© The Author 2006. Published by Oxford University Press.
ARTICLE |
Evaluation of Prostate-Specific Antigen Declines for Surrogacy in Patients Treated on SWOG 99-16
Affiliations of authors: Columbia University, New York, NY (DPP, MCB); Southwest Oncology Group Statistical Center, Seattle, WA (DPA, CSJ, CMT); University of Michigan Comprehensive Cancer Center, Ann Arbor, MI (MHAH); University of California–Davis, Sacramento, CA (PNL); Baylor College of Medicine, Houston, TX (JAJ); University of Massachusetts Medical Center, Worcester, MA (MET); Mayo Clinic, Rochester, MN (PAB); University of Arkansas for Medical Science, Little Rock, AR (MK); University of California San Francisco Cancer Center, San Francisco, CA (EJS); Cleveland Clinic Foundation, Cleveland, OH (DR); University of Colorado Health Science Center, Denver, CO (EDC)
Correspondence to: Daniel P. Petrylak, MD, Columbia University, 161 Ft. Washington Ave., Rm. 919, New York, NY 10032 (e-mail: dpp5{at}columbia.edu).
 |
ABSTRACT |
|---|
 TopNotes Abstract IntroductionSubjects and methodsResultsDiscussionReferences |
|---|
Background: The identification of surrogate endpoints that can replace true outcome endpoints is crucial to the rapid evaluation of new cancer drugs. Retrospective analyses of phase II and III trials in metastatic androgen-independent prostate cancer have shown associations between declines in serum prostate-specific antigen (PSA) levels and survival. We evaluated PSA changes as potential surrogate markers for survival by using data from a clinical trial. Methods: Men with androgen-independent prostate cancer were randomly assigned to either docetaxel/estramustine (D/E) or mitoxantrone/prednisone (M/P) treatment on Southwest Oncology Group Protocol 99-16. Of 674 eligible patients, 551 had a baseline PSA measurement and at least one PSA measurement during the first 3 months on protocol. PSA level declines of 5%–90% and PSA velocity at 1, 2, and 3 months were tested for surrogacy by using three statistical criteria: Prentice's criteria, the proportion of treatment effect explained, and the proportion of variation explained. All statistical tests were two-sided. Results: Three-month PSA level declines of 20%–40%, a 2-month PSA decline of 30%, and PSA velocity at 2 and 3 months met all three surrogacy criteria. For example, a 3-month PSA decline of at least 30% was associated with a more than 50% decrease in the risk of death compared with the lack of such a decline (hazard ratio [HR]= 0.43, 95% confidence interval [CI] = 0.34 to 0.55; P<.001), and the increased risk of death for men treated with M/P compared with D/E (HR = 1.24, 95% CI = 1.02 to 1.51; P = .032) lost statistical significance after adjustment for this surrogate, whereas the decrease in risk of death associated with a 3-month 30% PSA decline remained statistically significant after adjustment for treatment. PSA level declines of 50%, commonly reported in clinical trials, did not meet the criteria for surrogacy. Conclusions: Several PSA measures satisfied the surrogacy criteria for survival in a retrospective analysis of data from SWOG 99-16. However, these measures await prospective validation in future clinical trials of chemotherapy in men with androgen-independent prostate cancer.
|
INTRODUCTION |
|---|
|
TopNotesAbstractIntroductionSubjects and methodsResultsDiscussionReferences |
|---|
Phase III studies of novel chemotherapeutic agents for men with androgen-independent prostate cancer that use survival as a primary endpoint can take years to produce mature results. A prospectively validated surrogate endpoint could accelerate the evaluation and potential approval of new drugs for this disease. Prostate-specific antigen (PSA), a kallikrein-like serine protease, is a reliable serum marker for detection of relapse following androgen blockade (1). In retrospective studies, declines in PSA level of 50% or greater have been associated with improved survival in men treated for hormone-refractory prostate cancer. However, controversy exists about the optimal extent of PSA level decline relative to baseline PSA, the optimal time for PSA evaluation relative to the initiation of treatment, and the optimal duration of the PSA decline that are necessary to define a valid surrogate endpoint (2,3). A consensus conference organized by the Medicine Branch of the National Cancer Institute focused on PSA declines of at least 50% as a measure to report antitumor activity. However, the conclusions were meant to serve only as a guide to common terminology for reporting potential antitumor activity for confirmation in future phase II studies, not to address surrogacy (4).
