J Clin Oncol 23: by American Society of Clinical Oncology INTRODUCTION

Transcription

1 VOLUME 23 NUMBER 28 OCTOBER JOURNAL OF CLINICAL ONCOLOGY O R I G I N A L R E P O R T Predictors of Prostate Cancer Specific Mortality After Radical Prostatectomy or Radiation Therapy Ping Zhou, Ming-Hui Chen, David McLeod, Peter R. Carroll, Judd W. Moul, and Anthony V. D Amico From the Department of Radiation Oncology, Brigham and Women s Hospital and Dana-Farber Cancer Institute, Boston, MA; Department of Statistics, University of Connecticut, Storrs, CT; Department of Surgery and Urology Service, Center for Prostate Disease Research, Uniformed Services University and Walter Reed Army Medical Center, Bethesda, MD; Department of Urology, University of California, San Francisco, CA; and Department of Urology, Duke University, Durham, NC. Submitted January 21, 2005; accepted June 9, Supported by the Department of Defense Center for Prostate Disease Research funded by the US Army Medical Research and Materiel Command, Fort Detrick, MD. Authors disclosures of potential conflicts of interest are found at the end of this article. Address reprint requests to Ping Zhou, MD, PhD BROF, 375 Longwood Ave, Boston, MA 02215; pzhou@ partners.org by American Society of Clinical Oncology X/05/ /$20.00 DOI: /JCO A B S T R A C T Purpose We evaluated predictors of prostate cancer specific mortality (PCSM) after prostate-specific antigen (PSA) failure after radical prostatectomy (RP) or radiation therapy (RT). Patients and Methods A total of 1,159 men with clinically localized prostate cancer treated with RP (n 498) or RT (n 661) developed PSA failure, and they formed the study cohort. Competing risk regression analyses were used to evaluate whether previously identified predictors of time to metastasis, including post-treatment PSA doubling time (PSA-DT), Gleason score, and interval to PSA failure, could also predict time to PCSM after PSA failure. The cumulative incidence method was used to estimate PCSM after PSA failure. Results A post-rp PSA-DT of less than 3 months (hazard ratio [HR], 54.9; 95% CI, 16.7 to 180), a post-rt PSA-DT of less than 3 months (HR, 12.8; 95% CI, 7.0 to 23.1), and a biopsy Gleason score of 8 to 10 (HR, 6.1; 95% CI, 3.4 to 10.7) for patients treated with RT were significantly associated with PCSM. Post-RP estimated rates of PCSM 5 years after PSA failure were 31% (95% CI, 17% to 45%) v 1% (95% CI, 0% to 2%) for patients with PSA-DT of less than 3 months v 3 months. Post-RT estimated rates of PCSM 5 years after PSA failure were 75% (95% CI, 59% to 92%) v 35% (95% CI, 24% to 47%) for patients with a biopsy Gleason score of 8 v 7, respectively, and PSA-DT of less than 3 months; these rates were 15% (95% CI, 0.8% to 28%) v 4% (95% CI, 1% to 6%), respectively, for patients with a PSA-DT 3 months. Conclusion Patients at high risk for PCSM after PSA failure can be identified based on post-rp PSA-DT or post-rt PSA-DT and biopsy Gleason score. These parameters may be useful in identifying patients for a randomized trial evaluating hormonal therapy with or without docetaxel. J Clin Oncol 23: by American Society of Clinical Oncology INTRODUCTION Prostate-specific antigen (PSA) defined recurrence as many as 10 years after radical prostatectomy (RP) 1 or external-beam radiation therapy (RT) 2 for patients with clinically localized prostate cancer occurs in up to 30% of patients. However, only a small subset of these patients will develop clinical evidence of metastases, and even a smaller cohort will die of prostate cancer because of competing causes of mortality. 3 To identify patients for whom a PSAdefined recurrence is likely to translate into death from prostate cancer, investigators have tried to identify factors predictive of recurrence and prostate cancer specific mortality (PCSM) after PSA failure. Specifically, Pound et al 4 evaluated the natural history of progression to metastases after postoperative PSA failure and found that time to PSA relapse, Gleason score, and posttreatment PSA doubling time (PSA-DT) were predictive of the time to the development of 6992

