a high-risk category for devel- opment of distant dissemination. Efficacy of this technique will be established only when long-term follow-up evaluations have been performed for determination of whether the control of local disease and distant metastases is similar to results achieved with stan- dard radical prostatectomy. To aid in this comparison, we have been careful not to use adjuvant radiation therapy or hormonal therapy, so that we will be able to evaluate the true impact of radical prostatectomy alone on the control of prostate cancer. Of the 320 men who have been fol- lowed up for 1-5 years (table 1), 10 have developed dis- tant metastases as their first sign of failure. All 10 had extensive tumor in periprostatic soft tissue, and 8 under- went excision of their neurovascular bundles. Five of the 10 patients had involvement of the seminal vesicles, and 2 had positive pelvic lymph nodes. Three other patients devel- oped local recurrence as their first sign of treatment failure. Two patients had clinical stage B2 disease with extensive periprostatic involvement of soft tissue and positive surgical margins, and 1 patient with stage Bl disease had extensive periprostatic tumor with involvement of the seminal vesi- cles. Based on our experience with radical prostatectomy prior to the development of the nerve-sparing technique, we assumed that approximately 10% of the patients will develop local recurrence as the first sign of failure in the first 5 years postoperatively. Schellhammer (/8) has also reported a 10% failure rate in the first 5 years in patients with clinical stages A and B disease. With longer follow-up in that group, the total failure rate increased to only 15%. Thus most local failures are seen within the first 5 years. We are monitoring this closely in our group of patients, but more time must elapse to determine whether the total local recurrence rate will exceed 10% (29 more patients) over the next 4 years. COST OF SURGICAL CARE The cost of medical care is an important and legitimate concern for all. Indeed, the total medical costs for the sur- gical treatment of prostate cancer have not been described well in the past. We have recently evaluated the total med- ical costs (preoperative consultation, total hospital bill, and professional fees) for 12 patients who underwent radical retropubic prostatectomy with a 10-day hospital stay. The average total expense was $9,000 (range, $8,500-$9,500). If the hospital stay were reduced to 7 days, this cost could be reduced by $1,000. These medical expenses are com- parable to the total charges for external-beam radiotherapy for the treatment of localized prostate cancer at this hospi- tal, which average $9,000. FUTURE PROSPECTS Radical prostatectomy, like other cancer operations, is best suited for patients with early localized disease. How- ever, in the past, it was thought that the morbidity of this procedure was too high for the treatment of a disease that was detected so early in its onset and had a prolonged natu- ral history when treated conservatively. Refinements in the surgical anatomy of the prostate have made possible the reduction of the morbidity of this operation. By reducing blood loss, a surgeon can perform a more meticulous opera- tion with reduced blood transfusion requirements, improved visualization, wider excision of more advanced tumors than was previously possible, decreased frequency of pulmonary emboli, and improved sexual function and urinary conti- nence postoperatively. Because the morbidity of the procedure has been reduced, we can now offer this operation with confidence to a larger group of patients. Today it is clear that many patients with low-volume stage A disease may develop disease progres- sion without treatment, and thus radical prostatectomy is a more acceptable option for exactly the same group of young patients who found it unacceptable in the past (19). Through the use of improved screening techniques such as ultrasonography, more patients with early localized prostate cancer can be identified. Furthermore, through the use of better staging techniques, including improved imaging and tumor markers such as prostate-specific antigen, those pa- tients who are likely to benefit most from radical prosta- tectomy can be identified. REFERENCES (1) REINER WG, WALSH PC. An anatomical approach to the surgical management of the dorsal vein and San- torini’s plexus during radical retropubic surgery. J Urol 1979;121:198-200. (2) PETERS CA, WALSH PC. Blood transfusion and anesthetic practices in radical retropubic prostatectomy. J Urol 1985;134:81-83. NCI MONOGRAPHS, NUMBER 7, 1988 (3) MobIG J. Thromboembolism and blood loss. Continuous epidural block versus general anesthesia with controlled ventilation. Reg Anaesth 1982;7(suppl):84-88. (4) WaLsH PC. Radical prostatectomy for the treatment of localized prostate carcinoma. Urol Clinics North Am 1980;7:583-591. (5) WALSH PC, DONKER PJ. Impotence following radical prosta- tectomy: Insight into etiology and prevention. J Urol 1982;128:492-497. (6) WALSH PC, LEPOR H, EGGLESTON JC. Radical prostatectomy with preservation of sexual function: Anatomical and pathological considerations. Prostate 1983;4:473-485. (7) WALSH PC. Radical retropubic prostatectomy and cysto- prostatectomy: Surgical technique for preservation of sex- ual function (a film produced in 1984). Norwich, NY: Nor- wich Eaton Pharmaceuticals, Inc. (8) WALSH PC. Radical prostatectomy with preservation of sex- ual function. In Evolution of a Surgical Procedure, Update Ser, vol V. Bellaire, TX: Office of Education, Am Urol As- soc, 1985, pp 1-7. (9) WALSH PC. Radical retropubic prostatectomy. /n Campbell’s Textbook of Urology (Walsh PC, Gittes RF, Perlmutter AD, et al, eds), vol 3, Sth ed. Philadelphia: Saunders, 1986, pp 2754-2775. (10) LEPOR H, GREGERMAN M, CROSBY R, ET AL. Precise local- ization of the autonomic nerves from the pelvic plexus to the corpora cavernosa: A detailed anatomical study of the adult male pelvis. J Urol 1985;133:207-212. MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 137 (11) WaLsH PC, EpsTEIN JI, LOWE FC. Potency following rad- ical prostatectomy with wide unilateral excision of the neurovascular bundle. J Urol 1987;138:823-827. (12) HINnMAN F. Male incontinence: Relationship of physiology to surgery. J Urol 1976;115:274-276. (13) FOWLER JE JR, CLAYTON M, ROOHALLAH S, ET AL. Early experience with Walsh technique of radical retropubic prostatectomy. Urology 1987;29:242-246. (14) O'DONNELL PD, FINAN B. Urinary continence fol- lowing nerve sparing radical prostatectomy. J Urol 1987;137:225A. (15) MCLAUGHLIN AP III. Radical retropubic prostatectomy. In Campbell’s Textbook of Urology (Harrison JH, Gittes RF, Perlmutter AD, et al, eds), 4th ed. Philadelphia: Saunders, 1979, pp 2318-2326. (16) PETERS PC. Radical retropubic prostatectomy. In Urologic Surgery (Glenn JF, ed), 3rd ed. Philadelphia: Lippincott, 1983, pp 949-955. (17) EGGLESTON JC, WALSH PC. Radical prostatectomy with preservation of sexual function: Pathological findings in the first 100 cases. J Urol 1985;134:1146-1148. (18) SCHELLHAMMER PF. Radical prostatectomy. Patterns of local failure and survival in 67 patients. Urology 1988;31:191- 197. (19) EPSTEIN JI, PAULL G, EGGLESTON JC, ET AL. Prognosis of untreated stage Al prostatic carcinoma: A study of 94 cases with extended follow-up. J Urol 1986;136:837-839. ih . er | IV. Adjuvant Therapy I | -—_— -— Radiation Therapy as Adjuvant Treatment After Radical Prostatectomy Paul H. Lange,* Pratap K. Reddy, Eitan Medini, Seymour Levitt, Elwin E. Fraley’ ABSTRACT —Between 1977 and 1984, adjuvant radiation ther- apy was administered after radical prostatectomy to 71 patients at high risk for recurrence of carcinoma of the prostate. In 35 patients, tumor remained at the surgical margin (stage C2 dis- ease) and/or the disease had invaded the seminal vesicles (stage C3). Thirty-six patients had microscopic metastases in the pelvic lymph nodes (stage D1a). Radiation therapy was administered only after full recovery from surgery, which included full recov- ery of continence. The average period between surgery and ini- tiation of radiation therapy was 3 months. Serious or long-term complications attributable to irradiation occurred in 7% of the patients. Tumor recurred locally in only 2 patients. Five-year actuarial survival, disease-related survival, and disease-free sur vival for patients with stages C2 and C3 disease were 86%, 96%, and 80%, respectively. These survival values for patients with stage Dla disease were 74%, 90%, and 69%, respectively. Our results suggest a greater therapeutic benefit from radical prosta- tectomy and adjuvant radiation therapy than from radical prostatectomy alone for stages C2 and C3 disease or from radi- cal prostatectomy alone or radiation therapy alone for stage Dla disease; however, the length of follow-up, number of patients treated, and problems in comparing our results with those from historical controls do not allow us to draw firm conclusions about the benefits of this combined therapy. Controlled, randomized studies clearly are required. The serum levels of prostate-specific antigen, but not prostatic acid phosphatase, were invariably ele- vated in patients at the time of clinical detection of disease recur- rence and predicted recurrence up to 4 years before the event. Measurement of serum levels of prostate-specific antigen should be required in future studies of the efficacy of therapies for ap- parently localized carcinoma of the prostate. —NCI Monogr 7:141- 149, 1988. Radical prostatectomy after pelvic lymphadenectomy is an effective and increasingly popular treatment for patients with apparently localized carcinoma of the prostate (/). However, when this operation is applied broadly to such pa- tients, pathologic analysis of resected tissue reveals that, in 25% to 50% of the men, the tumor has extended locally be- yond the prostate (pathologic stage C), or has involved the pelvic lymph nodes (pathologic stage D1), or both (2-4). Many of these patients are at increased risk for local and ABBREVIATIONS: VAMC = Veterans Administration Medical Cen- ter, UMHC = University of Minnesota Hospital and Clinic; PAP = prostatic acid phosphatase; PSA = prostate-specific antigen. ! Departments of Urologic Surgery (P. H. Lange, P. K. Reddy, E. E. Fraley) and Therapeutic Radiology (S. Levitt), University of Minnesota Hospital and Clinic; and Urology Section (P. H. Lange, P. K. Reddy) and Therapeutic Radiology Service (E. Medini), Veterans Administration Medical Center, Minneapolis, MN. * Reprint requests to: Paul H. Lange, M.D., Department of Urologic Surgery, University of Minnesota Hospital and Clinic, Mayo Memorial Bldg., Box 394, 420 Delaware St., S.E., Minneapolis, MN 55455-0321. distant recurrence of the disease and need an effective ad- juvant therapy. Over the last 10 years, we have conducted a phase I-II trial to test the efficacy of and tolerance of patients for adjuvant radiation therapy after pelvic lymph- adenectomy and radical prostatectomy. Patients whose dis- ease was upstaged after surgery to pathologic stage D1 or high-risk stage C were involved. We (5) have reported on patient tolerance and the early results for our patients at the VAMC in Minneapolis. In this article, we update the follow-up on these patients and report on several additional patients from the VAMC and from the UMHC. METHODS Adjuvant radiation therapy was begun in June 1977 at the VAMC and formally in January 1983 at the UMHC. All candidates for the therapy had undergone pelvic lymph- adenectomy and radical prostatectomy. In all but 12, rad- ical prostatectomies were performed by the retropubic ap- proach. Before 1983, frozen sections of suspicious nodes were evaluated during surgery, and the procedure was ter- minated if tumor was detected. After 1983, in rare in- stances, radical prostatectomy was performed on a patient if one or two lymph nodes were visibly diseased but patho- logic evaluation detected no further involvement. Therefore, in the majority of patients (94%), those with positive lymph nodes had microscopic disease only (i.e, stage D1a), and in this article all D1 patients are so designated. During the surgery and after the prostate was removed, circumferential biopsies of the urethral and bladder neck anastomotic ends were submitted for frozen section anal- ysis. If the biopsy specimens contained tumor, more tissue was excised and evaluated. This continued until the biop- sies were negative or until it seemed that further excision would jeopardize postoperative continence. In the latter cir- cumstances, the anastomosis was completed, and it was as- sumed that the patient would receive adjuvant radiation therapy. Biopsies were performed during surgery because after radical prostatectomy the short segment of membra- nous urethra attached to the prostate virtually disappears upon fixation, and pathologists often take tissue from the prostatic apex and call it the distal margin. Similarly, the margin of the bladder neck is often difficult for them to as- sess after fixation (3). Finally, after 1979, surgeons did not remove tissue lateral to the external iliac artery of most pa- tients during pelvic lymphadenectomy so as to lessen their chances of postoperative leg edema (6). After final pathologic analysis, adjuvant radiation ther- apy was considered for patients who had completely recov- ered from the operation. The pathologic staging categories given in table 1 were used. Because we think that capsu- lar extension alone is a relatively less adverse prognostic parameter (7,8), stage C candidates for adjuvant radiation 141 142 TABLE 1.—Pathologic staging categories Stage Category description Al Focal, well-differentiated cancer discovered incidentally after radical prostatectomy A2 Incidentally discovered cancer that is diffusely distributed (>5% of the prostatectomy specimen) and/or a Gleason histologic grade higher than 6 Bl Palpable disease shown by pathologic evaluation to be confined to less than one lobe of the prostate B2 Palpable disease shown by pathologic evaluation to involve one or more lobes Cl Minimal perforation or penetration through the prostatic capsule but no tumor at the surgical margins C2 Tumor at the surgical margins C3 Invasion of seminal vesicles with or without other capsular extension Dl Pathologically confirmed metastasis to the pelvic lymph nodes D2 Clinical evidence of distant metastasis to bone or soft tissue therapy were restricted to those with stage C2 or C3 dis- ease. Men with pelvic lymph node involvement were con- sidered for radiation therapy if they had 6 or fewer positive nodes. Each candidate also had to 1) present more than 1 month postoperatively, 2) be fully continent, 3) have a well-healed abdominal wound, and 4) have returned to ap- proximately the same state of well-being as had existed preoperatively. Radiation therapy was administered only at our 2 institu- tions. At the VAMC, it was delivered by a 4-million electron volt linear accelerator according to the following standard- ized protocol: For pathologic stage C disease, 60 Gy was delivered to the prostatic bed (including the urethral-vesical anastomosis) in 6'2 weeks with a 4-field isocentric tech- nique through ports averaging 10 X 10 cm. Patients with positive lymph nodes received 45 Gy in 5 weeks to the pelvis by anterior/posterior fields averaging 14 X 15 cm, after which the radiation to the prostatic bed was boosted to 60 Gy by a reduced 3-field technique averaging 10 X 10 cm anteriorly and 9 X 10 cm bilaterally. At the UMHC, radiation was delivered by a 10-million electron volt linear accelerator. For pathologic stage C disease, 45 Gy was delivered to the prostatic bed through anterior and posterior fields of 12 X 12 cm. This was followed by a boost through 1 anterior and 2 lateral fields of approximately 10 X 10 cm and 9 X 10 cm, respectively, to bring the total dose to 60 Gy. In patients with positive nodes, the entire pelvis received 45 Gy through 15- X 16- cm anterior and posterior fields; prostatic boosts were then given through 3 fields to bring the total dose to 60 Gy. After radiation therapy, all patients were followed at 3- to 6-month intervals for 1 year and then at 6-month intervals. Evaluation for distant recurrence included iso- topic bone scans and serum PAP determinations. Local and pelvic recurrence was assessed by rectal examination and, when appropriate, by excretory urography or computed ax- ial tomography. Disease-free status was confirmed in all appropriate patients in June 1987. No endocrine therapy was given until recurrence. The PSA levels were retrospectively determined from frozen serum samples that had been obtained from many of the patients at their regular clinic visits and measured with the Tandem PSA assay (Hybritech Inc., San Diego, CA). For comparison, we also measured PAP levels from the same stored serum using the Tandem PAP assay (Hybritech Inc.). The upper limit of normal is 4 ng/ml for PSA and 3 ng/ml for PAP. RESULTS Of the patients who underwent radical prostatectomy be- fore December 1984, we administered adjuvant radiation therapy to 71. At the VAMC, from June 1977 to Decem- ber 1984, of the 199 patients who underwent radical prosta- tectomy, 37 had the appropriate pathologic stage C disease and 31 of these received radiation therapy. Of the 6 remain- ing patients, 2 refused therapy, 2 were lost to follow-up, and for 2 patients radiation therapy was ruled out to preserve postoperative potency. During the same period, 42 patients had stage D1 disease, 29 of whom received radiation ther- apy. Of the remaining 13 patients, 2 were lost to follow-up, 4 had more than 6 positive nodes, 2 were not totally conti- nent, 1 had had a rectal perforation at surgery, 1 received radiation therapy elsewhere, and 3 refused therapy. At the UMHC, the adjuvant radiation protocol was for- mally adopted in 1983. Before then, appropriate patients of only one of the authors (PHL) were irradiated (2 pa- tients with stage C and 2 with stage Dla disease). In 1983 and 1984, 40 patients underwent radical prostatectomy at the UMHC; 9 had the appropriate pathologic stage C dis- ease and 2 of them received radiation therapy. Of the re- maining 7, 3 refused therapy, 1 was lost to follow-up, and 3 received radiation therapy elsewhere. Nine patients had stage D1 disease and 5 of them received radiation therapy. Of the 4 remaining patients, 1 had more than 6 positive nodes, | was not totally continent, 1 received radiation ther- apy elsewhere, and 1 refused therapy. At neither institution were any patients excluded from radiation therapy because of medical deterioration, recurrent tumor, or perioperative death. The years the radical prostatectomies were performed for patients who received radiation therapy are presented in TABLE 2.—Year of prostatectomy in patients participating in adjuvant radiation therapy trials No. of patients Stage 1977 1978 1979 1980 1981 1982 1983 1984 Total C2 and 3 4 2 2 4 4 4 6 9 35 Dla 1 1 6 3 4 5 6 10 36 Total 5 3 8 7 8 9 12 19 71 NCI MONOGRAPHS, NUMBER 7, 1988 TABLE 3.