Operational criteria suggested by Prentice (5) require that the endpoint be associated with clinical outcome and fully capture the net effect of treatment on clinical outcome. In addition, they can be established only on a dataset that has a statistically significant effect of treatment on survival in the first place. The randomized controlled clinical trial remains the best mechanism to ensure that a treatment effect is due to treatment and not to other confounding variables. Southwest Oncology Group (SWOG) 99-16, an Intergroup randomized trial, demonstrated a 20% improvement in survival in patients treated with docetaxel and estramustine (D/E) compared with patients treated with mitoxantrone and prednisone (M/P) (6). In that trial, PSA level declines of more than 50% were observed in 50% and 27% of patients in the D/E and M/P arms, respectively. Because statistically significant differences in rates of PSA declines and survival were observed between the two arms, SWOG 99-16 provides an ideal dataset in which to evaluate whether PSA declines after the initiation of therapy could serve as surrogate endpoints for survival in clinical trials for men with androgen-independent prostate cancer. To this end, we evaluated several PSA changes, including time-dependent changes, as possible surrogate markers for survival in the context of SWOG 99-16.
|
SUBJECTS AND METHODS |
|---|
|
TopNotesAbstractIntroductionSubjects and methodsResultsDiscussionReferences |
|---|
Patient Selection, Treatment, Evaluation, and Follow-up
To be eligible for SWOG 99-16, patients had to have been diagnosed with pathologically confirmed adenocarcinoma of the prostate, stage N1 or M1 (7), and have progressive disease after surgical or medical castration. Progressive disease was defined by at least one of the following: 1) progression of a bidimensionally measurable lesion assessed within 28 days of study registration; 2) progression of evaluable but not measurable disease (e.g., bone scan) assessed within 42 days of registration; or 3) rising serum PSA level, with at least two consecutive increasing PSA measurements over baseline obtained over at least 7-day intervals. All patients provided written informed consent, and the protocol was approved by institutional review boards of each participating institution in SWOG, Cancer and Leukemia Group B (CALGB), North Central Cancer Treatment Group (NCCTG), the Clinical Trials Support Unit (CTSU), and the extended participation project (EPP) program through the National Cancer Institute. A total of 770 patients were accrued from October 1999 through January 2003, of whom 674 were eligible. The primary study endpoints have been reported elsewhere (6).
Participants in SWOG 99-16 were randomly assigned to the D/E arm (estramustine, 280 mg orally three times per day, 1 hour before or 2 hours after meals, on days 1–5; docetaxel 60 mg/m2 intravenously on day 2 every 21 days; and decadron 60 mg orally in three divided doses, starting the night before docetaxel) or the M/P arm (mitoxantrone, 12 mg/m2 intravenously, combined with prednisone 5 mg given orally twice a day). Treatment continued until disease progression or toxic effects occurred or until a maximum of 12 cycles of docetaxel/estramustine or 144 mg/m2 of mitoxantrone had been administered. Patients provided a complete medical history at baseline and were requested to have a physical examination and PSA measurement every 3 weeks while in the study. All patients were required to have a PSA measurement within 28 days of study entry, and the PSA measure most recent to but before the beginning of treatment was taken as the first PSA measure for analysis, the baseline PSA measure. In addition, patients were required to have at least one PSA measurement after the baseline PSA but during the first 3 months of the study to be eligible for this analysis. Patients who died during the first 3 months were also included in the analysis so long as they had a baseline PSA measurement and at least one additional PSA measurement before death. There were 15 such participants, seven in the D/E arm and eight in the M/P arm; exclusion of these participants did not change the results (data not shown). All patients included in this analysis were followed up for a maximum of 3 years.