2 Predictors of Prostate Cancer Mortality metastatic disease. Other investigators have also shown that PSA-DT after RP or RT is statistically significantly associated with the time to distant disease recurrence. 5-9 However, because of the protracted course of prostate cancer and competing causes of mortality in this patient population, 3 defining the determinants of PCSM after PSAdefined recurrence can be difficult A recent update of the Johns Hopkins University experience suggested that PSA-DT was significantly associated with PCSM after RP. 15 A multi-institutional study of patients treated with RP or RT demonstrated that a post-treatment PSA-DT of less than 3 months seemed to be a surrogate for PCSM after PSA-defined recurrence. 12 Specifically, men with a PSA-DT of less than 3 months had a median survival of 6 years after PSA failure. More recently, a multi-institutional study of RT-treated patients found that both a PSA-DT of less than 10 months and an interval to PSA failure of 2 years were significantly associated with distant metastasis-free survival after PSA failure. 10 In this study, we evaluated whether previously identified predictors of time to metastasis, including the posttreatment PSA-DT, Gleason score, and interval to PSA failure, could also predict time to PCSM after PSA failure. Identifying clinical factors associated with PCSM will serve as the basis for patient selection for randomized trials of hormonal therapy with or without additional systemic therapy in the setting of an increasing PSA after RP or RT. PATIENTS AND METHODS Patient Selection and Treatment Two multi-institutional databases, Cancer of the Prostate Strategic Urologic Research Endeavor 16 and the Center for Prostate Disease Research, 17 which contain baseline, treatment, and follow-up information on 8,669 patients treated with RP or RT between January 1, 1988, and January 1, 2002, for patients with clinical stage T1c-4 NX or N0 M0 (ie, localized or locally advanced, nonmetastatic) prostate cancer formed the database on which this study was based. One thousand one hundred fifty-nine patients, including 498 patients treated surgically and 661 patients treated with RT, developed PSA-defined recurrence and had complete information available for analysis; these patients formed the study cohort. An approved and signed internal review board informed consent form was obtained for each patient before study entry. Patients treated surgically were permitted to have received up to 3 months of neoadjuvant androgen suppression therapy (AST) because the 5-year results of a randomized trial 18 showed no statistically significant impact on PSA outcome when 3 months of neoadjuvant AST was added to RP. All RP-treated men had an undetectable PSA level ( 0.2 ng/ml) after surgery, and all RTtreated men had a PSA nadir before the subsequent increase. All patients in our study received hormonal therapy at some point after PSA-defined recurrence and before the bone scan was positive. The median PSA when hormonal therapy was initiated was 9.4 and 9.8 ng/ml for the RP and RT groups, respectively. The median age at presentation for the 498 patients treated with RP and the 661 patients treated with RT was 64.3 years (range, 34.3 to 96.8 years) and 71.1 years (range, 43.7 to 87.9 years), respectively. The pretreatment clinical characteristics of these 1,159 patients stratified by the treatment received are listed in Table 1. Staging and Follow-Up In all patients, staging evaluation involved a history and physical examination, including a digital rectal examination (DRE), serum PSA, and a transrectal ultrasound-guided needle biopsy of the prostate to determine Gleason score histologic grading. 19 The prostate biopsy was generally performed transrectally with an 18-gauge Tru-Cut needle (Travenol Laboratories, Deerfield, IL). Before 1996, patients generally had a computed tomography scan of the pelvis and a bone scan. After 1996, patients with PSA less than 10 ng/ml and a biopsy Gleason score 6 usually did not undergo radiographic staging because of the less than 1% chance that these studies would reveal metastatic disease. 20 The clinical stage was obtained from the DRE findings using the 2002 American Joint Committee on Cancer staging system. 21 Radiologic and biopsy information was not used to determine clinical stage. PSA levels were commonly measured with assays from Hybritech (San Diego, CA), Tosoh (Foster City, CA), or Abbott Lab (Chicago, IL). The median follow-up time for the 498 patients treated with RP and the 661 patients treated with RT was 6.8 years (range, 0.6 to 13.5 years) and 6.5 years (range, 1.0 to 14.3 years), respectively; follow-up started on the first day of treatment. The median Table 1. Pretreatment Clinical Characteristics of the 498 Surgically Treated and 661 Radiation-Treated Patients Who Developed PSA-Defined Recurrence Clinical Characteristic Surgery No. of Patients % Radiation No. of Patients % PSA level, ng/ml Median Biopsy Gleason score AJCC stage T1c T2a T2b T2c T3, Age, years Median Abbreviations: PSA, prostate-specific antigen; AJCC, American Joint Committee on Cancer