—Results of pathologic analysis of prostatectomy specimens Stage Pathology of specimens C Dla Surgical margins only positive 11 12 Seminal vesicles only positive 4 h Seminal vesicles positive and margins positive 18 16 Capsular penetration only 2° 0 No capsule penetration, seminal vesicles negative 0 3 One node positive 14 Two to six nodes positive” 22 “ These 2 patients had substantial capsular penetration, a large volume of disease, and a pathologic grade >7 but no definite involvement of the surgical margins. For the purposes of analysis, these 2 patients are included in the margin-positive category. b Four of these patients had one or two visibly positive nodes at surgery, but for the purposes of this analysis they are included in the Dla category (microscopic involvement of pelvic lymph nodes). table 2. Twenty patients with stage C and 20 with stage Dla disease received radiation therapy more than 5 years before this analysis, and all patients received radiation therapy at least 30 months before the June 1987 follow-up deadline. The median length of follow-up for patients with stage C disease was 51 months (range of 30 to 119) and for stage Dla it was 48 months (range of 8 to 113). Only 2 patients who received radiation therapy were lost to follow-up at 48 and 56 months after surgery, respectively. The median age of the patients was 64 years (range of 53 to 73). The specific pathologic findings for patients with stages C and D1a disease who received radiation therapy are given in table 3. Sixty-three percent of the patients with stage C and 58% of those with stage Dla disease had seminal vesicle involvement. The prostatic specimens of all but 3 patients with stage Dla disease were pathologic stage C, and 61% had more than 1 positive node. All but 2 of the patients who began radiation therapy completed the planned course. One patient suffered from severe nausea and vomiting and treatment was stopped at 52 Gy. The other patient developed a cutaneous herpes zoster infection in the irradiated field at 36 Gy and therefore therapy was discontinued. Neither patient had any subse- quent sequelae and both were included in all analyses. The median interval between radical prostatectomy and the be- ginning of radiation therapy was 3 months and the range was 1 to 12 months. The reasons for a delay of more than 4 months were delayed wound healing in 3 patients, slow recovery of continence in 5, a postoperative stricture in 1, patient preference in 10, and scheduling errors in 6. The complications of postoperative radiation therapy have been acceptable. During and immediately after treat- ment, many patients experienced diarrhea (32%), skin desquamation (14%), anal pain (4%), and dysuria and fre- quency (4%). For those with stage C disease, notable acute reactions included transient leg edema in 2 patients and a urinary tract infection in 1 patient. Only 3 patients with stage C disease had any complications after 6 months. One patient had a mild bladder neck stricture that was present before radiation therapy (but which became more severe after therapy) and required 3 dilations during 18 months. MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 143 Eventually, he was treated elsewhere by internal urethrot- omy, after which he immediately became incontinent for the first time. An inflatable urethral cuff was inserted, and the patient remains continent after more than 5 years. One other patient suffered from prolonged constipation and an- other had irritable bowel syndrome; both conditions grad- ually resolved. Among patients with stage D1a disease, 2 had significant acute symptoms of severe nausea and vomiting and a herpes zoster infection; they did not complete the planned course of therapy, as previously mentioned. Long-term compli- cations among patients with stage Dla disease included significant unilateral leg edema in 1, for which he wears support stockings (this patient did not have the modified lymphadenectomy); a transient episode of hematuria in an- other, and persistent edema of the penis in 2 patients. No patients suffered from significant persistent dysuria or frequency, incontinence, other voiding difficulties, or rec- tal troubles. All patients were impotent after surgery (the potency-saving radical prostatectomy was only attempted in 3 of these patients, and none had had a return of erec- tion at the time radiation therapy was begun). Seventeen patients requested a penile prosthesis; the prostheses were implanted without complications 6 months or more after completion of radiation therapy. In 5 patients, a trial of intracorporal drug injections to induce erections was initi- ated; in 2 of them the response was sufficient and they are maintaining potency by self-injection at home (9). Disease has recurred locally in 2 patients who received postoperative radiation therapy. In 1 patient with stage C disease, local recurrence was detected at 35 months, and in another, who originally had stage D1a disease, local re- currence was detected simultaneously with distant metas- tasis at 81 months. Bone metastasis developed in 6 pa- tients with stage C disease at 19, 26, 42, 60, 67, and 102 months, respectively, after surgery. All these patients ini- tially had pathologically proven seminal vesicle involve- ment. All were treated with endocrine therapy; 3 have died of their disease 3, 19, and 24 months, respectively, after this treatment, and 6 died of unrelated causes 32, 33, 48, 67, 94, and 119 months, respectively, after surgery. Among the 36 patients with stage D1 disease who re- ceived radiation therapy, distant metastasis occurred in 9 patients at 11, 12, 34, 36, 38, 42, 48, 62, and 81 months, respectively, after surgery. All men were treated with en- docrine therapy, and 6 have either died of their disease or have been lost to follow-up 13, 24, 40, 48, 79, and 82 months, respectively, after this treatment. Five additional patients have died of unrelated causes 8, 33, 46, 58, and 91 months, respectively, after surgery. To assess the effect of adjuvant radiation therapy on disease progression, we constructed actuarial curves for survival free of disease, total survival, and cancer-related survival by the Kaplan-Meier method (10). At 5 years, the disease-free survival was 80% for patients with stage C and 69% for patients with stage D1a disease (fig. 1). The 5-year total survival was 86% for patients with stage C disease and 74% for patients with stage D1a disease (fig. 2). Survival related only to deaths due to cancer (cancer-related survival) at 5 years was 96% for patients with stage C and 90% for patients with stage D1a disease (fig. 3). 144 100F=~ 774 bree mil a n @ © 9 3 50 bri Stage C phd —- age o 71 © TAT T — MTT NN 3 | 2 Kprmsmrpeyd 1 tal di = 60 I FIGURE |.—Five-year, actuarial disease-free = Stage D1a T T survival for patients with stage C and 2 i ETT TTT 7 stage D1a disease. Encircled numbers rep- 2 ol resent patients at risk 60 mo after radi- s 4 cal prostatectomy. Vertical lines represent n censored data. E 3) 5 20 a Stage C=80%18(2 SE) Stage D1a=69%18 0 | | | | | | | | 0 12 24 36 48 60 72 84 96 108 120 Time (months) TABLE 4.—Number of patients with an elevated level of PSA before diagnosis of recurrence No. of patients with elevated PSA/total Months before diagnosis evaluated 0-5 3/3 6-11 7/7 12-17 3/4 18-23 3/5 24-29 4/6 30-35 2/4 36-41 =P 42-47 1/3 4PSA was > 4 ng/ml. b No serum samples were obtained during this period. A serum sample was taken at the follow-up visit at which a recurrence was determined clinically in 9 of the 17 pa- tients whose disease recurred. In all 9, PSA was elevated (median of 100 ng/ml, range of 19 to 870 ng/ml), whereas PAP was elevated in only 3 (at 3, 15, and 47 ng/ml, respec- tively). Serum samples were available before the diagnosis of recurrence in 10 patients. The PSA, but never PAP, was elevated always at least 12 months before recurrence, and, in most patients, up to 36 months before recurrence and sometimes up to 4 years before recurrence (table 4). In 5 of these patients, the PSA values had returned to normal after radical prostatectomy and/or radiation therapy before becoming elevated; in the others, it could not be determined if PSA reached normal levels after surgery. Finally, among the patients who were free of disease at the last clinic visit, sera were available for 23; PSA (but not PAP) was elevated in 3 (15%). 100 80 60 40 Percent Surviving FIGURE 2.—Five-year, actuarial overall sur- vival for patients with stage C and stage Dla disease. 201 Stage C=86 +13(2 SE) Stage D1a=74%18 0 | | | | | | | | 0 12 24 36 48 60 72 84 96 108 120 Time (months) NCI MONOGRAPHS, NUMBER 7, 1988 145 HOT —— (13) -—mpar! A 2 TTEATOA NT TEC = | 3 oy an | Stage C oc n b wn © | Stage D1a 8 oof hpi a | FIGURE 3.—Five-year, cancer-related sur- 2 vival for patients with stage C and stage £ Dla disease. 2 401 5 a c S oor [o) Stage C=96+8 (2 SE) a Stage D1a=90+12 0 | 1 | | | | | 0 12 24 36 48 60 72 84 96 108 120 Time (months) at least some patients whose disease is upstaged to stage C2 DISCUSSION or C3 after surgery have persistent though still only local- Pelvic lymphadenectomy followed by radical prostatec- tomy is becoming an increasingly popular therapy for ap- parently localized carcinoma of the prostate, in part be- cause the morbidity, especially with regard to potency and incontinence, has been significantly reduced (1,6,11). How- ever, when this therapy is applied to all appropriate pa- tients whose disease is apparently confined to the prostate by rectal examination, a large percentage of them, after fi- nal pathologic staging, have disease outside the prostatic capsule, in the pelvic lymph nodes, or both (2-4). A great majority of these patients are at high risk for local or distant recurrence and need an adjuvant treatment. There appears to be a wide spectrum of severity of dis- ease labeled as pathologic stage C (7,8). Evidence is accu- mulating that the prognosis for patients with involved sem- inal vesicles is much worse than for those with no disease outside the prostate or for those with only capsular pen- etration (7,12,13). Also, patients with only positive surgi- cal margins and no seminal vesicle involvement fare worse than those with capsular penetration but negative margins (14). Indeed, in our series and in those of others (/2), most patients with documented metastases to the pelvic lymph nodes also have involved seminal vesicles and/or positive surgical margins and therefore may be at high risk for lo- cal recurrence (3). Thus we did not include pathologic stage C disease with only capsular penetration as a criterion for adjuvant radiation therapy. We made an exception to this criterion for 2 patients who had equivocally positive surgi- cal margins but in addition had a large volume of tumor with a Gleason combined grade higher than 7. One major theoretical choice for adjuvant therapy in pa- tients at high risk for recurrence after radical prostatectomy is radiation therapy. Of course, radiation therapy is of no benefit to high-risk patients who already have systemic, al- beit undetectable, disease after surgery. This might be true for a patient whose PSA level does not decline significantly after radical prostatectomy. However, we can assume that MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER ized disease. It is likely that some residual disease is present in patients with positive surgical margins, especially if dis- ease is at the anastomosis of the bladder neck and urethra. Also, an en bloc pelvic lymph node dissection for carcinoma of the prostate is not possible without removing the en- tire bladder. Thus tumor may commonly remain within the lymphatics. In such circumstances, radiation therapy may be efficacious because it seems to be curative as a primary treatment for some localized prostate cancers, especially if the volume of disease is small. It is less likely that adjuvant radiation therapy is benefi- cial for patients with stage D1 disease. Many experts be- lieve that most, if not all, patients with stage D1 have per- sistent disease beyond the pelvis at the time of surgery. This would seem to be the case for patients in whom there is evi- dence of recurrence a short time after surgery. Several stud- ies have indicated that radiation therapy to the pelvis after only pelvic lymphadenectomy is of little benefit (15,16). Although patients with minimal stage D1 disease have sur- vived 5 or more years, current data suggest that in most the disease recurs later (17,18). Even though such phenomena are consistent with occult distant metastasis at the time of radical prostatectomy, it is possible that some patients with stage D1 disease initially have only minimal persistent re- gional disease. This residual disease could serve as a source of subsequent systemic disease, and it might be sterilized by regional adjuvant radiation therapy. Thus we believed it was reasonable to test the therapeutic efficacy of adju- vant radiation therapy in these patients provided that the treatment was proved safe. The safety of radiation therapy after radical prostatec- tomy has not been extensively studied. The results of some pertinent studies are listed in table 5. Ray et al. (19) admin- istered adjuvant radiation therapy to 13 patients with stages C2 and C3 disease and to 19 patients who had locally re- current disease. Radiation dose was usually more than 70 Gy. This therapy was considered safe, but the complica- tions seemed significant: Five of 32 patients experienced 146 TABLE 5.—Safety of adjuvant radiation therapy after radical prostatectomy Radiation Principal No. of Chronic Incontinence, to prostatic investigator Reference patients complications % bed, Gy Ray 19) 32 16% severe 7 >704 Gibbons (20) 22 14% severe _— 46-70 Pilepich 21) 18 17% moderate/severe 6 =65 (78%)“ Forman (22) 16 38% total 13 65¢ Anscher (23) 46 13%-20% severe 4 60-70¢ Present study 71 61% total 20 60¢ 4 Most or all patients also received approximately 45 Gy to the pelvis. b One patient became incontinent after internal urethrotomy performed at another institution. 10 complications, which included exacerbation or onset of urinary incontinence (2), fecal incontinence (1), ure- thral stricture (2), rectal perturbations requiring colostomy (1), and leg edema (1). Gibbons et al. (20) observed se- vere complications in 3 of 22 patients who had undergone radical perineal prostatectomy and radiation therapy and in only 1 of 23 similar patients who did not receive ra- diation therapy. Complications after adjuvant radiation therapy included chronic proctitis and cystitis and rectal incontinence. One permanent colostomy and one urinary diversion were performed in this group of patients. All the complications occurred in patients given cobalt-60 in doses exceeding 65 Gy. Gill and associates (24) gave radiation therapy to 15 patients before and after radical prostatec- tomy. The total dosage was 75 Gy to the prostate and 50 Gy to the pelvis. Complications included stress incontinence in 40% of the patients and rectal injury at operation in 13%. Recently, Kaufman and co-workers (25) warned of the dangers of combined radiation therapy and radical prosta- tectomy and presented 4 patients in whom severe compli- cations occurred. Only 1 of the patients received radiation therapy after radical prostatectomy, and the total dose de- livered was more than 70 Gy. In Anscher and Prosnitz’s (23) recent report on complications in 40 patients given adjuvant radiation therapy, those attributable to radiation included severe cystitis in 3 patients and incontinence in 1 patient. Pilepich et al. (27) and Forman and associates (22) also reported a similar incidence and severity of complica- tions. Our incidence of complications after radical prostatec- tomy and radiation therapy is low compared with many reports in the literature. Significant long-term complica- tions occurred in 4 patients, but major surgery or a change in life-style was not required. This low incidence may partly reflect differences in reporting but may also be at- tributable to the following factors: 1) use of the modified lymphadenectomy technique to prevent leg edema; 2) lim- itation of dosage to 60 Gy; and 3) delivery of radiation therapy only after complete recovery from surgery, includ- ing recovery of continence. Radiation therapy given before radical prostatectomy is clearly fraught with more prob- lems. The effect of adjuvant radiation therapy on the incidence of local recurrence among our patients is difficult to assess. It is reassuring that disease has recurred locally in only 2 of 66 patients who had positive surgical margins, positive seminal vesicles, or both (table 3). However, in the absence of an internal control group, we must judge the significance of our results by comparison with the literature. It is unfor- tunate that most of the numerous reports on this subject are either unclear or involve patients who are not compa- rable to ours. For example, many series included patients who also received endocrine therapy (26). However, some reports on local recurrence after radical prostatectomy are pertinent to our series. The results of some of these series are listed in table 6. This subject was recently reviewed extensively by Robey and Schellhammer (26). Several important points emerge from a consideration of these series: 1) Local recurrence can often be ignored if the patient is asymptomatic or if it occurs simultaneously with the more symptomatic distant metastasis. After radical prostatectomy and radiation therapy, many patients have some rectal in- duration, which makes rectal examination less than totally accurate until disease is far advanced. Finally, the endocrine therapy for a distant metastasis attenuates the appearance of local progression. 2) The time until local recurrence is often prolonged. Jewett et al. (31) reported an interval of 6 months to 11 years to recurrence, whereas Culp (27) reported a median interval to recurrence of 4.5 years, although 15% of the recurrences were detected more than 15 years postopera- tively. In a more recent series, Gibbons et al. (20) reported a median interval to recurrence of 51 months (range of 4 to 147). In 1 of our patients, the local recurrence was detected 81 months after surgery. 3) It appears that the final incidence of recurrence in long-term studies approaches 30%. Perhaps in response to these local recurrence incidences, several investigators gave interstitial radiation therapy at the time of radical prostatectomy or external-beam therapy after surgery. Flocks and associates (32) used a radioac- tive colloidal gold solution interstitially during surgery and reported a low incidence of recurrence (4.4%) among 335 patients with clinical stage C cancer. A similar therapy was used more recently at the same institution by Rosenberg and co-workers (33). They treated 25 patients with stage C and 12 with stage Dla disease by pelvic lymphadenec- tomy, radical prostatectomy, and periprostatic radioactive gold seed implantation. With a median follow-up exceed- ing 45 months, local recurrence was observed in 2 of 25 patients with stage C and in 1 of 12 patients with stage Dla disease. NCI MONOGRAPHS, NUMBER 7, 1988 147 TABLE 6.—Local recurrence after radical prostatectomy Principal No. of Disease Recurrence, Follow-up, investigator Reference Treatment patients stage? % yr Culp 27) RP 123 B 26 1-14 McCullough (28) RP 14 C 43 1-14 Tomlinson (29) RP 24 C 8 2-14 Walsh (12) RP 57 B 12 15 Zincke (30) RP 80 D1 14 3 (median) Robey (26) RP 13 C 31 15 Gibbons (20) RP 23 Ci1-3 30 9 (mean) RP+RT 22 Cl1-3 5 Pilepich 2h RP+RT 18 Cl1-3 D1 0 3.4 (median) Forman (22) RP+RT 16 C1-3 D1 0 4 (median) Present study RP+RT 71 C2-3 3 25-10 DI “Treatment included: RP = radical prostatectomy; RT = external-beam radiation therapy. b All staging was pathologic, except that of McCullough and Leadbetter (28), which was clinical. Several reports cite the incidence of local recurrence after radical prostatectomy and subsequent external-beam radiation therapy (table 6). For example, Gibbons et al. (20) reported a 5% incidence of local recurrence in 22 patients with stages C1 to C3 disease after a median follow-up of 9 years compared with an incidence of 30% in a similar, although nonrandomized, group of 23 patients who did not receive radiation therapy. Pilepich and co-workers (27) and Forman et al. (22) found no local recurrence among a group of patients with stage C or D1 disease, but follow-up was short in both reports. Others (19,23) also cite a low incidence of local recurrence, but some of the patients simultaneously received endocrine and radiation therapy. Our incidence of recurrence after radical prostatectomy and radiation therapy seems lower than that reported af- ter radical prostatectomy alone and similar to that reported after adjuvant radiation therapy. However, all the reported series were uncontrolled, and the number of patients and length of follow-up was not sufficient to ensure that ra- diation therapy definitely reduced the number of local re- currences. Nonetheless, postoperative radiation therapy has demonstrated benefit. It is safe. Also, regression of docu- mented local postoperative recurrence has occurred with radiation therapy (/9,22). Thus the fear of leaving some cancer at the anastomotic ends need not compel a surgeon to continue removing tissue at the risk of continence. The 5-year actuarial survival free of disease among our patients with pathologic stages C2 and C3 disease was 80% after a median follow-up period of 51 months (range of 19 to 119). Although it was difficult to find reported data for patient groups comparable to ours, the reports by Gibbons and associates (20) and Zincke et al. (34) cited a 5-year disease-free survival of approximately 60% (table 7). Similarly, Middleton et al. (7) cited an actuarial 5-year survival in patients with positive seminal vesicles of 50%. Other reports on patients with pathologic stage C disease cited high 5-year, disease-free survival incidences of 70% to 82%, but many of the patients had capsular perforation only or did not undergo lymphadenectomy (35-37). Thus our S-year disease-free survival of 80% for patients with stages C2 and C3 disease may be an improvement and indicate that postoperative radiation is beneficial. However, the 95% confidence intervals (£2 SD) are broad (18%) and overlap significantly with the reported figures for radical prosta- tectomy alone. Our actuarial 5-year survival incidence for patients with stage Dla disease is 69% after a median follow-up of 48 months (range of 8 to 113). Again, comparisons with the literature are difficult, but reports on several series of patients with stage D1 cancer after radical prostatectomy seem comparable to ours. Zincke and Utz (30) reported on 51 patients with stage D1 disease and no treatment other than radical prostatectomy until progression. The 5-year actuarial survival free of disease was 18% overall and was slightly higher, at 26%, for patients with only one positive node. The mean interval to recurrence was 40 TABLE 7.