Abstract
Objective
To compare the prognostic performance of the International Federation of Gynecology and Obstetrics (FIGO) 2023 staging system (FIGO 2023) and the 2025 European Society of Gynaecological Oncology (ESGO), European Society for Radiotherapy and Oncology, and European Society of Pathology (ESGO 2025) risk classification in endometrial cancer (EC).
Methods
We retrospectively analyzed 296 patients who underwent primary surgical treatment for EC at a single tertiary institution. All cases were reclassified according to FIGO 2023 and ESGO 2025 using clinicopathological and molecular data. Recurrence-free survival (RFS) and disease-specific survival (DSS) were analyzed using the Kaplan-Meier method. Prognostic performance was assessed using the Akaike information criterion (AIC) and Harrell’s concordance index (C-index) with bootstrap-derived 95% confidence intervals (CIs).
Results
Kaplan-Meier analysis demonstrated significant differences in RFS and DSS according to both ESGO 2025 and FIGO 2023 (log-rank P<0.001). For recurrence prediction, ESGO 2025 showed better model fit than FIGO 2023, with a lower AIC value (410.849 vs. 428.644), and demonstrated a numerically higher C-index (0.8657; 95% CI, 0.8266–0.9114 vs. 0.8519; 95% CI, 0.8110–0.9005). In contrast, for disease-specific mortality prediction, FIGO 2023 showed better model fit than ESGO 2025, with a lower AIC value (143.758 vs. 148.109), and demonstrated a numerically higher C-index (0.9360; 95% CI, 0.9028–0.9769 vs. 0.9063; 95% CI, 0.8624–0.9628).
Conclusion
ESGO 2025 and FIGO 2023 demonstrated distinct prognostic characteristics according to the clinical endpoint, suggesting complementary roles for molecularly integrated risk stratification and anatomical staging in the management of EC.
Keywords: Endometrial cancer, Prognosis, Survival
Introduction
Endometrial cancer (EC) is one of the most common malignancies affecting women, with more than 417,000 new cases diagnosed worldwide in 2020, making it the sixth most frequently diagnosed cancer among women [1]. In particular, the incidence of EC has steadily increased in many countries over recent years [2], largely attributable to the rising prevalence of obesity and other risk factors, such as early menarche, late menopause, nulliparity, and diabetes mellitus [3]. This epidemiological trend is of particular concern because EC encompasses a heterogeneous group of tumors with diverse clinical behaviors [4,5]. While many patients present with early-stage, low-grade (LG) disease that is potentially curable with surgery, a substantial proportion present with high-risk features or aggressive histological subtypes associated with poor prognosis [6,7]. Therefore, effective prognostic stratification of EC is essential for guiding appropriate treatment strategies.
A few years ago, the International Federation of Gynecology and Obstetrics (FIGO) revised its staging system to maintain its clinical relevance and accuracy. The most recent FIGO staging system, introduced in 2023, represents a substantial departure from its traditional role, which was based solely on anatomical extent [8]. For the first time, the FIGO 2023 staging system (FIGO 2023) incorporates molecular markers, such as polymerase epsilon (POLE) and TP53 mutations, alongside histopathological features. This shift positions FIGO 2023 as a more comprehensive tool for risk assessment in EC. However, despite its potential utility in risk stratification, the FIGO staging system is not directly used to guide adjuvant therapy decisions in clinical practice, unlike the guidelines developed by the European Society of Gynaecological Oncology (ESGO), European Society for Radiotherapy and Oncology (ESTRO), and European Society of Pathology (ESP).
In 2025, the ESGO-ESTRO-ESP updated the guidelines published in 2021 by incorporating the revised FIGO 2023 staging system and the substantial body of evidence on the management of EC [9]. These updated guidelines reflect an improved understanding of the complex nature of the various types of EC and their underlying biological behavior. Importantly, additional factors beyond those used in FIGO 2023 (namely, estrogen receptor [ER] expression and mismatch repair [MMR] protein status) further define the risk categories. In the 2025 ESGO-ESTRO-ESP risk classification (ESGO 2025), prognostic risk groups are defined by the estimated overall 5-year risk of recurrence as follows: low-risk (<8%), intermediate-risk (8–14%), high-intermediate-risk (15–24%), and high-risk (≥25%).
