Probability of detecting germline BRCA1/2 pathogenic variants in histological subtypes of ovarian carcinoma. A meta-analysis.

Abstract

BACKGROUND\nHistology restricted genetic predisposition testing of ovarian carcinoma patients is a topic of debate as the prevalence of BRCA1/2 pathogenic variants (PVs) in various histological subtypes is ambiguous. Our primary aim was to investigate the proportion of germline BRCA1/2 PVs per histological subtype. Additionally, we evaluated (i) proportion of somatic BRCA1/2 PVs and (ii) proportion of germline PVs in other ovarian carcinoma risk genes.\n\n\nMETHODS\nPubMed, EMBASE and Web of Science were systematically searched and we included all studies reporting germline BRCA1/2 PVs per histological subtype. Pooled proportions were calculated using a random-effects meta-analysis model. Subsets of studies were used for secondary analyses.\n\n\nRESULTS\nTwenty-eight studies were identified. The overall estimated proportion of germline BRCA1/2 PVs was 16.8% (95% CI 14.6 to 19.2). Presence differed substantially among patients with varying histological subtypes of OC; proportions being highest in high-grade serous (22.2%, 95% CI 19.6 to 25.0) and lowest in clear cell (3.0%, 95% CI 1.6 to 5.6) and mucinous (2.5%, 95% CI 0.6 to 9.6) carcinomas. Somatic BRCA1/2 PVs were present with total estimated proportion of 6.0% (95% CI 5.0 to 7.3), based on a smaller subset of studies. Germline PVs in BRIP1, RAD51C, RAD51D, PALB2, and ATM were present in approximately 3%, based on a subset of nine studies.\n\n\nCONCLUSION\nGermline BRCA1/2 PVs are most frequently identified in high-grade serous ovarian carcinoma patients, but are also detected in patients having ovarian carcinomas of other histological subtypes. Limiting genetic predisposition testing to high-grade serous ovarian carcinoma patients will likely be insufficient to identify all patients with a germline PV.