Analysis of PSA Level Declines
A PSA level decline of 50% over 3 months was determined to have occurred if the lowest PSA measurement within the first 3 months was less than or equal to 50% of the baseline value. In addition to a 50% decline, PSA declines ranging from 5% to 90% were similarly evaluated for surrogacy. PSA velocity, i.e., the rate of change in PSA level, during the first 3 months of the study was computed by means of linear regressions (for patients with two or more PSA measurements in addition to the baseline measurement) and by the ratio of change in the logarithm of PSA by change in time (for patients with only one PSA value in addition to the baseline measurement). The linear regression fit a line through the logarithm of the PSA value on the y-axis and the time since baseline on the x-axis, with the baseline PSA value included at time zero. The resulting slope is the PSA velocity and is evaluated as a continuous surrogate measure. PSA declines and velocities were also computed for periods of 1 month and 2 months.
Statistical Criteria for Surrogacy
Three statistical criteria were required for a change in PSA level to qualify as a surrogate endpoint. These included the criteria suggested by Prentice (5); the proportion of treatment effect explained (PTE) (8,9), which is the most stringent criterion; and the proportion of variation explained (R2) (10).
To investigate Prentice's operational criteria for surrogacy, we first used logistic regression (for PSA decreases) and linear regression (for PSA velocity) to confirm that there was a statistically significant effect of treatment (at the P = .05 level) on the surrogate measure (data not shown). We then carried out a series of three hypothesis tests based on Cox proportional hazards regression models for survival; the assumptions underlying the Cox model were evaluated by using martingale residuals. These tests are as follows. Test 1: survival = treatment indicator (indicator for experimental arm); test 2: survival = PSA surrogate candidate; test 3: survival = treatment indicator + PSA surrogate candidate. Surrogacy is suggested if the treatment and surrogate effects are statistically significant (P<.05) in tests 1 and 2, respectively, and if the surrogate remains statistically significant but the treatment is no longer statistically significant in test 3.
The second surrogacy statistical criterion, the PTE, was calculated as 1 minus the ratio of the hazard ratio (HR) for treatment in test 3 to that for treatment in test 1. The PTE gives a measure of degree of surrogacy, with 1 indicating a perfect surrogate and 0 a failed surrogate (8,9). Confidence intervals for the PTE were calculated both by the delta method of Lin et al. (8) and by the nonparametric bootstrap (11). A strong degree of surrogacy is suggested if the lower bound of the 95% confidence interval (CI) exceeds 0.5.
The third statistical criterion, the R2 measure of proportion of variation explained, was calculated by the method of O'Quigley, using weighted sums of squared Schoenfeld residuals (10). Confidence intervals for R2 and comparisons of R2 from different surrogates were calculated with the nonparametric bootstrap (11). Because it is only an exploratory measure, the R2 statistic is checked for increases from the first to third Cox regressions of the Prentice criterion.
All regression models were adjusted for baseline PSA level and for whether a previous prostatectomy had been performed. All statistical tests were two-sided.
|
RESULTS |
|---|
|
TopNotesAbstractIntroductionSubjects and methodsResultsDiscussionReferences |
|---|
Patient Characteristics
Of the 674 eligible patients from SWOG 99-16, 551 (82%), 532 (79%), and 456 (68%) had at least one PSA measurement recorded after baseline to allow surrogacy analysis at 3 months, 2 months, and 1 month, respectively. (The sample size decreased as the time interval decreased because there was less time to have multiple PSA measures.) The baseline characteristics of the 551 patients eligible for the 3-month PSA surrogacy analysis (Table 1) did not differ statistically significantly from those of the 674 patients included in the final survival analysis of SWOG 99-16 (all P>.05). Among the 551 patients included in the 3-month PSA surrogacy analysis, there were also no statistically significant differences between the two treatment arms for any of the baseline characteristics (all P>.05). Also, the median survival by treatment arm and the reduction in the risk of death in the D/E arm relative to the M/P arm for these patients (Fig. 1, A) were also similar to those of all patients included in the final report of SWOG 99-16 (6). Table 2 summarizes the observed PSA dynamics for the two treatment arms. Patients on the D/E arm were more likely to exhibit a 50% decrease in PSA level than patients on the M/P arm, and their PSA velocities indicated bigger declines during each period.
