Molecular classification and adjuvant treatment in endometrioid endometrial cancer with microcystic elongated and fragmented (MELF) invasion pattern

Abstract

Objective

To assess the characteristics, molecular classification, and role of adjuvant treatment in patients with endometrioid endometrial cancer (EEC) and microcystic elongated and fragmented (MELF) myometrial invasion pattern.

Methods

This study included patients who were diagnosed with EEC with a MELF invasion pattern and underwent surgery from January 2019 to December 2023. We analyzed molecular classification, clinicopathological characteristics, and prognostic outcomes, including recurrence and adjuvant therapy.

Results

Out of 529 patients, 84 (15.9%) exhibited the MELF pattern, with 1 (1.2%) classified as POLE-mutation, 19 (22.6%) as mismatch repair deficient, 53 (63.1%) as no specific molecular profile, and 11 (13.1%) as p53-mutant subtype. Fifty (59.5%) patients with MELF invasion pattern received adjuvant chemotherapy (CT) ± external beam radiation therapy (EBRT), 19 (22.6%) received EBRT only, and 15 (17.9%) received no adjuvant treatment. Receiving adjuvant therapy was significantly associated with the risk level defined by the ESMO guideline for endometrial cancer (p = 0.002). With a median follow-up of 26 months (range: 1–59), the progression-free survival at 3-years for the MELF invasion patients was 91%. Seven patients with the MELF pattern experienced recurrence, of whom one was in stage IA (low risk, local recurrence) and did not receive any additional treatment, two were in stage IB (intermediate / high-intermediate risk, distant recurrence) and received EBRT only and the remaining four were in stage III to IV and had distant recurrence despite receiving adjuvant chemotherapy with or without EBRT. Among 43 intermediate- and high-intermediate-risk EEC patients, receiving CT ± EBRT was significantly associated with better DFS than without CT (p = 0.047).

Conclusion

The presence of the MELF pattern in EEC should be incorporated into decision-making regarding adjuvant therapy. The use of adjuvant treatment should be tailored based on histology and molecular type.

Introduction

Endometrial cancer is the fourth most prevalent malignancy among women in the United States, with its incidence increasing by approximately 1% annually since the mid-2000s [1]. Endometrioid endometrial carcinoma (EEC), constituting 80% of cases, generally has a favorable prognosis due to early-stage diagnosis, lower grade, and favorable molecular subtypes [2]. However, a subset of patients with low-grade EEC faces difficult-to-treat recurrences. The presence of microcystic, elongated, and fragmented (MELF) patterns of myometrial invasion is frequently observed in these low-grade cases. Researchers have established that MELF patterns are significantly associated with adverse histological features, including deep myometrial invasion, lymph node metastasis, and lymphovascular space invasion (LVSI). While the prognostic implications of MELF patterns in EEC are still debated, some studies suggest a correlation with poor outcomes [3,4,5]. Nonetheless, the necessity of additional treatment for EEC patients exhibiting MELF invasion remains uncertain.

Recent advancements in molecular classification have enhanced risk stratification and guided adjuvant treatment in endometrial cancer [6]. The Cancer Genome Atlas (TCGA) study delineated four distinct subtypes, each with unique prognostic implications [7]. This led to the development of an integrated genomic-pathologic classification algorithm, adaptable to varying resource settings, thereby facilitating the practical application of molecular testing in endometrial cancer management [8]. The World Health Organization (WHO) now categorizes endometrial cancer into four molecular subgroups: POLE-mutant (POLEmut), mismatch repair-deficient (dMMR), no specific molecular profile (NSMP), and p53-mutant (p53mut). However, limited research has explored the correlation between the MELF pattern and these molecular subgroups.

Despite advances in adjuvant therapy for EEC, a consensus on the optimal treatment strategy remains elusive, with protocols varying based on tumor stage, histology, and molecular classification. While risk stratification, informed by histopathological and molecular profiling, is acknowledged for its effectiveness in guiding adjuvant therapy, the MELF pattern has not been incorporated into these frameworks [6, 9]. The association of the MELF pattern with adverse pathological features and its prognostic significance suggests a potential impact on therapeutic outcomes. Therefore, the inclusion of MELF in risk assessment models warrants further investigation. This study aimed to investigate the characteristics, molecular classification, and response to adjuvant treatment in patients with EEC exhibiting the MELF pattern.

