Risk Factors Predicting Relapse 10 Years Following Extreme Dose-Escalated SBRT for Intermediate-Risk Prostate Cancer: Is De-Intensification Feasible in Unfavorable Subgroups?

Abstract

Adding short-term androgen-deprivation therapy (ADT) to radiotherapy is recommended in unfavorable intermediate-risk (UIR) prostate cancer (PCa). Data supporting the addition of ADT to stereotactic body radiotherapy (SBRT) in intermediate-risk (IR) patients are limited. Given ADT has well-documented toxicities, we sought to identify the UIR prognostic factors that predict for relapse following SBRT without ADT. We combined results from two mature multicenter trials to determine if extreme dose-escalated SBRT yielded UIR subgroups in which the omission of ADT may be considered. Between 2008 and 2011, two prospective national trials enlisted 39 centers to enroll 285 patients with IR PCa: 182 had Memorial Sloan Kettering (MSK) favorable intermediate-risk (FIR) and 103 had MSK UIF PCa. All were treated with a non-coplanar robotic SBRT platform using real-time tracking of implanted fiducials. Two dose regimens were used: 40Gy in 5 fractions of 8Gy, and 38Gy in 4 fractions of 9.5Gy. ADT was not allowed. Univariate and multivariate analyses using a Cox proportional hazards model was performed for relapse free survival (RFS): relapse included pathologic or radiographic failure, initiation of salvage or systemic therapy, or biochemical relapse by the nadir + 2 definition. Insufficient events prevented similar analyses for local control, metastasis-free and PCa-specific survival. Examined risk factors included dose regimen, clinical T-stage, Gleason score, pre-treatment PSA, % positive biopsy cores, and number of unfavorable risk factors (URFs). All reported rates are actuarial, using Kaplan-Meier method. Median follow-up was 8.1 years. 71 patients were followed 10 years. There were no statistically significant differences in rates of toxicity, local failure, RFS, overall nor metastasis-free survival between the two dose regimens. For the entire group, 10-year overall survival was 82.9%, RFS was 83.2%, and the local failure rate was 3.6%. On univariate analysis, primary Gleason pattern 4 (PGP4) and >2 URFs predicted for RFS. On multivariate analysis, only PGP4 (hazard ratio: 3.71, p = 0.0053) was statistically significant. Examining the UIR subgroup, the only predictor for RFS was PGP4 (HR: 3.64, p = 0.0253). 10-year RFS was 88.4% for FIR and 89.1% for UIR without PGP4; this fell to 58.5% in patients with a PGP4. Following dose-escalated SBRT monotherapy, 10-year RFS rates were favorable in UIR patients without PGP4. Extreme dose escalation appeared to effectively address UIR factors correlating with tumor bulk (i.e., CS T2b, >50% cores+, PSA 10-20), but was less effective with biologically aggressive pathology (PGP4). A randomized trial, ideally including genomic classification, would be necessary to determine if dose-escalated SBRT allows de-intensification (omission of ADT) in such biologically less aggressive UIR subgroups. The high relapse rate observed in UIR patients with PGP4 affirms the need for adjuvant ADT in this subgroup.