Abstract
BACKGROUND: The management of biochemical failure (BF) following external beam radiotherapy (EBRT) for prostate cancer is controversial, due to both the heterogeneous disease course following a BF and a lack of clinical trials in this setting. OBJECTIVE: We sought to characterize the natural history and predictors of outcome for patients experiencing BF in a large cohort of men with localized prostate cancer undergoing definitive dose-escalated EBRT. DESIGN, SETTING, AND PARTICIPANTS: This retrospective analysis included 2694 patients with localized prostate cancer treated with EBRT at a large academic center. Of these, 609 experienced BF, defined as prostate-specific antigen (PSA) nadir + 2 ng/ml. The median follow-up was 83 mo for all patients and 122 mo for BF patients. INTERVENTION(S): All patients received EBRT at doses of 75.6-86.4 Gy. OUTCOME MEASUREMENTS AND STATISTICAL ANALYSIS: The primary objective of this study was to determine predictors of distant progression at the time of BF. Cox proportional hazards models were used in univariate and multivariate analyses of distant metastases (DM), and a competing risks method was used to analyze prostate cancer-specific mortality (PCSM). RESULTS AND LIMITATIONS: From the date of BF, the median times to DM and PCSM mortality were 5.4 yr and 10.5 yr, respectively. Shorter posttreatment PSA doubling time, a higher initial clinical tumor stage, a higher pretreatment Gleason score, and a shorter interval from the end of radiotherapy to BF were independent predictors for clinical progression following BF. Patients with two of these risk factors had a significantly higher incidence of DM and PCSM following BF than those with zero or one risk factor. The main limitations of this study are its retrospective nature and heterogeneous salvage interventions. CONCLUSIONS: Clinical and pathologic factors can help identify patients at high risk of clinical progression following BF. PATIENT SUMMARY: In this report, we look at predictors of outcome for patients with prostate cancer recurrence, as determined by prostate-specific antigen (PSA) levels, following radiation treatment. We found that the approximate median times to distant metastasis and death from prostate cancer for patients in this situation were 5 yr and 10 yr, respectively. Furthermore, we found that patients with a rapid increase in PSA levels following treatment, a short time to PSA recurrence, invasion of extraprostatic organs, or a high Gleason score had worse outcomes.