Causality and functional relevance of BRCA1 and BRCA2 pathogenic variants in non‐high‐grade serous ovarian carcinomas

Abstract

AbstractThe identification of causal BRCA1/2 pathogenic variants (PVs) in epithelial ovarian carcinoma (EOC) aids the selection of patients for genetic counselling and treatment decision‐making. Current recommendations therefore stress sequencing of all EOCs, regardless of histotype. Although it is recognised that BRCA1/2 PVs cluster in high‐grade serous ovarian carcinomas (HGSOC), this view is largely unsubstantiated by detailed analysis. Here, we aimed to analyse the results of BRCA1/2 tumour sequencing in a centrally revised, consecutive, prospective series including all EOC histotypes. Sequencing of n = 946 EOCs revealed BRCA1/2 PVs in 125 samples (13%), only eight of which were found in non‐HGSOC histotypes. Specifically, BRCA1/2 PVs were identified in high‐grade endometrioid (3/20; 15%), low‐grade endometrioid (1/40; 2.5%), low‐grade serous (3/67; 4.5%), and clear cell (1/64; 1.6%) EOCs. No PVs were identified in any mucinous ovarian carcinomas tested. By re‐evaluation and using loss of heterozygosity and homologous recombination deficiency analyses, we then assessed: (1) whether the eight ‘anomalous’ cases were potentially histologically misclassified and (2) whether the identified variants were likely causal in carcinogenesis. The first ‘anomalous’ non‐HGSOC with a BRCA1/2 PV proved to be a misdiagnosed HGSOC. Next, germline BRCA2 variants, found in two p53‐abnormal high‐grade endometrioid tumours, showed substantial evidence supporting causality. One additional, likely causal variant, found in a p53‐wildtype low‐grade serous ovarian carcinoma, was of somatic origin. The remaining cases showed retention of the BRCA1/2 wildtype allele, suggestive of non‐causal secondary passenger variants. We conclude that likely causal BRCA1/2 variants are present in high‐grade endometrioid tumours but are absent from the other EOC histotypes tested. Although the findings require validation, these results seem to justify a transition from universal to histotype‐directed sequencing. Furthermore, in‐depth functional analysis of tumours harbouring BRCA1/2 variants combined with detailed revision of cancer histotypes can serve as a model in other BRCA1/2‐related cancers. © 2023 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.