Evaluation of quantitative polymerase chain reaction for detecting BRCA1 or BRCA2 copy number loss in high-grade serous ovarian cancer.

Copy number loss (CNL) of tumour suppressors genes, such as BRCA1/2, are frequent but under-researched tumour driver events. Accurate identification of BRCA1/2 CNL has potential therapeutic implications, particularly in high grade serous tubo-ovarian carcinoma (HGSOC), and cost-effective detection methods are desirable. This study assessed quantitative PCR (qPCR) as a method of BRCA1/2 CNL detection in HGSOC, in two patient cohorts (cohort 1 n = 355, cohort 2 n = 86) and a cell line panel (n = 10) using sequencing-based copy number (CN) assessment as a comparator (panel-based sequencing in cohort 1; whole genome sequencing [WGS] in cohort 2 and the cell line panel). The qPCR sensitivity for detecting BRCA1/2 CNL in both clinical cohorts was poor (BRCA1; 0-11.5%, BRCA2; 10-36.8%). There was frequent co-occurrence of BRCA1 and BRCA2 CNL by qPCR (cohort 1; OR 71.8), which was not mirrored in matched sequencing data. There was a significantly higher CN value for the reference gene RPPH1 in cases with qPCR-identified BRCA1/2 co-loss (cohort 1; RPPH1 median CN 5.2 ± 4.8 versus 2.5 ± 1.3, p < 0.0001). In a cell line panel, there was poor agreement between BRCA1/2 CN measured by qPCR versus WGS, and an additional reference assay (TERT) did not improve agreement. Overall, qPCR did not effectively detect BRCA1/2 CNL in HGSOC, which may be due to CN variation at the reference assay sites. This work underscores the need for caution when utilising qPCR for detecting CNL events, particularly in genomically unstable cancers.