—Five-year disease-free survival in patients with stage C disease after radical prostatectomy Five-year survival Principal No. of No evidence Disease investigator Reference Treatment patients of disease, % stage Follow-up, yr Zincke (34) RP 80 60 C3 <5 (60%) Middleton (7) RP 10 50 C3 5 (minimum) Gibbons (20) RP 23 60 C1-3 9 (mean) RP+RT 21 71 Present study RP+RT 71 80 C2-3 4.3 (median) MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 148 TABLE 8.—Five-year disease-free survival after therapy in patients with stage D1 disease Five-year survival Principal No. of No evidence investigator Reference Treatment patients of disease, % Follow-up Zincke (30) RP 51 18 S yr (mean) Paulson (15) RP 11 ? 18 mo (mean time to recurrence) Smith (16) RT 25 35 (N1) S yr (minimum) Bagshaw (38) RT 59 15 26-48 mo Scardino 17) Gold-198, RT* 37 33 (NI) 4.5 yr (mean) 7 8 yr Present study RP+RT 36 69 4 yr (median) “In addition to the radiation, the patients received gold-198 by interstitial implantation. months. Paulson et al. (15) reported even more dismal results for patients with stage D1 disease, but they also observed that after radiation therapy or after no therapy other than lymphadenectomy the median time to recurrence was similar to that after radical prostatectomy. Thus it seems pertinent that we examine the reported incidence of disease-free survival for comparable patients with stage D1 disease who received only radiation therapy. The results of several studies of patients with stage D1 disease who received radiation therapy are presented in ta- ble 8. Smith and Middleton (/6) reported on 40 patients with only microscopic nodal disease after lymphadenec- tomy alone, 25 of whom also received radiation therapy. The actuarial 5-year survival free of disease was 35%, and radiation therapy did not appear to affect recurrence. Among 18 patients with only one positive node, the actu- arial 5-year disease-free survival was 44%, but one-half of these patients developed a recurrence during further follow-up. Scardino et al. (/7) reported similar survival rates for patients with minimal nodal disease (i.e., N1) after pelvic lymphadenectomy, interstitial gold implanta- tion, and external-beam radiotherapy. The actuarial 5-year disease-free survival incidence for patients with limited mi- croscopic nodal disease was only 33%, and this decreased significantly after 5 years: At 8 years the disease-free sur- vival for patients with N1 disease was only 7%. On casual observation, our actuarial 5-year survival in- cidence of 69% for patients with stage D1a disease seems to be an improvement over other reported results. Nev- ertheless, for a variety of reasons we believe the efficacy of adjuvant radiation therapy for stage Dla disease is un- proved. Our 95% confidence levels are broad (+18%), and recurrences are being detected among our patients after 5 years. Finally, there is a wide spectrum of severity of stage D1 disease; it ranges from minimal microscopic dis- ease to gross nodal involvement. Miller and Catalona (/8) recently reported that 9 of 12 patients with very minimal nodal disease survived 5 years. In these patients, frozen sec- tions showed apparently normal tissue, but some metastasis was detected in permanent sections. However, at 7, 9, and 2 The PSA levels after radical prostatectomy by the Hybritech assay are abnormal if =0.4 ng/ml (manuscript in preparation). Nonetheless, for this analysis, we used the traditional value of 4 ng/ml as the upper limit of normal. 10 years, the disease-free survival steadily decreased to 5 of 12, 4 of 9, and 1 of 3 patients, respectively (Catalona WIJ: Personal communication). Thus it would appear that for both high-risk stage C and stage D1a disease random- ized testing of the efficacy of adjuvant therapies is sorely needed. An assay for serum PSA has recently become widely available, and many studies have now documented that this serum marker is extremely useful in monitoring patients with prostate cancer after initial therapies. Indeed, PSA is significantly superior to PAP in this regard. Elevated lev- els of PSA occur in over 95% of patients with stage D2 disease (39-42). More importantly, in those patients whose disease recurs after radical prostatectomy and adjuvant ra- diation therapy, PSA levels are often elevated for many years before a recurrence is manifest? What therapeutic strategies may be exploited with this insight is unknown, but it is probable that disease-free survival analysis can use serum PSA levels to great advantage. For example, the time physicians need to evaluate therapies might be signif- icantly reduced. Also, if the first unequivocal rise in PSA is used as the time to recurrence, the uncertainties of our pinpointing the time of recurrence might be reduced. If and when multi-institutional randomized studies of the value of adjuvant radiation therapy are begun, serum PSA determi- nations should be an essential part of patient evaluation. REFERENCES (1) LANGE PH. Management of localized prostatic cancer. Post- grad Med 1986;66:235-245. (2) CATALONA WJ, STEIN Al. Staging errors in clinically local- ized prostatic cancer. J Urol 1982;127:452-456. (3) LANGE PH, NARAYAN P. Understaging and undergrading of prostate cancer. Argument for postoperative radiation as adjuvant therapy. Urology 1983;21:113-118. (4) PoNTES JE, WAISMAN Z, HUBEN RP, ET AL. Prognos- tic factors in localized prostatic carcinoma. J Urol 1985;134:1137-1139. (5) LANGE PH, MOON TD, NARAYAN P, ET AL. Radiation therapy as adjuvant treatment after radical prostatectomy: Patient tolerance and preliminary results. J Urol 1986;136:45-49. (6) LANGE PH, REDDY PK. Technical nuances and surgical re- sults of radical retropubic prostatectomy in 150 patients. J Urol 1987;138:348-352. (7) MIDDLETON RG, SMITH JA JR, MELZER RM, ET AL. Patient survival and local recurrence rate following NCI MONOGRAPHS, NUMBER 7, 1988 radical prostatectomy for prostatic carcinoma. J Urol 1986;136:422-424. (8) McNEeAL JE, Bostwick DG, KINDRACHUK RA, ET AL. Patterns of progression in prostatic cancer. Lancet 1986;1:60-63. (9) Sib AA, CAMERON JS, DUFFY LM, ET AL. Intracavernous drug-induced erections in the management of male erec- tile dysfunction: Experience with 100 patients. J Urol 1986;135:704-706. (10) KapLAN EL, MEIER P. Nonparametric estimation from in- complete observations. J] Am Stat Assoc 1958;53:457- 481. (11) EGGLESTON JC, WALSH PC. Radical prostatectomy with preservation of sexual function: Pathological findings in first 100 cases. J Urol 1985;134:1146-1148. (12) WaLsH PC, JEWETT HJ. Radical surgery for prostatic cancer. Cancer 1980;45:1905-1911. (13) MIDDLETON RG. How to treat patients with margin-positive disease following radical prostatectomy. In Problems in Prostatic Cancer Control (Paulson DF, ed). Philadelphia: Lippincott, 1987, pp 117-123. (14) PAULSON DF, STONE AR, WALTHER PJ, ET AL. Radical prostatectomy: Anatomical predictors of success or fail- ure. J Urol 1986;136:1041-1043. (15) PAULSON DF, URO-ONCOLOGY RESEARCH Group. Radi- cal surgery for the management of prostatic carcinoma. World J Urol 1983;1:29-34. (16) SMITH JA JR, MIDDLETON RG. Implications of volume of nodal metastasis in patients with adenocarcinoma of the prostate. J Urol 1985;133:617-619. (17) SCARDINO PT, GERVASI L, MATA LA. The prognostic signifi- cance of the extent of nodal metastasis in prostatic cancer. J Urol. In press. (18) MILLER DR, CATALONA WI. Five- and seven-year sur- vival results of radical prostatectomy in clinically under- staged stage A and B prostatic cancer patients. J Urol 1987;137:abstr 481. (19) RAY GR, BAGSHAW MA, FrREIHA F. External beam radiation salvage for residual or recurrent local tumor following radical prostatectomy. J Urol 1984;132:926-930. (20) GiBBONS RP, COLE BS, RICHARDSON RG, ET AL. Adjuvant radiotherapy following radical prostatectomy: Results and complications. J Urol 1986;135:65-68. (21) PiLEPICH MV, WALZ BJ, BAGLAN RI. Postoperative irradia- tion in carcinoma of the prostate. Int J Radiat Oncol Biol Phys 1984;10:1869-1873. (22) FORMAN JD, WHARAM MB, LEE DJ, ET AL. Definitive radio- therapy following prostatectomy: Results and complica- tions. Int J Radiat Oncol Biol Phys 1986;12:185-189. (23) ANSCHER MS, PROSNITZ LR. Post-operative radiotherapy for patients with carcinoma of the prostate undergoing radi- cal prostatectomy with positive surgical margins, seminal vesicle involvement and/or penetration through the cap- sule. J Urol 1987;138:1407-1412. (24) GILL WB, SCHOENBERG HW, BANNO JJ, ET AL. Sandwich ra- diotherapy (3,000 and 4,500 rad) around radical retropu- bic prostatectomy for stage C prostatic carcinoma. Urol- ogy 1980;16:470-475. MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 149 (25) KAUFMAN JJ, SMITH RB, RAZ S. Radiation therapy in car- cinoma of the prostate: A contributing cause of urinary incontinence. J Urol 1984;132:998-999. (26) ROBEY EL, SCHELLHAMMER PF. Local failure after definitive therapy for prostatic cancer. J Urol 1987;137:613-619. (27) Curr OS. Radical perineal prostatectomy: Its past, present and possible future. J Urol 1968;98:618-626. (28) McCuLLoUuGH DL, LEADBETTER WF. Radical pelvic sur- gery for locally extensive carcinoma of the prostate. J Urol 1972;108:939-943. (29) ToMmLINSON RL, Currie DT, Boyce WH. Radical prosta- tectomy: Palliation for stage C carcinoma of the prostate. J Urol 1977;117:85-87. (30) ZiNncke H, Utz DC. Observations on surgical management of carcinoma of prostate with limited nodal metastases. Urology 1984;24:137-145. (31) JEWETT HJ, EGGLESTON JC, YAWN DH. Radical prostatec- tomy in the management of carcinoma of the prostate: Probable causes of some therapeutic failures. J Urol 1972;107:1034-1040. (32) Frocks RH, O'DONOGHUE EN, MILLEMAN LA, ET AL. Surgery of prostatic carcinoma. Cancer 1975;36:705-717. (33) ROSENBERG SJ, LOENING SA, HAWTREY CE, ET AL. Radi- cal prostatectomy with adjuvant radioactive gold for pro- static cancer: A preliminary report. J Urol 1985;133:225- 227. (34) ZINCKE H, Utz DC, BENSON RC JR, ET AL. Bilateral pelvic lymphadenectomy and radical retropubic prosta- tectomy for stage C adenocarcinoma of prostate. Urology 1984;24:532-539. (35) GiBBONS RP, CORREA RJ JR, BRANNEN GE, ET AL. To- tal prostatectomy for localized prostatic cancer. J Urol 1984;131:73-76. (36) ELDER JS, JEWETT JH, WALSH PC. Radical perineal prosta- tectomy for clinical stage B2 carcinoma of the prostate. J Urol 1982;127:704-706. (37) PUNTENNEY RED, LAMSON BG. Cancer of the prostate gland: A clinical and pathological evaluation of patients treated by radical prostatectomy. J Urol 1961;85:649-658. (38) BAGSHAW MA. Radiation therapy of prostatic carcinoma. In Genitourinary Cancer Surgery (Crawford EB, Borden TA, eds). Philadelphia: Lea & Febiger, 1982, pp 405-411. (39) KILLIAN CS, YANG N, EMERICH LJ, ET AL. Prognostic im- portance of prostatic-specific antigen for monitoring pa- tients with stages B2 to D1 prostatic cancer. Cancer Res 1985;45:886-891. (40) AHMANN FR, ScHIFMAN RB. Prospective comparison be- tween serum monoclonal prostatic specific antigen and acid phosphatase measurements in metastatic prostatic cancer. J Urol 1987;137:431-434. (41) ERCOLE CJ, LANGE PH, MATHISEN M, ET AL. Prostatic- specific antigen and prostatic acid phosphatase in moni- toring and staging of patients with prostatic cancer. J Urol 1987;138:1181-1184. (42) StAMEY TA, YANG N, HAY AR, ET AL. Prostatic-specific antigen is the serum marker for adenocarcinoma of the prostate. N Engl J Med 1987;317:909-916. Chemotherapy for Prostate Carcinoma Mario A. Eisenberger' ABSTRACT —We have evaluated the role of chemotherapy for the treatment of prostate carcinoma. The data of patients with endocrine-resistant stage D2 disease indicate that clinical benefits in such patients are at best marginal. Despite the controversies involved in the assessment of response in this disease, in this review we show that in over 3,000 patients eligible for evaluation, less than 10% had complete or partial responses to various treatment regimens. Survival evaluation on all prospective ran- domized clinical trials showed no advantages in favor of any treatment tested and, moreover, in 2 of such studies involving various single agents, survival was not better than a “no chemotherapy” control arm. Because of these data, we conclude that chemotherapy is not indicated as an adjuvant treatment for patients with localized prostate cancer. Although patients with prostate cancer frequently respond to androgen deprivation procedures, preclinical and clinical data strongly suggest the existence of endocrine-independent cell clones, which supports further testing with nonhormonal cytotoxic treatment. A close multidisciplinary interaction is a prerequisite for development of new effective systemic treatment in this disease. —NCI Monogr 7:151-163, 1988. Our task for this important consensus development conference is to evaluate the role of nonhormonal cytotoxic chemotherapy as an adjuvant treatment for patients with localized prostate cancer. Although a significant proportion of patients with clinically localized disease can be cured or have their disease-free intervals prolonged by local modalities of treatment, most present with evidence of extracapsular tumor extension or disseminated disease, which generally implies incurability by these forms of treatment. Similarly, despite the increased sophistication of local forms of treatment and their widespread application, a significant proportion of patients initially approached with curative intent continue to die of metastatic cancer. Prostate cancer represents a prime model of endocrine-dependent tumors in males, and though most of the patients respond, both objectively and subjectively and sometimes dramatically, to a variety of androgen deprivation procedures, this effect is usually palliative and temporary. Most patients will eventu- ally develop effective and irreversible resistance to this form of treatment in an almost predictable fashion. Current data support a biclonal origin of prostate cancer cells, which include both androgen-sensitive and -resistant cell clones (1-4). Resistance to endocrine manipulations in this disease may represent the expansion of previous insensitive clones, or a somatic mutation of sensitive cells, or both. In preclinical studies in various animal prostate tumor models, ABBREVIATION: NPCP = National Prostatic Cancer Project. ! Division of Medical Oncology, Department of Medicine, University of Maryland Cancer Center, 22 South Greene St., Baltimore, MD 21201. the combined use of chemotherapy and endocrine ablation has been shown at times to be superior to either modality alone (5). Although extrapolation of data with nonhormonal cytotoxic treatment to other species (i.e., rat tumor models to humans) has not proved useful in this disease (5), both laboratory experiments and clinical data support the exis- tence of an androgen-independent tumor growth phenom- enon that provides a strong rationale for the application of nonhormonal chemotherapy in prostate cancer. The evaluation of chemotherapy in prostate cancer is obscured by significant methodologic problems. Many of the difficulties are related to the disease itself. The most common metastatic site is in bone, manifested by diffuse sclerotic (osteoblastic) lesions that cannot be measured reliably to allow for assessments of therapeutic benefits. Serum markers, such as acid phosphatase, have not been shown to correlate well with the status of disease in most situations, and the evaluation of unidimensionally measured indicator lesions, such as prostate size by digital examina- tion or radiologic methods, is controversial and is the subject of significant bias. Several response criteria that apply a variety of relatively imprecise and usually subjective param- eters have been developed in the past, most of which are frequently the focus of significant criticisms (6-15). Despite these controversies, chemotherapy studies in this disease rely mostly on response rates as the primary evidence of therapeutic benefits. Because of the unsettled nature of this problem, we attempted to determine the impact of chemotherapy on survival of patients with endocrine-resistant disease. We have analyzed all prospective randomized studies reported in the English literature over the past decades, focusing on survival as the main end point. We also provide a detailed description of results reported with a wealth of treatment regimens that have been primarily tested in patients with endocrine-resistant disease and indicate the specific re- sponse criteria used. These data represent the bulk of experience with chemotherapy in this disease and provide the background on which we base our recommendations with regard to its possible use in patients with localized disease. UNCONTROLLED CLINICAL TRIALS The data reported with single agents, including the response criteria utilized in individual studies, are illustrated in table 1. These studies should be carefully scrutinized because of their differences in criteria for evaluation, patient selection factors, and study design. In some instances, they represent broad trials including patients with various pri- mary tumors in addition to prostate cancer (broad phase II trials), without disease-specific criteria for response, and frequently included insufficient numbers of patients in each 151 152 TABLE 1.—Single-agent phase II trials with patients with hormone-resistant prostate carcinoma Total No. of a responders/No. Response of patients Complete Principal eligible for and Stable Response Drug investigator Reference evaluation partial disease Improvement” criteria Doxorubicin O’Bryan 16) 2/9 2 0 0 Broad phase 11° O’Bryan 17) 5/15 5 0 0 Torti 18) 21/25 44 17 0 NPCP (13) Blum 19) 7/51 NR NR NR Not specified or unclear Scher (20) 6/39 2 1 3(5) MSKCC (6) Carmustine Carter? 20 2/15 NR NR NR Not specified or unclear Lomustine Carter? 20 2/19 NR NR NR Not specified or unclear Cyclophosphamide Carter 2n 8/57 NR NR NR Not specified or unclear Cisplatin Yagoda 9) 4/25 3 1 0(8) MSKCC (6) Merrin (22) 24/54 17 7 0020) Not specified or unclear Rossof (23) 4/21 4 0 0 Broad phase I1¢ Qazi (24) 0/17 0 0 0 Standard for solid tumors, decrease in markers Moore (25) 3/29 0 0 3f Standard for solid tumors, decrease in markers Estracyt Mittleman (26) 9/44 9 0 oT) NPCP (13) Foss (27) 6/17 NR NR 6(3) Not specified or unclear Jonsson (28) 28/91 NR NR 28(24) Not specified or unclear Kuss (29) 3/15 3 0 02) Not specified or unclear Leistenschneider (30) 8/23 NR NR 8(10) Not specified or unclear Edsmyr an 19/908 NR NR NR(49) Not specified or unclear Nilsson (32) 28/91 NR NR NR(24) Not specified or unclear Veronesi 33) 20/27 3 17 0 NPCP (13) 5-Fluorouracil Moore (34) 7/7 4 3 0 Broad phase 11° Ansfield 35) 1/7 0 0 NR Broad phase I1¢ Weiss (36) 1/4 NR NR NR Broad phase I1¢ Hall 37) 3/6 NR NR NR(3) Broad phase I1¢ Hydroxyurea Lerner 38) 19/30 15 4 0 Not specified or unclear Mithramycin Kofman 39) 2/6 NR NR NR Broad phase I1¢ Carter? @2n 2/36 NR NR NR Not specified or unclear disease category. Table 1 also includes broad reviews of multiple studies and specific aspects. These basic differences in methodology account for a disturbing variability in response reported with the same agent by a different investigator. For example, as reported by O'Bryan et al. (16,17) and Torti and associates (/8), doxorubicin appeared to have significant activity. However, use of stricter criteria for evaluation in patients with bidimensionally measurable disease, as tested by Scher and co-workers (20), resulted in disappointing response rates, e.g., 2 of 39 patients had partial responses (5%, 95% confidence interval 0%-12%). Similar experience was reported with single agents, such as cisplatin, 5-fluorouracil, and estramustine phosphate (see table 1). Several of these NCI MONOGRAPHS, NUMBER 7, 1988 153 TABLE 1.—Single-agent phase II trials with patients with hormone-resistant prostate carcinoma (continued) Total No. of R a responders/No. Ssponse of patients Complete Principal eligible for and Stable Response Drug investigator Reference evaluation partial disease Improvement? criteria Mitomycin Humphrey (40) 0/4 0 04) Broad phase I1¢ Melphalan Houghton “41 1/15 0 0 1(1) Not specified or unclear Nitrogen mustard Karnofsky (42) 0/3 0 0 02) Broad phase I1¢ Carter® 21) 12/31 NR NR NR Not specified or unclear Prednimustine Catane 43) 5/23 0 0 5(8) Not specified or unclear Vincristine Carter 2n 2/22 NR NR NR Not specified or unclear m-Amsacrine Drelichman (44) 10/21 0 10 0 NPCP (13) Natale 45) 0/19 0 0 0 MSKCC (6) Aziridinylbenzoquinone Nichols (46) 16/36 0 13 3 Broad phase I1¢ Dihydroxyanthracenedione Drelichman 47) 7/35 2 5 0 Not specified or unclear Hexamethylmelamine Drelichman (48) 0/14 0 0 0 NPCP (13) Mitoguazone Scher 49) 6/29 6 0 03) MSKCC (6) Moore (50) 0/19 0 0 0 Standard for solid tumors, decrease in markers Neocarzinostatin Natale Sh 0/14 0 0 0 MSKCC (6) Vindesine Jones (52) 16/27 5 11 Standard for solid tumors, decrease in markers VP-16-213 Nissen (53) 2/5 1 0 1 Broad phase 11° Walther (54) 1/23 1 4 0 Standard for solid tumors, decrease in markers Spirogermanium Dexeus (55) 0/13 0 0 0 Logothetis et al. 4) “NR = not reported or responses reported but not quantitatively classified as complete or partial, stable disease, or improvement; MSKCC = Memorial Sloan-Kettering Cancer Center. b Improvement indicated objective evidence of response but less than a partial response. Values = numbers of patients; values in parentheses indicate subjective improvement. © Broad phase II indicates that no specific criteria were listed or description was unclear. Patients (4 of 12 with complete and partial responses) had bidimensionally measurable disease. ¢ Work was a review of multiple dose schedules. / All 3 patients had normalization of acid phosphatase levels only; 9 patients with measurable disease had no regression of tumor. 