Although ESGO 2025 and FIGO 2023 are largely aligned, key differences can be identified. Importantly, FIGO 2023 adopts a more granular framework, dividing EC into 21 stages, including two molecular substages, compared with the five risk categories used in ESGO 2025. Furthermore, FIGO 2023 incorporates only the polymerase epsilon-mutated (POLEmut) and p53-abnormal (p53abn) molecular subtypes. In contrast, ESGO 2025 includes not only POLEmut and p53abn tumors but also further subdivides the no-specific-molecular-profile (NSMP) group according to histologic grade and ER expression, while also incorporating the mismatch repair-deficient (MMRd) group. In this context, a direct comparison between ESGO 2025 and FIGO 2023 is both feasible and clinically warranted.
In this study, we examined the relationship between ESGO 2025 and FIGO 2023 in EC. We further evaluated whether ESGO 2025 offers superior prognostic discrimination to FIGO 2023, with the aim of informing the refinement of risk stratification.
Materials and methods
1. Study population
The study protocol was reviewed and approved by the Institutional Review Board (IRB) of Korea University Guro Hospital (IRB number: 2024GR0367). The requirement for informed consent was waived because of the retrospective nature of the study and patient data were anonymized to ensure confidentiality.
The study population consisted of women who underwent surgical treatment for EC at Korea University Guro Hospital between January 2010 and March 2023. The standard surgical procedure involved total hysterectomy and bilateral salpingo-oophorectomy. Ovarian preservation was performed in selected premenopausal patients with stage IA1–2, grade 1, endometrioid-type disease, and grossly normal-appearing ovaries. Pelvic and/or para-aortic lymphadenectomy was performed in most patients, except those with stage IA1, grade 1–2, endometrioid-type disease. Adjuvant treatment was administered according to the National Comprehensive Cancer Network guidelines until 2020. Following the publication of the ESGO-ESTRO-ESP guidelines in 2021, treatment decisions were guided by these recommendations.
2. Molecular classification
All patients were classified into four molecular subgroups-POLEmut, p53abn, MMRd, and NSMP-according to the Proactive Molecular Risk Classifier for Endometrial Cancer (ProMisE) classification [10]. In addition, the NSMP group was further stratified into NSMP/LG, estrogen receptor-positive, and NSMP/high-grade, estrogen receptor-negative subgroups based on histologic grade and ER expression, in accordance with ESGO 2025.
Immunohistochemistry was performed on biopsy and surgical specimens to determine MMRd, p53abn, and ER status. Immunohistochemical staining for MMRd status was performed using a BOND-III Automated Staining System (Leica, Wetzlar, Germany) with the following antibodies: MLH1 (1:100, ES05, Novocastra), MSH2 (1:400, G219–1129, Novocastra), PMS2 (1:100, MRQ-28, Cell Marque), and MSH6 (1:200, 44, Cell Marque). MMR staining results were classified as either retained or lost nuclear expression. Retained nuclear expression was considered mismatch repair-proficient, whereas loss of nuclear expression was considered MMRd. Immunohistochemical staining for p53 (1:200, DO-7, Novocastra) was also performed. p53 staining was classified as abnormal (mutant-type) when more than 80% of tumor cells exhibited diffuse strong nuclear staining, unequivocal cytoplasmic staining, or complete absence of nuclear and cytoplasmic staining. Cases were classified as normal (wild-type) if tumor cells exhibited any degree of nondiffuse nuclear staining (<80%). ER expression was also assessed by immunohistochemistry. Positive expression was defined as nuclear staining in at least 10% of tumor cells. To identify POLEmut status, the QX200 ddPCR System (Bio-Rad Laboratories, Hercules, CA, USA), in conjunction with a commercially available mutation assay kit (Droplex POLE Mutation Test; Gencurix, Seoul, Korea), was employed. This assay is designed to detect five recurrent pathogenic point mutations in the POLE exonuclease domain: P286R, S297F, V411L, A456P, and S459F. These hotspot mutations account for the great majority of known POLE-ultramutated cases in EC and are the key targets recommended for molecular testing in current EC classification guidelines. A sample was interpreted as POLE mutation-positive in the ddPCR assay (Bio-Rad Laboratories) if the specific mutation probe yielded at least the minimum number of positive droplets above the assay’s limit of blank (generally corresponding to ≥6 mutant copies per 20 μL or a mutation index ≥0.3%) [11–13].
3. Clinicopathological data
Histopathological data-including histologic type, tumor grade, depth of myometrial invasion, cervical stromal invasion, lymphovascular space invasion (LVSI), lymph node metastasis, and parametrial invasion-were recorded following a review of all slides by a gynecologic oncology pathologist. Substantial LVSI was defined as the presence of ≥4 LVSI-involved vessels on at least one hematoxylin and eosin-stained slide [14]. Clinical data were obtained through a review of the institutional medical records. Each case was assigned a molecularly integrated classification according to FIGO 2023 and ESGO 2025. The ESGO 2025 risk classification was applied using clinicopathological and molecular variables, including histologic type, tumor grade, depth of myometrial invasion, LVSI, FIGO 2023 stage, POLE mutation status, MMR status, p53 status, and ER expression.