Methods

Patients

Between January 2019 and December 2023, this study encompassed all patients diagnosed with EEC exhibiting the MELF invasion pattern at Peking University First Hospital, who underwent surgical staging. Exclusion criteria were applied to cases with incomplete clinical data or synchronous malignancies at additional sites. Molecular classification was conducted for the study cohort using the PromiseE protocol, a routine procedure in our center for endometrial cancer patients [9]. Endometrial carcinoma cases were assessed and classified based on the 2014 WHO criteria, with tumor staging following the 2009 International Federation of Gynecology and Obstetrics (FIGO) guidelines. Molecular classification was performed for all participants using the integrated histo-molecular diagnostic algorithm as described in the WHO Classification of Tumors, 5th edition, Female Genital Tumors [10]. Written informed consent was obtained from all participants, and the study protocol was approved by the institutional ethics committee.

Histopathology and definition of the MELF pattern

Two pathologists conducted a comprehensive review of cases to assess MELF pattern invasion, histologic subtype, tumor grade, lymphovascular space invasion, myometrial invasion, and lymph node metastasis. The MELF pattern was identified according to the criteria established by Murray et al. and evaluated as previously described [11, 12]. The histological features indicative of the ‘microcystic, elongated, and fragmented’ (MELF) glands included invasive glands with cystic-dilated or slit-like structures, lined by flattened, endothelial-like epithelium or squamous tumor cells with eosinophilic cytoplasm. These glands often contained intraluminal tufts or fragmented cells, as well as small groups of isolated tumor cells.

Statistical analysis

Patient data, including age, body mass index (BMI), tumor size, pathological stage, molecular subtype, and treatment details, were extracted from digital medical records. Disease-free survival (DFS), defined as the interval from initial diagnosis to the detection of recurrence, was analyzed using the Kaplan-Meier method. Statistical significance was set at p < 0.05. Categorical variables were compared using the chi-square test. Data analysis was performed with SPSS software, version 24.0 (IBM, Armonk, NY). The ggplot2 and base plot package of R software (R Core Team, New Zealand) were used for graphing.

Results

Tumor characteristics and molecular classification

A total of 529 patients with endometrioid endometrial cancer were included in this study. Among them, 84 patients (15.9%) exhibited MELF invasion, the population we aimed to characterize in detail. As shown in Table 1, the mean age was 59.8 ± 8.27 years, the median BMI was 25.9 (range 24.1–28.4), and the median tumor size was 3.51 cm (range 2.5–4.8 cm). All patients had endometrioid carcinoma, which was predominantly low grade, with 84.2% classified as grade 1 or 2, and 15.5% as grade 3. Lymphovascular invasion was present or suspected in 50% of cases. Regarding uterine risk factors, 51.2% of patients had ≥ 50% myometrial invasion, and 11.9% had cervical stromal invasion. All patients underwent surgical staging. The FIGO stage distribution was as follows: 41.7% stage IA, 34.5% stage IB, 8.3% stage II, 13.1% stage III, and 2.4% stage IV. Lymph node metastases were identified in 9 patients (10.7%), with macrometastasis, micrometastasis, and isolated tumor cells (ITCs) accounting for 7.1%, 2.3%, and 1.2%, respectively. Notably, patients with ITCs were not considered node-positive and therefore were not classified as stage IIIC1.

POLE mutations were identified in one case of endometrioid endometrial carcinomas (EECs) with the MELF pattern, specifically a POLE exon 9 mutation c.857 C > G (p.P286R). Among the 19 cases of EEC classified as mismatch repair deficient (dMMR), 15 exhibited co-deletion of MLH1 and PMS2, and one displayed co-deletion of MSH2 and MSH6. One case showed isolated PMS2 loss. Two cases were mismatch repair proficient, and PCR analysis confirmed microsatellite instability-high (MSI-H) status, with allelic shifts in two or more loci (D2S123, D5S346, BAT25, BAT26, and D17S250). Abnormal p53 expression was detected in 11 cases: six cases presented with diffuse nuclear overexpression, while five cases displayed a null phenotype, indicating a p53 aberrant mutation. According to the fifth edition of the WHO molecular classification, the 84 EEC cases with the MELF pattern were classified as follows: 1.2% POLEmut, 22.6% dMMR, 13.1% p53mut, and 63.1% NSMP (Table 1).

Table 1 Patients and tumor characteristics of patients with MELF pattern

MELF pattern and adjuvant treatment

As illustrated in Figs. 1 and 50 patients (59.5%) with MELF-positive endometrioid endometrial carcinoma received adjuvant chemotherapy with or without pelvic radiation therapy, 19 patients (22.6%) received external beam radiation therapy alone, and 15 patients (17.9%) did not receive any adjuvant treatment.

Fig. 1

Treatment given to patients with the MELF pattern

Our institution tailored adjuvant chemotherapy recommendations based on patients’ molecular classifications and other risk factors. Patients with higher ESMO-risk levels received significantly more chemotherapy with or without radiation therapy (p = 0.002) compared to low-risk patients (Fig. 2). Specifically, 90.9% of patients with p53mut, 58.5% of patients with NSMP, and 47.3% of patients with dMMR subtype received adjuvant chemotherapy with or without radiation therapy. No patients with POLEmut received any adjuvant therapy. Additionally, 74.2% of stage IA patients, 82.7% of stage IB patients, and 95% of patients with stage III to IV disease received adjuvant chemotherapy and/or external beam radiation therapy (Table 2).