8 Numbers included 26 of 90 patients with no prior hormonal treatment. Actual results for patients with hormone-resistant disease are unclear. agents alone and in combination were subsequently tested in prospective randomized trials with discouraging results, despite preliminary evidence of antitumor activity derived from phase II studies. This divergence of results and lack of reproducibility are not unique to prostate cancer, but these data illustrate that they are particularly prevalent in this disease. Despite these difficulties, the data illustrated with the various single agents in table 1 suggest that their antitumor activity is at best marginal and of limited clinical significance. MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER Table 2 illustrates the data available with various drug combinations tested in prostate cancer therapy. Not sur- prisingly, combinations of agents with modest activity have not yielded satisfactory results. RANDOMIZED CLINICAL TRIALS Because current standard methodology of clinical drug testing looking at response rates as the major study end point is not feasible for the majority of patients with prostate 154 TABLE 2.—Phase II trials with combination chemotherapy Total No. of R a responders/No. Ssponse of patients Complete Drug Principal eligible for and Stable Response combination investigator Reference evaluation partial disease Improvement? criteria Cyclophosphamide + Izbicki (56) 8/20 3 5 (8) Standard for solid doxorubicin tumors, decrease in markers Ihde (8) 11/22 7 4 See (8) Merrin 57) 5/19 0 5 ®) Not specified or unclear Lloyd (58) 2/11 2 0 Broad phase II Soloway 59) 12/21 0 12 (5) NPCP (13) Cyclophosphamide + Merrin (ST 2/13 1 1 07) Not specified or S-fluorouracil unclear Estramustine phosphate + Kennealey (60) 3/25 0 0 3(8) Not specified or 5-fluorouracil unclear Chlorambucil + Beckley 61) 2/11 0 2 0 NPCP (13) prednisolone Carmustine + Presant (62) 11/27 7 4 2 Standard for solid cyclophosphamide + tumors, decrease doxorubicin in markers Doxorubicin + Citrin (63) 10/21 NR NR NR See (63) cisplatin Perloff (64) 9/17 9 0 2) Not specified or unclear Cyclophosphamide + Anderson 65) 7/83 NR NR 7(55) Not specified or prednisolone unclear Cyclophosphamide + Seifter (66) 721° 7 6 0 See (8) doxorubicin + cisplatin Cyclophosphamide + Berry (67) 10/22 0 10 0 NPCP (13) cisplatin + prednisone Doxorubicin + Logothetis (14) 30/624 NR NR NR See (14) 5-luorouracil + Kasimis (68) 7/16 0 7 009) NPCP (13) mitomycin Hsu 69) 9/14 1 8 0(10) NPCP (13) Melphalan + Paulson (70) 51/84¢ 3/ 0 NR(40) See (70) methotrexate + S-1luorouracil + vincristine + prednisone Cyclophosphamide + Buell (71) 6/16 5 1 (11) Standard for solid methotrexate + 5-fluorouracil + vincristine + prednisone tumors, decrease in markers 4 See footnote a, table 1. b See footnote b, table 1. © All stage D2 patients were treated before initiation of endocrine treatment. Sites and types of response are unclear. 4 Responses were seen in 18 of 41 patients with bone metastasis only and 12 of 21 with bone and visceral sites (8 in the lung). ¢ Twenty-four patients had at least a 50% decrease in acid phosphatase, 13 had normalization, and 11 had a 50% reduction without normalization. / Three of 7 patients with bidimensionally measurable disease had complete and partial responses. carcinoma, evaluation of survival, independent of response status, in prospective trials may be fundamental to the evalua- tion of treatment efficacy in this disease. The evaluation of survival in phase Il studies, in which the survival of responders versus nonresponders is compared, is not a valid method of assessment of therapeutic efficacy (72). Investigators involved in the NPCP have evaluated several chemotherapeutic agents in a series of prospective randomized studies in hormone-resistant patients (table 3). On their initial 2 trials, NPCP investigators tested various chemotherapeutic agents against a standard treatment control arm. Standard treatment represented a group of NCI MONOGRAPHS, NUMBER 7, 1988 TABLE 3.—Randomized trials for prostate carcinoma: NPCP studies 155 R No. of patients eponse eligible for Complete evaluation/No. and Stable Study No. Reference Treatment entered partial disease Survival, wk? 100 (73) Cyclophosphamide 41 4 20 47 5-Fluorouracil 33 4 14 44 Standard” 36 0 7 38 200 (74) Estramustine phosphate 46/54 3 11 26 Streptozocin 38/46 0 12 25 Standard? 21/25 0 4 24 300 (73) Cyclophosphamide 35/39 0 9 27 Dacarbazine 55/68 2 13 40 Procarbazine 39/58 0 5 31 400 (76) Estramustine phosphate + 54 1 6 37 prednimustine Prednimustine 62 0 8 36 700 (77) Cyclophosphamide 43/47 3 12 41 Semustine 27/38 1 7 22 Hydroxyurea 28/40 2 2 19 800 (78) Estramustine phosphate 27/38 1 6 26 Vincristine 29/42 1 4 22 Estramustine phosphate + 4/41 0 7 32 vincristine 1100 (79) Estramustine phosphate 50/63 1 16 43 Methotrexate 58/67 3 21 37 Cisplatin 50/59 2 16 33 1200 (80) Estramustine phosphate 40/50 0 7 38 Cisplatin 42/51 0 9 28 Estramustine phosphate + 42/48 0 14 40 cisplatin “Median survival was calculated for all studies except study 400. b Standard therapy consisted of radiation, prednisone, chlorotrianisene, dexamethasone, testosterone, diethylstilbestrol, stilphostrol, spironolactone, cryosurgery, dicorvin, or Estinyl. patients randomly allocated to receive various palliative measures including radiation therapy and/or alternative hormonal approaches such as corticosteroids, chlorotri- anisene, testosterone, diethylstilbestrol, stilphostrol, aldac- tone, dicorvin, and Estinyl. For their first study (NPCP 100), patients were randomized to receive 1 of the single agents, i.e., cyclophosphamide, 5-fluorouracil, or standard treat- ment (73). In their second study (NPCP 200) for patients who had received pelvic radiation (=20 Gy) previously, the random allocation included streptozocin, estramustine phosphate, or standard treatment (74). Based on higher percentages of responses (complete and partial responses and stable disease) observed on the chemotherapy arms in both studies compared with standard treatment, researchers concluded that chemotherapy was superior to standard treatment. This served as the background for further testing of various single agents and combinations in a similar fashion. The NPCP studies include stabilization of disease in their response criteria as a reflection of therapeutic benefits. According to the group’s definition, stable disease reflects no evidence of progression for 12 weeks. Patients with stable disease have a median survival comparable to those classified as achieving a partial response by their criteria and MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER have a longer survival than do those with evidence of progression during this interval (/3). However, what re- mains unproved is whether stable disease resulted from treatment. It may be logical for one to assume that patients with more aggressive disease demonstrating objective pro- gression during the first 12 weeks of treatment live shorter lives than do those with a more indolent biology, but without evidence of progression during that same interval. Thus, because it remains possible that stabilization is a function of the disease itself rather than treatment, its inclusion is not acceptable and may inflate response rates falsely. Of the 79 “responders” in studies 100 and 200, only 11 (14%) could be classified as objective responders, i.e., had complete and partial responses, whereas the remaining 68 (86%) were included in the stable disease category. Survival results (fig. 1 and 2) indicated no differences among the various study arms (87). These results would indicate that chemotherapy has not been shown to be superior to standard treatment and that a “no chemotherapy” treatment continues to be the most appropriate control arm with which new treatments should be compared (table 3, fig. 1-3). Several other groups conducted additional randomized studies in endocrine-resistant patients, and these are shown 156 100 vo N Nn ES ly Standard 36 3 nC ———5FU 33 s 80 Nail, seereraen Cytoxan 41 2 i oc 3 ? gol Oo So E = 2 a0} < a Oo & 200 1 1 1 L 1 1 1 J 0 10 20 30 40 50 60 70 80 WEEKS ON STUDY FIGURE 1.—NPCP Protocol 100: Probability of survival for randomized treatment groups. in table 4 and figure 4. Like the NPCP studies, these other trials failed to demonstrate a clear superiority of one treatment regimen over another. Figure 5 illustrates the similarities between survival curves observed with various treatment programs tested in prospective randomized clin- ical trials including 20 or more adequately treated patients per arm. It also stresses the point that survival of patients with endocrine-resistant prostate carcinoma entering che- motherapy clinical trials is short and limited to a few months. DISCUSSION We described the overall experience with chemotherapy in prostate cancer. Because reproducible measures of antitumor effects are usually lacking, the conduct of phase II studies for screening new agents is mostly unsatisfactory. We have analyzed the randomized clinical trials focusing on survival as the main end point, and the data on all such studies reported in the English literature indicate that in no instance was any treatment program proved superior to another. Furthermore, in 2 studies conducted by the NPCP, the evaluation of 4 chemotherapeutic agents (cyclophospha- mide, 5-fluorouracil, estramustine phosphate, and strep- tozocin) did not result in any prolongation of survival compared with a no chemotherapy control arm (standard treatment). Regardless of the controversies associated with response criteria, of the 3,184 patients treated with the various regimens illustrated in this review, only 202 (6.5%) satisfied the criteria for a complete or partial response. Four hundred eighty-five patients (15%) had stabilization of disease at 12 weeks that cannot be interpreted as a beneficial drug effect. We have focused on the difficulties involved in testing chemotherapeutic agents in the treatment of this disease, stressing various methodologic aspects inherent to the clinical presentation of prostate carcinoma. An equally important factor relates to the patient population usually referred for treatment in clinical trials with cytotoxic drugs. Such patients have far advanced disease and present with a multitude of debilitating symptoms reflected by their limita- 100 a N Mi N es, Standard 21 o 3 — — —Estramustine 46 < 80 Neg ree ene Streptozotocin 38 S 105 oc 3 2 60 Oo >= = 3 40} < wm o & 20 1 l 4 1 1 1 dl — 0 10 20 30 40 50 60 70 80 WEEKS ON STUDY FIGURE 2.—NPCP Protocol 200: Probability of survival for randomized treatment groups. tions in performance status and poor tolerance to therapy. The lack of hematologic tolerance in these patients is primarily due to widespread bone marrow replacement by tumor and is frequently aggravated by prior palliative radiation therapy to marrow-harboring areas; these two aspects result in signs and symptoms of adverse prognostic significance (91). Such aspects undoubtedly limit the chances for a more favorable response to treatment and may at least partially account for the refractory nature of this disease. Similarly, it remains possible that the mechanisms involved in the development of resistance to endocrine and cytotoxic chemotherapy in prostate cancer have common elements and that failure with one modality of treatment reflects resistance to the other. Several investigators have reported attempts to test chemo- therapy in conjunction with standard endocrine approaches or prior to such treatment. As previously mentioned, current data support a biclonal origin of prostate cancer cells that includes both androgen-dependent and -independent clones (1-4). This hypothesis offers an attractive rationale for testing nonhormonal cytotoxic treatment with conventional endocrine manipulations. Most randomized studies in which a combined chemohormonal therapy is tested against hormonal therapy alone have been done with either cyclophosphamide or estramustine phosphate (92-94). This latter compound (a conjugated estradiol derivative and mechlorethamine linked to carbon 3 of the steroid) was tested either alone, in combination with a standard endo- crine treatment, or with cyclophosphamide. None of the studies reported thus far has shown a clear-cut advantage of responses, time to progression, and survival for combined approaches compared with standard endocrine treatment (92-94). However, this method of study may be the most appropriate way for further development of newer agents that demonstrate preliminary evidence of antitumor activity in phase II studies. Figure 6 illustrates the survival curves of NPCP studies 500 and 600. The administration of exogenous testosterone for priming prostate cancer cells and increasing their susceptibility to chemotherapy has also been explored (95-97). In a rela- tively small randomized study, patients with endocrine- NCI MONOGRAPHS, NUMBER 7, 1988 157 10 aw gs —— Standard (36) s Standard (21) > 08 Ni -—-5-FU (33) z 0.8 -—- Bavraimustine (46) = + OME wwe Cytoxan (41) 2 seenenes Streptozotocin (38) wn w 06 w 06 o o E04 Z 04 = = m om g 02 & 02 TEE Q « a — J a — 1 J 20 40 60 80 100 20 40 60 80 100 WEEKS WEEKS 10 L100 = —— Cyclophosphamide (43) 5 “2 —— Cytoxan (36) > 08 = = = MethylCCNU (27) S 08 Qe --- DTIC (55) = ow VL mes Hydroxyurea(28) = ens eee Procarbazine (39) on »n w 06 o 06 o Oo E 0.4 i 0.4 a = Cel om & 0.2 < 02 . 0 oc no T= & 1 | | 1 1 | | ie 1 J 20 40 60 80 100 20 40 60 80 100 WEEKS WEEKS 3 1.08 10= = “ ——— Estramustine (27) = —— Estramustine (50) > — — — Estramustine + VCR (34) > 08 == = Methotrexate (58) 2 VCR (29) 2 <++v-w0- CIS - Platinum (50) o o > cosa 0 gal om © Nm we g g 0.2 re Nmmamm 0 0 oc oc a a 1 J 20 40 60 80 100 WEEKS —— Estramustine (40) — — — Estramustine + DDP (42) mwas DDP (42) PROBABILITY OF SURVIVAL WEEKS FIGURE 3.—Probability of survival in NPCP randomized studies. 5-FU == 5-fluorouracil; methylCCNU = semustine; DTIC = dacarbazine; VCR = vincristine; CIS-Platinum = cisplatin; DDP = cisplatin. MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 158 TABLE 4.—Other randomized trials with drug therapy for patients with prostate carcinoma“ No. of patients R e eligible for SSpons evaluation/No. Complete Drug and/or Principal of patients and Stable Median Response combination investigator Reference entered partial disease survival? criteria“ 5-Fluorouracil Smalley (82) 32/49 2 5 34 wk See (82) Cyclophosphamide + 39/52 2 4 25 wk doxorubicin + S-fluorouracil Adriamycin Eagan 10) 19 — — NR Ancillary scoring system [including crossed-over patients (/10)] Cyclophosphamide + 18 — _— NR 5-fluorouracil Cyclophosphamide Chlebowski (83) 15 0 8 7.2 mo NPCP (13) Cyclophosphamide + 12 0 6 8.9 mo doxorubicin + S-fluorouracil Cyclophosphamide Muss (84) 17 9 8 mo NPCP (13) Cyclophosphamide + 15 1 7 5 mo methotrexate + S-fluorouracil Cyclophosphamide + Herr (85) 20 3 4 26 wk Standard for solid methotrexate + tumors, decreased 5-fluorouracil acid phosphatase Lomustine 20 0 6 24 wk 5-Fluorouracil Tejada (86) 8 2 1 NR Standard for solid tumors, decreased acid phosphatase Lomustine 10 4 2 NR Doxorubicin Pavone-Malacuso (87) 11/22 0 3 NR NPCP (13) Procarbazine 14/24 1 0 NR Doxorubicin DeWys 1) 96/112 15/614 0 29 wk Standard for solid tumors, decreased acid phosphatase 5-Fluorouracil 51/54 3/424 0 24 wk Cyclophosphamide + Stephens (88) 68 6/19¢ 18 27 wk Standard for solid doxorubicin tumors, decreased acid phosphatase Hydroxyurea 69 1/244 9 28 wk Doxorubicin Torti (89) 20 113% 8 48 wk See (89) Doxorubicin + 17 2/10¢ 9 43 wk cisplatin Cyclophosphamide Kasimis (90) 16 0 8 7.9 mo NPCP (13) 5-Fluorouracil + 14/15 1 5 8.9 mo doxorubicin + mitomycin “ Improvement was reported but not quantitated or some evidence was given that was regarded as a treatment benefit, e.g., decrease in marker values or decrease in prostate size, by most investigators. Eagan et al. (/0) reported improvement in 5 patients given Adriamycin and 2 with the combination therapy. b NR = not reported. ¢ Criteria used were the same for both trials. 4 Only values for patients with measurable disease are given, including those with bidimensionally measurable disease, elevated markers, or with bony lesions that could be evaluated. Values include crossed-over patients. ¢ Objective responses are recorded separately according to their category of measurable versus those with evaluation potential. Ancillary responses included regarding improvement do not allow for a determination of an actual denominator. NCI MONOGRAPHS, NUMBER 7, 1988 — Hydroxyurea (69) —— — Adria + CTX (68) PROBABILITY OF SURVIVAL PROBABILITY OF SURVIVAL 159 —— CAF (39) -=-=5FU (32) 1 1 ep 60 80 100 WEEKS —— CMF (20) === CON) FIGURE 4.—Probability of survival in other randomized studies. Adria = doxorubicin; CTX = cyclophosphamide; CAF = cy- clophosphamide, doxorubicin, 5-fluoroura- cil; SFU = 5-fluorouracil; CMF = cyclo- phosphamide, methotrexate, 5-fluoroura- — — sg —c™ (an g > ~~~ CMF (15) > ac ac 5 = @ @ w w o o > > = E = = © «@ < < @ © o Q oc ac a La , a 40 60 80 100 WEEKS - = —— CAF (12) 3 -—=CTX (15) 5 w w oc = m < © o oc a lee Lo 4 40 60 80 WEEKS resistant disease were treated with a combination of drugs with or without prior stimulation with exogenous androgens; preliminary results indicated no advantages for this new approach (97). These negative preliminary findings may be explained by the lack of effective chemotherapy for this disease and by the fact that the investigators chose to test this concept in patients with endocrine-resistant disease, who may not be susceptible to exogenous androgenic stimula- tion. Additional studies are necessary if we are to determine optimal ways to combine chemotherapy and hormonal treatment. Seifter et al. (66) recently reported their preliminary results with the combination of cyclophosphamide, doxo- rubicin, and cisplatin in newly diagnosed stage D2 patients before they were given endocrine treatment. Although their results with cytotoxic treatment were essentially negative, it — < 2 > oc 3 z = 2 Ra ry E04 Ry 3 AR TEN S02 > 2 oa 20 40 60 $0 00 WEEKS FIGURE 5.—Composite figure of survival curves of treatments with 20 or more patients who could be evaluated per arm. MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER cil; CCNU = lomustine. appeared that such an approach did not prevent subsequent responses to endocrine treatment, which suggests that chemotherapy does not change the biology of prostate cancer relative to its sensitivity to androgen deprivation procedures. This approach is ethically justifiable because delaying endocrine treatment for stage D2 patients until they become symptomatic has not been shown to affect survival adversely (98-101). Patients with newly diagnosed —— DES/ORCH (83) =) TI SE 3 ~~~ DES +CTX (77) z 08} CTX +Estramustine (86) zZ + uw 06 ~~ Sc i > * E 04r 5 3 I I 02f o | oc a 1 Lot) fed db 20 40 60 80 100 120 140 160 180 200 WEEKS 10 S —— DES (52) S os} — — — DES + Cytoxan (49) 5 DES + Estramustine (60) w o 0.6F £9 0 Mae Z 0a} - 3 | I 02f o | oc a feito oe el el el eid 20 40 60 80 100 120 140 160 180 200 WEEKS FIGURE 6.— Chemotherapy and hormonal therapy used in NPCP studies 500 (A) and 600 (B). DES/ORCH = diethylstilbestrol/orchiectomy. 