Application of a Prognostic Stratification System for High-risk Prostate Cancer to Patients Treated With Radiotherapy: Implications for Treatment Optimization

Objectives: We applied an established prognostic model to high-risk prostate cancer (HRPC) patients treated with radiotherapy (RT) and evaluated the influence of clinical and treatment variables on treatment outcomes. Methods: In total, 1075 HRPC patients undergoing definitive radiotherapy (RT) between 1995 and 2010 were retrospectively reviewed. Median follow-up was 62.3 months. Patients received either dose-escalated external beam radiotherapy (n=628, EBRT) or combined-modality radiotherapy (n=447, pelvic RT and low-dose rate brachytherapy boost, CMRT). 82.9% received androgen-deprivation therapy (ADT). A prognostic model stratified patients into predefined groups (good, intermediate, and poor). Kaplan-Meier methods and Cox proportional hazards regressions assessed biochemical failure (BF), distant metastasis (DM), prostate cancer-specific mortality (PCSM) and overall mortality (OM). C-indices analyzed predictive value. Results: The model was prognostic; C-indices for BF, DM, PCSM and OM were: 0.62, 0.64, 0.61, and 0.57. On multivariate analysis, CMRT and longer ADT (≥24 mo) were associated with improved BF, DM, and PCSM. Gleason score (GS) 9-10 was the strongest predictor of PCSM. C-indices for BF, DM, PCSM, and OM using a 4-compartment model incorporating GS 9-10 were: 0.62, 0.65, 0.68, and 0.56. In poor-prognosis patients (GS 8-10+additional risk factors), CMRT+LTADT (>12 mo) had 10-year PCSM (3.7%±3.6%), comparing favorably to 25.8%±9.2% with EBRT+LTADT. Conclusions: The model applies to high-risk RT patients; GS 9-10 remains a powerful predictor of PCSM. Comparing similar prognosis patients, CMRT is associated with improved disease-specific outcomes relative to EBRT. In poor-prognosis patients, CMRT+LTADT yields superior 10-year PCSM, potentially improving RT treatment personalization for those with HRPC. B.F., W.J. and D.H. contributed to statistical analyses and all authors were instrumental in manuscript review and editing. The authors declare no conflicts of interest. Reprints: Daniel Hamstra, MD, PhD, 18101 Oakwood Boulevard, Lower Level, Beaumont Hospital, Department of Radiation Oncology, Dearborn, MI 48124. E-mail: daniel.hamstra@beaumont.org. Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.

http://bit.ly/2WMASPN