Patterns of Poly(ADP-Ribose) Polymerase Inhibitor Use and Overall Survival in United States Veterans With Prostate Cancer.

PARP inhibitors are life-prolonging therapies for men with homologous recombination repair (HRR) deficient mCRPC. Estimates of PARPi benefit in diverse patient populations are limited. Here we describe patterns of PARPi use and outcomes in mCRPC in the nationwide Veterans Health Administration (VA), which includes a high proportion of Black patients, otherwise underrepresented in published PARPi studies. This retrospective cohort included mCRPC patients prescribed a PARPi in the VA. Demographic, clinical, pathologic, and genomic data were collected. Descriptive statistics and survival analyses were conducted, using inverse probability of treatment weighting (IPTW) to adjust for baseline differences by race (Gleason score, prostatectomy, stage, number of therapies, age, prostate-specific antigen (PSA) at diagnosis and at PARPi start). The primary outcome was overall survival (OS), defined as time from PARPi initiation to death or censoring. Other outcomes included percent PSA change, time on PARPi, and genomics-stratified OS. Among 597 Veterans, 79.4% were White and 20.6% Black. Compared to White men, Black men were younger (62 vs. 68 years), had higher PSA at diagnosis (19.0 vs. 8.8 ng/mL), longer time from diagnosis to PARPi initiation (90.6 vs. 69.4 mos), and more commonly received PARPi as ≥4th line therapy (43.9 vs. 31.4%). Median OS from PARPi initiation was 13.7 months (95% CI, 12.6-15.8), with no statistically significant difference between Black and White patients after IPTW (P = .95). The most common HRR variants were BRCA2, ATM, or CDK12. In exploratory analyses, BRCA altered patients had longer OS than non-BRCA (15.9 vs. 12.6 mos). Black patients with CDK12 alterations had longer OS than those without (21.4 vs. 12.2 mos). In this real-world cohort of Veterans with mCRPC, Black patients had higher-risk disease and later PARPi use, pointing to the need to address equity in practice patterns. After adjustment for baseline differences, survival from PARPi start was similar by race. Further research on CDK12 mutations and PARPi response, especially in Black patients, is warranted.