The association of germline HSD3B1 genotype with outcomes in metastatic hormone-sensitive prostate cancer (mHSPC) treated with androgen deprivation therapy (ADT) with or without enzalutamide (ENZA) [ARCHES]

5022 Background: A common missense-encoding germline polymorphism in HSD3B1(1245A→C) stabilizes the 3βHSD1 enzyme and increases the rate-limiting step for metabolic flux from adrenal precursors to potent androgens. This adrenal-permissive HSD3B1(1245C) allele mechanistically drives more rapid progression after ADT in nonmetastatic and low-volume (LV) mHSPC. It is unknown whether upfront treatment with ENZA could improve these poor outcomes. We sought to determine the association between the HSD3B1 genotype and outcomes in ARCHES, a randomized, placebo (PBO)-controlled, phase 3 global study of ADT with ENZA or PBO in mHSPC. Overall, 31.3% of men on placebo crossed over to ENZA prior to progression. Methods: Germline DNA from 660 men (243 LV and 417 high volume [HV]) in ARCHES was genotyped for HSD3B1. Radiographic progression-free survival (rPFS), prostate-specific antigen (PSA) outcomes, and overall survival (OS) were compared between men who inherited 0 (adrenal restrictive) vs. 1–2 copies of the HSD3B1(1245C) allele (adrenal permissive) using log-rank tests, Kaplan-Meier estimates, and Cox proportional hazards models for hazard ratios (HRs). Results: In ARCHES, 340 and 320 men had the adrenal-restrictive and adrenal-permissive HSD3B1 genotype, respectively. The adrenal-permissive genotype was not associated with worse rPFS, time to PSA progression (TTPP), or OS. Cox model HRs for death in the adrenal-permissive compared with adrenal-restrictive men in the ENZA + ADT and PBO + ADT arms were 1.21 (95% confidence interval [CI] 0.84, 1.74) and 1.48 (95% CI 0.98, 2.25), respectively. Three-year OS was 82% and 73% in men treated with ENZA + ADT with restrictive vs. permissive genotypes, respectively, and 68% and 70% in men treated with PBO + ADT. ENZA significantly improved rPFS, TTPP, and OS, irrespective of genotype. HRs for ENZA + ADT vs. PBO + ADT in models adjusted for genotype were: rPFS, 0.64 (p < 0.001); TTPP, 0.27 (p < 0.001), and OS, 0.62 (p < 0.001). Of OS events in the HSD3B1-genotyped men, 48 (23%) and 163 (77%) deaths occurred in the LV and HV groups, respectively. Conclusions: In men with mHSPC, ENZA improved OS, rPFS, and TTPP over PBO + ADT, irrespective of HSD3B1 genotype. Survival analysis of HSD3B1-genotyped men in this study is largely driven by men with HV disease. The analysis of men with LV disease was limited by the small number of events. Clinical trial information: NCT02677896.