4. Statistical analysis
Stage distribution according to the individual risk groups was summarized using cross-tabulations and visualized using a Sankey diagram. The median follow-up duration was calculated from the date of surgery to the date of last follow-up or death. Patients without events were censored at the date of their last clinical follow-up. Recurrence-free survival (RFS) was defined as the interval from surgery to documented disease recurrence or last follow-up. Disease-specific survival (DSS) was defined as the interval from surgery to death from EC. Disease-specific death was defined as death attributable to EC or its complications. Survival curves were estimated using the Kaplan-Meier method and differences between the curves were assessed using the log-rank test. Statistical significance was set at P<0.05.
To evaluate the prognostic performance of ESGO 2025 versus FIGO 2023, the Akaike information criterion (AIC) and Harrell’s concordance index (C-index) were calculated using R software version 4.3.2 (R Foundation for Statistical Computing, Vienna, Austria). Bootstrap resampling analyses were performed to calculate 95% confidence intervals (CIs) for the C-index. Because the uncertain-risk classification category does not correspond to a definitive prognostic group, these patients were excluded from Kaplan-Meier survival analyses and comparisons of prognostic performance.
Results
1. Patient characteristics
Baseline clinicopathological and molecular characteristics of the study cohort are summarized in Table 1. Among the 298 patients initially identified, two were excluded from the analysis: one due to an inadequate tissue specimen for analysis and the other because it was not possible to determine whether the uterine tumor was primary or metastatic. Consequently, a total of 296 patients were included in the final analysis. Among these patients, the majority had endometrioid histology, whereas 43 (14.6%) had non-endometrioid histology. Forty-seven (15.9%) patients had substantial LVSI, and 29 (9.8%) had lymph node metastasis. More than half of the patients did not receive postoperative adjuvant treatment.
Table 1.
Characteristics of the study population
| Clinicopathological factor | Value |
|---|---|
| Age (yr) | |
| >50 | 223 (75.3) |
| ≤50 | 73 (24.7) |
| Parity | |
| Nullipara | 49 (16.5) |
| ≥1 | 247 (83.5) |
| Body mass index (kg/m2) | |
| <25 | 146 (49.3) |
| ≥25 | 150 (50.7) |
| Histology | |
| Endometrioid | 253 (85.4) |
| Non-endometrioid | 43 (14.6) |
| Grade | |
| 1–2 | 225 (76.0) |
| 3 | 71 (24.0) |
| Depth of myometrial invasion | |
| No | 91 (30.7) |
| <50% | 114 (38.6) |
| ≥50% | 91 (30.7) |
| Cervical stromal invasion | |
| No | 246 (83.1) |
| Yes | 50 (16.9) |
| Parametrial invasion | |
| No | 284 (95.9) |
| Yes | 12 (4.1) |
| Lymphovascular space invasion | |
| No | 222 (75.0) |
| Focal | 27 (9.1) |
| Substantial | 47 (15.9) |
| Lymph node metastasis | |
| None | 251 (84.7) |
| Pelvic lymph node | 18 (6.1) |
| Para-aortic lymph node | 11 (3.7) |
| Not conducted | 16 (5.5) |
| Adjuvant therapy | |
| None | 149 (50.3) |
| Radiation | 72 (24.3) |
| Chemoradiation | 41 (13.8) |
| Chemotherapy | 34 (11.6) |
Values are presented as number (%).
Complete molecular classification was achieved in all patients. The distribution of molecular subtypes was as follows: 37 (12.5%) had POLEmut, 64 (21.6%) had MMRd, 49 (16.6%) had p53abn, and 146 (49.3%) had NSMP. Immunohistochemical staining for ER showed that 46 patients (15.6%) were ER-negative (Supplementary Table 1).