3 Zhou et al follow-up time for the 464 living patients at the end of follow-up in the RP group and the 552 living patients in the RT group was 6.9 years (range, 0.6 to 13.2 years) and 6.6 years (range, 1.0 to 14.3 years), respectively. Before PSA-defined recurrence, as specified by the American Society for Therapeutic Radiology and Oncology consensus criteria, 22 patients generally had a serum PSA level evaluation and a DRE every 3 months after RT for 2 years, then every 6 months for an additional 3 years, and then annually thereafter. The median follow-up time after PSA-defined recurrence for the 498 RP-treated patients and the 661 RT-treated patients was 4.0 years (range, 0.3 to 11.8 years) and 3.6 years (range, 0.02 to 12.0 years), respectively; the corresponding follow-up time for living patients was 4.0 years (range, 0.3 to 11.8 years) and 3.7 years (0.02 to 12.0 years), respectively. Overall, there were 102 prostate cancer specific deaths, including 25 in the RP group and 77 in the RT group. Determination of the cause of death was based on death certificate. Statistical Methods Calculation of the PSA-DT has been previously described. 12 PSA-DT was calculated by assuming first-order kinetics and by using a minimum of three PSA measurements closest in time to the initiation of hormonal therapy, each separated by a minimum of 3 months and each with a PSA increase of more than 0.2 ng/ml. By using PSA values closest in time to the start of hormonal therapy, the PSA-DT calculated should represent the worst-case scenario in the setting of an accelerating PSA increase or a shortening PSA-DT over time. For patients treated with RT, the nadir PSA level was subtracted from the post-rt PSA level before the PSA-DT was determined. This way, the magnitude of the PSA-DT would be the same for RP-treated patients and RT-treated patients who experienced the same absolute increase in PSA level. Competing risks methodology was used for the entire analysis of this article to take into account the competing causes of mortality. Competing risk regression analyses (based on Gray s R-package cmprsk test 23 ) were used to evaluate the ability of post-treatment PSA-DT, interval to PSA failure, and Gleason score to predict time to death from prostate cancer after PSA-defined recurrence in 1,159 patients treated with RP (498 patients) or RT (661 patients) who had experienced PSA failure. The biopsy Gleason score was used for men treated with RT, whereas the prostatectomy Gleason score was used for men treated with RP. Time 0 was considered the day of PSA-defined recurrence, which was defined as the midpoint between the PSA nadir and first increase. 22 The PSA-DT, interval to PSA failure, and Gleason score were treated first as continuous variables and then as categoric variables in separate competing risk regression analyses. The categories selected for the PSA-DT and interval to PSA failure were 3 months and 2 years, respectively. The categories selected for Gleason score were 6, 7, and 8 to 10. Baseline groups were defined as PSA-DT 3 months, interval to PSA failure more than 2 years, and Gleason score 6. These breakpoints for PSA-DT, interval to PSA failure, and Gleason score were suggested in previous studies to be clinically useful categories for predicting time to distant metastases and time to PCSM after PSA failure. 4,7,11,12,14 The hazard ratios (HRs) of PCSM for PSA-DT, interval to PSA failure, and Gleason score were calculated for the individual cohorts of men treated with RP or RT on the basis of the coefficients from the competing risk regression model and reported with 95% CIs. For the purpose of illustration, estimates of PCSM after PSA-defined recurrence were calculated using the cumulative incidence method 24 and were graphically displayed. Comparisons of the estimates of cumulative incidence of PCSM after PSA failure between groups were made using the K-sample tests to take into consideration the competing risks of mortality. 