Mismatch repair status in endometrial cancer: detection with immunohistochemical markers and its relationship with clinical outcome

1. Introduction

Endometrial cancer (EC) is the most common gynecological malignancy, with its incidence steadily increasing in developed countries [1]. The majority of patients diagnosed with EC are postmenopausal with a median age at diagnosis of 60 years [2]. Endometrioid endometrial cancer (EEC) is the most prevalent histologic subtypes, accounting for 75–80% of all ECs [3]. The majority of cases diagnosed with EEC are presented as stage I or II, with a 5-year survival rate of around 90%, reflecting a favorable prognosis [4]. Nevertheless, recurrence occurs in a subset of patients with early-stage disease; in a large multicenter cohort of 2974 patients with stage I–II endometrial cancer reporting a recurrence rate of 6.4%, and significantly poorer overall survival among patients who developed recurrence [5]. The prognosis is less favorable in advanced-stage EEC, with a 5-year overall survival rate of 70.4%, and a median progression-free survival of 43 months reported among patients with stage III–IV disease [6].

The standard care for endometrial cancer typically involves surgery, including total hysterectomy, bilateral salpingo-oophorectomy, and lymph node assessment [7]. The conventional approach to prognostic stratification is based on clinicopathologic parameters (e.g., histotype, grade, stage), which are also used to define the therapeutic strategy [8]. A significant diagnostic challenge lies in differentiating patients with early-stage EC who have low-risk disease and a limited likelihood of recurrence from those with high-risk disease who require adjuvant therapy [9]. Consequently, numerous studies have been conducted to identify new prognostic biomarkers that enable the early identification of patients at high risk of recurrence [10].

The majority of endometrial cancer cases are sporadic. Nevertheless, a small subset of hereditary cases arises from germline mutations, predominantly affecting the mismatch repair (MMR) genes [11]. Microsatellite instability (MSI) results from functional inactivation of MMR genes and is characteristic of patients harboring germline mutations associated with Lynch syndrome [12]. Since 2018, the National Comprehensive Cancer Network (NCCN) guidelines have recommended universal testing of all ECs for MSI/MMR-deficient tumors [13]. MSI has been detected in approximately 20–40% of all ECs [14]. A significant number of oncology centers utilize immunohistochemistry (IHC) for MSI/MMR testing, as it is cost-effective and exhibits high sensitivity, specificity, and reproducibility [14, 15]. According to the European Society for Medical Oncology (ESMO) guidelines, MMR-IHC testing is recommended for all patients with EC, regardless of histologic subtype [16]. Furthermore, with the advent and regulatory approval of immuno-oncologic therapies for tumors exhibiting MSI-high or MMR-deficient (MMRd) phenotypes, the identification of these biomarkers has become increasingly important as predictive and therapeutic indicators [17]. However, the prognostic significance of MMR deficiency remains controversial [17]. Several studies have reported improved survival outcomes in MSI-high tumors compared with microsatellite-stable tumors receiving the same adjuvant therapy [18], whereas others have shown no difference or even poorer outcomes in MSI-high cases [19, 20, 21].

The primary objective of the present study is to comprehensively evaluate the relationship between MMR status and clinicopathological variables, with a particular focus on their impact on survival and recurrence outcomes in patients with EEC.

2. Materials and methods

2.2 Immunohistochemistry

No comprehensive molecular classification was performed for the cases; only p53 and MMR-IHC analyses were conducted. The haematoxylin and eosin (H&E) stained slides were re-evaluated by an expert pathologist (GS) for detecting MMR status. Tissue microarray (TMA) blocks were prepared from formalin-fixed, paraffin-embedded tumor samples, and 4 μm-thick sections were immunohistochemically stained using antibodies against MutL Homolog 1 (MLH-1), MutS Homolog 2 (MSH-2), MutS Homolog 6 (MSH-6), and PostMeiotic Segregation increased 2 (PMS-2) (Fig. 1) [23]. Loss of protein expression for MLH-1, MSH-2, MSH-6, or PMS-2 was defined as staining in fewer than 5% of tumor cell nuclei, provided that internal positive controls (normal lymphocytes and/or stromal cells) demonstrated retained nuclear staining [24]. Tumor samples were divided into two groups. (1) MMR-deficient (MMRd) was defined as the loss of at least one of four MMR proteins was lost. (2) MMR-proficient (MMRp) was defined as intact expression of all four MMR proteins in tumors. The monoclonal antibodies used for IHC detection of MMR proteins during routine pathological assessment are shown in Table 1.

Fig. 1.Immunohistochemical staining demonstrating loss of MLH-1 and PMS-2 expression, with retained nuclear expression of MSH-2 and MSH-6 (×200). Detection of all 4 MMR proteins was performed on serial sections obtained from the same tumor tissue. MLH-1: MutL Homolog 1; MSH-2: MutS Homolog 2; MSH-6: MutS Homolog 6; PMS-2: PostMeiotic Segregation increased 2.

Table 1.Monoclonal antibodies for mismatch repair proteins.

