Clinicopathological characteristics of endometrial carcinomas according to DNA mismatch repair protein status

Abstract

DNA mismatch repair protein deficiency (MMRd) in endometrial carcinoma is associated with the risk of Lynch syndrome and response to immune checkpoint inhibitors. It is also related to microsatellite instability and corresponds to a molecular subtype of endometrial tumor with an unclear prognosis. Here, we evaluated the clinicopathological characteristics and prognosis of 312 consecutive endometrial carcinoma cases submitted to complete surgical staging at a single institution. We compared MMRd and mismatch repair protein–proficient (MMRp) tumors and examined the effects of the MMR protein loss type (MLH1/PMS2 vs. MSH2/MSH6) and influence of L1CAM and p53 expression. The median follow-up period was 54.5 (range, 0–120.5) months. No difference was observed between MMRd [n = 166 (37.2%)] and MMRp [n = 196 (62.8%)] cases in terms of age, body mass index, FIGO stage, tumor grade, tumor size, depth of myometrial infiltration, or lymph node metastasis. More MMRd than MMRp tumors had endometrioid histology (87.9% vs. 75.5%) and despite MMRd had more lymphovascular space invasion (LVSI; 27.2% vs. 16.9%), they presented fewer recurrences and no difference in lymph node metastasis and disease-related death. Relative to those with MLH1/MSH6 loss, tumors with MSH2/MSH6 loss were diagnosed at earlier FIGO stages, were smaller, and had less ≥50% myometrial invasion, LVSI and lymph node metastasis. Outcomes, however, did not differ between these groups. L1CAM positivity and mutation-type p53 expression were more common in MMRp than in MMRd tumors and did not differ between the MLH1/PMS2 and MSH2/MSH6 loss groups. In the entire cohort, L1CAM and mutation p53 expression were associated with worse prognosis, but only non-endometrioid histology, FIGO stage III/IV, and deep myometrial infiltration were significant predictors. In the subgroup of endometrioid carcinomas, only FIGO stage III/IV was associated with poor outcomes. The risk of lymph node metastasis was associated with tumor size, non-endometrioid histology, and multifocal LVSI. For MMRd tumors, only tumor size and myometrial invasion depth were predictive of lymph node involvement. In our cohort, MMRd tumors were associated with greater recurrence-free, but not overall, survival. The precise identification of MMRd status, present in a substantial proportion of endometrial cancer cases, is a challenge to be overcome for proper patient management. MMRd status serves as a marker for Lynch syndrome, and a significant number of these tumors are high risk and candidate to immunotherapy.