Clinical and molecular biomarkers predicting response to PARP inhibitors in ovarian cancer.

OBJECTIVE:To determine the useful biomarker for predicting the effects of poly-(ADP ribose)-polymerase (PARP) inhibitors in Japanese patients with ovarian cancer. METHODS:We collected clinical information and performed molecular biological analysis on 42 patients with ovarian, fallopian tube, and primary peritoneal carcinomas who received PARP inhibitors. RESULTS:Among the analyzed patients with ovarian cancer, 23.8% had germline BRCA mutation (gBRCAm), 42.9% had homologous recombination repair-related gene mutation (HRRm), and 61.1% had a genomic instability score (GIS) of ≥42. Patients with HRRm had a significantly longer progression-free survival (PFS) than those without HRRm (median PFS 35.6 vs. 7.9 months; p=0.009), with a particularly marked increase in PFS in patients with gBRCAm (median PFS 42.3 months). Similarly, among patients with recurrent ovarian cancer, those with HRRm had a longer PFS than those without HRRm (median PFS 42.3 vs. 7.7 months; p=0.040). Multivariate Cox proportional hazards regression analysis found that performance status and gBRCAm status were independent factors associated with prolonged PFS with PARP inhibitors. In recurrent ovarian cancer, multivariate regression analysis identified platinum-free interval (PFI) in addition to performance status as a significant predictor of PFS. On the contrary, no significant association was observed between PFS and a GIS of ≥42 used in clinical practice. CONCLUSION:We found that HRRm can be a useful biomarker for predicting the effects of PARP inhibitors in treating ovarian cancer and that the PFI can also be useful in recurrent ovarian cancer.