Abstract
PURPOSE: Strict clinical criteria used by Medicare for germline testing for Lynch syndrome (LS) could lead to missed diagnoses of hereditary cancer syndromes given variable individual and family phenotypes. The aim of this study was to compare rates and spectrum of pathogenic or likely pathogenic (P/LP) variants in LS and other hereditary cancer genes on the basis of meeting Medicare LS testing criteria. METHODS: Retrospective review of Medicare beneficiaries who had multigene panel testing with an indication of personal or family history of colorectal cancer (CRC) was performed. Ordering providers determined if Medicare LS criteria were met. The results of genetic testing were compared on the basis of whether or not Medicare testing criteria were met. RESULTS: Among 639 Medicare beneficiaries, 495 (77.5%) met testing criteria. Overall rates of P/LP variant identification were similar between those meeting and not meeting testing criteria (18.4% v 11.8%; P = .06). LS was diagnosed more frequently among those meeting testing criteria (10.1% v 4.9%; P = .05). No statistical differences were found in rates of P/LP variant identification for non-LS CRC genes (5.3% v 5.6%; P = .89) or non-CRC genes (4.2% v 2.1%; P = .23). PMS2, MUTYH, and ATM P/LP variants were found at higher rates among those outside of criteria. CONCLUSION: Among Medicare beneficiaries undergoing genetic testing for suspected LS, rates of P/LP variants in actionable cancer genes were similar regardless of whether testing criteria were met. Current testing criteria fail to identify individuals with P/LP variants in PMS2 and other actionable cancer genes. Relaxing LS testing criteria could improve identification of individuals with hereditary cancer syndromes among Medicare beneficiaries.