2. Distribution of FIGO 2023 and ESGO 2025
Patient distribution according to FIGO 2023 and ESGO 2025 is presented in Table 2. Among the 237 patients whose disease was confined to the uterus, 31 (13.1%) were classified as stage IAmpolemut. Among these patients, nine had stage IA1 disease, 10 had stage IA2 disease, one had stage IB disease, four had stage IIA disease, one had stage IIB disease, and six had stage IIC disease. In addition, among patients with advanced-stage disease, POLE mutations were identified in one patient with stage IIIB1 disease, four patients with stage IIIC1ii disease, and one patient with stage IVB disease. The POLE hotspot mutations identified in these patients were as follows: 10 patients had V411L, 20 had P286R, three had A456P, and three had S297F. Nineteen (6.4%) patients were classified as stage IICmp53abn. Among these patients, three had stage IA1 disease, five had stage IC disease, and 11 had stage IIC disease. Consequently, the incorporation of molecular classification resulted in shifts in risk classification in 20 cases (6.7%): 12 (4.1%) underwent downward risk reclassification based on POLE mutations, whereas eight (2.6%) underwent upward risk reclassification based on p53 abnormality. Based on ESGO 2025, 153 patients (51.6%) were classified as low risk, 30 (10.1%) as intermediate risk, 19 (6.4%) as high-intermediate risk, 64 (21.6%) as high risk, and 13 (4.3%) as advanced/metastatic. A small proportion of patients (6.0%) were categorized as having an uncertain-risk classification according to ESGO 2025. These cases were not considered missing data but rather represented a predefined category within ESGO 2025. Substantial overlap between the two classification systems was observed, particularly within stage IIC (Fig. 1). Even within stage IIC, ESGO 2025 stratified patients into different risk groups: two were classified as intermediate risk, five as high-intermediate risk, and six as high risk. Among the 13 patients with stage IIC disease, the two patients classified as intermediate risk were both MMRd, and all five patients classified as high-intermediate risk were also MMRd. Among the remaining six patients classified as high risk, three were p53abn, and three were classified as NSMP with ER-negative status. Although RFS analysis was performed, no statistically significant differences were observed because of the extremely small number of patients in each group (Supplementary Fig. 1). Nevertheless, a trend toward an increasing number of recurrences across progressively higher risk groups was observed. No recurrences occurred in the intermediate-risk group, whereas three of the six patients in the high-risk group experienced recurrence (50.0%). In contrast, two recurrences occurred in the high-intermediate- risk group (40.0%).
Table 2.
Patient distribution according to FIGO 2023 stages and the 2025 ESGO-ESTRO-ESP risk classification
| FIGO 2023 staging | |
| IAmPOLEmut | 31 (10.4) |
| IA1 | 58 (19.5) |
| IA2 | 66 (22.2) |
| IA3 | 1 (0.5) |
| IB | 19 (6.4) |
| IC | 4 (1.3) |
| IIA | 14 (4.7) |
| IIB | 12 (4.0) |
| IIC | 13 (4.3) |
| IICmp53abn | 19 (6.4) |
| IIIA1 | 10 (3.3) |
| IIIA2 | 8 (2.7) |
| IIIB1 | 2 (0.9) |
| IIIB2 | 1 (0.5) |
| IIIC1i | 8 (2.7) |
| IIIC1ii | 9 (3.0) |
| IIIC2i | 4 (1.3) |
| IIIC2ii | 1 (0.5) |
| IVA | 3 (1.1) |
| IVB | 8 (2.7) |
| IVC | 5 (1.6) |
| ESGO-ESTRO-ESP 2025 classification | |
| Low risk | 153 (51.6) |
| Intermediate risk | 30 (10.1) |
| High-intermediate risk | 19 (6.4) |
| High-risk | 64 (21.6) |
| Advanced-metastatic | 13 (4.3) |
| Uncertain risk | 17 (6.0) |
Values are presented as number (%).
FIGO, International Federation of Gynecology and Obstetrics; ESGO, European Society of Gynaecological Oncology; ESTRO, European Society for Radiotherapy and Oncology; ESP, European Society of Pathology.
Fig. 1.
Distribution of FIGO 2023 stages across 2025 ESGO-ESTRO-ESP risk groups. FIGO, International Federation of Gynecology and Obstetrics; ESGO, European Society of Gynaecological Oncology; ESTRO, European Society for Radiotherapy and Oncology; ESP, European Society of Pathology.
Stage IA2 (n=66) and stage IB (n=19) were the most prevalent stages in the low-risk and intermediate-risk groups, accounting for 43.1% and 63.3% of patients in each group, respectively. The stage distribution was most heterogeneous in the high-risk group, with stage IICmp53abn (n=11) being the most prevalent.
Among the 146 patients classified as NSMP, 131 (89.7%) were ER-positive, and 15 (10.3%) were ER-negative. Among the ER-positive patients, the distribution of risk groups was as follows: 90 patients were classified as low risk, 25 as intermediate risk, five as high-intermediate risk, and 10 as high risk, with the low-risk group accounting for the largest proportion (68.7%). In contrast, among the 15 ER-negative patients, 13 were classified as high risk and two as advanced/metastatic risk, suggesting an association between ER-negative status and an increased risk of recurrence.