|
|
|
Criteria for Surrogacy for all PSA Level Declines Over 3 Months
PSA level declines of 5%–90% during the first 3 months on treatment were evaluated for the three criteria for surrogacy. Table 3 shows results for the most stringent criterion, the PTE. The point estimate for the PTE was 1.0 for all decreases of 5%–60%, but only decreases of 20%–40% achieved a PTE with sufficient precision to be considered a surrogate—that is, only for these decreases did the lower bound of the 95% confidence interval exceed 0.5. The 30% PSA decline had the optimal PTE in terms of having the largest lower confidence bound (95% CI = 0.73 to 1.0). The commonly reported 50% decline just missed the PTE criterion for surrogacy, with a lower bound of the 95% CI equal to 0.45. The Prentice and R2 criteria for surrogacy were also met for all 3-month PSA declines between 20% and 40%.
|
Criteria for Surrogacy for the Optimal 30% PSA Level Decline Over 3 Months
The 3-month 30% PSA decline that was identified as optimal by virtue of best satisfying the PTE criterion was attained by 76% of patients in the D/E arm but only 40% of patients in the M/P arm (P<.001), consistent with the superior performance of D/E in terms of the survival endpoint. For the Prentice surrogacy criterion, analysis of survival by whether or not a 30% decline in PSA was achieved (Fig. 1, B) resulted in a greater separation of survival curves than analysis by treatment (Fig. 1, A). In Cox proportional hazards analysis (Table 4), the treatment effect in favor of D/E, which was initially statistically significant (hazard ratio [HR] of death in the M/P arm compared with the D/E arm = 1.24, 95% CI = 1.02 to 1.51; P = .032), lost statistical significance after adjustment for the PSA surrogate (HR for risk of death on M/P compared with D/E adjusted for 3-month 30% PSA decline = 0.86, 95% CI = 0.69 to 1.07; P = .16). However, the association of a 30% or greater PSA decline with improved survival (HR of death for 3-month PSA decline compared with no such decline = 0.43, 95% CI = 0.34 to 0.55; P<.001), which was also initially highly statistically significant, remained statistically significant and of the same magnitude after adjustment for treatment (adjusted HR = 0.40, 95% CI = 0.32 to 0.50). The R2 criterion increased from .03 for test 1 of the Prentice criterion to .14 for test 2 and .15 for test 3, satisfying the third surrogacy criterion (Table 4). This series of findings thus indicates that the effect of treatment is modulated through the PSA surrogate.
|
Criteria for Surrogacy by 30% and 50% PSA Declines Over 1 and 2 Months
We investigated whether the optimal 30% PSA decline and the commonly used 50% PSA decline would satisfy the conditions for surrogacy at 2 months or even 1 month on treatment. The 30% PSA level decline satisfied the three criteria for surrogacy at 2 months (e.g., PTE = 1.0, 95% CI = .62 to 1.0) but not at 1 month (PTE = 0.50, 95% CI = 0.18 to 1.0). The 50% PSA level decline failed the PTE criterion at both 2 months and 1 month, with PTEs of 0.64 (95% CI = 0.26 to 1.0) and 0.15 (95% CI = 0 to 0.69), respectively.
