Endometrial Cancer Characteristics and Risk of Recurrence

The Role of Epigenetic Silencing of MLH1

Lindsay E. Borden, MD; Tonja M. Locklear, PhD; Douglas J. Grider, MD; Janet L. Osborne, MD; Erin J. Saks, MD; Fidel A. Valea, MD; David A. Iglesias, MD


Am J Clin Pathol. 2022;157(1):90-97. 

In This Article


In total, 533 women met initial study inclusion criteria, and the following were excluded: 90 for nonendometrioid histology, 102 for incomplete MMR immunohistochemistry screening, 12 for missing confirmatory MLH1 hypermethylation testing, 5 for other synchronous malignancies, 2 for concurrent chemotherapy or hormone therapy, and 6 for both synchronous malignancies and concurrent chemotherapy or hormone therapy. A total of 316 women were then categorized into one of three groups based upon the results of MMR testing. The MMR+ group had 235 (74.4%) patients, MMR– had 10 (3.1%) patients, and hMLH1+ had 71 (22.5%) patients.

Demographic characteristics of each group were compared and are summarized in Table 1. The hMLH1+ patients were significantly older (mean age = 66.9 years) compared with the MMR+ (60.7 years) and MMR– (57.1 years) patients (P < .001). Demographic characteristics were otherwise well matched among groups, with no significant difference in race or tobacco use. Mean body mass index (BMI; kg/m2) trended toward significance, with hMLH1+ and MMR– patients showing a lower BMI than MMR+ patients (hMLH1+ 34.0, MMR– 33.8, MMR+ 37.1; P = .05). hMLH1+ patients were more likely to have a lower BMI category than the other two groups (P = .01).

Pathologic characteristics are also presented and compared in Table 1. Compared with MMR+ and MMR– patients, tumors from hMLH1+ patients were significantly more likely to exhibit grade 3 histology (4.3% vs 10.0% vs 15.5%, respectively, P < .001). The hMLH1+ and MMR– patients were more likely to exhibit presence of LVSI than MMR+ patients (22.9% vs 20.0% vs 9.1%, respectively, P = .007). There was no difference in the depth of myometrial invasion, stage, or lymph node status among the three groups (P = .14, .14, and .07, respectively).

The hMLH1+ patients were more likely to receive adjuvant treatment than MMR+ patients (43.7% vs 26.4%, P = .02), shown in Table 1. There was no significant difference among the type of adjuvant treatment received (P = .41).

Recurrence rates were then compared, shown in Table 2. Early stage tumors were significantly more likely to recur in hMLH1+ patients (hMLH1+ 15.3% vs MMR+ 2.3% vs MMR– 12.5%, respectively, P < .001). Table 2 also describes sites of recurrence. In the hMLH1+ group, four of nine recurrences were distant, as opposed to one of five in the MMR+ group. There was no difference in the rate of recurrence among advanced-stage (III and IV) endometrioid-type tumors for the three groups (MMR+ 23.8% vs MMR– 0% vs hMLH1+ 16.7%, P = 1.00).

A multiple regression analysis was performed to evaluate the contribution of MMR status and several other variables to recurrence in early stage patients Table 3. The adjusted odds ratios (ORs) for recurrence provide the OR for the specific variable while holding the other variables constant. After adjusting for age, grade, LVSI, and adjuvant treatment, hMLH1+ status was significantly associated with recurrence compared with MMR+ status (OR, 5.09; 95% confidence interval [CI], 1.54–16.86; P = .008). MMR– status was not significant in multivariable analysis (P = .12). Age and LVSI were also significant predictors of recurrence (P = .04 and .04, respectively). Grade and adjuvant treatment were not significant predictors of recurrence (P = .50 and .41, respectively).

Since only one early stage MMR– patient had a recurrence, MMR– was not included in the analysis for time to recurrence. The median time to recurrence was not statistically different for either the early stage patients (MMR+ 8.3 months vs hMLH1 + 16.3 months, P = .13) or the advanced-stage patients (MMR+ 22.3 months vs hMLH1 + 22.2 months, P = .58).

Early stage endometrial cancers were stratified into risk categories to compare recurrence rates among the three MMR groups, shown in Table 4. The recurrence rate for the low-risk patients did not differ significantly by group (MMR+ 0% [n = 40] vs MMR– 0% [n = 3] vs hMLH1+ 11.1% [n = 9], P = .23). The low-intermediate-risk hMLH1 + patients had a significantly higher recurrence rate compared with MMR+ and MMR– patients (12.9% [n = 31] vs 2.1% [n = 144] vs 0% [n = 4], respectively, P = .03). The recurrence rate was significantly different for the high-intermediate risk patients as well (MMR+ 6.7% [n = 30] vs MMR– 100% [n = 1] vs hMLH1+ 21.1% [n = 19], P = .04).

Finally, recurrence-free survival (RFS) was significantly reduced in the early stage MLH1+ patients, shown in Figure 3A (P < .001). This is associated with a hazard ratio of 7.40 of recurrence or death compared with the MMR+ group (95% CI, 2.80–21.62; P < .001). The 3-year RFS probability for each group was MMR+ 97.2%, MMR– 87.5%, and hMLH1+ 76.7%.

Figure 3.

Recurrence-free survival, stratified by mismatch repair (MMR) class. A, Early stage. B, Advanced stage.

There was no statistically significant difference in the RFS probability of the MMR+ and hMLH1+ advanced-stage patients, shown in Figure 3B (P = .79). This is associated with a hazard ratio of 0.81 of recurrence or death in the advanced-stage patients compared with the MMR+ group (95% CI, 0.12–3.52). The 3-year RFS probability for each group was MMR+ 66.7% and hMLH1+ 66.7%. In the advanced-stage patients, none of the MMR– patients had a recurrence or death, so they were not included in the survival analysis.