Final results from ZZFIRST: A randomized phase 2 trial of enzalutamide (EZ) and talazoparib (TALA) in metastatic hormone-naïve prostate cancer (mHNPC).

5006 Background: ADT plus AR pathway inhibition (ARPI) improves outcomes in mHNPC, yet responses remain heterogeneous. AR–PARP cotargeting has shown enhanced clinical activity in prostate cancer, but the underlying mechanisms are incompletely defined. ZZFIRST evaluated enzalutamide (EZ) plus talazoparib (TALA) in mHNPC, including the study of tumor adaptation response to ADT+ARPI through biopsies obtained on therapy. Methods: ZZFIRST (NCT04332744) is an open label, randomized academic phase 2 trial that included patients (pts) with high-volume (HV)-mHNPC (CHAARTED), ECOG PS ≤1, PSA≥4 ng/mL. All pts received standard ADT and started EZ 160 mg/day. After two 28-day cycles, pts were randomized 2:1 to EZ+TALA 0.5 mg/day or continue EZ alone until progressive disease (PD), unacceptable toxicity, or withdrawal. Randomization was stratified by DDR mutation status. Primary endpoint was confirmed PSA < 0.2 response rate (PSA < 0.2-RR) at 12 months (mo) for EZ+TALA. Key secondary endpoints included rPFS, time to PSA-PD, time to castration resistance, and safety. The study was powered (80%, one-sided α = 5%) to detect improved PSA < 0.2-RR (H0: 20%; H1≥40%) for EZ+TALA. Between-arm comparisons were exploratory. Spatial transcriptomics (Visum HD) was performed on baseline (BL) and Cycle 2-Day 1 (C2D1) on-treatment tumor biopsies. Results: 54 pts were enrolled (37 EZ+TALA, 17 EZ). 7 pts harbored HRR alterations by tumor tissue NGS (13%; 5 EZ+TALA, 2 EZ). Visceral metastases were present in 17 pts (32%); 51 pts (94%) had de novo M1 disease. Median follow-up was 3.6 years. The primary endpoint was met, with a 73% 12-mo PSA < 0.2-RR for EZ+TALA (95%CI, 55.9%-86.2%, p < 0.001); the rate was 64.7% for EZ. Median rPFS was 45.3 mo (29.4-NR) for EZ+TALA vs 31.1 mo (13.5-NR) for EZ (HR = 0.62; 95%CI, 0.28-1.37). Median time to PSA-PD was not reached (30.4-NR) for EZ+TALA vs 30.7 mo (6.6-NR) for EZ (HR = 0.47; 95%CI, 0.21-1.04). Median TTCR was 30.4 mo (23.0-NR) for EZ+TALA vs 22.8 mo (6.3-45.4) for EZ (HR = 0.62; 95%CI, 0.30-1.26). In TALA+EZ, common treatment-emergent AEs were fatigue (any grade: 83.8%; G≥3: 13.5%) and anemia (67.6%; G≥3: 40.5%). 14 (38%) pts required TALA dose reductions. Two pts on EZ+TALA developed acute myeloid leukemia (AML) 26.0 and 32.8 mo after study entry. Transcriptomics analysis on C2D1 vs BL biopsies showed induction of cell cycle arrest, with downregulation of AR and DDR signatures and upregulation of IFN-related and inflammatory response signatures upon ADT+EZ. Conclusions: EZ+TALA demonstrated promising antitumor activity in HV-mHNPC. The toxicity profile was consistent with prior AR+PARP trials, although two cases of AML were observed with prolonged TALA exposure. Paired tumor biopsies revealed mechanisms of tumor adaptation to ADT+EZ inhibition, supporting a biologically informed rationale for AR+PARP treatment intensification strategies in mHNPC. Clinical trial information: NCT04332744 .