Criteria for Surrogacy for PSA Velocity
PSA velocity at both 3 months and 2 months also satisfied the three criteria for surrogacy. For example, for PSA velocity at 3 months, the PTE was 1.0 (95% CI = 0.67 to 1.0) (Table 3), and for PSA velocity at 2 months the PTE was 1.0 (95% CI = 0.56 to 1.0). However, for PSA velocity at 1 month, the PTE was 0.54 (95% CI = .21 to 1.0) and did not satisfy this criterion for surrogacy.
|
DISCUSSION |
|---|
|
TopNotesAbstractIntroductionSubjects and methodsResultsDiscussionReferences |
|---|
Our analysis of the PSA data in patients entered in SWOG 99-16 demonstrates that PSA level declines of at least 20%–40% and PSA velocity at 3 months after beginning treatment satisfied the stringent requirements of statistical surrogacy for survival. The optimal biochemical surrogate found in this study was a 30% PSA decline 3 months after treatment initiation. However, because this surrogate was identified through a data-driven approach, it must be validated in independent prospective trials. A statistical surrogate does not constitute a clinical surrogate, which must undergo evaluation in relation to other clinical benefit parameters by clinicians and regulatory agencies before it can be used in trials.
The most commonly reported percentage PSA decline in clinical trials, a decline of greater than 50%, failed—albeit barely—the most stringent surrogacy criterion, the PTE. At first glance, this finding appears paradoxical; it appears intuitively obvious that greater PSA declines should be consistent with a stronger antitumor response. However, the inability to establish surrogacy at higher rates of PSA decline may reflect the fact that fewer patients achieved higher-percentage PSA declines, leading to a reduction in statistical power. For example, 76% and 40% of patients in the D/E and M/P arms, respectively, had PSA declines of at least 30%, whereas 61% and 28%, respectively, had PSA declines of at least 50% and only 15% and 7%, respectively, had declines of at least 90%. Although random chance may explain why the 50% decrease barely missed reaching statistical surrogacy in SWOG 99-16, the optimal percentage decrease (i.e., 30% decline within 3 months) is statistically unimodal (the lower bound of the 95% CI for the PTE increases monotonically from declines of 5% to 30%, reaches a peak at 30%, and then decreases monotonically) and well identified, and increasing the percentage decrease will not necessarily increase the likelihood of achieving surrogacy in other trials. Nevertheless, a 50% PSA level decline should not be ruled out for validation in future clinical trials.
Other Phase III studies have been analyzed to study the relationship between prognosis and PSA declines. In an analysis of PSA declines in patients treated on a randomized trial comparing mitoxantrone and prednisone with prednisone alone, Dowling et al. (12) found that those who achieved a PSA decline of at least 50% had better survival than those who did not, regardless of treatment. Similar observations were made in a CALGB study comparing mitoxantrone and hydrocortisone to hydrocortisone (13), as well as in a randomized trial comparing suramin/prednisone to prednisone in men with hormone-refractory prostate cancer (14). However, data from these trials could not be used to establish surrogacy because there were no statistically significant survival differences between treatment arms, as required by Prentice's criteria.
By contrast, a relationship between PSA declines and survival was not observed in the TAX 327 trial, in which 1006 men with androgen-independent prostate cancer were randomly assigned to receive docetaxel/prednisone, administered either every 3 weeks or weekly, or mitoxantrone plus prednisone (15). PSA declines of at least 50% were observed in 45% and 48% of patients treated with every-3-week or weekly docetaxel/prednisone, respectively; however, only those treated with docetaxel/prednisone every 3 weeks had better survival than patients treated with mitoxantrone and prednisone. An analysis of 50% PSA declines in patients treated on TAX 327 found that the 50% decline did not meet the Prentice criteria. It will be interesting to see whether the model derived from SWOG 99-16 is valid in a more rigorous retrospective analysis of data from this trial. Prospective validation of the PSA surrogacy model using 30% declines during the first 3 months of treatment has been incorporated in the next randomized trial in SWOG, which will compare docetaxel/prednisone with docetaxel/prednisone combined with atresentan, a novel orally administered entothelin 1 antagonist, in men with androgen-independent prostate cancer. The CALGB is also planning to collect PSA data to evaluate PSA level declines as a surrogate in a randomized trial comparing docetaxel/prednisone to docetaxel/ prednisone combined with avastin, an antibody against vascular endothelial growth factor, in men with androgen-independent prostate cancer.