25 RESULTS Patient Characteristics One thousand one hundred fifty-nine patients developed PSA-defined recurrence, including 498 patients treated with RP and 661 patients treated with RT. The clinical characteristics of the 1,159 patients before RP or RT are listed in Table 1. Post-RP Predictors of PCSM When evaluated in univariable analyses, only PSA-DT (P.003) was a significant predictor of time to PCSM after Table 2. Predictors of Time to Prostate Cancer Specific Mortality After PSA-Defined Recurrence in Patients Treated With Surgery or Radiation Therapy Predictor UVA P Surgery Radiation MVA MVA HR 95% CI P UVA P HR 95% CI P Continuous variable model PSA-DT to to Gleason score to to Interval to PSA failure to to Categoric variable model PSA-DT 3 months to to Gleason score to to to to Interval to PSA failure 2 years to to Abbreviations: UVA, univariable analysis; MVA, multivariable analysis; HR, hazard ratio; PSA, prostate-specific antigen; DT, doubling time. Gleason score refers to prostatectomy or biopsy Gleason score for men treated surgically or with radiation therapy, respectively. In the categoric variable model, the baseline groups are PSA-DT 3 months, Gleason score 6, and interval to PSA failure 2 years JOURNAL OF CLINICAL ONCOLOGY

4 Predictors of Prostate Cancer Mortality PSA-defined recurrence (Table 2). Neither Gleason score nor interval to PSA failure was significantly associated with time to PCSM (Table 2). Similarly, on multivariable analyses, only PSA-DT as a continuous variable or a categoric variable was significant (Table 2). The HR of PCSM was 54.9 (95% CI, 16.7 to 180) for patients with a PSA-DT of less than 3 months compared with 3 months. The estimated rates of PCSM 5 years after PSA failure were 31% (95% CI, 17% to 45%) v 1% (95% CI, 0% to 2%) for patients with a PSA-DT of less than 3 months v 3 months, respectively. Of note, 13% of men treated surgically who experienced a PSA-defined recurrence in this study had a PSA-DT of less than 3 months. A significant association with PCSM was not noted for RP Gleason score (P.16) or interval to PSA failure (P.22), as illustrated in Figures 1 and 2, respectively, and summarized in Table 2. Post-RT Predictors of PCSM When evaluated in univariable analyses, all three predictors, including PSA-DT (P.002), biopsy Gleason score (P.0001), and interval to PSA failure (P.004), were significantly associated with time to PCSM after PSA failure (Table 2). On multivariable analysis, PSA-DT as a continuous and categoric variable was a significant independent predictor of time to PCSM, resulting in an HR of 12.8 (95% CI, 7.0 to 23.1) for patients with a PSA-DT of less than 3 months compared with 3 months (Table 2). In contrast to the findings in patients treated with RP, biopsy Gleason score as a continuous and categoric variable (Gleason score Fig 2. Cumulative incidence estimates of prostate cancer specific mortality after radical prostatectomy stratified by prostate-specific antigen (PSA) doubling time (PSA-DT) and interval to PSA failure. Pairwise, two-sided, K-sample test showed P values as follows: for PSA-DT of less than 3 months (interval to PSA failure 2 v 2 years), P.95; for PSA-DT of 3 months (interval 2 v 2 years), P.41. of8to10v 6) was significantly associated with PCSM in men treated with RT (Table 2 and Fig 3). The HR of PCSM was 6.1 (95% CI, 3.4 to 10.7) for patients with a biopsy Gleason score of 8 to 10 compared with 6. The estimated rates of PCSM 5 years after PSA failure were 75% (95% CI, Fig 1. Cumulative incidence estimates of prostate cancer specific mortality after radical prostatectomy stratified by prostate-specific antigen (PSA) doubling time (PSA-DT) and prostatectomy Gleason score. Pairwise, twosided, K-sample test showed P values as follows: for PSA-DT of less than 3 months (prostatectomy Gleason score 8 v 7), P.37; for PSA-DT of 3 months (prostatectomy Gleason score 8 v 7), P.58. Fig 3. Cumulative incidence estimates of prostate cancer specific mortality after radiation therapy stratified by prostate-specific antigen (PSA) doubling time (PSA-DT) and biopsy Gleason score. Pairwise, two-sided, K-sample test showed P values as follows: for PSA-DT of less than 3 months (biopsy score 8 v 7), P.0001; for PSA-DT of 3 months (biopsy Gleason score 8 v 7), P

