Background: Advances in next-generation sequencing facilitate the simultaneous evaluation of large numbers of cancer predisposition genes (CPGs) in patients with cancer irrespective of family history or tumour phenotype. Several studies have reported the frequent occurrence of germline mutations in CPGs in patients with discordant cancer types raising the possibility of novel gene-cancer associations. The current study aimed to evaluate the significance of germline mutations in monogenic endocrine tumour genes in individuals with such atypical cancers.
Methods: Twelve monogenic endocrine tumour genes were selected including 11 tumour suppressor genes (e.g. MEN1, VHL, SDHX, NF1) and RET proto-oncogene. The typical tumour spectrum for each gene was defined based on current literature. Pathogenic/likely pathogenic (P/LP) germline variants (as defined using ACMG criteria) were identified in each of the 12 genes in patients with discordant tumour phenotypes from previously reported cancer cohorts (>20 000 patients) and compared to their frequency in a large control population (non-cancer GnomAD cohort (n=134 187 individuals)).
Results: P/LP variants were observed in 10 of the 12 genes evaluated, with SDHA, RET, and NF1 most frequently implicated. However, the frequency of germline P/LP variants in individuals with discordant cancers did not differ significantly to that observed in the control population, although it was notable that for the majority of genes the frequency of P/LP variants in the control cohort exceeded estimates of disease prevalence. Loss of heterozygosity data, where available, did not support an aetiological role in the majority of discordant cancers.
Conclusion: Although P/LP germline variants in monogenic endocrine tumour genes may be observed in patients with discordant cancer phenotypes, in most instances these do not appear to be causal. The higher than expected frequency of pathogenic germline variants in the control cohort suggests either reduced disease penetrance, inaccurate estimates of disease prevalence, or possible variant misclassification.