The validation of PSA declines as surrogate endpoints in future trials has the potential to substantially alter the design and duration of future phase III studies. In SWOG 99-16, for example, accrual occurred over 3 years, and the analysis required 1 year of follow-up to achieve 80% power to detect the desired survival improvement. If PSA declines had been used as a primary endpoint in SWOG 99-16, events in the course of the clinical trial would have occurred sooner, in this case no longer than 3 months after the start of treatment. Thus, by essentially equating the duration of the trial to that of enrollment, the use of the PSA surrogate would have decreased the time of the study by several years. Even in trials in which survival remains the primary endpoint, interim analyses could be informed by PSA declines as a surrogate, and such trials would benefit from the increased power to define the activity of a treatment earlier.
The definition of a surrogate is treatment specific, and certain PSA declines as well as PSA velocity, which served as a surrogate in SWOG 99-16, may be applicable only for treatments that induce cell death. Percent declines in PSA during the first 3 months of treatment may not be applicable as valid surrogates for patients treated with cytostatic drugs rather than cytotoxic drugs. That is, one would expect PSA stabilization rather than declines for a cytostatic drug, and in this context smaller changes in PSA or stabilization of PSA over longer periods may be more likely to be observed than rapid, large declines (16). Thus, it may be necessary to perform a similar analysis to that performed on the results from SWOG 99-16 for separate drug classes.
|
NOTES |
|---|
|
TopNotesAbstractIntroductionSubjects and methodsResultsDiscussionReferences |
|---|
Dr. D. Petrylak is a member of the speaker's bureau for Sanofi-Aventis, Dr. M. Hussain holds stock in Sanofi-Aventis, and Dr. Primo N. Lara Jr. is a member of the speaker's bureau and of the advisory board of Sanofi-Aventis.
Supported by the following PHS Cooperative Agreement grant numbers awarded by the National Cancer Institute, Department of Health and Human Services: CA38926, CA32102, CA37135, CA25224, CA46441, CA37981, CA45808, CA27057, CA12644, CA68183, CA22433, CA35261, CA58861, CA20319, CA46113, CA58882, CA76447, CA04919, CA16385, CA35090, CA03096, CA67663, CA45450, CA35431, CA45807, CA58416, CA14028, CA45377, CA63845, CA42777, CA46136, CA11083, CA35119, CA58658, CA46282, CA76129, CA46368, CA35176, CA86780, CA46462, CA35192, CA35178, CA67575, CA63844, CA12213, CA74647, CA35128, CA35996, CA58686, CA13612, CA45461, CA58723, CA63848, CA35281, CA63850, CA76132, CA74811, and was supported in part by Aventis. The study sponsors had no role in the design, analysis, or decision to report the study.
Reprints: Southwest Oncology Group (S9916), Operations Office, 14980 Omicron Dr., San Antonio, TX 78245-3217.
|
REFERENCES |
|---|
|
TopNotesAbstractIntroductionSubjects and methodsResultsDiscussionReferences |
|---|
(1) Eisenberger MA, Blumenstein BA, Crawford ED, Miller G, McLeod DG, Loehrer PJ, et al. Bilateral orchiectomy with or without flutamide for metastatic prostate cancer. N Engl J Med 1998;339:1036–42.
