Abstract
PURPOSE: We sought to evaluate the contribution of mismatch repair (MMR) status to traditional risk stratification algorithms used to predict nodal involvement and recurrence in a large single-institution cohort. METHODS: Endometrioid endometrial cancer (EC) cases from 2014-2020 were evaluated. MMR immunohistochemistry (IHC) was performed universally. Uterine factors assessed in the Mayo criteria were used to retrospectively classify patients as low or high risk for lymphatic spread. Patients were classified according to risk for recurrence using GOG 99 and PORTEC criteria. Associations were evaluated using chi-square and t-tests and contributing factors assessed using logistic regression models. RESULTS: 1,514 endometrioid EC were evaluated; 392 (25.9%) were MMR (MMR) deficient of which 80.4% of MMR defects were associated with epigenetic silencing of MLH1. Epigenetic MMR defects were significantly more likely to be high risk for lymph node (LN) metastasis based on Mayo criteria (74.9% vs 60.6%, p=<0.001) and with the presence of LN metastasis (20.3 vs 10.5%, p=0.003) compared to MMR proficient tumors. Tumors with epigenetic MMR defects were significantly more likely to be classified as high or high intermediate risk using GOG99 and PORTEC criteria. Furthermore, cases with epigenetic MMR defects classified as low or low intermediate risk were significantly more likely to recur (GOG99 p=0.013; PORTEC p=0.008) and independently associated with worse disease-free survival (DFS). MMR status was found to be independently associated with worse DFS (HR 1.90; 95% CI 1.34-2.70; p=0.003) but not overall survival. CONCLUSION: While MMR deficient EC has been associated with poor prognostic features in prior reports; we demonstrate that only epigenetic MMR defects have poorer outcomes. Epigenetic MMR defect were independently associated with lymph node metastasis after controlling for risk criteria. Epigenetic MMR deficiency was found to be an independent predictor of recurrence beyond the factors considered in traditional risk stratification algorithms. Traditional uterine-based risk stratification algorithms may not fully reflect the risk for recurrence in MMR deficient tumors. Consideration should be given to implementing MMR status and MLH1 hypermethylation alongside traditional risk stratification algorithms. Performing MMR IHC on preoperative pathologic specimens may aid in risk stratification and patient counseling.