Biopsy-based Basal-luminal Subtyping Classifier in High-risk Prostate Cancer: A Combined Analysis of the NRG Oncology/RTOG 9202, 9413, and 9902 Phase 3 Trials.

Background and objective: Long-term (LT) androgen deprivation therapy (ADT) has been found to be beneficial to patients with high-risk prostate cancer (PCa). However, administration of LT-ADT to all patients with high-risk PCa may lead to overtreatment. Enhanced risk stratification using genomic classifiers (such as the recently developed prostate subtyping classifier [PSC]) might be useful. This study aims to characterize the prognostic and predictive ability of the PSC in patients with high-risk PCa undergoing radiotherapy long-term (LT; 24-28 mo) versus short-term (ST; 4 mo) ADT.

Methods: Biopsy samples from three randomized, phase 3 trials-NRG/RTOG 9202, 9413, and 9902-were classified as either PSC basal or luminal. The prognostic and predictive values of PSC for each oncologic endpoint (biochemical failure [BF], distant metastasis [DM], metastasis-free survival [MFS], PCa-specific mortality [PCSM], overall survival [OS]) and other cause-mortality (OCM) were assessed with Cox proportional hazards (MFS, OCM, and OS), Fine-Gray (BF, DM, and PCSM), and restricted mean survival time (RMST) models.

Key findings and limitations: On a multivariable analysis, the basal subtype was found to have a worse prognosis for MFS (hazard ratio [HR] 1.8 [1.3-2.5], p < 0.001), PCSM (subdistribution HR 2.4 [95% confidence interval {CI} 1.4-4.1], p = 0.001), and OS (HR 1.8 [1.3-2.6], p < 0.001). Ten-year PCSM was 15% better for the luminal subtype than for the basal subtype (11% [95% CI 6-15%] vs 26% [95% CI 17-35%]). A significant interaction between ADT duration (LT vs ST) and PSC subtype (basal vs luminal) was observed for PCSM (p_interaction = 0.008), leading to the observation that 10-yr PCSM was improved with LT-ADT only in patients with basal-type tumors (5% [95% CI 0-11%] vs 42% [29-56%], p < 0.001). Improvements in 10-yr RMST with LT-ADT were greater for basal tumors for oncologic endpoints with the exception of OCM.

Conclusions and clinical implications: PSC is both a prognostic and a predictive biomarker for patients who benefit from LT-ADT. PSC subtypes may be used to personalize ADT recommendations for patients with high-risk PCa, pending further validation in a prospective study.

Patient summary: In this study, we tried to understand the usefulness of a new genomic test in patients with high-risk, nonmetastatic prostate cancer who underwent radiation therapy and hormonal therapy (HT). We found that this test can help determine a patient's prognosis (eg, a patient's chance of having the cancer return) and, more importantly, personalize treatment decisions by understanding which patients may benefit from long-term HT. This has the potential to save many patients who may not benefit from prolonged HT from "overtreatment" or the unnecessary side effects of such treatment.