(2) Kelly WK, Scher HI, Mazumdar M, Vlamis V, Schwartz M, Fossa SD. Prostate-specific antigen as a measure of disease outcome in metastatic hormone-refractory prostate cancer. J Clin Oncol 1993;11:607–15.[Abstract]
(3) Smith DC, Dunn RL, Strawderman MS, Pienta KJ. Change in serum prostate-specific antigen as a marker of response to cytotoxic therapy for hormone-refractory prostate cancer. J Clin Oncol 1998;16:1835–43.[Abstract]
(4) Bubley GJ, Carducci M, Dahut W, Dawson N, Daliani D, Eisenberger M, et al. Eligibility and response guidelines for phase II clinical trials in androgen-independent prostate cancer: recommendations from the Prostate-Specific Antigen Working Group. J Clin Oncol 1999;17:3461–7.
(5) Prentice RL. Surrogate endpoints in clinical trials: definition and operational criteria. Stat Med 1989;8:431–4.[ISI][Medline]
(6) Petrylak DP, Tangen CM, Hussain MH, Lara PN Jr, Jones JA, Taplin ME, et al. Docetaxel and estramustine compared with mitoxantrone and prednisone for advanced refractory prostate cancer. N Engl J Med 2004;351:1513–20.
(7) Greene FL, Page DL, Fleming ID, Fritz AD, Balch CM, Haller DG, et al. AJCC Cancer Staging Manual. 6th ed. New York (NY): Springer-Verlag; 2002.
(8) Lin DY, Fleming TR, De Gruttola V. Estimating the proportion of treatment effect explained by a surrogate marker. Stat Med 1997;16:1515–27.[CrossRef][ISI][Medline]
(9) Freedman LS, Graubard BI, Schatzkin A. Statistical validation of intermediate endpoints for chronic diseases. Stat Med 1992;11:167–78.[ISI][Medline]
(10) O'Quigley J, Xu R. Explained variation in proportional hazard hazards regression. In Handbook of statistics in clinical oncology, ed. John Crowley. New York: Marcel-Dekker; 2001. p. 397–409.
(11) Schemper M. The relative importance of prognostic factors in studies of survival. Stat Med 1993;12:2377–82.[ISI][Medline]
(12) Dowling AJ, Czaykowski PM, Krahn MD, Moore MJ, Tannock IF. Prostate-specific antigen response to mitoxantrone and prednisone in patients with refractory prostate cancer: prognostic factors and generalizability of a multicenter trial to clinical practice. J Urol 2000;163:1481–5.[CrossRef][ISI][Medline]
(13) Kantoff P. Recent progress in management of advanced prostate cancer. Oncology (Williston Park) 2005;9:631–6.
(14) Small EJ, McMillan A, Meyer M, Chen L, Slichenmyer WJ, Lenehan PF, et al. Serum prostate-specific antigen decline as a marker of clinical outcome in hormone-refractory prostate cancer patients: association with progression-free survival, pain end points, and survival. J Clin Oncol 2001;19:1304–11.
(15) Tannock IF, de Wit R, Berry WR, Horti J, Pluzanska A, Chi KN, et al. Docetaxel plus prednisone or mitoxantrone plus prednisone for advanced prostate cancer. N Engl J Med 2004;351:1502–12.
(16) Scher HI, Eisenberger M, D'Amico AV, Halabi S, Small EJ, Morris M, et al. Eligibility and outcomes reporting guidelines for clinical trials for patients in the state of a rising prostate-specific antigen: recommendations from the Prostate-Specific Antigen Working Group. J Clin Oncol 2004;22:537–56.
Manuscript received May 27, 2005; revised February 7, 2006; accepted March 13, 2006.
Editorial about this Article
CiteULike 
Connotea 
Del.icio.us What's this?
J Natl Cancer Inst 2006 98: 502-503.