160 stage D2 disease may be a more suitable group for testing new drugs because they have less extensive disease, better performance status, less prior treatment, and more fre- quently present with evidence of measurable soft tissue disease. The selection of patients with bidimensionally measurable disease for new drug development has been the subject of significant criticism. These patients are considered by many as a subgroup with distinct biologic features and not representative of the usual patients with prostate cancer presenting with bone metastasis only. However, this has not been substantiated by any solid clinical evidence. Patients with soft tissue metastases respond well to first-line endocrine treatments and frequently demonstrate marked regression of measurable disease. Data on file at the Southwest Oncology Group (Blumenstein B, Crawford ED: Personal communication) of patients with endocrine-resis- tant prostate cancer who entered various chemotherapy studies indicate that time-to-treatment failure and survival are the same for patients with measurable disease versus bone disease only. At present, we favor the conduct of phase II studies with patients with reproducible evidence of bidimensionally measurable disease. Whereas this approach allows for a more reliable assessment regarding the pre- liminary activity of new agents, it is frequently confounded by the fact that even in this situation the dominant site of disease is still in bones, thus potentially providing frag- mentary evaluation of the patient. Because of this, we suggest that more definitive evidence of therapeutic efficacy should derive from phase III studies in which survival is the main study end point. Because results with chemotherapy have thus far been generally disappointing, the most appropriate control with which new treatments should be compared continues to be a no chemotherapy arm, consist- ing of a uniformly applied second-line, endocrine manipula- tion, or symptomatic care. In summary, we have described and discussed the extensive experience with chemotherapy for prostate cancer. The results indicate that, regardless of the controversies, the clinical benefits associated with this form of treatment have been marginal and that routine use of this modality should be reserved to an investigational setting or for the clinical situation in which other forms of less toxic palliative treatments have failed. At this time, there is no indication that cytotoxic chemotherapy is likely to provide additional benefits for the treatment of patients with localized prostate cancer. REFERENCES (1) IsAACSJT, WAKE N, COFFEY DS, ET AL. Genetic instability coupled to clonal selection as a mechanism for tumor progression in the Dunning R-3327 rat prostatic adeno- carcinoma system. Cancer Res 1982;42:2353-2371. (2) Isaacs JT, Isaacs WB, COFFEY DS. Models for develop- ment of nonreceptor methods for distinguishing andro- gen-sensitive and -insensitive prostatic tumors. Cancer Res 1979;39:2652-2659. (3) IsAAcs JT, COFFEY DS. Adaptation versus selection as the mechanism responsible for the relapse of prostatic cancer to androgen ablation therapy as studied in the Dunning R-3327-H adenocarcinoma. Cancer Res 1981;41:5070- 5075. (4) ISAACSIT, ISAACS WB, FEITZ WFJ, ET AL. Establishment and characterization of seven Dunning rat prostatic cancer cell lines and their use in developing methods for predicting metastatic abilities of prostatic cancer. Prostate 1986; 9:261-281. (5) Brock NL, CANuzzi F, DENEFRIO J, ET AL. Chemotherapy of the transplantable adenocarcinoma (R-3327) of the Copenhagen rat. Oncology 1977;34:110-113. (6) YAGODA A. Response in prostate cancer: An enigma. Semin Urol 1984;1:311-319. (7) ScHMIDT JD, GIBBONS RP, JOHNSON DE, ET AL. Chemo- therapy of advanced prostatic cancer. Evaluation of response parameters. Urology 1976;7:602-610. (8) InpE DC, BUNN PA, COHEN MH, ET AL. Effective treatment of hormonally-unresponsive metastatic carci- noma of the prostate with Adriamycin and cyclophos- phamide. Methods of documenting tumor response and progression. Cancer 1980;45:1300-1310. (9) YAGODA A, WATSON RC, NATALE RB, ET AL. A critical analysis of response criteria in patients with prostatic cancer treated with cis-diamminedichloride platinum IL Cancer 1979;44:1553-1562. (10) EAGAN RT, HAHN RG, MYERS RR. Adriamycin (NSC- 123127) versus 5-fluorouracil (NSC-19893) and cyclo- phosphamide (NSC-26271) in the treatment of meta- static prostatic cancer. Cancer Treat Rep 1976;60: 115-117. (11) DEWYS WD, BEGG CB, BRODOWSKY H, ET AL. A compara- tive clinical trial of Adriamycin and 5-fluorouracil in advanced prostatic cancer: Prognostic factors and re- sponse. Prostate 1983;4:1-11. (12) CitRIN DI, COHEN AF, HARBERG J, ET AL. Systemic treatment of advanced prostatic cancer. Development of a new system for defining response. J Urol 1981;125: 224-2217. (13) SLACK NH, MITTLEMAN A, BRADY MF, ET AL. The importance of the stable category for chemotherapy treated patients with advanced and relapsing prostate cancer. Cancer 1980;46:2393-2402. (14) LOGOTHETIS CJ, SAMUELS ML, VON ESCHENBACK AC, ET AL. Doxorubicin, mitomycin-C and 5-fluorouracil (DMF) in the treatment of metastatic hormonal refractory adenocarcinoma of the prostate, with a note on the staging of metastatic prostate cancer. J Clin Oncol 1983; 1:368-378. (15) Torti FM, CARTER SK. The chemotherapy of prostatic adenocarcinoma. Ann Intern Med 1980;92:681-689. (16) O'BRYAN RM, LUCE JK, TALLEY RW, ET AL. Phase II evaluation of Adriamycin in human neoplasia. Cancer 1973;32:1-8. (17) O’BRYAN RM, BAKER LH, GOTTLIEB JF, ET AL. Dose- response evaluation of Adriamycin in human neoplasia. Cancer 1977;39:1940-1948. (18) TORTI F, ASTON D, LUM BL, ET AL. Weekly doxorubicin in endocrine refractory carcinoma of the prostate. J Clin Oncol 1983;1:477-482. (19) BLUM RH. An overview of studies with Adriamycin (NSC- 123127) in the United States. Cancer Chemother Rep 1975;Part 3, 6(2):247-251. (20) SCHER H, YAGODA A, WATSON R, ET AL. Phase II trial of doxorubicin in bidimensionally measurable prostatic ade- nocarcinoma. J Urol 1984;13:1099-1102. (21) CARTER SK, WASSERMAN TH. The chemotherapy of urologic cancer. Cancer 1975;36:729-747. (22) MERRIN CE. Treatment of genitourinary tumors with cis- dichlorodiammineplatinum(Il): Experience in 250 pa- tients. Cancer Treat Rep 1979;63:1579-1589. NCI MONOGRAPHS, NUMBER 7, 1988 (23) ROSSOF AH, TALLEY RW, STEPHENS R, ET AL. Phase II evaluation of cis-dichlorodiammineplatinum(Il) in ad- vanced malignancies of the genitourinary and gynecologic organs: A Southwest Oncology Group study. Cancer Treat Rep 1979;63:1557-1564. (24) Qazi R, KHANDEKAR J. Phase II study of cisplatin for metastatic prostatic carcinoma. An Eastern Cooperative Oncology Group study. Am J Clin Oncol 1983;6:203- 205. (25) MOORE MR, TRONER MB, DESIMONE P, ET AL. Phase II evaluation of weekly cisplatin in metastatic hormone- resistant prostate cancer: A Southeastern Cancer Study Group trial. Cancer Treat Rep 1986;70:541-542. (26) MITTLEMAN A, SHUKLA SK, MURPHY GP. Extended therapy of stage D carcinoma of the prostate with oral estra- mustine phosphate. J Urol 1976;115:409-412. (27) FossA DS, MILLER A. Treatment of advanced carcinoma of the prostate with estramustine phosphate. J Urol 1976; 115:406-408. (28) JoNssON G, HOGBERG B, NILSSON T. Treatment of ad- vanced prostatic carcinoma with estramustine phosphate (Estracyt®). Scand J Urol Nephrol 1977;11:231-238. (29) Kuss R, KHOURY S, RICHARD F, ET AL. Estramustine phosphate in the treatment of advanced prostatic cancer. Br J Urol 1980;52:29-33. (30) LEISTENSCHNEIDER W, NAGEL R. Estracyt® therapy of advanced prostatic cancer with special reference to control of therapy with cytology and DNA cytopho- tometry. Eur Urol 1980;6:111-115. (31) EDSMYR F, ESpPoOSTI PL, ANDERSON L. Estramustine phos- phate therapy in poorly differentiated carcinoma of the prostate. Scand J Urol Nephrol [Suppl] 1980;55:135-138. (32) NiLssON T. Estracyt® —clinical experiences. Scand J Urol Nephrol [Suppl] 1980;55:135-138. (33) VERONESI A, ZATTONI F, FRUSTACCI S, ET AL. Estramustine phosphate (Estracyt®) treatment of T3-T4 prostatic carcinoma. Prostate 1982;3:159-164. (34) MOORE GE, Bross IDJ, AUSMAN R, ET AL. Effects of S-fluorouracil (NSC-19893) in 389 patients with cancer. Eastern Clinical Drug Evaluation Program. Cancer Chemother Rep 1968;52:641-653. (35) ANSFIELD FJ, SCHROEDER J, CURRERI AR. Five years clinical experience with 5-fluorouracil. JAMA 1962; 181:295-299. (36) WEISS Al, JACKSON LG, CARABASI R. An evaluation of 5-fluorouracil in malignant disease. Ann Intern Med 1961;55:731-741. (37) HALL BE, Goop JW. Treatment of far-advanced cancer with 5-fluorouracil used alone and in combination with irradiation. Incidence and duration of remission and survival-time data in 223 patients. Cancer Chemother Rep 1962;16:369-386. (38) LERNER HJ, MALLOY TR. Hydroxyurea in stage D car- cinoma of the prostate. Urology 1977;10:35-38. (39) KOFMAN S, EISENSTEIN R. Mithramycin in the treatment of disseminated cancer. Cancer Chemother Rep 1963;32: 77-96. (40) HUMPHREY EW, HYMES AC, AUSMAN RK, ET AL. An evaluation of actinomycin-D and mitomycin-C in patients with advanced cancer. Surgery 1961;50:881-885. (41) YORKSHIRE UROLOGICAL GROUP: HOUGHTON AL, ROBIN- SON MRG, SMITH PH. Melphalan in advanced prostatic cancer: A pilot study. Cancer Treat Rep 1977;61:923-924. (42) KARNOFSKY DA, ABELMAN WH, CRAVER LF, ET AL. The use of nitrogen mustards in the palliative treatment of carcinoma with particular reference to bronchogenic carcinoma. Cancer 1948;1:634-656. MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 161 (43) CATANE R, KAUFMAN JH, MADAJEWICZ S, ET AL. Pred- nimustine therapy for advanced prostatic cancer. Br J Urol 1978;50:29-32. (44) DRELICHMAN A, DECKER DH, AL-SARRAF M, ET AL. m-AMSA in disseminated prostatic carcinoma: A phase II study. Cancer Treat Rep 1982;66:1993-1994. (45) NATALE RB, YAGODA A, WATSON RC. Phase II trial of AMSA in prostatic cancer. Cancer Treat Rep 198266: 208-209. (46) NICHOLS WC, KvOLS LK, RICHARDSON RL, ET AL. A phase IT study of aziridinylbenzoquinone in advanced genito- urinary cancer. Proc ASCO 1982;1:117. (47) DRELICHMAN A, OSBORNE CK, VON HOFF DD. A phase II clinical investigation of dihydroxyanthracenedione (DHAD) in patients with advanced prostatic cancer. Proc ASCO 1982;1:120. (48) DRELICHMAN A., BROWNLEE R, AL-SARRAF M. Phase II study of hexamethylmelamine for disseminated prostatic carcinoma. Cancer Clin Trials 1981;4:309-312. (49) SCHER H, YAGODA A, TAUSEEF A, ET AL. Methylglyoxal bis(guanylhydrazone) MGBG: An active drug in prostatic cancer. J Clin Oncol 1985;3:224-228. (50) MOORE M, GRAHAM SD, BIRCH R, ET AL. Phase II evaluation of mitoguazone in metastatic hormone-resis- tant prostate cancer: A Southeastern Cancer Study Group trial. Cancer Treat Rep 1987;71:89-90. (51) NATALE RB, YAGODA A, WATSON RC, ET AL. Phase II trial of neocarzinostatin in patients with bladder and prostatic cancer. Cancer 1980;45:2836-2842. (52) JONES WG, FossA SD, DENIS L, ET AL. An EORTC phase II study of vindesine in advanced prostate cancer. Eur J Cancer Clin Oncol 1983;19:583-588. (53) NisseN NI, PAaJAK TF, LEONE LA, ET AL. Clinical trial of VP-16-213 (NSC 141540) IV twice weekly in advanced neoplastic disease. A study by the Cancer and Leukemia Group B. Cancer 1980;45:232-235. (54) WALTHER P, WILLIAMS SD, TRONER M. A phase II study of etoposide for carcinoma of the prostate. Cancer Treat Rep 1986;70:771-772. (55) DEXEUS FH, LOGOTHETIS C, SAMUELS ML, ET AL. Phase 11 study of spirogermanium in metastatic prostate cancer. Cancer Treat Rep 1986;70:1129-1130. (56) IzBiICKI RM, AMER MH, AL-SARRAF M. Combination of Adriamycin and cyclophosphamide in the treatment of metastatic prostatic carcinoma. Cancer Treat Rep 1979; 63:999-1001. (57) MERRIN C, ETRA W, WAISMAN Z, ET AL. Chemotherapy of advanced carcinoma of the prostate with 5-fluorouracil, cyclophosphamide, and Adriamycin. J Urol 1976;115: 86-88. (58) LLoYDp RE, JONES SE, SALMON SE, ET AL. Combination chemotherapy with Adriamycin (NSC-123127) and cy- clophosphamide (NSC-26271) for solid tumors: A phase II trial. Cancer Treat Rep 1976;60:77-83. (59) SOLOWAY MS, SHIPPEL RM, IKARD M. Cyclophosphamide, doxorubicin hydrochloride, and 5-fluorouracil in ad- vanced carcinoma of the prostate. J Urol 1979;122: 637-639. (60) KENNEALEY GT, MARSH JC, WALSH DA, ET AL. Treatment of advanced carcinoma of the prostate with estramustine and 5-fluorouracil (FU). Proc Am Assoc Cancer Res, ASCO 1978;19:394. (61) BECKLEY S, WAISMAN Z, SLACK N, ET AL. Chemotherapy in metastatic, hormone refractory prostatic cancer using chlorambucil in combination with prednisolone versus conjugate prednimustine (LEO 1031). Urology 1981; 17:446-448. 162 (62) PRESANT CA, VAN AMBURG A, KLAHR C, ET AL. Chemo- therapy of advanced prostatic cancer with Adriamycin, BCNU, and cyclophosphamide. Cancer 1980;46:2389- 2392. (63) CITRIN DL, HOGAN TF. A phase II evaluation of Adria- mycin and cisplatinum in hormone resistant prostate cancer. Cancer 1982;50:201-206. (64) PERLOFF M, OHNUMA T, HOLLAND JF, ET AL. Adriamycin (ADM) and diammine dichloroplatinum (DDP) in ad- vanced prostatic cancer. Proc Am Assoc Cancer Res, ASCO 1977;18:333. (65) ANDERSON L, EDSMYR F, EsposTi PL, ET AL. Cyclophos- phamide, prednisolone therapy in advanced prostatic carcinoma. Scand J Urol Nephrol [Suppl] 1980;55:169- 171. (66) SEIFTER E, BUNN P, COHEN M, ET AL. A trial of combination chemotherapy followed by hormonal therapy for previ- ously untreated metastatic carcinoma of the prostate. J Clin Oncol 1984;4:1365-1373. (67) BERRY J, MACDONALD RN. Cisplatin, cyclophosphamide, and prednisone therapy for stage D prostatic cancer. Cancer Treat Rep 1982;66:1403-1404. (68) KasiMIS BS, MORAN EM, MILLER JB, ET AL. Treatment of hormone-resistant metastatic carcinoma of the prostate with 5-FU, doxorubicin, and mitomycin (FAM’): A preliminary report. Cancer Treat Rep 1983;67:937-939. (69) Hsu DS, BABAIAN RJ. 5-Fluorouracil, Adriamycin, mito- mycin-C (FAM) in the treatment of hormonal-resistant stage D adenocarcinoma of the prostate. Proc ASCO 1983,2:133. (70) PAULSON DF, BERRY WR, Cox EB, ET AL. Treatment of metastatic endocrine-unresponsive carcinoma of the pros- tate gland with multiagent chemotherapy: Indicators of response to therapy. INCI 1979;63:615-622. (71) BUELL GV, SAIERS JH, SAIKI JH, ET AL. Chemotherapy trial with COMP-F regimen in advanced adenocarcinoma of the prostate. Urology 1978;11:247-250. (72) ANDERSON JR, CAIN KC, GELBER RD. Analysis of survival by tumor response. J Clin Oncol 1983;1:710-719. (73) Scott WW, GIBBONS RP, JOHNSON DE, ET AL. The continued evaluation of the effects of chemotherapy in patients with advanced carcinoma of the prostate. J Urol 1976,116:211-213. (74) MurPHY GP, GIBBONS RP, JOHNSON DE, ET AL. A comparison of estramustine phosphate and streptozotocin in patients with prostatic carcinoma who had extensive irradiation. J Urol 1977;118:288-291. (75) ScHMmiDT JD, ScoTT WW, GIBBONS RP, ET AL. Comparison of procarbazine, imidazole-carbamide and cyclophos- phamide in relapsing patients with advanced carcinoma of the prostate. J Urol 1979;121:185-189. (76) MURPHY GP, GIBBONS RP, JOHNSON DE, ET AL. The use of estramustine and prednimustine versus prednimustine alone in advanced metastatic prostatic cancer patients who have received prior irradiation. J Urol 1979;121: 763-765. (77) LOENING SA, Scott WW, DEKERNION J, ET AL. A comparison of hydroxyurea, methyl-chloroethyl-chloro- hexyl-nitrosourea and cyclophosphamide in patients with advanced prostate cancer. J Urol 1981;125:812-816. (78) SoLOowAY MS, DEKERNION JB, GIBBONS RP, ET AL. Comparison of estramustine phosphate and vincristine alone or in combination for patients with advanced hormone refractory previously irradiated carcinoma of the prostate. J Urol 1981;125:664-667. (79) LOENING SA, BECKLEY S, BRADY MF, ET AL. Comparison of estramustine phosphate, methotrexate, and cis-platinum in patients with advanced, hormone refractory prostate cancer. J Urol 1983;129:1001-1006. (80) SOLOWAY MS, BECKLEY S, BRADY MF, ET AL. A com- parison of estramustine phosphate, versus cis-platinum alone versus estramustine phosphate plus cis-platinum in patients with advanced hormone refractory prostate cancer who had extensive irradiation to the pelvis or lumbosacral area. J Urol 1983;129:56-61. (81) EISENBERGER M, SIMON R, O’DWYER P, ET AL. A reevalua- tion of nonhormonal cytotoxic chemotherapy in the treatment of prostatic carcinoma. J Clin Oncol 1985;3: 827-841. (82) SMALLEY RV, BARTOLUCCI A, HEMSTREET G, ET AL. A phase II evaluation of a three drug combination of cyclophosphamide, doxorubicin and 5-fluorouracil and of 5-fluorouracil in patients with advanced bladder carci- noma or stage D prostatic carcinoma. J Urol 1981;125: 191-194. (83) CHLEBOWSKI RT, HESTORFF R, SARDOFF L, ET AL. Cyclophosphamide (NSC-26271) versus the combination of Adriamycin (NSC-123127), 5-fluorouracil (NSC- 19893) and cyclophosphamide in the treatment of meta- static prostatic cancer—a randomized trial. Cancer 1978; 42:2546-2552. (84) Muss HB, HOWARD V, RICHARDS F, ET AL. Cyclophos- phamide versus cyclophosphamide, methotrexate and 5-fluorouracil in advanced prostatic cancer—random- ized trial. Cancer 1981;47:1949-1953. (85) HERR HW. Cyclophosphamide, methotrexate and 5-fluo- rouracil combination chemotherapy versus chloroethyl- cyclohexy-nitrosourea in the treatment of metastatic prostatic cancer. J Urol 1982;127:462-465. (86) TEJADA F, EISENBERGER MA, BRODER LA, ET AL. 5-Flu- orouracil versus CCNU in the treatment of metastatic prostatic cancer. Cancer Treat Rep 1977;6:1589-1590. (87) PAVONE-MALACUSO M, LuND F, MULDER JF, ET AL. EORTC protocols in prostatic cancer. An interim report. Scand J Urol Nephrol [Suppl] 1980;55:163-168. (88) STEPHENS RL, VAUGHN C, LANE M, ET AL. Adriamycin and cyclophosphamide versus hydroxyurea in advanced pros- tatic cancer. A randomized Southwest Oncology Group study. Cancer 1984;53:406-410. (89) TORTI FM, SHORTLIFFE L, CARTER SK, ET AL. A randomized study of doxorubicin versus doxorubicin plus cisplatin in endocrine-unresponsive metastatic prostatic carcinoma. Cancer 1985;56:2580-2586. (90) KASIMIS B, MILLER JB, KANESHIRO CA, ET AL. Cyclophos- phamide versus 5-fluorouracil, doxorubicin, and mito- mycin C (FAM) in the treatment of hormone-resistant metastatic carcinoma of the prostate: A preliminary report of a randomized trial. J Clin Oncol 1985;3: 385-392. (91) EMRICH LJ, PRIORE RL, MURPHY GP, ET AL. Prognostic factors in patients with advanced stage prostatic cancer. Cancer Res 1985;45:5173-5179. (92) MURPHY GP, BECKLEY S, BRADY MF, ET AL. Treatment of newly diagnosed metastatic prostate cancer patients with chemotherapy agents in combination with hormones versus hormones alone. Cancer 1983;51:1264-1277. (93) GIBBONS RP, BECKLEY S, BRADY MF, ET AL. The addition of chemotherapy to hormonal therapy for treatment of patients with metastatic carcinoma of the prostate. J Surg Oncol 1983;23:133-142. (94) TANNOCK IF. Is there evidence that chemotherapy is of benefit to patients with carcinoma of the prostate? J Clin Oncol 1985;3:1013-1021. (95) SUAREZ AJ, LAMM DL, RADWIN HM, ET AL. Androgen NCI MONOGRAPHS, NUMBER 7, 1988 priming and cytotoxic chemotherapy in advanced pros- tatic cancer. Cancer Chemother Pharmacol 1982;18: 261-265. (96) KEDIA KR, KELLERMEYER RW, PERSKY L, ET AL. Hormonal stimulation followed by multiagent chemotherapy in estrogen unresponsive prostatic carcinoma. In Proceed- ings of the American Urological Association, Boston, MA, May 10-14, 1981, abstr 191. (97) MANNI A, SANTEN RJ, ET AL. Androgen priming and response to chemotherapy in advanced prostatic cancer. J Urol 1986;136:1242-1246. (98) BYAR DP. The Veterans Administration Cooperative Uro- MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 163 logical Research Group's studies of cancer of the prostate. Cancer 1973;32:1126-1130. (99) BYAR DP. Review of the Veterans Administration studies of cancer of the prostate and new results concerning treatment of stage I and II tumors. In Bladder Tumors and Other Topics in Urological Oncology (Pavone-Malacuso M, Smith PH, Edsmyr F, eds). New York: Plenum Press, 1980, pp 471-492. (100) HURST KS, BYAR DP. An analysis of the effects of changes from the original treatment in a clinical trial of treatment for prostatic cancer. J Chronic Dis 1973;26:311-324. (101) BYAR DP. VACURG studies of conservative treatment. Scand J Urol Nephrol [Suppl] 1980:;55:99-102. Hormone Therapy for Prostate Cancer: Results of the Veterans Administration Cooperative Urological Research Group Studies’ David P. Byar,* Donald K. Corle? ABSTRACT —Between 1960 and 1975, the Veterans Adminis- tration Cooperative Urological Research Group conducted a con- secutive series of 3 major randomized clinical trials comparing various endocrine treatments for newly diagnosed prostate cancer patients. Six major conclusions concerning hormonal treatment emerged from these studies: 1) increased hazard of cardiovascu- lar death after therapy with 5 mg diethylstilbestrol (DES); 2) or- chiectomy plus DES no better than orchiectomy or DES alone; 3) equivalent effect of 1.0 and 5.0 mg DES on cancer; 4) reduced cardiovascular hazard from therapy with 1.0 mg DES; 5) Prem- arin and Provera no better than 1.0 mg DES at doses studied; 6) decisions about hormone treatment at diagnosis dependent on patient characteristics, mainly age and Gleason grade. In this pa- per, these studies are reviewed briefly and data are presented to support these conclusions. Some tentative treatment recommen- dations are proposed.—NCI Monogr 7:165-170, 1988. The VACURG was organized in 1959 to investigate treatment of prostate and bladder cancer in randomized clinical trials. Over the next 15 years, approximately 4,000 patients with newly diagnosed prostate cancer were ran- domized to 1 of the 3 consecutive VACURG protocols (1-3). These studies came to an end in 1975 when a review committee of the National Cancer Institute disapproved a grant request for further studies, presumably believing that the era of endocrine treatment for prostate cancer was over and that further progress in treatment by endocrine manip- ulations seemed unlikely. Recent studies with the luteinizing hormone-releasing hormone agonists have reawakened interest in the VACURG studies because they were large, well-conducted, randomized clinical trials and because a careful review of these studies might provide answers to questions needed for designing new protocols such as: 1) Has endocrine therapy been shown to increase overall survival? 2) If DES is used, what dose schedule should be studied? 