Comparison of quantitative functional RAD51 foci and genomic alterations in predicting PARPi response in mCRPC: A multicenter real-world study.

Abstract

114 Background: Currently, the clinical outcomes of poly(ADP-ribose) polymerase (PARP) inhibitors varied widely in metastatic castration-resistant prostate cancer (mCRPC) patients with different HRR gene alterations (HRRalt). It is urgent and necessary to develop a novel biomarker for predicting the clinical efficiency of PARP inhibitors in the mCRPC patients. Tissue-based immunofluorescence (IF) assay for the functional RAD51 foci is emerging as a promising biomarker for assessing homologous recombination deficiency (HRD). However, the predictive value of the RAD51 foci using the IF assay for mCRPC patients following PARP inhibitors remains uncertain. This multicenter real-world study was to evaluate the predictive value of the RAD51 foci in the IF assay in mCRPC patients with HRRalt and not-altered HRR after the PARP inhibitor treatment. Methods: The multicenter real-world study involved 217 consecutive patients who had received the next-generation sequencing (NGS) and PARP inhibitor treatment from three tertiary referral hospitals. RAD51 foci were quantified in formalin-fixed, paraffin-bedded, untreated tumor specimens by IF. We assessed progression-free survival (PFS) and overall survival (OS) using the Kaplan-Meier method. Potential factors influencing PFS and OS were compared between the treatment arms using Cox proportional-hazards models. Prediction models were established to compare their predictive value in PFS and OS. The prostate-specific antigen (PSA) response and the treatment effect across subgroups were also evaluated. Results: RAD51 foci had significantly better predictive efficacy than the BRCA1/2alt and HRRalt, with a significantly higher predictive power in PFS and OS (RAD51: area under the curve [AUC]: PFS: 0.795; OS: 0.765, BRCA1/2alt: AUC: PFS: 0.669; OS: 0.628, HRRalt: AUC: PFS: 0.635; OS: 0.566). In the overall cohort, median PFS and OS were significantly longer in the RAD51-Low group than the RAD51-High group (PFS: 9.0 vs 2.0 mo; hazard ratio [HR] 0.25, 95% confidence interval [CI] 0.18–0.35; p < 0.01; OS: 25.0 vs 11.0 mo; HR 0.34, 95% CI 0.22–0.43; p < 0.01). In the not-altered HRR cohort, the RAD51-Low group also had significantly longer survival than the RAD51-High group (PFS: 6.5 vs 1.0 mo; HR 0.21, 95% CI 0.11–0.40; p < 0.01; OS: 24.0 vs 8.0 mo; HR 0.32, 95% CI 0.17–0.56; p < 0.01). PSA responses were also significantly higher in the RAD51-Low group than the RAD51-High group. Limitations include its retrospective design and small sample size. Conclusions: In mCRPC patients treated with PARP inhibitors, RAD51 foci demonstrates a superior predictive value for clinical outcomes compared to HRRalt and BRCA1/2alt. RAD51 foci also serve as a pivotal biomarker for extending the PARP inhibitor efficacy to not-altered HRR patients. A large-scale prospective randomized controlled trial is needed to confirm these results.