Antibody	Clone	Dilution	Company
MLH-1	m1	ready-to-use solution	Ventana
PMS-2	G219-1129	ready-to-use solution	Ventana
MSH-2	sp93	ready-to-use solution	Ventana
MSH-6	A16-4	ready-to-use solution	Ventana
MLH-1: MutL Homolog 1; MSH-2: MutS Homolog 2; MSH-6: MutS Homolog 6; PMS-2: PostMeiotic Segregation increased 2.

3. Results

A total of 100 patients with EEC were included in this study. Overall, MMRp was detected in 71 patients (71%), while loss of at least one MMR protein was observed in 29 patients (29%) (Table 2). Among the 29 MMRd cases, the most common pattern of protein loss was concurrent loss of MLH-1 and PMS-2 in 23 cases (79.3%), followed by loss of MSH-2 and MSH-6 in 2 cases (6.8%), isolated loss of MSH-6 in 2 cases (6.8%), combined loss of MLH-1, PMS-2, and MSH-6 in 1 case (3.4%), and loss of all four MMR proteins in 1 case (3.4%). No cases exhibited isolated loss of PMS-2 expression.

The relationship between MMR protein status and clinicopathologic parameters was shown in Table 3. At diagnosis, the mean age was 57.7 years for MMRd group and 58.3 for MMRp group (p = 0.775). Four patients in MMRd group and ten patients in MMRp group had family history for any cancers (p = 1.000). Deep MI was found higher in MMRp group compared to MMRd group (49.3% vs. 13.8%, p = 0.002). On the other hand, patients with MMRd tumors tend to have had more frequent LVSI (55.2% vs. 28.2%, p = 0.020).

However, no statistically significant differences were observed between the two groups in terms of FIGO stage, tumor grade, tumor size, adnexal involvement, or LN metastasis (p > 0.05 for all comparisons). A total of 97 patients received adjuvant treatment, including chemotherapy and radiotherapy or radiotherapy only; however, this result had no statistically significant for both groups (MMRd: n = 28, (96.6%) vs. MMRp: n = 69, (97.2%); p = 1.000). At the last follow-up, 15 recurrences and 15 deaths were recorded across the entire cohort, with no statistically significant difference in recurrence or mortality between MMRd and MMRp groups (p = 0.219).

The relationships between MMR protein status and survival outcomes were evaluated, as presented in Fig. 2. The median OS was 117 months in the MMR-deficient group and 130 months in the MMR-proficient group, with corresponding 5-year OS rates of 93% in both groups. The median PFS was 112 months for patients with MMRd tumors and 122 months for those with MMRp tumors. Neither OS nor PFS was significantly affected by MMR protein loss (p = 0.154 and p = 0.423, respectively).

Fig. 2.Overall survival (A) and progression-free survival (B) according to MMR protein status. MMR: mismatch repair; MMRp: mismatch repair proficient; MMRd: mismatch repair deficient.

In the subgroup analysis limited to patients with FIGO stage I–II disease (n = 91), deep MI was significantly more frequent in the MMR-proficient group than in the MMR-deficient group (29/64 (45.3%) vs. 4/27 (14.8%); p = 0.008). No significant differences were observed for other clinicopathologic parameters (p > 0.05 for all comparisons).

4. Discussion

In the present study, the MMR protein status was retrospectively assessed using IHC, and its effects on the clinicopathologic features of 100 patients with EEC were examined. Our results demonstrated that MMRd was found in 29% of all patients and MMRd was significantly associated with LVSI (p = 0.020). However, there were no effects on survival outcomes.

Despite the generally favorable prognosis of patients with EEC, approximately 20% present with high-risk disease, which is associated with an increased likelihood of recurrence [10]. Consistent with these reports, the present study identified a high-risk subgroup comprising 24% of cases. The defining features of high-risk EC include high tumor grade, advanced stage, and/or non-endometrioid histology [25]. In recent years, the molecular characterization of EC has developed. Defects in the DNA-MMR system represent one of the most frequent molecular alterations observed in EEC [26, 27]. Tumors with these defects can be easily identified by analysing MSI or loss of MMR protein expression [26, 27].

In the current study, the frequency of MMR-deficient tumors was comparable to previous findings, with MMRd identified in 29% of cases, aligning with the 20–40% range reported in earlier studies [28, 29, 30]. Among the characteristic expression patterns of MMR deficiency, concurrent loss of MLH-1 and PMS-2 is the most prevalent, accounting for approximately 60–70% of cases, followed by combined MSH-2/MSH-6 loss (20–25%), isolated MSH-6 loss (5–10%), and isolated PMS-2 loss (~5%) [17, 30, 31].

