180 Background: RW data comparing first-line (1L) androgen receptor pathway inhibitors for mCSPC treatment (tx) are limited and misclassification of disease state can complicate interpretation. This RW study compared outcomes in patients (pts) with mCSPC who initiated 1L enza or apa. Methods: Flatiron Health EHR data (January 2013–May 2025) were analyzed for males in the US with chart-abstracted evidence of mCSPC who initiated index tx (1L enza or apa) on/after January 1, 2020, and had no evidence of metastatic castration-resistant prostate cancer (mCRPC) within 3 months post 1L tx initiation. Outcomes included duration of tx (DoT), time to next tx (TTNT), and time to mCRPC. Cox proportional hazards models with inverse probability of tx weighting (IPTW) compared outcomes between enza and apa in the overall population, adjusting for baseline characteristics. Sensitivity analyses were restricted to pts who received 1L tx in the mCSPC setting only (i.e., no evidence of mCRPC while on 1L tx). Results: Overall, 864 pts had enza and 410 pts had apa as 1L mCSPC tx. Median follow-up was 23 (enza) and 24 (apa) months. Baseline characteristics were balanced after IPTW: mean age was 73 years, 80% of pts received tx in community oncology practices, and 68% had de novo metastatic disease. All IPTW-adjusted outcomes were comparable for enza and apa (Table). Similar proportions of pts in the enza and apa cohorts discontinued index tx (56% vs 58%), initiated next tx (30% vs 31%), and progressed to mCRPC (25% vs 25%). Sensitivity analyses yielded consistent results. Conclusions: In pts with mCSPC (defined stringently to minimize bias) who were treated in US oncology practices, 1L enza and apa showed comparable DoT, TTNT, and time to mCRPC, which supports similar effectiveness of enza and apa. Disclosure: A genAI tool (09/23/25; Pfizer; GPT-4o) developed the first draft; authors assume content responsibility. IPTW-adjusted outcomes in pts with mCSPC. Overall Sensitivity analysis Enza(N ‡ = 861) Apa † (N ‡ = 407) Enza(N ‡ = 738) Apa † (N ‡ = 355) Outcome § Median (95% CI), mo Median (95% CI), mo aHR (95% CI) Median (95% CI), mo Median (95% CI), mo aHR (95% CI) Time to tx discontinuation ║ 22 (19, 25) 19 (16, 24) 0.95 (0.81, 1.12) 25 (22, 29) 22 (17, 30) 0.90 (0.75, 1.08) Time to next tx ¶ NR (46, NE) NR (51, NE) 0.98 (0.78, 1.22) NR (NE, NE) NR (53, NE) 0.93 (0.72, 1.21) Time to mCRPC # NR (NE, NE) NR (NE, NE) 1.03 (0.81, 1.32) NR (NE, NE) NR (NE, NE) 0.87 (0.61, 1.24) † Apa was the reference group. ‡ N for IPTW-weighted sample. § IPTW-adjusted by baseline characteristics. ║ Time from index date to the discontinuation of index tx (next tx, tx gap >90 days, death). ¶ Time from index date to the earliest initiation of next tx. #
Morgans et al. (2026) studied this question.