5 Zhou et al 59% to 92%) v 35% (95% CI, 24% to 47%) for patients with a biopsy Gleason score of 8 v 7, respectively, and a PSA-DT of less than 3 months; these estimates were 15% (95% CI, 0.8% to 28%) v 4% (95% CI, 1% to 6%), respectively, for patients with a PSA-DT of 3 months. Of note, 21% of men treated with RT in this study who experienced a PSAdefined recurrence had a PSA-DT of less than 3 months. In addition, 9% of men had a PSA-DT of 3 months and a biopsy Gleason score of 8. A significant association between PCSM and interval to PSA failure was not noted (P.94), as illustrated in Figure 4. DISCUSSION An increasing PSA level after RP 1 or external-beam RT 2 for men with clinically localized prostate cancer is a significant clinical problem in the management of prostate cancer. Because of the heterogeneity of this patient population with regard to the risk of dying from prostate cancer as opposed to competing causes of mortality, it is important to identify the group of patients at high risk of PCSM because these are the patients who need effective systemic treatment beyond hormonal therapy. Therefore, such patients would be ideally suited for a randomized study comparing the use of AST alone with AST plus a chemotherapy that targets hormone-refractory prostate cancer. Consistent with the published studies, the results of this study demonstrated that the post-rp and post-rt Fig 4. Cumulative incidence estimates of prostate cancer specific mortality after radiation therapy stratified by prostate-specific antigen (PSA) doubling time (PSA-DT) and interval to PSA failure. Pairwise, two-sided, K-sample test showed P values as follows: for PSA-DT of less than 3 months (interval to PSA failure 2 v 2 years), P.12; for PSA-DT of 3 months (interval 2 v 2 years), P.31. PSA-DT was significantly associated with the time to PCSM after PSA failure. In addition, biopsy Gleason score was an independent predictor of time to PCSM in men treated with RT. Of note, competing risks methodology was used for the entire statistical analysis of this article to take into account the competing causes of mortality. Of clinical importance, 13% of RP-treated men in this study who experienced a PSA-defined recurrence had a PSA-DT less than 3 months. In addition, 30% of RT-treated men with PSA failure in this study had a PSA-DT of less than 3 months or a combination of PSA-DT of 3 months and a biopsy Gleason score 8. These patients were at high risk for PCSM after PSA failure. The rapid increase in PSA level after primary local therapy suggests that the recurrent and/or residual prostate cancer cells responsible for the PSA increase are, in part, androgen independent. This is suggested by the high estimates of PCSM 5 years after PSA failure in patients with a post-treatment PSA-DT of less than 3 months or a biopsy Gleason score 8 in the case of RT-treated patients (Figs 1 to 4) despite the use of hormonal therapy initiated after PSA failure. The role of chemotherapy in treating hormonerefractory prostate cancer has been evaluated in several randomized clinical trials, and only two recent phase III studies have shown significant survival benefit with the addition of chemotherapy. 30,31 Both studies used docetaxelbased regimens administered together with prednisone 31 or estramustine 30 and compared the effects to mitoxantrone and prednisone, which is an accepted standard therapy in treating hormone-refractory prostate cancer. These studies demonstrated that the regimen including docetaxel was associated with significant improvement in median survival, PSA response, and quality of life compared with the standard treatment. Therefore, docetaxel-based chemotherapy could be the therapy of choice to study in a randomized clinical trial comparing AST alone with AST plus chemotherapy in a group of men at high risk of PCSM after PSA failure after RP or RT. The current study has demonstrated that up to 30% of RT-treated patients and up to 13% of RP-treated patients who experience PSA failure are at increased risk of PCSM. Therefore, this group of patients may be the optimal patient population in a proposed randomized study comparing AST with or without docetaxelbased chemotherapy. Potential limitations of this study need to be considered. First, the exact timing of salvage hormonal therapy was not considered in this study. Specifically, we calculated PSA-DT based on at least the last three PSA values closest in time to the initiation of hormonal therapy, which should represent the worst-case scenario after PSA failure in the setting of an accelerating PSA increase or a shortening PSA-DT. Therefore, the results of this study apply to patients in whom hormonal therapy is initiated at a PSA level of approximately 10 ng/ml after PSA-defined recurrence and with a negative 6996 JOURNAL OF CLINICAL ONCOLOGY

