Validation of a 22-gene Genomic Classifier in the NRG Oncology/RTOG 9202, 9413 and 9902 Phase III Randomized Trials: A Biopsy-Based Individual Patient Meta-Analysis in High-Risk Prostate Cancer.

PURPOSE/OBJECTIVE(S) Decipher is a prognostic 22-gene genomic classifier (GC) prospectively validated post-prostatectomy. Herein, we validate the performance of the GC in pre-treatment biopsy samples within the context of three randomized phase III high-risk definitive radiotherapy trials. MATERIALS/METHODS Following a pre-specified and approved CTEP-CCSC analysis plan (NRG-GU-TS006), we obtained all available formalin-fixed paraffin-embedded tissue from biopsy specimens from the NRG biobank from patients enrolled on NRG/RTOG 9202, 9413, and 9902 phase III randomized trials. After central review, the highest-grade tumors were profiled on clinical-grade whole-transcriptome arrays and GC scores were obtained. The primary objective was to validate the independent prognostic ability of GC for distant metastases (DM), and secondary was prostate cancer-specific mortality (PCSM) and overall survival (OS), with Cox multivariable analyses (MVA). RESULTS GC scores were obtained on 385 samples (n = 90 on 9202, n = 172 on 9413, and n = 123 on 9902), of which 265 passed microarray quality control (69%) and had a median follow-up of 11 years (interquartile range, 9, 13). On MVA, the GC (per 0.1 unit) was independently associated with DM (HR 1.24, 95% CI 1.11-1.39), PCSM (HR 1.27, 95% CI 1.13-1.43), and OS (HR 1.12, 95% CI 1.05-1.20) after adjusting for age, PSA, Gleason score, cT-stage, trial, and randomized treatment arm. For categorical GC, on MVA, GC score ≥ 0.45 (representing the intermediate and high GC categories) had worse DM (HR 2.18, 95% CI 1.25-3.80), PCSM (HR 2.34, 95% CI 1.31-4.16), and OS (HR 1.45, 95% CI 1.03-2.04) outcomes as compared to those with low GC. Cumulative incidence of distant-metastasis at 10-years was 29% (95% CI 20-38%) for intermediate/high GC vs 13% (95% CI 7-18%) for low GC. For the subset with GC > 0.85, the threshold for inclusion in the intensification study of NRG GU009 (PREDICT-RT), at 5-years and 10-years DM was 29% (95% CI 7-52%) and 41% (95% CI 17-66%). GC had similar prognostic ability in patients receiving short-term or long-term androgen-deprivation therapy (ADT). CONCLUSION This is the first validation of any gene expression biomarker on pre-treatment biopsy samples from prospective randomized trials and demonstrates an independent association of GC score with DM, PCSM, and OS. High-risk prostate cancer is a heterogeneous disease state and GC can improve risk stratification to help personalize shared decision-making. NRG-GU009/PREDICT-RT will further determine the optimal therapy based on GC score. NCT04513717.