Background: Poly ADP-ribose polymerase inhibitors (PARPi) are being studied as a treatment option for men with mCRPC and deficient homologous recombination DNA repair. While the FDA has granted “breakthrough” designation status to olaparib for use in mCRPC patients with either BRCA1/2 or ATM mutations, it is not known if responsiveness to PARPi may be influenced by the mutation type. Because ATM functions as a sensor of DNA damage rather than a mediator of DNA repair, we hypothesized that patients with ATM mutations may not show robust responses to PARPi therapy compared to those with BRCA1/2 mutations. Methods: We conducted a retrospective analysis of mCRPC patients treated with off-label olaparib at Johns Hopkins Hospital and Mayo Clinic Scottsdale who had somatic or germline mutations in BRCA1/2 or ATM. The primary endpoint was PSA₅₀ response. Secondary endpoints included overall survival (OS) and radiographic/clinical progression-free survival (PFS). Results: 40 men received olaparib for mCRPC of which 17 had pathogenic mutations (5 ATM-mutant, 12 BRCA1/2-mutant). Eleven (65%) had germline and 6 (35%) had somatic alterations. Five (29%) had Gleason sum ≤7 at diagnosis, 4 (24%) had Gleason sum 8, and 8 (47%) had Gleason sum ≥9. Three (18%) had metastatic disease at diagnosis. Mean age at start of olaparib therapy was 65.6 (range, 52-76) years. Prior chemotherapy was received by 59% of men, and prior abiraterone/enzalutamide by 88% of men. Eleven (65%) had extra-osseous metastases at start of therapy. PSA₅₀ responses were achieved by 10/12 men (83%) with BRCA1/2 mutations and 0/5 men (0%) with ATM mutations (Fisher’s exact test, P= 0.003). Men with BRCA1/2 mutations had a median PFS of 12.3 mo versus 3.3 mo for those with ATM mutations (HR 0.14, 95%CI 0.03-0.60; P= 0.008). There was no difference in OS (HR 0.24, 95%CI 0.03-1.173; P= 0.16). Conclusions: Men with mCRPC harboring mutations in BRCA1/2 appear to have better PSA responses and longer PFS with olaparib compared to men with ATM mutations. ATM-mutant mCRPC patients may require alternative therapies (e.g. chemotherapy, ATR inhibitors). Larger prospective studies are needed to confirm or refute these preliminary findings.