Real life use of biomarkers of homologous recombination deficiency (HRD) status beyond BRCA to predict the effectiveness of PARP inhibitors in ovarian cancer patients.

10592 Background: Testing for BRCA mutations (BRCAm) and genomic instability can identify epithelial ovarian cancer (OC) patients most likely to benefit from PARP-inhibitors (PARPi). However, current biomarkers of non- BRCA Homologous Recombination Repair (HRR) mutations are insufficient for guiding use of PARPi in the clinic. Despite non- BRCA HRR pathway gene mutations are rare, these patients may benefit from PARPi. Furthermore, recent preclinical findings showed that sensitivity to PARPi could be associated also with mutations in mismatch repair (MMR) genes, although sensitivity in the clinic is not proven. Methods: We report our real-life experience to assess the HRD status beyond BRCAm in newly-diagnosed OC patients, with the use of HRR gene panel and HRD genomic instability tests. After primary diagnosis, tumor and germline BRCA (g BRCA) status were assessed and, if BRCA WT, tumor HRR deficiency status were centrally determined by myChoiceCDx (Myriad) assay. NGS panel evaluating 20 MMR and HRR-genes beyond BRCA1/2 was proposed to patients with significant personal and/or family history of cancer, resulted negative to g BRCA testing. Results: From January 2017 to January 2023, 540 unselected epithelial OC patients, aging 27 to 81, were tested for tumor and g BRCA status; 109 were carriers of germline pathogenic or likely pathogenic variants (PVs) in BRCA1/2 genes (20.2%): 72 in BRCA1 (66.1%) and 37 in BRCA2 gene (33.9%). Additional 19 patients showed somatic PVs confined to tumor samples (3.5%).In the population of 70 BRCA WT patients tested by multigene panel testing, 14 germline PVs in HRR-associated (7.1%) or MMR-associated genes (12.8%) were found, including 5 in MUTYH (35.7%), 2 in ATM (14.3%), 2 in MLH1 (14.3%), 2 in PMS2 (14.3%), 1 in RAD51C (7.14%), 1 in RAD51D (7.14%), and 1 in CHEK2 gene (7.14%).Out of 72 samples analyzed by HRD genomic instability test, 23 cases were identified as HRD positive (31.9%), with a median GIS score of 65.5 (44-83). Median GIS were 35 (1-72) in the 6 cases of non- BRCA mutated tumors (8.3%), 40 (4-82) in the 17 tumor with non- BRCA variant of uncertain significance (VUS) (23.6%), and 27 (2-83) in the WT samples. Median GIS were significantly higher in tumors with PVs/VUS of RAD51 than BRIP1 (55.5 vs 35). Conclusions: HRRm gene panel and HRD genomic instability tests are not interchangeable to study HRR deficiency beyond BRCAm. HRD genomic instability test is effective to identify patients most likely to gain benefit from PARPi, but not for predicting familial risk of cancer. In our population, NGS panel evaluating HRR genes, beyond BRCA1/2, improved the detection rate of HRRm by 7%, with additional finding of MMR germline mutations, and important clinical implications for family members. Prospective data are expected to evaluate the effectiveness of PARPi in these population.