Purpose To describe outcomes of salvage radiotherapy (SRT) for men with detectable prostate-specific antigen (PSA) after radical prostatectomy for prostate cancer and identify associations with outcomes. Patients and Methods A total of 1,106 patients received SRT between January 1987 and July 2013, with median follow-up 8.9 years. Outcomes were estimated using Kaplan-Meier for overall survival (OS) and cumulative incidence for biochemical recurrence (BcR), distant metastases (DM), and cause-specific mortality (CSM). Variable associations with outcomes used Cox or Fine-Gray methods, as appropriate. Multiple variable analyses used backward selection with P < .05 for retention. Results In multiple variable analyses, pathologic tumor stage, Gleason score, and pre-SRT PSA were associated with BcR, DM, CSM, and OS; androgen suppression and SRT doses > 68 Gy were associated with BcR; and age was associated with OS. Each pre-SRT PSA doubling increased significantly the relative risk of BcR (hazard ratio [HR], 1.30; P < .001), DM (HR, 1.32; P < .001), CSM (HR, 1.40; P < .001), and all-cause mortality (HR, 1.12; P = .02). Using a pre-SRT PSA cutoff ≤ 0.5 versus > 0.5 ng/mL, 5-year and 10-year cumulative incidences for BcR were 42% versus 56% and 60% versus 68% ( P < .001), DM 7% versus 14% and 13% versus 25% ( P< .001), CSM 1% versus 4% and 6% versus 13% ( P < .001), and OS of 94% versus 92% and 83% versus 73% ( P > .05). Conclusion SRT outcomes are in part affected by factors associated with prostatectomy findings but may be positively affected by using SRT at lower PSA levels, including reductions in BcR, DM, CSM, and all-cause mortality. These findings argue against prolonged monitoring of detectable postprostatectomy PSA levels that delay initiation of SRT.
