LAUSR

BACKGROUND Intermediate clinical endpoints (ICEs) prognostic for overall survival (OS) are needed for men receiving postprostatectomy radiation therapy (PORT) to improve clinical trial design. OBJECTIVE To identify a potential ICE for men receiving PORT. DESIGN, SETTING, AND PARTICIPANTS We performed an institutional review board-approved multi-institutional retrospective study of 566 men consecutively treated with PORT at tertiary care centers from 1986 to 2013. The median follow-up was 8.2 yr. OUTCOME MEASUREMENTS AND STATISTICAL ANALYSIS Biochemical failure (BF), distant metastases (DM), and castrate-resistant prostate cancer (CRPC) were evaluated for correlation with OS and assessed as time-dependent variables in a multivariable Cox proportional hazards model and in landmark analyses at 1, 3, 5, and 7 yr after PORT. Cross-validated concordance (c) indices were used to assess model discrimination. RESULTS AND LIMITATIONS OS at 1, 3, 5, and 7 yr after PORT was 98%, 95%, 90%, and 82%, respectively. In a time-varying model controlling for clinical and pathologic variables, BF (hazard ratio [HR] 2.32, 95% confidence interval [CI] 1.45-3.71; p<0.001), DM (HR 6.52, 95% CI 4.20-10.1; p<0.001), and CRPC (HR 2.47, 95% CI 1.56-3.92; p<0.001) were associated with OS. In landmark analyses, 5-yr DM had the highest c index when adjusting for baseline covariates (0.78), with 5-yr DM also providing the greatest increase in discriminatory power over a model only including baseline covariates. These findings require validation in prospective randomized data. CONCLUSIONS While limited by the retrospective nature of the data, 5-yr DM is associated with lower OS following PORT, outperforming the prognostic capability of BF and CRPC at 1, 3, 5, or 7 yr after treatment. Confirmation of this ICE as a surrogate for OS is needed from randomized trial data so that it can be incorporated into future clinical trial design. PATIENT SUMMARY We assessed potential intermediate clinical endpoints prognostic for overall survival in a cohort of men receiving radiotherapy after prostatectomy. We identified the development of metastatic disease within 5 yr after treatment as the strongest predictor of overall survival.