6 Predictors of Prostate Cancer Mortality bone scan. If administering salvage hormonal therapy at the time of PSA failure, as opposed to later, is shown to prolong survival, then the predictors of PCSM determined in this report would need to be re-evaluated in a study in which all men received salvage hormonal therapy at the time of PSA failure. Second, the median follow-up time of this study was almost 7 years, which is still relatively short given the long natural history of prostate cancer. Perhaps with longer follow-up, Gleason score 7 disease and longer PSA-DT of possibly up to 6 months would also be associated with increased risk of PCSM. Third, the lack of significance of a prostatectomy Gleason score of 8 to 10 to predict PCSM in surgically managed patients may be explained by the lack of power because only 25 patients died of prostate cancer in the surgical group. Therefore, with a larger surgical cohort and longer follow-up, prostatectomy Gleason score might also become significant for PCSM after PSA failure. Fourth, the current database does not contain baseline information on the testosterone levels of the patients. It has been recognized that testosterone can be low in approximately 15% of men who are beginning hormonal therapy 32,33 ; thus, some men may have been functionally castrated before all therapy. As a result, PSA failure in these men may effectively be a hormone-refractory relapse. Whether adjuvant chemotherapy will prolong survival in these men is unclear. In conclusion, patients with clinically localized prostate cancer who are at high risk for PCSM after post-rp or post-rt PSA failure have been identified as those men with a post-rp or post-rt PSA-DT of less than 3 months or, in the case of RT-treated patients, a biopsy Gleason score of 8. Enrollment of these men onto a randomized clinical trial evaluating AST with or without docetaxel in the setting of an increasing PSA level after RP or RT should be considered. Authors Disclosures of Potential Conflicts of Interest The authors indicated no potential conflicts of interest. REFERENCES 1. Han M, Partin AW, Zahurak M, et al: Biochemical (prostate specific antigen) recurrence probability following radical prostatectomy for clinically localized prostate cancer. J Urol 169: , Shipley WU, Thames HD, Sandler HM, et al: Radiation therapy for clinically localized prostate cancer: A multi-institutional pooled analysis. JAMA 281: , Albertsen PC, Hanley JA, Gleason DF, et al: Competing risk analysis of men aged 55 to 74 years at diagnosis managed conservatively for clinically localized prostate cancer. JAMA 280: , Pound CR, Partin AW, Eisenberger MA, et al: Natural history of progression after PSA elevation following radical prostatectomy. JAMA 281: , Patel A, Dorey F, Franklin J, et al: Recurrence patterns after radical retropubic prostatectomy: Clinical usefulness of prostate specific antigen doubling times and log slope prostate specific antigen. J Urol 158: , Ward JF, Blute ML, Slezak J, et al: The long-term clinical impact of biochemical recurrence of prostate cancer 5 or more years after radical prostatectomy. J Urol 170: , Lee WR, Hanks GE, Hanlon A: Increasing prostate-specific antigen profile following definitive radiation therapy for localized prostate cancer: Clinical observations. J Clin Oncol 15: , Sartor CI, Strawderman MH, Lin XH, et al: Rate of PSA rise predicts metastatic versus local recurrence after definitive radiotherapy. Int J Radiat Oncol Biol Phys 38: , Hanlon AL, Diratzouian H, Hanks GE: Posttreatment prostate-specific antigen nadir highly predictive of distant failure and death from prostate cancer. Int J Radiat Oncol Biol Phys 53: , Kuban DA, Thames H, Horwitz E, et al: Predicting outcome after PSA failure in prostate cancer patients treated by radiation: Who needs salvage