Table 2 Clinicopathological characteristics and adjuvant treatment of EECs with MELF pattern

Fig. 2

Adjuvant treatment in relation to the ESMO-risk level (CT, chemotherapy; RT, radiotherapy)

Recurrence and survival

With a median follow-up period of 26 months (range: 1–59), the disease-free survival (DFS) rate at 36 months for patients with MELF invasion was 91%. Disease recurred in 7 patients (8.3%), as detailed in Table 3. One patient (14.3%) experienced an isolated vaginal vault recurrence, while the remaining 6 patients (85.7%) had distant recurrences. The median time to recurrence was 34 months (range: 3–56 months). Among the seven patients with recurrent disease, two had the p53 abnormal type, four had the NSMP type, and one had the MMR deficient type. A trend towards a reduced DFS was observed in patients with p53 mutations, although the difference was not statistically significant (Fig. 3A). Four of these patients were classified as high-risk, while the remaining three were categorized as low-risk, intermediate-risk, or high-intermediate-risk. All high-risk patients received chemotherapy with or without radiotherapy. Despite undergoing adjuvant radiotherapy, both intermediate- and high-intermediate-risk patients experienced relapses. The low-risk patient with local recurrence did not receive adjuvant therapy. Among the 43 intermediate- and high-intermediate-risk EEC patients, those who received chemotherapy with or without radiotherapy had significantly better DFS compared to those who did not receive chemotherapy (p = 0.047). (Fig. 3B)

Table 3 Characteristics of patients with recurrence

Fig. 3

disease-free survival (DFS) of EEC with MELF invasion in (A) all cases according to molecular subtype and (B) intermediate-and high-intermediate-risk cases according to adjuvant chemotherapy

Discussion

We present the clinical outcomes of a real-world cohort consisting of 84 patients with MELF invasion endometrioid endometrial carcinoma (EEC). Our study identifies four molecular subtypes of EEC with MELF patterns based on current pathological methods, with most cases classified as NSMP. Approximately 63% (27 out of 43) of patients with intermediate- or high-intermediate-risk endometrioid endometrial carcinoma received chemotherapy, either alone or in combination with radiotherapy. The addition of adjuvant chemotherapy significantly improved disease-free survival (DFS) in these patients.

Multiple studies have identified that MELF pattern is prevalent in endometrioid endometrial carcinoma (EEC) and is significantly correlated with adverse prognostic indicators. In a cohort study conducted by Okcu et al., the MELF pattern was observed in 29.86% of the patients, showing a significant association with lymphovascular space invasion (LVSI), advanced stage, infiltrative growth patterns, and necrosis [3, 12]. Similarly, research by van den Heerk et al. demonstrated that MELF invasion in endometrial cancer is linked with aggressive features such as deep myometrial invasion, LVSI, and high tumor grade, suggesting a more advanced disease state [4]. Additionally, Altindag et al.‘s findings indicated that MELF is associated with LVSI, deep myometrial invasion, and lymph node metastasis (LNM), underscoring its association with poorer clinical outcomes in EEC [13]. Consistent with the existing literature, our study revealed that up to 51.2% of patients with MELF infiltration exhibited deep myometrial invasion, and 50% had LVSI. Notably, the incidence of lymph node metastasis in our cohort was low, with only 9 patients (10.74%), of whom two had micrometastases and one had isolated tumor cells (ITCs). This may be attributable to the fact that about 50% of the patients in our study underwent sentinel lymph node biopsy only. Other investigations have reported that micrometastases and ITCs are common in lymph node metastases in patients with MELF, recommending additional sectioning and immunohistochemical (IHC) analysis of the resected lymph nodes. Furthermore, Rabe et al. concluded that cytokeratin IHC staining could effectively detect malignant cells in the sentinel lymph nodes of patients with endometrial cancer and MELF [14].

Previous studies have suggested an association between MELF and dMMR [15]. However, Baohui et al. reported that the most common subtype among 77 patients with MELF positivity was NSMP (no specific molecular profile), representing approximately 61.0% (47 of 77 cases), while dMMR accounted for only 20.8% [16]. Another study based on PORTEC-1, and − 2 trials found that 85.6% of MELF invasion-positive tumors were NSMP endometrial carcinoma, with 92.2% being CTNNB1 wild type and 24.4% exhibiting KRAS mutations [4]. These findings align with our results, where 63.1% of MELF-positive patients were classified as NSMP, and 22.6% were identified as dMMR.

