Prostate Cancer–Specific Mortality and Survival Outcomes for Salvage Radiation Therapy After Radical Prostatectomy

Abstract

Early salvage radiation therapy (SRT) following radical prostatectomy (RP) has been shown to reduce biochemical recurrence and distant metastases. Using a consortium database including data from 10 academic institutions, we assessed the impact of SRT initiation at lower prostate specific antigen (PSA) levels on prostate cancer-specific mortality (PCSM) and all-cause mortality (ACM). In this IRB-approved retrospective study, 2,454 node-negative patients (pts) with detectable post-prostatectomy PSA (≥0.01 ng/mL) treated with SRT ± neoadjuvant/concurrent androgen deprivation therapy (n/c ADT) were included. Exclusion criteria included previous ADT or incomplete treatment, staging, or follow up details. Cumulative incidence method of Fine-Gray and Kaplan-Meier methods were used to estimate rates of PCSM and ACM, respectively. Univariate and multivariable analyses (MVA) were performed by competing risks regression (CRR) and Cox proportional hazards (CPH) methods for PCSM and ACM, respectively. Median follow-up was 5.1 years following SRT end-date; 597 pts (24%) had pathologic Gleason score (GS) of ≤6, 1383 (56%) GS 7, and 474 (19%) GS ≥8. There were 1365 (56%) with extraprostatic extension, 451 (18%) seminal vesicle invasion, 1430 (58%) positive surgical margins, and 390 (16%) received n/c ADT for a median of 6 months. Median age at RP and SRT were 61 years (IQR = 56-66) and 64 years (59-69), respectively. Median SRT dose to the prostate bed was 66 Gy (IQR = 64.8-68) and median pre-SRT PSA was 0.5 ng/mL (IQR = 0.3-1.1). The 5 and 10-year PCSM rates were 3% and 6%, respectively. The 10 year PCSM rate was 5% for pre-SRT PSA ≤0.2 ng/mL, 6% for 0.21-0.50 ng/mL, 8% for 0.51-1.0 ng/mL, 18% for 1.01-2.0 ng/mL, and 22% for >2.0 ng/mL, P < 0.0001. The 5 and 10-year ACM rates were 7% and 23%, respectively, and at 10 years was 14% for pre-SRT PSA ≤0.2, 16% for 0.21-0.50, 23% for 0.51-1.0, 30% for 1.01-2.0, and 38% for >2.0, P < 0.0001. On MVA, higher pre-SRT PSA (HR = 2.13, P < 0.0001), higher GS (GS 7 vs. ≤6: HR = 2.01, P = 0.0012; GS ≥8 vs. 6: HR = 3.34, P < 0.0001), seminal vesicle invasion (HR = 2.48, P < 0.0001), and year of SRT (2000-2004, 1995-1999, 1985-1994 vs. 2005-2012; HR = 2.87, P = 0.021, HR = 2.50, P = 0.0097, HR = 3.58, P = 0.0016, respectively) were significantly associated with higher PCSM, while extraprostatic extension, surgical margins, ADT use, SRT dose, age at SRT, and age at RP were not. These same variables were significantly associated with higher ACM on MVA, in addition to advanced age at RP (HR = 1.06, P < 0.0001). Initiation of early SRT at low PSA levels compared to higher PSA levels following RP is associated with reduced risk of PCSM and ACM. Other factors significantly associated with PCSM include higher GS, seminal vesicle invasion, and earlier year of SRT.