Utilization of Salvage Radiation Therapy for Biochemical Recurrence After Radical Prostatectomy

Abstract

For men with biochemical recurrence after radical prostatectomy (RP), salvage radiation therapy (SRT), especially "early" SRT (PSA level ≤0.5ng/mL), is a potentially curative option; however, its utilization is not well defined. We sought to determine factors associated with SRT utilization as well as variation in its administration. Patients with localized prostate cancer undergoing RP at 33 practices participating in the statewide Michigan Urological Surgery Improvement Collaborative between 2012 and 2016 were prospectively followed. Eligible patients had at least 1 post-RP PSA level ≥0.1ng/mL with ≥6months of follow-up after the first detectable PSA level. Patients undergoing adjuvant radiation therapy were excluded. SRT utilization and clinical and pathologic patient characteristics were examined. Of 1010 eligible patients with a detectable PSA level, 29.5% underwent SRT. Of patients who received SRT, 46.9% either reached a PSA ≥0.2ng/mL or were treated before reaching that PSA level. A total of 30.6% of patients had a PSA level ≥0.5ng/mL without undergoing prior SRT; of this group, 42.1% later received SRT. After adjusting for patient and practice level factors, positive surgical margins, higher T stage, and higher grade group were all associated with receipt of SRT (P<.05). Even after adjusting for patient and tumor characteristics, significant variation remained in the adjusted rate of SRT utilization across practices sites, ranging from 7% (95% confidence interval, 3%-17%) to 73% (95% confidence interval, 45%-90%, P<.001). Practices were grouped into tertiles based on SRT utilization, and those practices that used SRT more frequently overall were more likely to administer SRT across all patient-based predictors of SRT utilization. SRT utilization is low among men with a detectable post-RP PSA level, with significant variation in practice-level SRT utilization that cannot be explained by patient factors alone. Factors suggesting higher-risk disease were predictors of SRT administration. These data support the potential to expand the use of SRT, particularly among sites with low utilization.