MELF infiltration has been linked to worse survival outcomes. A comprehensive examination of 464 cases of FIGO Grade 1 endometrioid endometrial adenocarcinomas aimed to elucidate the relationship between the MELF invasion pattern, lymph node metastases, and recurrence rates. The findings indicated a significant correlation between MELF invasion and both lymph node metastases and shorter time to recurrence, underscoring its predictive value for disease progression [17]. Alndag’s analysis further revealed that cases of EEC with the MELF pattern exhibited notably lower overall survival (OS) and disease-free survival (DFS) rates compared to those without the pattern [13]. Other researchers have identified the MELF pattern as an independent prognostic factor for progression-free survival and overall survival in endometrial cancer, reinforcing its predictive importance [4]. Recent studies by Qi et al. demonstrated that the combination of MELF and tumor budding serves as a histological marker for tumor aggressiveness and can independently predict adverse outcomes [5].

Adjuvant therapy recommendations for early-stage endometrial cancer have traditionally been based on clinicopathologic factors. However, the integration of molecular markers such as POLE mutations and microsatellite instability has refined risk assessment. The 2020 ESGO/ESTRO/ESP consensus introduced different prognostic risk groups by incorporating molecular alterations alongside clinicopathologic factors, aiming to tailor adjuvant therapy for endometrial cancer patients [6]. According to these guidelines, low-risk patients may require no adjuvant therapy, high-risk patients need chemotherapy, and intermediate- or high-intermediate-risk patients may benefit from radiotherapy. Nonetheless, there are no specific recommendations for adjuvant therapy in MELF-positive patients. Our study suggests that intermediate- or high-intermediate-risk patients with MELF-positive endometrial cancer may benefit from chemotherapy, likely due to their propensity for distant metastasis. Supporting this, one study found a significantly higher risk of distant metastasis in MELF-positive patients compared to those without this pattern, indicating a poorer prognosis [4]. Another study reported a higher incidence of distant recurrences in patients with the MELF pattern of myometrial invasion (71.4%) compared to those without it (42.9%) [18]. Our findings also indicate that most recurrences in MELF-positive patients were distant metastases, suggesting that systemic chemotherapy could be associated with better outcomes.

Additionally, our study identified a low-risk patient who developed a localized recurrence without receiving any adjuvant therapy. Previous studies have also noted a significantly higher recurrence risk in low-risk patients when combined with MELF. A large cohort study by He et al. revealed that participants with POLE-mutated tumors and MELF pattern invasion had a 15.1-fold increase in tumor recurrence or progression risk compared to those with POLE-wild type tumors [11].

Four high-risk patients in our study developed distant metastatic recurrence despite adjuvant chemotherapy and radiotherapy. Molecular analysis has revealed that MELF-positive tumors are associated with alterations in the PI3K/AKT pathway, which is implicated in cancer progression and treatment resistance [4]. Additionally, overexpression of L1CAM in MELF glands has been found to predict lymph node spread in low-grade endometrioid carcinomas, providing valuable insights for patient management [19]. Tahara et al. reported strong expression of programmed death-ligand 1 (PD-L1) at the invasive front of the MELF pattern and concluded that programmed cell death protein 1/PD-L1 immunotherapy has demonstrated significant therapeutic efficacy in managing patients with grade 1 endometrial cancer and the MELF pattern [20]. Therefore, a combination of molecular testing may be necessary for high-risk MELF-positive patients to identify additional therapeutic targets.

This study is pioneering in exploring the impact of adjuvant treatment on endometrial cancer patients with MELF invasion, addressing a critical gap in the literature. Additionally, we provide valuable insights into the distribution of molecular subtypes among these patients, contributing to a more comprehensive understanding of the disease’s heterogeneity and potentially informing more targeted therapeutic strategies. However, several limitations must be acknowledged. The retrospective design introduces potential biases and limits causal inference. The small sample size and abbreviated follow-up period could have compromised the generalizability of our findings and limited the statistical power. Furthermore, the lack of standardization in adjuvant treatment for patients with MELF invasion may have led to variability in treatment outcomes. To bolster the reliability of our conclusions, we intend to extend the follow-up period in subsequent studies and to initiate well-designed prospective cohorts, thereby amassing a more extensive dataset that will facilitate a more rigorous analysis.

Patients with endometrial cancer exhibiting the MELF pattern should be closely monitored, and MELF status should be routinely included in pathology reports to better understand its significance. The necessity of adjuvant chemotherapy for MELF-positive patients requires validation in future prospective studies.

Conclusion

The incorporation of the MELF pattern in Endometrioid Endometrial Carcinoma (EEC) into the decision-making process concerning adjuvant therapy is imperative. The application of adjuvant treatment ought to be customized in accordance with histological and molecular types.