Case-case analysis addressing ascertainment bias for multigene panel testing implicates BRCA1 and PALB2 in endometrial cancer

Hereditary endometrial cancer (EC) is most commonly attributed to pathogenic variants in mismatch repair genes. Evidence supports the existence of additional genetic risk factors in the context of multiple cancer diagnoses and/or family history of EC. EC patients (n = 5292) referred for diagnostic multigene cancer panel testing were annotated for presence of a pathogenic gene variant; personal history of prior, concurrent, or subsequent cancer of another type; reported family history of Lynch syndrome or EC. The Pearson chi(2) test was used to assess differences in gene variant prevalence between case sub-groups defined by personal and/or family history of cancer/s, using cases with no family history of Lynch/EC as reference. Another cancer diagnosis was reported for 55% of EC cases. EC cases with a prior and reported family history of Lynch cancer were enriched for variants in MLH1 (p = 3.5 x 10(-7)), MSH2 (p = 3.1 x 10(-7)), and PMS2 (p = .02). Consistent with expectations for a breast cancer gene also predisposing to EC, the variant frequency was increased in EC patients with prior BC and family history of EC for BRCA1 (p = 1.7 x 10(-5)) and PALB2 (p = .0002). Strategic case-case analyses to address cohort ascertainment bias have provided a rationale to direct future studies of candidate hereditary EC genes.