3) Is it still permis- sible to have an untreated control group? These questions should be kept in mind as we review the 3 major VACURG studies. In all VACURG studies, the patients were newly di- agnosed, and the following staging system was used (fig. ABBREVIATIONS: VACURG = Veterans Administration Coopera- tive Urological Research Group; DES = diethylstilbestrol. ! Mention of trade names does not imply endorsement of the products by the United States Government. 2 Biometry Branch, Division of Cancer Prevention and Control, National Cancer Institute, Bethesda, MD. * Reprint requests to: David P. Byar, M.D., Biometry Branch, Division of Cancer Prevention and Control, National Cancer Institute, Executive Plaza North, Rm. 344, National Institutes of Health, Bethesda, MD 20892. 1): stage I, incidentally found microscopic cancer; stage II, palpable cancer by rectal examination not extended beyond the prostatic capsule; stage III, local extension beyond the prostatic capsule; and stage IV, elevated prostatic acid phos- phatase (determined by central laboratory) or demonstrated metastases. No patients had staging laparotomies and bone scans were not used in staging. Note that unlike many other staging systems, the VACURG put all patients with elevated acid phosphatase levels into stage IV. The data presented in table 1 provide justification for the use of acid phos- phatase in staging prostate cancer patients. Note that the death rates from cancer were about the same for patients who only had elevated phosphatase levels and those who only had metastases. Both death rates were distinctly higher than that for stage III patients. These data strongly sug- gest that elevated acid phosphatase levels represent nonde- tectable distant metastasis. This Consensus Development Conference is concerned with the management of clinically localized prostate can- cer. According to the VACURG staging system, this would refer at most to stages I, II, and III as just presented. How- ever, it is well-known today that many patients with these apparently clinically localized tumors actually have metas- tases in the regional lymph nodes. Because endocrine ther- apy is a systemic rather than a regional or localized form of treatment, we will not limit our discussion of the VACURG studies to patients in these 3 stages but rather present the data from all stages that we believe are pertinent to under- standing the action, toxicity, and role of endocrine treatment in the management of patients with prostate cancer. Study 1 was undertaken by the VACURG to find out how 5.0 mg DES daily compared with orchiectomy and to determine whether the 2 treatments together were better than either one alone (4). Accordingly, from 1960 until 1967, patients with stages III and IV prostate cancer were randomized into 4 treatment groups: placebo, 5.0 mg DES daily by mouth, orchiectomy plus placebo, and orchiectomy plus 5.0 mg DES. By the end of the study, about 475 patients were in each of the 4 groups. An additional 120 patients with stage I disease and 179 with stage II disease were randomized to either prostatectomy plus placebo or prostatectomy plus 5.0 mg DES daily by mouth. The principal result of study 1 was that the 5.0-mg dose of DES was associated with an increased risk of death from cardiovascular causes (5). This effect is illustrated in figure 2, in which cumulative deaths from heart or vas- cular disease are plotted versus time for the 4 treatment groups of patients with stages III and IV disease. The ex- cess cardiovascular toxicity of estrogen appears within the first year. A similar plot was observed for deaths from pul- 165 166 PROSTATIC ACID EVIDENCE OF METASTASES STAGE RECTA ¢ CIAL EXAMINATION PHOSPHATASE X-RAY OR BIOPSY | No Induration < 10 KAU. 0 . . QQ) FIGURE 1.—Staging system used in all VACURG studies; KAU = tartrate- I Goce NEALE <10KAU. 0 inhibitable portion of the serum acid phosphatase measured in King- Armstrong units. nm Extra-Prostatic <10KAU 0 Extension ToT vo | Any O92 > 1.0 KAU. OR + Findings monary emboli. This result surprised both the investigators and the general medical public and was not believed for some time by many. This effect was also seen (fig. 3) in stage | DES-treated patients who had a distinctly worse overall survival (P < .05). The cardiovascular hazard of 5 mg DES daily was not seen in patients with stage II disease. A possible explanation for this may be that stage I patients were usually diagnosed following transurethral resection, whereas stage II patients were diagnosed because of ab- normal findings on rectal examination. Because all patients in stages I and II were to be treated by radical prostatec- tomy, most stage I patients had two major operations, the transurethral resection followed by the radical prostatec- tomy in a relatively short period, whereas this was not true for the stage II patients whose initial operation was usually only a needle biopsy of the prostate. In fact, we found that almost all the excess deaths that occurred among patients in stage I during the first year were confined to that group who had two operations within 2 months. Analyses of time to progression for stage III patients (defined as time until first metastases or first increase in acid phosphatase or death from prostate carcinoma) showed that the 3 endocrine treatment arms had distinctly less rapid progression than did the placebo arm (fig. 4), but curves for overall survival showed no clear survival advantage for initial endocrine therapy in either stage III or stage IV, partly because of the excess cardiovascular deaths in the DES-treated arms. Orchiectomy plus 5.0 mg DES was no better than either treatment alone with respect to overall survival, but DES was more effective than orchiectomy TABLE 1.—Study 3: death rates for patients in stage III and various categories of stage IV Deaths/1,000 patient-months No.of All Prostate ~~ Cardiovascular Category patients causes cancer disease Stage III patients 531 11.8 21 5.1 Stage IV patients Increased acid 289 19.7 8.8 5.0 phosphatase only Metastases only 25 19.5 9.0 4.5 Both metastases and 150 327 213 5.6 increased acid phosphatase in preventing cancer deaths (6). When interpreting these results, it is important for one to note that in all VACURG studies the physician could change the patient’s treatment if indicated because of symptoms or advancing disease. In fact, 44% of patients in stages III and IV assigned to placebo had their treatments changed later, and thus comparisons with the placebo group may more properly be interpreted as evaluations of immediate endocrine treatment at diagnosis versus delayed endocrine treatment, if a change was later required. After the completion of study 1, 506 patients in stages III and IV were randomized to the second VACURG study in which placebo, 0.2 mg, 1.0 mg, and 5.0 mg DES, all daily and by mouth, were compared. This study was stopped early because once again a pattern of excess cardiovascular deaths in the 5.0-mg DES arm was beginning to emerge. This pattern was particularly prominent in stage III patients who were at less risk from cancer death than were stage IV patients (fig. 5). The principal finding of this study was that 1.0 mg DES had the same beneficial effects in retarding cancer progression as did 5.0 mg DES (fig. 6). The equivalence of the 1.0- and 5.0-mg doses of DES was also seen in plots of cause-specific survival for stage IV patients when cancer deaths only were used. 0 7 27 n T 100 zs 2° : op 0 J Le I 7 oc w m => L = 50 Z e Placebo S o Estrogen Ek BI a Orch + Placebo 3 a Orch + Estrogen = 2 © 1 I I 1 deer 0 1 2 3 4 5 5+ YEARS FIGURE 2.—Cumulative No. of deaths from heart or vascular disease for stage III and IV patients in study 1. Orch = orchiectomy. NCI MONOGRAPHS, NUMBER 7, 1988 D oS % SURVIVING SH o e Pros + Pcb, Stage | Fo Pros + Des, Stage | a Pros + Pcb, Stage Il 20 + a Pros + Des, Stage I fg Id IA AU I mde lll] 01 2 3 45 6 7 8 9 10112 YEARS FIGURE 3.—Actuarial survival curves for all causes of death for patients in stages I and II in study 1. Pros + Pcb = prostatectomy and placebo. Study 2 also showed that treatment with 1.0 mg DES beginning at diagnosis increased overall survival in stages IIT and IV patients compared with placebo (fig. 7). In ad- dition, an analysis of study 2 data with a mathematical model incorporating covariate information (7) suggested that placebo or no treatment would be preferable for older patients with low-grade tumors, but immediate estrogen therapy might increase survival for younger patients with high-grade tumors. Study 3 was begun in 1969. Patients with stage IIl or IV disease were randomized to 1 of 4 oral treatments given daily: 2.5 mg Premarin, 30 mg Provera, 30 mg Provera plus 1.0 mg DES, and 1.0 mg DES alone. Altogether, 1,112 pa- tients were entered before the study was closed in 1975. Although there were no significant differences in overall survival between treatments, 1.0 mg DES was somewhat more effective than 2.5 mg Premarin or 30 mg Provera in retarding progression from stage III to IV, so that the overall nN oO S S xD o eo Placebo Io Estrogen a Orch + Placebo a Orch + Estrogen % PROGRESSED TO STAGE IV & beers Ried 01 2 3 4 5 6 7 8 9 1011 12 YEARS FIGURE 4.—Actuarial curves for progression from stage III to IV for patients in study 1. Orch = orchiectomy. MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 167 Q 2 2 > oc ? 40 2 e Placebo I a Des 0.2 mg 20 | o Des 1.0 mg a Des 5.0 mg 0 I 1 1 1 1 | | 0 1 2 3 4 5 6 7 YEARS FIGURE 5.—Actuarial survival curves for deaths due to cardiovascular disease only for patients in stage III in study 2. conclusion was that the other 3 endocrine treatment arms (at the doses studied) were not superior to 1.0 mg DES. Patients in study 3 were older than patients in studies 1 or 2. Results of exploratory data analysis suggested that the moderate benefit observed in the 2 treatment groups receiving 1.0 mg DES was most apparent in younger pa- tients (<75 yr) with high-grade tumors (Gleason sum 7-10) and that there could be increased hazard of cardiovascular death for these 2 treatment groups compared with those re- ceiving either Premarin or Provera alone for patients aged 75 or older. From 1967 to 1975, patients with stage I or II dis- ease were randomized to placebo or prostatectomy plus placebo. No significant benefit for radical prostatectomy was demonstrated (8). However, this result must be inter- preted with caution because the study was small and mod- ern staging methods were not used (9). In conjunction with study 3, another clinical trial desig- e Placebo I aDes0.2mg 80 | oDes 1.0 mg | a Des 50mg % PROGRESSED TO STAGE IV 1 . 1 1 1 J 0 1 2 3 4 5 6 YEARS FIGURE 6.—Actuarial curves for progression from stage III to IV for patients in study 2. 168 100 80 40 % SURVIVING | e Placebo a Des 0.2 mg = o Des 1.0 mg I a Des 5.0 mg 0 1 1 1 de 1 1 J 0 1 2 3 4 5 6 7 YEARS FIGURE 7.—Actuarial survival curves for all causes of death for patients in stages III and IV in study 2. nated “Study 3, Phase II” was conducted for stage I and II patients who were either too old, too ill, or unwilling to be randomized to the trial just described in which radical prostatectomy was 1 treatment option. In this study, about 100 stage I and about 50 stage II patients were randomized to each of 2 treatments given daily by mouth, i.e., placebo or 1.0 mg DES. The results were puzzling because the estrogen treatment was found to be distinctly harmful in stage I pa- tients (especially producing excess cardiovascular deaths), but estrogen treatment appeared to be protective in stage II patients. Summary data for this study are given in table 2. The 5-year survival rate for all causes of death in stage I was 61% for placebo- versus only 45% for DES-treated pa- tients (P = .026). The corresponding rates in stage II, how- ever, were 48% for placebo and 75% for DES (P = .058). The results were even more striking for survival curves con- structed for cardiovascular causes only. The same reversal was noted with P<<.002 favoring placebo in stage [ but P< .019 favoring 1.0 mg DES in stage II. A recently proposed test for such cross-over or qualitative interactions (10) was significant at P < .05 for all causes of death and P < .025 for cardiovascular deaths. Despite this evidence for a statis- tically significant interaction, our overall impression is that the results in stage II are anomalous because the pattern of causes of death shown in table 2 seems implausible in light of the results from the 3 main studies. What this study does suggest is that even 1.0 mg DES can be associated with increased hazard of cardiovascular death in older patients. DISCUSSION AND CONCLUSIONS We have tried to review briefly the highlights of about 15 years of research on over 4,000 patients, with emphasis on what can be said about hormonal treatment of prostate cancer. Other important findings from this series of studies, such as the development of the Gleason grading system, the general importance and refinement of knowledge about prognostic factors (/1), and changes in laboratory values associated with treatment for prostate cancer, have been described elsewhere (2). Concerning hormonal treatment, six main conclusions can be derived from the VACURG studies. Cardiovascular Hazard of 5.0 mg Diethylstilbestrol This is probably the finding for which the VACURG studies are best known. It is almost certain that this finding would never have emerged if a large randomized clinical trial had not been conducted because it is commonplace for older men such as those who get prostate cancer to die of cardiovascular causes. For several decades, such deaths were regarded as triumphs of treatment because the patients were not dying of prostate cancer! For several years after the first publication of this finding, many physicians did not believe it, but now it appears to be accepted generally. Orchiectomy Plus Diethylstilbestrol No Better Than Orchiectomy or Diethylstilbestrol Alone The VACURG studies were originally set up to deter- mine whether orchiectomy plus estrogen was better than either used alone. In a large observational study published in 1950, Nesbit and Baum (/2) suggested that the combina- tion was superior. Our conclusion that they are equivalent refers to survival from all causes of death. However, data from study 1 suggest that 5.0 mg DES daily are some- what more effective than orchiectomy alone in retarding the growth of the cancer. This is seen both in plots of time to progression for stage III patients (fig. 4) and in survival curves constructed for cancer deaths only (not shown). Equivalent Effect of 1.0 and 5.0 mg Diethylstilbestrol on Cancer The data from study 2 clearly showed that the 1.0-mg dose of DES was equivalent in its effect on the cancer to TABLE 2.—Summary results for study 3, phase II No. of deaths Five-yr No. of Prostate Other survival, No. who Progression Stage patients cancer Cardiovascular (unknown) % progressed rate’ Stage | Placebo 98 1 10 17 (3) 61 5 1.34 1 mg DES 107 4 29 12 (1) 45 7 2.06 Stage 11 Placebo 45 3 16 30) 48 11 7.41 1 mg DES 48 4 7 3(D 75 7 3.58 4 Progression rate is expressed as the number of patients progressing/1,000 patient-months of observation. NCI MONOGRAPHS, NUMBER 7, 1988 that of the 5.0-mg dose. However, it is known that 1.0 mg DES daily is not sufficient to suppress serum testosterone to castrate levels (13,14). Because of this, many urologists currently use 3.0 mg DES (1.0 mg three times a day) in treating prostate cancer. We suspect strongly that this dose is too large. To our knowledge, 1.0 and 3.0 mg DES have never been compared in a randomized trial. We think it is more important to treat the prostate cancer than to treat the serum testosterone. Nevertheless, one must wonder how a dose of estrogen insufficient to suppress completely the serum testosterone could be as effective as the larger dose, which does suppress testosterone to castrate level. We sug- gest that the equivalence of the 1.0- and 5.0-mg doses plus the apparent superiority of 5.0 mg DES over orchiectomy alone in retarding cancer growth indicate that DES acts directly on the cancer cells in addition to inhibiting testos- terone secretion. In an electron microscopic radiographic study, Sinha et al. (1/5) detected in vitro nuclear binding of radiolabeled estradiol in prostate cancer cells. Reduced Cardiovascular Hazard for 1.0 mg Diethylstilbestrol The data from study 2 clearly showed that 1.0 mg DES had a lower cardiovascular hazard than did the 5.0-mg dose and, in fact, that hazard did not appear to be elevated compared with that for the group initially treated with placebo. However, as we have noted, study 2 was stopped early because again the cardiovascular toxicity of estrogen was detected; therefore, it is a smaller study than either study 1 or 3. For this reason, conclusions are less certain. The data from study 3 described in this article, both the main study and the special study (phase II) for stages I and II patients, suggest that even a daily 1.0-mg dose of DES is associated with increased hazard of cardiovascular death in susceptible patients, particularly older men. Premarin or Provera No Better Than 1.0 mg Diethylstilbestrol at Doses Studied Study 3 was a large study, yet we were unable to demon- strate that either Premarin or Provera was superior to 1.0 mg DES in its effect on overall survival in either stage III or stage IV patients. These results clearly depend on the doses chosen for these other agents, but we are unaware of any randomized clinical trials demonstrating that any en- docrine treatment is superior to 1.0 mg DES in its effect on the prostate cancer. The possibility of cardiovascular side effects cannot be ignored, and in making a choice of ther- apy, a physician should take both efficacy and toxicity into account. Decisions About Hormone Treatment at Diagnosis Dependent on Patient Characteristics Whether endocrine treatment for patients with prostate cancer should begin at diagnosis remains an important question. The results presented in figure 7 suggest that patients in stages III and IV begun on treatment with 1.0 mg DES by mouth at diagnosis survive longer than patients begun on placebo, but the difference in overall survival is unimpressive for the first 3 years. We are tempted to say nothing further because these are the results of a properly conducted randomized clinical trial. However, in 2 papers MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 169 (7,16), we applied sophisticated mathematical models to determine which patients were most likely to benefit from treatment, and recently an independent analysis suggested results similar to ours (/7). We admit that conclusions based on such exploratory data analyses are not as firm as those based on random- ized comparisons. However, we believe that such analy- ses may help clinicians interpret randomized clinical trial results and suggest hypotheses to be examined in future studies. Accordingly, it is our overall impression that the question of whether to treat newly diagnosed patients with prostate cancer depends on the characteristics of the pa- tients. In our analyses, the most important factors for choos- ing whether to treat at diagnosis were the patient’s age and the histologic grade of his tumor. In the simplest terms, we believe that younger patients with high-grade tumors are the most likely to benefit from the therapy with DES, whereas older patients with lower grade tumors may actu- ally be harmed. These results may well not apply to other forms of endocrine therapy. RECOMMENDATIONS Based on these six main conclusions and our understand- ing of them as just presented, we tentatively put forth some treatment recommendations that might be considered ap- propriate either for treatment of individual patients or in the design of future randomized clinical trials. Even with the data from randomized clinical trials, treatment recom- mendations are difficult because the question of the gen- eralizability of the results almost always requires specu- lation beyond what was actually observed. We wish to dis- tinguish clearly these tentative treatment recommendations from the conclusions we have just presented. However, we believe strongly in the first treatment recommendation. In summary, our tentative recommendations are the following: 1) If DES is used, the initial dose should not exceed 1.0 mg daily. 2) Patients in stages I-III with low-grade tumors (Gleason score 2-6) probably do not need hormone therapy. 3) Patients with higher grade tumors (Gleason score 7-10) may benefit from hormone therapy begun at diagnosis. 4) Endocrine therapy other than DES should probably be considered for patients over age 75. REFERENCES (1) BYAR DP. The Veterans Administration Cooperative Uro- logical Research Group's studies of cancer of the prostate. Cancer 1973;32:1126-1130. (2) ByArR DP. VACURG studies on prostatic cancer and its treat- ment. /n Urologic Pathology: The Prostate (Tannenbaum M, ed). New York: Lea & Febiger, 1977, pp 241-267. (3) CouNnE A. Carcinoma of the prostate. In Randomized Trials in Cancer: A Critical Review by Sites (Staquet MJ, ed), vol 4. New York: Raven Press, 1978, pp 389-409. (4) VETERANS ADMINISTRATION COOPERATIVE UROLOGICAL RESEARCH GROUP. Carcinoma of the prostate: A contin- uing co-operative study. J Urol 1964;91:590-594. 