DNA damage repair gene defects combined predict the efficacy of immune checkpoint inhibitors in metastatic castration-resistant prostate cancer: a retrospective cohort study.

BACKGROUND: Immune checkpoint inhibitors (ICIs) yield heterogeneous benefit in metastatic castration-resistant prostate cancer (mCRPC), underscoring the unmet need for robust predictive biomarkers. DNA damage repair (DDR) gene defects encompass mechanistically distinct subtypes-homologous recombination repair (HRR) deficiency, mismatch repair deficiency/microsatellite instability-high (MMRd/MSI-H), and CDK12 inactivation-that may differentially modulate anti-tumor immunity. This study evaluated the combined predictive value of DDR mutation profiles, tumor mutational burden (TMB), and PD-L1 expression for ICI efficacy in mCRPC.

METHODS: In this single-center retrospective cohort study conducted at PKUCare CNOOC Hospital (Tianjin, China), 152 patients with mCRPC who received ICI-containing therapy following ≥2 prior lines of systemic treatment were enrolled (January 2018-December 2023). Of these, 127 (83.6%) underwent tissue-based profiling with a validated 520-gene targeted NGS panel, and 25 (16.4%) were profiled by a validated 168-gene liquid biopsy panel. Patients were classified into four DDR subgroups: HRR-deficient, MMRd/MSI-H, CDK12-mutant, and DDR wild-type. Primary endpoints were objective response rate (ORR) and progression-free survival (PFS). Secondary endpoints included overall survival (OS), disease control rate (DCR), and prostate-specific antigen (PSA) response. Kaplan-Meier analysis, multivariate Cox regression, and ROC analysis were performed.

RESULTS: Among 152 patients, 101 (66.4%) harbored DDR alterations: 58 HRR-deficient (38.2%), 24 MMRd/MSI-H (15.8%), and 19 CDK12-mutant (12.5%). ORR was highest in MMRd/MSI-H (50.0%), followed by CDK12-mutant (26.3%), HRR-deficient (22.4%), and DDR wild-type (9.8%; p < 0.001). Median PFS was significantly longer across all DDR-mutant groups versus DDR wild-type (MMRd/MSI-H: 8.1 months; HR 0.31, 95% CI 0.18-0.53; p < 0.001). On multivariate analysis, DDR mutation subtype, TMB ≥10 mut/Mb, and liver metastasis were independent predictors of PFS. A composite biomarker integrating DDR status, TMB, and PD-L1 achieved the highest predictive accuracy (AUC 0.83, 95% CI 0.75-0.91). irAE rates did not differ significantly across subgroups (p = 0.614).

CONCLUSION: DDR gene mutation profiling, integrated with TMB and PD-L1 expression, provides clinically actionable stratification of mCRPC patients for ICI therapy. A composite DDR/TMB/PD-L1 biomarker panel warrants prospective validation in dedicated clinical trials.