Comparing clinicopathological parameters between MMR-proficient and MMR-deficient tumors revealed that MMRd was significantly associated with the presence of LVSI. Similarly, Kim et al. [17] evaluated 163 patients with EEC and reported that MMRd tumors were significantly more likely to exhibit LVSI, which was also correlated with an increased risk of LN metastasis. Consistently, the NRG Oncology/Gynecologic Oncology Group, along with two additional studies, demonstrated a significant association between MMR deficiency and advanced FIGO stage, higher tumor grade, LVSI, and deep MI [20, 32, 33]. In a separate cohort of 206 EC patients, MMRd tumors were likewise associated with LVSI and deep MI and were more likely to have received adjuvant therapy; however, MMR status did not correlate with LN metastasis [34]. In 2021, the Hong Kong study investigated 238 patients with FIGO 2009 EEC and was found that MMRd in EEC was associated with older age, greater tumor size, stage II and high-grade disease, as well as increased need of adjuvant pelvic radiotherapy [35]. In contrast, several studies have reported conflicting findings or no significant differences in prognostic factors based on MMR protein expression. For instance, Ruiz et al. [36] observed no association between MMRd and age, grade, stage, LVSI, and MI in 212 EEC patients. Similarly, Black et al. [37] and Arabi et al. [38] found no correlation with age, race and LVSI; but deep MI and advanced stage were significantly higher in MMRd patients. In line with these reports, the current study found no significant associations between MMR status and variables such as age, stage, grade, adnexal involvement, MI, or tumor size. Although LVSI was more frequently observed in MMRd tumors, the rate of LN metastasis was comparable between groups, and the use of adjuvant therapy did not differ significantly. Decisions regarding adjuvant treatment were made by a multidisciplinary tumor board and individualized for each patient, based on an integrated assessment of prognostic factors including tumor size, grade, MMR status, patient age, LN status, depth of myometrial invasion, and FIGO stage.

A comparative analysis of recurrence and mortality rates between MSI-positive and MSI-negative groups, as reported by Zighelboim et al. [21], revealed no statistically significant difference that aligns with the results of our study. Conversely, Cosgrove et al. [33] reported that recurrence rates were similar between MSI-positive and MSI-negative patients in early-stage EEC; however, a higher recurrence rate was observed in MSI-positive patients with advanced-stage disease. Moreover, the need for adjuvant therapy was greater among MSI-positive patients, even in the early-stage cohort. In a similar context, Lee et al. [35] found that patients with MMRd tumors experienced higher recurrence rates despite a greater likelihood of receiving adjuvant pelvic radiotherapy.

In the current study, we found no significant association between MMRd and survival outcomes in patients with EEC. These findings are in agreement with a previous study of 212 EEC patients by Ruiz et al. [36], which also reported no significant relationship between MMRd and survival. Similarly, two additional studies demonstrate no association between MMR protein status and OS, and PFS [21, 39]. In contrast, a previous meta-analysis suggested that MMRd was associated with poorer PFS and OS; however, the differences did not reach statistical significance [19]. Kim et al. [17] demonstrated that PFS was significantly shorter in patients with MMRd tumors, this result was attributed to the higher incidence of lymph node metastasis, which was identified as one of the independent risk factors. Conversely, Kato et al. [40] found that MMRd tumors were significantly associated with early-stage disease and low-grade (grade I–II) histology and concluded that MMRd could serve as a favorable prognostic marker for OS.

The retrospective design of this study, its single-institution setting, and the relatively small number of patients with MMR-deficient tumors represent key limitations. Additionally, the inclusion of cases with advanced-stage disease and the fact that the study sample spanned a somewhat older diagnostic period (2002–2017) may have influenced the generalizability of the findings. Despite these limitations, the study’s strength lies in its ability to demonstrate a significant association between MMR deficiency and LVSI, without a corresponding increase in LN metastasis.

5. Conclusions

In summary, our findings indicate that MMRd tumors, defined as the loss of at least one MMR protein by immunohistochemistry are significantly associated with LVSI, while showing a reduced incidence of deep MI, and no apparent impact on prognosis or survival outcomes. These results suggest the possibility of distinct biological pathways of tumor progression in MMRd endometrial cancers. The strong association between MMR deficiency and LVSI highlights a potentially aggressive tumor subset, which may benefit from intensified adjuvant therapy and could represent an ideal target population for future immunotherapy trials. These findings underscore the importance of integrating molecular classification, particularly MMR status, with traditional pathological features such as LVSI in order to enhance risk stratification models. Nonetheless, the prognostic value of MSI or MMR deficiency in EEC remains inconclusive, as evidenced by conflicting results across prior studies. Ongoing and future research may enable more accurate risk assessment and support the development of individualized treatment strategies based on both molecular and clinicopathological parameters.

Availability of data and materials

The data underpinning the finding of this study can be obtained from the corresponding author, LA, upon a reasonable request.

Author contributions

MCT—project administration, supervision. LA—writing–review & editing, validation. TS—writing–original draft, investigation, data curation. GS—conceptualization, methodology. GK—resources, investigation, formal analysis. İH—writing–review & editing, visualization. All authors read and approved the final manuscript.

Ethics approval and consent to participate

The study protocol was approved by the institutional Ethics Committee of Ege University Faculty of Medicine (approval number: 22-7T/45). Consents were obtained from all participants included in the study.

Acknowledgment

We would like to thank the members of gynaecological oncology, pathology, radiation oncology and medical oncology team of Ege University Faculty of Medicine for collaboration and supports.

Funding

This research received no external funding.

Conflict of interest

The authors declare no conflict of interest.