Matching-adjusted indirect comparisons (MAICs) of talazoparib plus enzalutamide (TALA+ENZA) versus olaparib plus abiraterone and prednisone/prednisolone (OLAP+AAP) for first-line (1L) therapy in patients with metastatic castration-resistant prostate cancer (mCRPC) and homologous recombination repair mutations (HRRm)/BRCAm.

5063 Background: Poly-ADP ribose polymerase inhibitors (PARPi) in combination with a novel hormonal therapy (NHT) have shown benefit for the 1L treatment of mCRPC in an all-comers population as well as HRRm and BRCAm subpopulations (TALA+ENZA TALAPRO-2; NCT03395197 and OLAP+AAP PROpel; NCT03732820). In the absence of head-to-head studies, their comparative efficacy is unknown. The relative efficacy of TALA+ENZA (n=200 HRRm; n=71 BRCAm) vs OLAP+AAP (n=111 HRRm; n=47 BRCAm) in these subpopulations were estimated using MAICs for radiographic progression-free survival (rPFS) based on blinded independent central review and overall survival (OS). Methods: Unanchored MAICs were conducted using individual patient data from TALAPRO-2 Cohort 2 (data cutoff DCO: 03/10/22 rPFS/OS) and published summary level data from PROpel (DCO: 30/07/21 rPFS; 12/10/22 OS). To align across the two trials, patients from TALAPRO-2 with specific HRRm/BRCAm (co-occurring or standalone) that were not assessed in PROpel were removed from the dataset for each analysis. TALAPRO-2 patients were also matched based on PROpel’s eligibility criteria and characteristics were adjusted for key prognostic factors identified in the literature and clinical expertise including prior taxane chemotherapy in castration sensitive prostate cancer (CSPC), visceral metastasis, bone only metastasis, Eastern Cooperative Oncology Group score, prostate-specific antigen levels, Gleason score, BRCA1 and BRCA2. Results: After the removal of patients with specific gene mutations not assessed in PROpel and those who received prior NHT in CSPC, 157 and 64 patients remained in the TALA+ENZA arm for HRRm and BRCAm, respectively. The comparative effect estimates for rPFS and OS are presented in the Table. None of the results were statistically significant. Conclusions: These analyses demonstrate numerically favorable results for TALA+ENZA compared to OLAP+AAP highlighting its therapeutic potential in 1L mCRPC for patients with HRRm and BRCAm. Limitations include inability to adjust for all characteristics and biases due to unobserved trial differences. Table: see text