PO-343 Targeted resequencing identifies novel and ultra-rare high-impact variants in breast cancer susceptibility genes in an Irish population

Abstract

IntroductionBreast cancer (BC) remains the most common female malignancy worldwide (incidence of 89.7/100,000 women). Pathogenic variants in BRCA1 and BRCA2 account for 3% of all BC cases. A further 25% of BCs demonstrate familial clustering that may be accounted for by variants of reduced penetrance in other BC susceptibility genes. Next-generation sequencing facilitates massively parallel sequencing of multiple genes in a cost-effective manner. Multiple expanded gene panels exist, including high- and moderate risk BC susceptibility genes, as well as loci with weak/putative association with disease. Expanded gene panels may increase diagnostic yield, but also increase identification of variants of uncertain significance.We aimed to investigate the frequency of high impact variants in known or putative BC susceptibility genes in an Irish population using a custom-designed multi-gene panel.Material and methodsTargeted resequencing of 168 gDNA samples (91 patients with BC; 77 unaffected ethnically-matched controls) was performed on an Illumina NextSeq using a Roche-Nimblegen custom 282-gene panel capture. ,GATK best practices, 2016 were implemented for data analysis. Plink1.9 was utilised for confirmation and removal of first-/second-degree relationships within our cohort. 1000Genomes data was used in population stratification to confirm ethnicity. VEP, SnpEff, and Annovar, were used for variant annotation.Results and discussions69 high-impact loss-of-function (LoF) variants were identified in 54 genes (including 51 not typically tested in diagnostic setting). 50 patients and 40 healthy controls were heterozygous for ≥one LoF variant. Frequency data for 34 variants was absent from population databases, while 23 variants were reported as ultra-rare (minor allele frequency ≤1.5 × 10-5). Two LoF variants were identified in BRCA1; 1 in ATM and 1 CHEK2 (c.1100delC). Considering all variants, enrichment in cases v- controls was not significant (p=0.761). 9 LoF variants were observed in genes that are recurrently somatically mutated in BC.ConclusionAfter controlling for ethnicity and relatedness, NGS identified a high-impact variant in 56% of cases and 58% of controls. However, variants in only 3 out of 54 genes in which variants were identified currently impact clinical management.Our results confirm that NGS increases diagnostic yield, but of variants in genes with moderate-weak association with BC. Further analyses are required to determine contribution and clinical utility of testing these variants in patients with BC.