Redefining the EPCAM Phenotype: Deletion Architecture Determines Lynch Syndrome Cancer Risk

Purpose To characterize the cancer phenotype and clinical presentation across molecularly distinct EPCAM deletion subgroups and determine how the phenotype varies with MSH2 involvement. Methods We retrospectively evaluated 662 unrelated probands undergoing hereditary cancer multigene panel testing (MGPT) with isolated EPCAM full-gene deletions (n = 22), isolated EPCAM 3' deletions (n = 96), contiguous EPCAM full-gene deletions extending into MSH2 (n = 64), contiguous EPCAM 3' deletions extending into MSH2 (n = 116), MSH2 gross deletions (n = 155), and MSH2 nonsense/frameshift (ns/fs) variants (n = 209). Lynch syndrome (LS)–associated cancer prevalence, tumor spectrum, and PREMM5 scores were compared across cohorts. Multivariable logistic regression estimated cancer-specific associations relative to a MGPT-negative reference cohort. Results Isolated EPCAM full-gene deletions, comprising 18.6% of observed isolated EPCAM deletions, were not significantly associated with LS-associated cancers. The LS-associated cancer prevalence was 13.6% in this cohort, compared with 54.6% with isolated EPCAM 3' deletions and 64.1%-72.2% in MSH2 -involving cohorts ( P < .001). Relative to MGPT-negative individuals, isolated EPCAM 3' deletions were strongly associated with colorectal cancer (odds ratio [OR], 17.59; P < .001) but not endometrial cancer (OR, 0.76; P = .790). MSH2 -involving cohorts demonstrated strong associations with both colorectal (ORs, 8.50-16.03; P < .001) and endometrial cancers (ORs, 17.31-30.47; P < .001). Conclusion These findings provide new insights into the molecular and clinical heterogeneity of germline deletions involving EPCAM . Expanded detection across the full EPCAM gene and its 3' region distinguishes isolated full-gene deletions, which do not disrupt downstream MSH2 and are not associated with LS phenotypes, from isolated 3' deletions with a colorectal-predominant phenotype, and contiguous EPCAM-MSH2 deletions with an MSH2 -like LS phenotype. Precise delineation of EPCAM deletion boundaries and MSH2 involvement define three clinically relevant risk groups and support more precise risk assessment, counseling, and surveillance.