Prostate-specific antigen decline during salvage radiation therapy following prostatectomy is associated with reduced biochemical failure

Purpose: To evaluate the prognostic value of prostate-specific antigen (PSA) decline during salvage radiation therapy (SRT) after prostatectomy. Methods and materials: We reviewed an institutional database and identified all prostate cancer patients who were treated with SRT between the years 2003 and 2010, had at least 1 PSA measurement during their SRT course, and had no history of androgen deprivation therapy use prior to or during SRT. Disease characteristics, treatment information, and clinical outcomes data were tabulated for each patient. The PSA response during SRT was defined as a PSA decline of at least 0.2 ng/mL compared with the pretreatment PSA level. Bivariate and multivariate analyses using Cox proportional hazards modeling were performed to identify predictors of biochemical recurrence. Results: Sixty-four patients met eligibility criteria for this analysis. Median PSA before SRT was 0.63 ng/mL (interquartile range: 0.42-1.00). With a median follow-up time of 70 months after SRT, 5-year actuarial rates for biochemical control and metastasis-free survival were 61% (95% confidence interval [CI], 48%-75%) and 88% (95% CI, 79%-97%), respectively. The median number of PSA measurements per patient during SRT was 3 (range, 1-5). On bivariate analysis, PSA response during SRT and positive surgical margins were significantly associated with a decreased risk of biochemical recurrence (BR), with hazard ratios of 0.160 (95% CI, 0.059-0.431, P < .001) and 0.396 (95% CI, 0.168-0.935, P = .035). On multivariate analysis, PSA response during SRT and positive surgical margin were independent, favorable predictors for BR, with hazard ratios of 0.171 (95% CI, 0.063-0.463, P < .001) and 0.411 (95% CI, 0.177-0.956, P = .039). The 5-year biochemical control rate for PSA responders was 81%, compared with 37% for nonresponders (. P < .001). Conclusions: Prostate-specific antigen decline during SRT may be a valuable prognostic factor for subsequent clinical outcomes. Future studies should investigate the value of monitoring PSA during SRT and how PSA response may be used to personalize therapy.