Enzalutamide-era mCRPC: ctDNA AR enhancer/AR copy gains sharpen rPFS risk stratification (Alliance A031201)

A031201 randomized chemotherapy-naïve mCRPC patients to the androgen receptor (AR) pathway inhibitor enzalutamide ± abiraterone acetate/prednisone (n=776 with baseline ctDNA profiled by the AR-ctDETECT assay; NCT01949337), and this analysis built a circulating tumor DNA (ctDNA) clinical-genetic (CG) prognostic model for radiographic PFS (rPFS). AR enhancer gain and AR gain were among the strongest adverse markers (median rPFS 13.7 vs 27.3 months and 13.5 vs 27.0 months with vs without the alteration, respectively); the random survival forest model selected gains (AR, AR enhancer, MYC, CCND1, FOXA1) and losses (PTEN, TP53, LRP1B) plus clinical labs (hemoglobin, PSA, alkaline phosphatase), improving discrimination vs a clinical-only model (mean itAUC 0.73 vs 0.66; NRI 0.30 at 22 months). The model stratified patients into low/intermediate/poor risk with median rPFS 39.2/25.2/12.8 months (HR low vs poor 0.25; intermediate vs poor 0.46); as a benchmark, first-line ARSI-era trials report rPFS on the order of ~20 months for enzalutamide vs placebo in PREVAIL and ~16.5 months for abiraterone in COU‑AA‑302 (investigator assessment), underscoring how much baseline biology can swing outcomes even within “ARSI-treated” mCRPC populations.