therapy. Int J Radiat Oncol Biol Phys 60:S167, 2004 (suppl 1) 11. Sandler HM, Dunn RL, McLaughlin PW, et al: Overall survival after prostate-specificantigen-detected recurrence following conformal radiation therapy. Int J Radiat Oncol Biol Phys 48: , D Amico AV, Moul JW, Carroll PR, et al: Surrogate end point for prostate cancer-specific mortality after radical prostatectomy or radiation therapy. J Natl Cancer Inst 95: , Albertsen PC, Hanley JA, Penson DF, et al: Validation of increasing prostate specific antigen as a predictor of prostate cancer death after treatment of localized prostate cancer with surgery or radiation. J Urol 171: , D Amico AV, Cote K, Loffredo M, et al: Determinants of prostate cancer-specific survival after radiation therapy for patients with clinically localized prostate cancer. J Clin Oncol 20: , Partin AW, Eisenberger MA, Sinibaldi VJ, et al: Prostate specific antigen doubling time (PSADT) predicts for distant failure and prostate cancer specific survival (PCSS) in men with biochemical relapse after radical prostatectomy (RP). Presented at the Annual Meeting of American Urologic Association, San Francisco, CA, May 8-13, Lubeck DP, Litwin MS, Henning JM, et al: The CaPSURE database: A methodology for clinical practice and research in prostate cancer CaPSURE Research Panel Cancer of the Prostate Strategic Urologic Research Endeavor. Urology 48: , Sun L, Gancarczy K, Paquette EL, et al: Introduction to the Department of Defense Center for Prostate Disease Research Multicenter National Prostate Cancer Database, and analysis of changes in PSA-era. Urol Oncol 6: , Soloway MS, Pareek K, Sharifi R, et al: Neoadjuvant androgen ablation before radical prostatectomy in ct2bnxmo prostate cancer: 5-year results. J Urol 167: , The Veterans Administration Cooperative Urological Research Group, Gleason DF: Histologic grading and staging of prostatic carcinoma, in Tannenbaum M (ed): Urologic Pathology. Philadelphia, PA, Lea & Febiger, 1977, pp Lee CT, Oesterling JE: Using prostatespecific antigen to eliminate the staging radionuclide bone scan. Urol Clin North Am 24: , Greene FL, Page DL, Fleming ID, et al: American Joint Committee on Cancer Manual for Staging Cancer (ed 6). New York, NY, Springer- Verlag, Cox JD: Consensus statement: Guidelines for PSA following radiation therapy American Society for Therapeutic Radiology and Oncology Consensus Panel. Int J Radiat Oncol Biol Phys 37: , Fine JP, Gray RJ: A proportional hazards model for the sub-distribution of a competing risk. J Am Stat Assoc 94: , Gaynor JJ, Feur EJ, Tan CC, et al: On the use of cause-specific failure and conditional failure probabilities: Examples from clinical oncology data. J Am Stat Assoc 88: , Gray RJ: A class of k-sample tests for comparing the cumulative incidence of a competing risk. Ann Stat 16: , Berry W, Dakhil S, Modiano M, et al: Phase III study of mitoxantrone plus low dose prednisone versus low dose prednisone alone in patients with asymptomatic hormone refractory prostate cancer. J Urol 168: , Ernst DS, Tannock IF, Winquist EW, et al: Randomized, double-blind, controlled trial of mitoxantrone/prednisone and clodronate versus mitoxantrone/ prednisone and placebo in patients with hormone refractory prostate cancer and pain. J Clin Oncol 21: , Kantoff PW, Halabi S, Conaway M, et al: Hydrocortisone with or without mitoxantrone in men with hormone-refractory prostate cancer:

7 Zhou et al Results of the Cancer and Leukemia Group B 9182 study. J Clin Oncol 17: , Tannock IF, Osoba D, Stockler MR, et al: Chemotherapy with mitoxantrone plus prednisone or prednisone alone for symptomatic hormone-resistant prostate cancer: A Canadian randomized trial with palliative end points. J Clin Oncol 14: , Petrylak DP, Tangen CM, Hussain MH, et al: Docetaxel and estramustine compared with mitoxantrone and prednisone for advanced refractory prostate cancer. N Engl J Med 351: , Tannock IF, de Wit R, Berry WR, et al: Docetaxel plus prednisone or mitoxantrone plus prednisone for advanced prostate cancer. N Engl J Med 351: , Harper ME, Pierrepoint CG, Griffiths K: Carcinoma of the prostate: Relationship of pretreatment hormone levels to survival. Eur J Cancer Clin Oncol 20: , Zhang PL, Rosen S, Veeramachaneni R, et al: Association between prostate cancer and serum testosterone levels. Prostate 53: , JOURNAL OF CLINICAL ONCOLOGY