170 (5) VETERANS ADMINISTRATION COOPERATIVE UROLOGICAL RESEARCH GROUP. Treatment and survival of patients with cancer of the prostate. Surg Gynecol Obstet 1967;124:1011-1017. (6) BLACKARD CE, BYAR DP, JORDAN WP Jr. Orchiectomy for advanced prostatic carcinoma—a reevaluation. Urology 1973;1:553-560. (7) Bar DP, GREEN SB. The choice of treatment for cancer patients based on covariate information: Application to prostate cancer. Bull Cancer (Paris) 1980;67:477-490. (8) ByArR DP, COrRLE DK, VETERANS ADMINISTRATION CO- OPERATIVE UROLOGICAL RESEARCH Group. VACURG randomized trial of radical prostatectomy for stages I and IT prostate cancer. Urology 1981;17(suppl):7-11. (9) ByAr DP, CorLE DK. Letter: Randomized study of prostatic cancer. Urology 1981;18:219-220. (10) GAIL M, SIMON R. Testing for qualitative interactions be- tween treatment effects and patient subsets. Biometrics 1985;41:361-372. (11) Byar DP, CorLE DK. Analysis of prognostic factors for prostatic cancer in the VACURG studies. In Controlled Clinical Trials in Urologic Oncology (Denis L, Murphy GP, Prout GR, et al, eds). New York: Raven Press, 1984, pp 147-169. (12) NesBIT RM, BAUM WC. Endocrine control of prostatic car- cinoma. Clinical and statistical survey of 1,818 cases. JAMA 1950;143:1317-1320. (13) KENT JR, BISCHOFF AJ, ARDUINO LJ, ET AL. Estrogen dosage and suppression of testosterone levels in patients with prostatic carcinoma. J Urol 1973;109:858-860. (14) SHEARER RJ, HENDRY WF, SOMMERVILLE IF, ET AL. Plasma testosterone: An accurate monitor of hormone treatment in prostatic cancer. Br J Urol 1973;45:668-677. (15) SINHA AA, BLACKARD CE, DOE RP, ET AL. The in vitro localization of Hj estradiol in human prostatic carcinoma: An electron microscopic autoradiographic study. Cancer 1973;31:682-688. (16) ByAar DP, CorLE DK. Selecting optimal treatment in clin- ical trials using covariate information. J Chronic Dis 1977,30:445-459. (17) KAY R. Treatment effects in competing-risks analysis of prostate cancer data. Biometrics 1986;42:203-211. Hormonal Therapy for Locally Advanced Prostate Cancer Albert B. Einstein, Jr.! ABSTRACT —A patient with locally advanced prostate cancer (stages C and D1) has a poor prognosis with a high risk of de- veloping and dying of distant metastases. Hormonal therapy is the major form of systemic therapy for metastatic (stage D2) prostate cancer. The most commonly used forms of hormonal therapy are orchiectomy, diethylstilbestrol, and luteinizing hor- mone releasing hormone, agonists that prevent the stimulation of tumor cells by testosterone. They produce a 60%-80% symp- tomatic or objective response rate, but their ability to prolong overall survival remains uncertain. Surgical adrenalectomy, hy- pophysectomy, and pharmacologic adrenal suppression prevent the clinically less significant adrenal androgen stimulation of tumor cells. Antiandrogens competitively inhibit the interaction between androgens and cytosolic androgen receptors. Complete androgen blockade (luteinizing hormone releasing hormone ag- onist and antiandrogen) was initially espoused to be superior to single-agent hormonal therapy, but preliminary results from a multigroup randomized trial suggest that it has only a minimal advantage. The benefit of hormonal therapy in stages C and D1 prostate cancer at the time of diagnosis has not been clearly es- tablished. Available studies are few, and most often they are un- controlled or include only small numbers of patients. However, they suggest that the early use of hormonal therapy prolongs disease-free survival but does not prevent ultimate disease pro- gression or prolong overall survival. Hormone receptor assays may be helpful in the selection of patients who would benefit from early hormonal therapy.—NCI Monogr 7:171-174, 1988. Locally advanced prostate cancer is a systemic dis- ease. Poor prognosis correlates with advanced clinical stage (large size of tumor, extracapsular extension, and regional node metastases), poor tumor differentiation, and elevation of serum acid phosphatase (/). Patients with clinical stage C and surgical stage D1 disease have a high risk of occult distant metastases which, if the patient lives long enough, will become clinically apparent and lead to death (2). Im- provement in the survival of patients with these stages of disease requires effective adjuvant systemic therapy in com- bination with effective local therapy that will either delay or prevent the appearance of the distant metastases. In 1941, Huggins and Hodges (3) first demonstrated that prostate cancer growth was dependent on androgenic stim- ulation. In addition, they reported that orchiectomy or phar- macologic doses of estrogens clinically benefited patients with metastatic prostate cancer. Subsequently, hormonal therapy designed to interfere with androgen stimulation ABBREVIATIONS: DHT = dihydrotestosterone; LH-RH = luteiniz- ing hormone releasing hormone; DES = diethylstilbestrol; VACURG = Veterans Administration Cooperative Urological Re- search Group. ! Division of Hematology and Medical Oncology, Virginia Mason Med- ical Center, 1100 Ninth Ave., Seattle, WA 98111. has been the standard form of systemic therapy for pros- tate cancer. Despite the early development of hormonal therapy, con- tinued controversy surrounds several basic issues: 1) the effect of hormonal therapy on the overall sur- vival of patients with metastatic disease; 2) the relative effectiveness of the different forms of hormonal therapy; and 3) the benefit of early application of hormonal therapy in stages C and D disease compared with utilization of this therapy to the patient who is symptomatic due to progressive disease. ANDROGEN HORMONE PHYSIOLOGY Testosterone, produced by the testis and the principal androgen that stimulates normal prostate and prostate can- cer cell growth, constitutes 95% of the androgen circu- lating in the serum (4). Androstenedione and dehydro- 3-epiandrosterone, both produced by the adrenal glands, compose the other 5%. Ninety-five percent of the serum testosterone is protein-bound to sex hormone-binding glob- ulin with less affinity for serum albumin. Only free serum testosterone is capable of entering the prostate cell to influ- ence its metabolism. It passes through the cell membrane passively and is converted enzymatically to DHT, which combines with a cytoplasmic androgen receptor. This com- plex enters the nucleus and interacts with the genetic ma- terial to cause an increase in messenger RNA synthesis and subsequently protein synthesis (fig. 1). Testosterone production is regulated by the hypotha- lamic pituitary gonadal feedback mechanism (fig. 2). The hypothalamus of the brain produces LH-RH, which stim- ulates the pituitary gland to secrete LH. The LH stimu- lates the Leydig cells of the testis to produce testosterone, which then reacts with hypothalamic receptors to prevent further release of LH-RH. In a similar feedback mecha- nism, the synthesis of adrenal androgens is controlled by the production of adrenocorticotropic hormone by the pitu- itary gland, which stimulates the adrenal gland to produce cortisol and androgen (fig. 2). Cortisol interacts with the hypothalamus to inhibit further production of adrenocorti- cotropic hormone releasing hormone. HORMONAL THERAPY FOR PROSTATE CANCER The objective of physicians in prescribing hormonal ther- apy for prostate cancer is to reduce the production of an- drogens or block the effect of the androgens on the cancer cell. The different types of hormonal therapy have been de- veloped with physicians’ increasing knowledge of the phys- iology of androgen synthesis and regulation and the effect of androgen on the cancer cell. 171 172 Protein Synthesis FIGURE 1.—Testosterone (T) stimulation of prostate cancer cells via re- ceptor (Rec) interaction. Bilateral orchiectomy is the hormonal treatment with which all others must be compared. It removes the source of testosterone, which accounts for 95% of the circulat- ing androgens (4). The procedure is simple and results in rapid symptomatic response. The major disadvantages of the treatment are the psychologic trauma associated with castration and the loss of potency (5). Palliative effects of orchiectomy are observed in 60%-80% of the patients and include subjective pain relief and objective reduction in pri- mary tumor size, soft tissue metastases, and ureteral dila- tion. The duration of the palliative response is limited to 12-18 months. Despite objective and subjective responses, the impact of orchiectomy on overall survival has been un- certain. Reported studies included nonrandomized or histor- ical untreated control groups (6), or randomized placebo or control groups; patients ultimately received hormonal ther- apy when their disease progressed (7). The synthetic compound DES directly inhibits the hy- pothalamic production of LH-RH and indirectly LH and testosterone (4). The VACURG studies have demonstrated that 1- to 3-mg daily doses of DES produce palliative re- sponses similar to orchiectomy (7). Possible side effects include gynecomastia, loss of libido, and loss of potency. However, a 5-mg daily dose is associated with an increased death rate due to cardiovascular complications, such as myocardial infarction, congestive heart failure, and throm- boembolic complications, which offset any therapeutic ad- vantage. The combination of orchiectomy and DES does not enhance the response rate. Patients who fail to respond to either orchiectomy or DES or progress after initial re- sponse to either treatment rarely respond to the alternative treatment. The definition of the amino acid sequence of the deca- peptide LH-RH and the subsequent synthesis of analogs with greater potency have provided an alternative method to suppressing testosterone production. Synthetic LH-RH agonists, such as leuprolide or Buserelin (Farbwerke Hoechst AG, Frankfurt, Federal Republic of Germany), initially stimulate the pituitary gland to produce in- creased levels of LH and subsequently testosterone dur- ing the first 2 to 4 weeks of administration. Contin- ued treatment subsequently causes a decrease in LH production and a decline in production to orchiectomy levels (8). The Leuprolide Study Group (9) has com- pared leuprolide and DES in patients with stage D disease. They (9) reported an 86% response rate for leuprolide compared with 85% for DES and a 1-year sur- vival rate of 87% compared with 78%, respectively. The LH-RH agonists are relatively nontoxic but can cause an eID Hypothalamus AS LHRH Pituitary ACTH Gland NN T Cortisol _ Adrenal \ LH / Gland Androgenic Hormones FIGURE 2.—Hypothalamic pituitary regulation of testosterone (T) and adrenal androgen hormone synthesis. increase in symptoms such as bone pain during the first 2 to 4 weeks due to the increased testosterone production. In addition, the initial forms of this drug required daily par- enteral injections or intranasal administration. A new depot form is now available that is administered monthly. Second-line hormonal therapy following progression af- ter orchiectomy or DES had been directed toward the suppression of adrenal androgen production by surgical adrenalectomy (1/0), hypophysectomy (10), or medical sup- pression of adrenal function with prednisone or amino- glutethimide (7/1). Unfortunately, they result only in brief subjective responses in 20% to 40% of the patients. This result suggests that the role of adrenal androgens in stimu- lating prostate cancer growth is minimal. Antiandrogens (flutamide, cyproterone acetate) are chemicals that compete with DHT for the cytoplasmic an- drogen receptor (1/2). The antiandrogen-receptor complex is inactive. These agents have clinical activity similar to DES with less toxicity. Control trials during which antian- drogens were compared with DES or orchiectomy have not been completed; therefore, response rates have not been carefully defined. The use of antiandrogens as a single agent is limited by a secondary increase in gonadotropin hormone and androgen production due to the neutralization of the in- hibitory feedback action of androgens at the hypothalamic level. Progressively higher doses of antiandrogens are re- quired to neutralize the rising level of testosterone. Complete androgen blockade has been advocated by Labrie et al. (13) to be the optimal hormonal therapy be- cause it eliminates the effect of both testicular and adrenal androgens. Labrie and his associates treated 47 patients with stage D prostate cancer with leuprolide to suppress the production of testosterone and with flutamide to block the effect of adrenal androgens. They reported a 100% positive response rate and only a 3.3% death rate after 1% years of follow-up. The combination hormonal ther- apy also produced a positive response in 36% of stage D patients previously treated with DES. Labrie believes that untreated prostate cancer is exquisitely sensitive to an- drogens and that the adrenal androgens do play a signifi- cant role in stimulating progression of disease in patients treated only with orchiectomy or DES. Moreover, he be- lieves that androgen-insensitive clones arise only when an- drogen blockade has been incomplete. Stimulated by these results, an intergroup study was initi- ated in the United States in which 617 patients with stage D prostate cancer were randomized to receive either leupro- lide plus flutamide or leuprolide alone (/4). Preliminary re- NCI MONOGRAPHS, NUMBER 7, 1988 sults after a 20-month follow-up indicate only a minimally significant difference in time to progression of disease in favor of the combined therapy. Thus it appears that the ad- dition of flutamide to an LH-RH agonist might offer some slightly increased advantage, but not to the degree origi- nally proposed by Labrie. EARLY USE OF HORMONAL THERAPY FOR LOCALLY ADVANCED DISEASE Controversy continues regarding whether hormonal ther- apy is best used early in asymptomatic patients with locally advanced prostate cancer or whether it should be reserved for the time when the patient develops symptomatic metas- tases. Unfortunately, good randomized controlled trials that address this specific question do not exist. In 1950, Nesbit and Baum (6) reported that patients with various stages of prostate cancer treated with hormone ther- apy had a 5-year survival of 20% compared with 9% for a historical control group that was followed before it was generally known that prostate cancer was a hormonally sensitive tumor. This information has been extrapolated to mean that patients would best be treated by hormonal ther- apy at the time of diagnosis for both locally advanced and metastatic disease whether or not they are symptomatic. This study obviously suffers from the difficulties of com- parisons of treatment results with a historical control group. In the first VACURG study, patients with various stages of prostate cancer were treated with placebo, DES, orchiec- tomy and placebo, or orchiectomy and DES. The daily dose of DES was 5 mg (15). Of the patients with stage C disease, the total deaths were similar in all 4 groups, which suggests that hormonal therapy did not prolong survival compared with the placebo group. However, further analysis indicated that many of the patients initially in the placebo group were treated with hormone therapy at a later date when their dis- ease progressed. Therefore, the control group was not truly untreated. With this in mind, the data suggest that early hor- monal therapy does not improve overall survival compared with delayed treatment. Patients treated with DES experi- enced fewer cancer-related deaths but a greater number of cardiovascular-related deaths than the placebo patients, a finding that offsets the advantage of the treatment. In the second VACURG study, patients were treated with either placebo, 0.2 or 1 mg, or 5S mg DES (/6). Patients with stage C disease had a decreased cancer-related death rate if they received 1 mg or 5 mg DES compared with placebo or 0.2 mg DES. Again, the high death rate due to cardiovascular problems was noted with the 5-mg but not with the 1-mg dose. In addition, 1 mg and 5 mg DES were reported to delay the progression of disease from stage C to stage D. The VACURG studies suggest that hormonal therapy delays progression of disease from early to advanced stage, but it does not prolong overall survival. Taylor and his colleagues (17), reporting on 221 patients with clinical stage C disease treated primarily by definitive external radiotherapy, found no difference in overall sur- vival between those patients treated with or without DES. The 5-year survival rate for both groups was 57%. Zincke and Utz (/8) have reported results of patients with stage D1 disease treated with radical prostatectomy and lymph node dissection, many of whom also had or- MANAGEMENT OF CLINICALLY LOCALIZED PROSTATE CANCER 173 chiectomies. In this nonrandomized study, castrated patients had a 5-year nonprogression rate of 95% compared with 18.5% for patients not treated with orchiectomy. The cas- trated patients also had higher (though not significant) sur- vival rates of 94% at 5 years and 85% at 10 years. In all treatment groups, the number of positive nodes adversely affected progression and survival. This experience suggests that early hormonal treatment is beneficial for locally advanced disease following extensive surgery. Unfortu- nately, the study is nonrandomized and contains a relatively small number of patients. TUMOR CELL HETEROGENEITY Tumors that respond to hormonal therapy contain a sig- nificant number of androgen-sensitive cells. Progression of disease after hormonal therapy results from the emergence of androgen-resistant cell clones, which may arise due to clonal selection or cell adaptation (19). If clonal selection is primarily operative, mutations causing androgen-resistant cells are more likely to occur the longer the tumor grows. These clones may arise before diagnosis and treatment are established or later, after treatment has been started. The early use of hormonal therapy before resistant cells are prevalent could be theoretically justified, particularly if an- drogen sensitivity could be predicted by simple convenient tests. ANDROGEN SENSITIVITY ASSAYS The ability of physicians to identify patients with androgen-sensitive tumors at high risk of occult distant metastases might lead to more rational early treatment. Gleason histologic grade of tumor, intracellular DHT lev- els, cytosol androgen receptor content, and nuclear andro- gen receptor content have been studied for their ability to predict hormone responsiveness (20-22). Trachtenburg and Walsh (20) reported a correlation between nuclear andro- gen receptor content and duration of subjective and objec- tive responses to hormone therapy. A significantly increased disease-free interval was observed in patients treated with hormonal therapy with DHT levels greater than 2 ng/g in prostate tissue compared with patients with values less than this (27). Benson and his colleagues (22) have recently found that the time to progression following hormonal ther- apy correlated with the nuclear and total androgen receptor binding, but not with cytosol androgen receptor binding. In addition, histologic grade 4 lesions were the least respon- sive to hormonal therapy. Unfortunately, these assays are not practical enough at this time for general applicabil- ity. With more practical useful assays, patients with locally advanced disease might be selected for further studies in which the potential benefit of hormone therapy in the ad- juvant setting would be evaluated. CONCLUSIONS The early use of traditional suppressing hormone therapy for clinical stages C and DO and pathologic stage DI disease may prolong disease-free survival, but there is no conclusive evidence that it prevents disease progression or prolongs overall survival. Unfortunately, the few studies in which investigators attempted to evaluate effect on disease- 174 free survival have been poorly designed. Prolongation of disease- and symptom-free survival for this elderly male population may be of considerable benefit and thus a goal worthy of achievement. Areas of future research in the early use of hormonal therapy in patients with locally advanced disease that physi- cians might consider include the: 1) effect of complete androgen blockage by the com- bination of orchiectomy or an LH-RH agonist plus an antiandrogen on the rate of progression of dis- ease and overall long-term survival, 2) degree of heterogeneity of early tumors regarding androgen sensitivity and insensitivity and identifi- cation of factors that influence the development of insensitivity; 3) correlation of predictors of hormone-sensitive tu- mors with clinical results of adjuvant hormonal trials; 4) benefit of adjuvant combination hormonal therapy, chemotherapy, and/or immunotherapy in delaying progression and prolonging survival, with the as- sumption of future development of effective agents; 5) development of new agents that take advantage of the hormonal responsiveness of