Importance and background: In high-grade serous ovarian cancer (HGSOC) bevacizumab (bev)/olaparib (ola) maintenance was approved for patients with homologous recombination DNA repair deficiency (HRD+) tumors. Although different methods exist to score genomic instability, DNA quality, tumor cell content, and costs may impair our ability to identify patients that will benefit from treatment.

Patients and method: We analyzed BRCA1 and RAD51C methylation as an HRD determination tool in patients newly diagnosed of HGSOC (n = 519) based on data from the PAOLA-1/ENGOT-ov25 trial phase III prospective trial. Methylation was analyzed using quantitative methylation specific PCR, correlated to HRD scores, PFS and OS.

Results: 67 (12.9 %) were BRCA1 and 25 (4.8 %) were RAD51C methylated. Of the 81 samples with a failed HRD score, 13 were methylated. Methylated samples were HRD+ (mean score [95 % CI]; 65.9 [62.7-69.1] and 53.3 [48.0-58.6]) and almost mutually exclusive of BRCA1&2 mutations. A significant PFS1 benefit independently of methylation ratios was observed in patients with methylated tumors with bev-ola maintenance compared to bev alone (HR=0.49, 95 % CI 0.29-0.83, P = 0.008). An OS benefit was shown for patients defined as "all-HRD" (including methylation) (HR=0.59, 95 % CI 0.41-0.86, P = 0.007).

Conclusions: This study confirms the feasibility and clinical value of BRCA1/RAD51C methylation for predicting response to ola-bev maintenance in newly diagnosed HGSOC. Assessment of methylation in parallel to mutation testing allowed the identification of nearly 85 % of HRD+ samples at low costs. This study suggests that methylation testing could be easily implemented to optimize the selection of patients that benefit from olaparib+bevacizumab maintenance.