340 Performance characteristics of screening strategies to identify Lynch syndrome in women with non-serous and non-mucinous ovarian cancer

Objectives The incidence of Lynch syndrome (LS) and the optimal screening strategy has not been determined for women with ovarian cancer (OC). We compared the performance characteristics between immunohistochemistry (IHC) for mismatch repair (MMR) proteins, microsatellite instability (MSI) testing and family history. Methods Women with non-serous/mucinous OC were prospectively recruited from three cancer centers in Ontario, Canada. Tumors were assessed for defects in MMR by IHC and MSI testing. All women completed a family history assessment and underwent germline testing for LS. The performance characteristics were compared between the screening strategies compared to germline result. Results Of 215 women, 185 had OC alone (86%) and 30 had synchronous OC and endometrial cancers (14%). Germline data was available for 189 women (88%). Twenty-eight were MMR deficient (MMRd) by IHC (13%; N = 215), one (0.5%) IHC equivocal and 19 (12%; N=162) were MSI-high (MSI-H). Of those with MMRd, 11 had germline mutation (39%; N=28). In total, thirteen women (7%; N = 189) had germline mutations: 3 MLH1, 7 MSH6, 1 MSH2 and 2 PMS2. Combined IHC and MSI testing after excluding MLH1 hypermethylated cases was the best screening strategy with sensitivity of 92.3%, specificity of 97.7%, PPV of 75% and NPV of 99.4%. Family history had the lowest performance characteristics with sensitivity of 55%. Conclusions The most superior screening strategy to identify women with LS in this population is combined IHC and MSI testing after MLH1 hypermethylation testing and should be considered as standard of care.