prostate cancer; and 6) potential value of androgenic stimulation of prostate cancer to enhance its susceptibilities to cycle-specific cytotoxic therapy. REFERENCES (1) HuBeN RP, MURPHY GP. Prostate cancer: An update. CA 1986;36:274-292. (2) CaTAaTONA WIJ, Scott WW. Carcinoma of the prostate: A review. J Urol 1978;119:1-8. (3) HuGaGIins C, HODGES CV. Studies on prostatic cancer. I. The effect of castration, estrogen and androgen injections on serum phosphatases in metastatic carcinoma of the prostate. Cancer Res 1941;1:293-297. (4) Scott WW, MENON M, WALSH PC. Hormonal therapy of prostatic cancer. Cancer 1980;45:1929-1936. (5) REsNICK MI. Hormonal therapy in prostatic carcinoma. Urol- ogy 1984;24(suppl): 18-23. (6) NEsBIT RM, BAuM WC. Endocrine control of prostatic can- cer. Clinical and statistical survey of 1,818 cases. JAMA 1950;143:1317-1320. (7) BYAR DP. The Veterans Administration Cooperative Uro- logical Research Group’s studies of cancer of the prostate. Cancer 1973; 32:1126-1130. (8) SCHALLY AV, KASTIN AJ, Coy DH. LH-releasing hormone and its analogues: Recent basic and clinical investigations. Int J Fertil 1976;21:1-30. (9) LEUPROLIDE STUDY GROUP. Leuprolide versus diethylstil- bestrol for metastatic prostatic cancer. N Engl J Med 1984;311:1281-1286. (10) REYNOSO G, MuRrPHY GP. Adrenalectomy and hypophy- sectomy in advanced prostatic carcinoma. Cancer 1972; 29:941-945. (11) SANFORD EJ, DRAGO JR, ROHNER TJ, ET AL. Amino- glutethimide medical adrenalectomy for advanced pros- tatic carcinoma. J Urol 1976;115:170-174. (12) WALSH PC, KORENMAN SG. Mechanism of antiandrogenic action: Effect of specific intracellular inhibitors. J Urol 1971;105:850-857. (13) LABRIE F, DUPONT A, BELANGER A. Complete androgen blockade for the treatment of prostate cancer. In Impor- tant Advances in Oncology (DeVita VT Jr, Hellman §S, Rosenberg SA, eds). Philadelphia: Lippincott, 1985, pp 193-217. (14) CRAWFORD ED, MCLEOD AD, SPAULDING J, ET AL. A com- parison of leuprolide with flutamide and leuprolide in pre- viously untreated patients with clinical stage D2 cancer of the prostate. J Urol 1987;137:abstr 256. (15) VETERANS ADMINISTRATION COOPERATIVE UROLOGICAL RESEARCH Group. Treatment comparisons. J Urol 1967;98:516-522. (16) VETERANS ADMINISTRATION COOPERATIVE UROLOGICAL RESEARCH GROUP. Estrogen treatment for cancer of the prostate—early results with three doses of diethylstilbes- trol and placebo. Cancer 1970;26:257-261. (17) TAYLOR WIJ, RICHARDSON RG, HAFERMANN MD. Ra- diation therapy for localized prostate cancer. Cancer 1979;43:1123-1127. (18) ZINCKE H, Utz DC. Observations on surgical management of carcinoma of prostate with limited nodal metastases. Urology 1984;24:137-145. (19) BENSON MC, COFFEY DS. Prostate cancer research: Current concepts and controversies. Semin Urol 1983;1:323-330. (20) TRACHTENBURG J, WALSH PC. Correlation of prostatic nu- clear androgen receptor content with duration of response and survival following hormonal therapy in advanced pros- tatic cancer. J Urol 1982;127:466-471. (21) GELLER J, ALBERT JD. Endocrine therapy: Predictors of response to prostatic cancer. Semin Urol 1983;1:291-298. (22) BENSON RC, GORMAN PA, O'BRIEN PC, ET AL. Relation- ship between androgen receptor binding activity in hu- man prostate cancer and clinical response to endocrine therapy. Cancer 1987;59:1599-1606. PREPUBLICATION NOTICE FROM THE NATIONAL CANCER INSTITUTE ONCOLOGY RECENT OVERVIEW REVIEWS oars H-.. Er EEE Gs Ee @® [J ONCOLOGY [] RECENT REVIEWS [J] NCI Monographs OVERVIEWS If you would like to be notified when the above and other cancer-related publications are issued, please complete the address information below and mail the entire page to the appropriate address listed below. Priority Announcement Request Yes, please put me on your free Priority Announcement List (N-569) to be notified when new RECENT REVIEWS, ONCOLOGY OVERVIEWS, NCI MONOGRAPHS and other cancer publications are issued. If you are located in the U.S., mail this form to: (Name) 10/87 guperintendent of Documents Mail Stop: SSOM (Organization) Washington, D.C. 20401-9374 (Address) If you are located outside the U.S., return this form to: ICIC/NCI/NIH (City, State, Zip Code, or Country) Bldg 82/Rm 103 Bethesda, MD 20892 PLEASE DO NOT COMPLETE THIS FORM IF YOU HAVE DONE SO PREVIOUSLY. THANK YOU. Please detach and mail entire page to the appropriate address. NCI MONOGRAPHS, NUMBER 7, 1988 175 ( NATIONAL CANCER INSTITUTE MONOGRAPHS ] Yes, please send me the following indicated publications: NCI MONOGRAPHS [ ( ( ( ( ( ] Number 1, 1986 - Proceedings of the NIH Consensus Development Conference on Adjuvant Chemotherapy and Endocrine Therapy for Breast Cancer Stock No. 017-042-00190-1, $8.50 ] Number 2, 1986 - Cancer Control Objectives for the Nation: 1985-2000 Stock No. 017-042-00191-9, $6.00 ] Number 3, 1987 - International Symposium on Labeled and Unlabeled Antibody in Cancer Diagnosis and Therapy Stock No. 017-042-00194-3, $9.50 ] Number 4, 1987 - First Conference on DNA Topoisomer- ases in Cancer Chemotherapy Stock No. 017-042-00208-7, $8.00 ] Number 5, 1987 - Development of Folates and Folic Acid Antagonists in Cancer Chemotherapy Stock No. 017-042-00215-0, $12.00 ] Number 6, 1988 - Conference on the Interaction of Radiation Therapy and Chemotherapy Stock No. 017-042-00234-6, $19.00 NATIONAL CANCER INSTITUTE MONOGRAPH SERIES * * Superintendent of Documents Publication Order Form 1. ] No. 54 - The Occurrence of Tumors in Domestic Animals, December 1980 Stock No. 017-042-00145-5, $9.50 ] No. 55 - Experimental Evaluation of Antitumor Drugs in the USA and the USSR and Clinical Correlations, December 1980 Stock No. 017-042-00144-7, $13.00 No. 57 - Surveillance, Epidemiology, and End Results Program: Incidence and Mortality Data: 1973-77, October 1981 Stock No. 017-042-00156-1, $26.00 No. 62 - Third Symposium on Epidemiology and Cancer Registries in the Pacific Basin, January 1983 Stock No. 017-042-00164-1, $13.00 No. 64 - Management Operations of the National Cancer Institute That Influence the Governance of Science, May 1984 Stock No. 017-042-00175-7, $9.00 No. 65 - Use of Small Fish Species in Carcinogenicity Testing, May 1984 Stock No. 017-042-00173-1, $16.00 No. 66 - Photobiologic, Toxicologic, and Pharmacologic Aspects of Psoralens, December 1984 Stock No. 017-042-00176-5, $12.00 — Total cost of my order is $ [ Order Processing Code: 6242 ] No. 67 - Selection, Follow-up, and Analysis of Prospective Studies: A Workshop, June 1985 Stock No. 017-042-00182-0, $14.00 ] No. 68 - Multiple Primary Cancers in Connecticut and Denmark, December 1985 Stock No. 017-042-00186-2, $23.00 ] No. 69 - Fourth Symposium on Epidemiology and Cancer Registries in the Pacific Basin, December 1985 Stock No. 017-042-00188-9, $16.00 ] No. 70 - Forty-five Years of Cancer Incidence in Connecticut, January 1986 Stock No. 017-042-00187-1, $37.00 CANCER TREATMENT SYMPOSIA ( ] Volume 2, 1983 - Proceedings of the Workshop on Patterns of Failure After Cancer Treatment Stock No. 017-042-00167-6, $15.00 Volume 1, 1984 - The Interdisciplinary Program for Radiation Oncology Research Stock No. 017-042-00171-4, $6.50 Volume 2, 1984 - Proceedings of the Conference on 2'-Deoxycoformycin: Current Status and Future Directions Stock No. 017-042-00172-2, $5.00 Volume 3, 1984 - Proceedings of the Workshop on Mitoxantrone Cardiotoxicity and the Workshop on Clinical Evaluation of Anthracycline Cardiotoxicity Stock No. 017-042-00180-3, $4.75 Volume 1, 1985 - Proceedings of the Third Annual Sanrocco International Cancer Symposium: Clinical and Biological Evaluation of the Immunomodifiers Stock No. 017-042-00181-1, $5.50 Volume 2, 1985 - Proceedings of the Workshop on Radiotherapy for Lung Cancer Stock No. 017-042-00183-8, $8.50 Volume 3, 1985 - Proceedings of the National Institutes of Health Consensus Development Conference on Limb-Sparing Treatment of Adult Soft Tissue Sarcomas and Osteosarcomas Stock No. 017-042-00185-4, $6.50 — — — ] Volume 4, 1985 - Compilation of Phase Il Results With Single Antineoplastic Agents Stock No. 017-042-00189-7, $22.00 * Limited number of copies available by calling (301) 496-6735 or writing to NCI Monographs, NIH/NCI, Bldg. 82, Room 227, Bethesda, MD 20892. Charge your order. It’s easy! oY |, . (International customers please add 25% to price.) Prices include regular domestic postage & handling and are good through 12/88. After this date, please call Order and Information Desk at 202-783-3238 to verify prices. Please Type or Print. Mail To: Superintendent of Documents, Government Printing Office, Washington, D.C. 20402-9325. 2. (Company or personal name) 10/87 (Additional address/attention line) (Street address) (City, State, ZIP Code) ( ) (Daytime phone including area code) NCI MONOGRAPHS, NUMBER 7, 1988 3. Please Choose Method of Payment: [] Check payable to the Superintendent of Documents [| apo Deposit Account [TTT TTT [] VISA or MasterCard Account HEEEEEEEEEEEEEEEERER Thank you for your order! (Credit card expiration date) (Signature) 177 me eellem— — - mallee = Elon ll ole — le LL a =e NSIS NEW NCI MONOGRAPH ON COMBINED MODALITY THERAPY The National Cancer Institute announces publication of the proceedings of the “Conference on the Interaction of Radiation Therapy and Chemotherapy” in NCI Monographs, Number 6, 1988. This conference, held in Williamsburg, Virginia, in the fall of 1986, was the first international meeting dedicated to this subject in more than 8 years. The meeting brought together laboratory researchers and clinical investigators in all of the oncologic specialties: chemotherapy, radiation therapy, medical oncology, and pediatric oncology. Topics included the results of clinical trials, the mechanisms of action of combined modality therapy, and the mechanisms of resistance and cross-resistance between the modalities. Included in these proceedings are summaries of the extensive discussions, which guided the preliminary planning of future directions for laboratory and clinical research in this field. Copies of this volume of NCI Monographs (Number 6, 1988) are available from the U.S. Government Printing Office ($19.00 per copy domestic). To order, request Stock No. 017-042-00234-6. Make checks payable to the Superintendent of Documents and mail to: Superintendent of Documents, U.S. Government Printing Office, Washington, D.C. 20402. i i ! Superintendent of Documents Subscriptions Order Form i 1 1 | Order Processing Code Charge your order. VISA i 1 It's easy! NY umm l ! 6242 ! 1 1 1 1 [YES 1 l 9 please send me _____ copy (copies) of the publication indicated above. ) i i i i 1 1. The total cost of my order is $ ; international customers please add an additional 25%. All prices include regular ! 1 y : please p : g 1 1 domestic postage and handling and are good through 12/88. After this date, please call Order and Information Desk at 1 : postag . fing g g p ! 1 202-783-3238 to verify prices. i i i ; Please Type or Print 2. __ 3. Please choose method of payment: | ! (Company oF persons] name) [] Check payable to the Superintendent of Documents : I 1 y o 1 : (Additional address/attention line) [] GPO Deposit Account [TITTTTT] [] : ' [] VISA or MasterCard Account : 1 : - 1 } Guesakien (TTT TTTTITTTTITITI ITI Ig 1 i ! (City, State, ZIP Code) rE Thank you for your order! | l ( ) (Credit card expiration date) I 1 1 ) (Daytime phone including area code) Sm ! } (Signature) | 1 | 4. Mail to: Superintendent of Documents, Government Printing Office, Washington, D.C. 20402-9325. A i i : i Please put me on your FREE Priority Announcement List (N-569) to receive notices of ] 1 new NCI Monographs and other publications of the National Cancer Institute as soon ) ! as they are printed. | 1 1 NCI MONOGRAPHS, NUMBER 7, 1988 179 NATIONAL CANCER INSTITUTE ONCOLOGY OVERVIEWS Produced by the National Cancer Institute, ONCOLOGY OVERVIEWS are specialized bibliographies with abstracts, each referencing 150-500 recent publications on a cancer topic of high current interest, drawn from over 3000 sources. Leading researchers review and select the most relevant and significant abstracts for each topic, resulting in a tightly focused, quick reference to the most recent cancer literature. [ ] YES, please send me the following indicated publications: DIAGNOSIS AND THERAPY SERIES [ | Bone Marrow Transplantation in Cancer, Stock No. 017-042-00201-0, $5.50 [ |] Diagnosis and Treatment of The Acquired Immunodeficiency Syndrome, Stock No. 017-042-00214-1, $5.50 [ ] Biological Markers of Central Nervous System Neoplasms, Stock No. 017-042-00211-7, $4.00 [ |] Diagnosis and Treatment of Esophageal Neoplasms, Stock No. 017-042-00213-3, $3.25 [ ] Cancer Prevention, Stock No. 017-042-00210-9, $4.25 [ |] Diagnosis and Therapy of Basal Cell Carcinoma of the Skin, Stock No. 017-042-00212-5, $2.75 [ ] Childhood Nervous System Neoplasms, Stock No. 017-042-00217-6, $4.50 [ 1 Recent Advances in the Systemic Therapy of Genitourinary Malignancies, Stock No. 017-042-00218-4, $4.50 [ |] Diagnosis and Therapy of Large Bowel Cancer, Stock No. 017-042-00216-8, $5.00 [ ] Carcinoids and Islet Cell Neoplasms, Stock No. 017-042-00230-3, $3.50 CARCINOGENESIS SERIES | | Translocation and Amplification of Oncogenes, Stock No. 017-042-0198-6, $6.00 | 1 Oncogene Protein Products, Stock No. 017-042-00205-2, $5.50 [ |] Chemopreventive Agents, Stock No. 017-042-00204-4, $6.50 | |] Potential Occupational Causes of Cancer II. The Urinary Tract, Stock No. 017-042-00203-6, $4.75 | | Dioxins and Dibenzofurans in Carcinogenesis, 1980-1986, Stock No. 017-042-00206-1, $5.50 [ |] Radioprotectors: Experimental Studies and Clinical Applications, Stock No. 017-042-00207-9, $6.50 Superintendent of Documents Publication Order Form 1. The total cost of my order $ Order Processing Code: *6284 CARCINOGENESIS SERIES (cont.) | |] Free Radicals and Peroxides in the Etiology of Cancer, Stock No. 017-042-00231-1, $9.50 [ ] The Role of Dietary Factors in Carcinogenesis I. Fats, Stock No. 017-042-00226-5, $8.50 [ |] Tumor Promoter Action in Cell Culture, Stock No. 017-042-00225-7, $10.00 [ ] Inherited Cancers and Predisposing Syndromes in Man, Stock No. 017-042-00233-8, $5.50 VIROLOGY, IMMUNOLOGY, AND BIOLOGY SERIES [ ] Viral Etiology of Cancer II. Papillomaviruses, Stock No. 017-042-00199-4, $5.50 [ | Viral Etiology of Cancer III. Hepatitis B Virus, Stock No. 017-042-00197-8, $6.50 [ | Idiotypes, Anti-Idiotypes and Network Regulation, Stock No. 017-042-00202-8, $6.00 [ | Transforming Growth Factors and Other Autocrine Growth Factors, Stock No. 017-042-00200-1, $5.00 [ ] Viral Etiology of Human Cancer IV. Epstein-Barr Virus, Stock No. 017-042-00228-1, $8.50 [ ] Anti-tumor Antibodies, Stock No. 017-042-00229-0, $7.50 [ |] Induced Differentiation of Neoplastic Cells, Stock No. 017-042-00232-0, $10.00 [ | Mechanisms of Resistance to Chemotherapeutic Agents and Strategies for Circumventing Resistance, Stock No. 017-042-00223-1, $9.50 [ |] Tumor Necrosis Factor, Stock No. 017-042-00224-9, $6.50 [ | Oncogenes of DNA Viruses, Stock No. 017-042-00227-3, $8.50 Charge your order. VISA It’s easy! NY wmmm— (International customers please add an additional 25%.) All prices include regular domestic postage and handling and are good through 12/88. After this date, please call Order and Information Desk at 202-783-3238 to verify prices. Please Type or Print 2 ’ (Company or personal name) (Additional address/attention line) (Street address) (City, State, ZIP Code) ( ) (Daytime phone including area code) 3. Please Choose Method of Payment: [] Check payable to the Superintendent of Documents GPO Deposit Account (TTITITIT1-0] VISA or MasterCard Account LIP rrr rrr y] Thank you for your order! (Credit card expiration date) (Signature) 4. Mail To: Superintendent of Documents, Government Printing Office, Washington, D.C. 20402-9325 181 REVIEW REVIEWS The International Cancer Information Center of the National Cancer Institute announces publication of the following three RECENT REVIEWS. RECENT REVIEWS are compilations providing abstracts and citations of major articles published during each calendar year. Each volume is internally categorized, and subject- and author-indexed for easy reference. RECENT REVIEWS: Carcinogenesis 1986 Stock No. 017-042-00219-2, $8.00 RECENT REVIEWS: Cancer Virology, Immunology and Biology 1986 Stock No. 017-042-00220-6, $10.00 RECENT REVIEWS: Cancer Diagnosis and Therapy 1986 Stock No. 017-042-00221-4, $8.50 Superintendent of Documents Publication Order Form Order Processing Code: Charge your order. rea — * 6284 It’s easy! WW) hie 1. Please Type or Print (Form is aligned for typewriter use.) All prices include regular domestic postage and handling and are good through 12/88. After this date, please call Order and Information Desk at 202-783-3238 to verify prices. ; Price Total ty. Stock Numb Titl Qty - umber He Each Price International customers please add an additional 25% Total for Publications Please Type or Print 2 3. Please Choose Method of Payment: ’ (Company or personal name) 0] Check payable to the Superintendent of Documents [] GPO Deposit Account (TII1TIT1T1] -] (Additional address/ attention line) (Street address) [] VISA, CHOICE or MasterCard Account (City, State, ZIP Code) HEEEEEEEEEEEEEEEEEEN ( ) pO a ETE Thank you for your order! (Daytime phone including area code) (Credit card expiration date) (Signature) 4. Mail To: Superintendent of Documents, Government Printing Office, Washington, D.C. 20402-9325 5/87 NCI MONOGRAPHS, NUMBER 7, 1988 183 Se TT TTT. === Superintendent of Documents Subscriptions Order Form Order Processing Code Charge your order. ceo) | VISA | * 6365 It's easy! N | eo—) | ES 9 please send me the following indicated subscription(s) at $60.00 each ($75.00 foreign). JOURNAL OF THE NATIONAL CANCER INSTITUTE 1. The total cost of my orderis$_______for ______ subscriptions. All prices include regular domestic postage and handling and are subject to change. Please Type or Print : 3. Please choose method of payment: (Company or personal name) [] Check payable to the Superintendent of Documents (Additional address/attention line) ] GPO Deposit Account ITI ~t 1 [] VISA or MasterCard Account Gieseratiresy rrr rr TT (City, State, ZIP Code) ( ) (Credit card expiration date) Thank you for your order! (Daytime phone including area code) (Signature) 11/87 4. Mail To: Superintendent of Documents, Government Printing Office, Washington, D.C. 20402-9371 INCI 2-E NCI MONOGRAPHS, NUMBER 7, 1988 185 RETURN PUBLIC HEALTH LIBRARY TO =» 42 Warren Hall » 642-2511 LOAN PERIOD 1 |2 3 2 HOURS 4 5 6 ALL BOOKS MAY BE RECALLED AFTER 7 DAYS ALL JOURNALS ARE NON-RENEWABLE Return to desk from which borrowed DUE AS STAMPED BELOW TPR 28 -4 PV REC'D PUBL APR ¢ & '98 UNIVERSITY OF CALIFORNIA, BERKELEY FORM NO. DD26-7 BERKELEY, CA 94720-6000 FEB 6 1989 NCI Monographs 1 Introduction 3 Consensus Statement: The Management of Clinically Localized Prostate Cancer 7 Overview: Historical and Contemporary I. Staging 15 Histologic Grade, Clinical Stage, and Patient Age in Prostate Cancer 19 Fine-needle Aspiration of the Prostate 25 Application of Flow Cytometry and Automated Image Analysis to the Study of Prostatc Cancer 3] Noninvasive Imaging for Staging of Prostate Cancer: Magnetic Resonance Imaging, Computed Tomography, and Ultrasound 37 Lymphography in Clinically Localized Prostate Cancer 41 Value of and Indications for Pelvic Lymph Node Dissection in the Staging of Prostate Cancer II. Radiation Therapy 47 Status of Radiation Treatment of Prostate Cancer at Stanford University - 61 Radiation Therapy Oncology Group Studies in Carcinoma of the Prostate 67 Radiation Therapy for Localized Prostate Carcinoma: Experience at the Massachusetts General Hospital (1973-1981) 75 External-beam Radiation Therapy for Clinically Localized Prostate Cancer: Patterns of Care Studies in the United States 85 Definitive Radiation Therapy in Carcinoma of the Prostate Localized to the Pelvis: Experience at the Mallinckrodt Institute of Radiology NIH Publication No. 88-3005 Number 7 1988 95 Local Control of Prostate Cancer With Radiotherapy: Frequency and Prognostic Significance of Positive Results of Postirradiation Prostate Biopsy III. Surgery 107 Selection Criteria for Radical Pros:atectomy Based on Morphometric Studies in Prosta e Carcinoma 109 Bilateral Pelvic Lymphadenectomy and Radical Retropubic Prostatectomy for Stage C or D1 Adenocarcinoma of the Prostate: Possible Beneficial Effect of Adjuvant Treatment 117 Long-term Results of Radical Prostatectomy in Clinically Localized-Prostate Cancer: Experience at The Johns Hopkins Hospital 123 Total Prostatectomy for Clinically Localized Prostate Cancer: Long-term Surgical Results and Current Morbidity 127 Randomized Series of Treatment With Surgery Versus Radiation for Prostate Adenocarcinoma 133 Radical Retropubic Prostatectomy With Reduced Morbidity: An Anatomic Approach IV. Adjuvant Therapy 141 Radiation Therapy as Adjuvant Treatment Ai. =r Radical Prostatectomy 151 Chemotherapy for Prostate Carcinoma 165 Hormone Therapy for Prostate Canczr: Resulis of the Veterans Administration Cooperative Urological Research Group Studies 171 Hormonal Therapy for Locally Advanced Prostate Cancer U.C. BERKELEY LIBRARIES AERC CO00500407y