This article has been cited by other articles:
|
|
 |
|
  G. Colloca, F. Checcaglini, and A. Venturino About sorafenib in castration-resistant prostate cancer Ann. Onc., August 9, 2008; (2008) mdn546v1. [Full Text] [PDF]
|
|
|
|
 |
|
 G. Sonpavde, M. T. Fleming, T. E. Hutson, and M. D. Galsky Trial Design for Metastatic Castration-Resistant Prostate Cancer J. Clin. Oncol., July 20, 2008; 26(21): 3647 - 3648. [Full Text] [PDF]
|
|
|
|
|
|
H. I. Scher, S. Halabi, I. F. Tannock, M. Morris, C. N. Sternberg, M. A. Carducci, M. A. Eisenberger, C. Higano, G. J. Bubley, R. Dreicer, et al. In Reply J. Clin. Oncol., July 20, 2008; 26(21): 3648 - 3649. [Full Text] [PDF]
|
|
|
|
 |
|
 C. A. Engstrom Hot Flashes in Prostate Cancer: State of the Science American Journal of Men's Health, June 1, 2008; 2(2): 122 - 132. [Abstract] [PDF]
|
|
|
|
 |
|
 D. R. Berthold, G. R. Pond, M. Roessner, R. de Wit, M. Eisenberger, a. I. F. Tannock, and on behalf of the TAX-327 investigators Treatment of Hormone-Refractory Prostate Cancer with Docetaxel or Mitoxantrone: Relationships between Prostate-Specific Antigen, Pain, and Quality of Life Response and Survival in the TAX-327 Study Clin. Cancer Res., May 1, 2008; 14(9): 2763 - 2767. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. M. Beer In Reply J. Clin. Oncol., April 10, 2008; 26(11): 1901 - 1902. [Full Text] [PDF]
|
|
|
|
|
|
K. N. Chi, S. L. Ellard, S. J. Hotte, P. Czaykowski, M. Moore, J. D. Ruether, A. J. Schell, S. Taylor, C. Hansen, I. Gauthier, et al. A phase II study of sorafenib in patients with chemo-naive castration-resistant prostate cancer Ann. Onc., April 1, 2008; 19(4): 746 - 751. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Ng, G. R. Pond, P. A. Tang, P. W. MacIntosh, L. L. Siu, and E. X. Chen Correlation of changes between 2-year disease-free survival and 5-year overall survival in adjuvant breast cancer trials from 1966 to 2006 Ann. Onc., March 1, 2008; 19(3): 481 - 486. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Goldman, M. LeBlanc, and J. Crowley Interim futility analysis with intermediate endpoints Clinical Trials, February 1, 2008; 5(1): 14 - 22. [Abstract] [PDF]
|
|
|
|
|
|
L. Collette, M. Buyse, and T. Burzykowski Are Prostate-Specific Antigen Changes Valid Surrogates for Survival in Hormone-Refractory Prostate Cancer? A Meta-Analysis Is Needed! J. Clin. Oncol., December 10, 2007; 25(35): 5673 - 5674. [Full Text] [PDF]
|
|
|
|
|
|
A. J. Armstrong, E. Garrett-Mayer, M. Eisenberger, I. Tannock, and R. de Wit In Reply J. Clin. Oncol., December 10, 2007; 25(35): 5674 - 5674. [Full Text] [PDF]
|
|
|
|
|
|
S. Oudard, E. Banu, F. Scotte, A. Banu, J. Medioni, P. Beuzeboc, F. Joly, J.-M. Ferrero, F. Goldwasser, and J.-M. Andrieu Prostate-specific antigen doubling time before onset of chemotherapy as a predictor of survival for hormone-refractory prostate cancer patients Ann. Onc., November 1, 2007; 18(11): 1828 - 1833. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. J. Armstrong, E. Garrett-Mayer, Y.-C. Ou Yang, M. A. Carducci, I. Tannock, R. de Wit, and M. Eisenberger Prostate-Specific Antigen and Pain Surrogacy Analysis in Metastatic Hormone-Refractory Prostate Cancer J. Clin. Oncol., September 1, 2007; 25(25): 3965 - 3970. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. G. Baker Surrogate endpoints: wishful thinking or reality? J Natl Cancer Inst, April 19, 2006; 98(8): 502 - 503. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||