3. Recurrence-free and DSS according to ESGO 2025
The median follow-up duration was 58.8 months (range, 0.4–161.5). RFS differed significantly among ESGO 2025 risk groups (log-rank P<0.001) (Fig. 2). No recurrences were observed in the intermediate-risk group, whereas four recurrences occurred in the low-risk group (2.6%). In contrast, recurrence rates were high in the high-intermediate-risk (52.6%), high-risk (43.8%), and advanced/metastatic-risk (41.7%) groups.
Fig. 2.
Kaplan-Meier curves according to 2025 ESGO-ESTRO-ESP risk groups. (A) Recurrence-free survival. (B) Disease-specific survival. ESGO, European Society of Gynaecological Oncology; ESTRO, European Society for Radiotherapy and Oncology; ESP, European Society of Pathology.
During follow-up, disease-specific deaths were observed exclusively in the high-intermediate-risk, high-risk, and advanced/metastatic-risk groups. Two disease-specific deaths occurred in the high-intermediate-risk group (10.5%), 12 deaths occurred in the high-risk group (18.8%), and five deaths occurred in the advanced/metastatic-risk group (38.5%). No disease-specific deaths were recorded among patients classified as low risk or intermediate risk.
Kaplan-Meier analysis demonstrated significant differences in DSS across ESGO 2025 risk groups (log-rank P<0.001) (Fig. 2). Survival curves for the low-risk and intermediate-risk groups overlapped at a probability of 1.0 throughout the follow- up period.
4. Recurrence-free and DSS according to FIGO 2023
RFS differed significantly according to FIGO 2023 stage (logrank P<0.001) (Fig. 3). Patients with early-stage disease demonstrated excellent outcomes, with no recurrences observed in stage IAmpolemut. In contrast, recurrence rates increased with advancing stage, particularly among patients with stage IIC and stage IICmp53abn disease. Patients with advanced- stage disease (stages III and IV) experienced the highest recurrence rates during follow-up.
Fig. 3.
Kaplan-Meier curves according to FIGO 2023 stages. (A) Recurrence-free survival. (B) Disease-specific survival. FIGO, International Federation of Gynecology and Obstetrics.
DSS also differed significantly according to FIGO 2023 stage (log-rank P<0.001) (Fig. 3). No disease-specific deaths occurred among patients with stage IAmpolemut, whereas mortality was primarily observed in patients with higher-stage disease, including stage IICmp53abn and stages III–IV. Kaplan-Meier curves demonstrated clear separation of survival probabilities across stages, indicating that FIGO 2023 provides strong discrimination of mortality risk.
5. Performance comparison between ESGO 2025 and FIGO 2023 for predicting outcomes
To compare the prognostic performance of ESGO 2025 and FIGO 2023, model performance was evaluated using the AIC and the C-index, with bootstrap-derived 95% CIs. For recurrence prediction, ESGO 2025 demonstrated a lower AIC value than FIGO 2023 (410.849 vs. 428.644). In addition, ESGO 2025 showed a numerically higher C-index than FIGO 2023 (0.8657; 95% CI, 0.8266–0.9114 vs. 0.8519; 95% CI, 0.8110–0.9005). In contrast, for disease-specific mortality prediction, FIGO 2023 demonstrated a lower AIC value than ESGO 2025 (143.758 vs. 148.109). FIGO 2023 also showed a numerically higher C-index than ESGO 2025 (0.9360; 95% CI, 0.9028–0.9769 vs. 0.9063; 95% CI, 0.8624–0.9628). Although differing patterns of prognostic discrimination were observed according to the clinical endpoint, the CIs of the C-indices substantially overlapped between the two systems.
Discussion
In the present study, we compared the prognostic performance of FIGO 2023 and the updated ESGO 2025 classification system in a real-world cohort of patients with EC. To the best of our knowledge, this is the first study to directly compare the prognostic performance of ESGO 2025 with that of FIGO 2023. Our findings demonstrated substantial overlap between the two systems while also revealing differing patterns of prognostic discrimination according to the clinical endpoint.
The introduction of molecular classification has fundamentally changed the prognostic assessment of EC. The Cancer Genome Atlas first demonstrated that EC can be divided into four molecular subtypes with distinct prognostic profiles [15]. Subsequent studies developed clinically applicable classifiers, such as the ProMisE algorithm, which enables molecular stratification using immunohistochemistry and targeted sequencing [16]. These molecular insights have been progressively incorporated into clinical guidelines and staging systems.
FIGO 2023 represents a major paradigm shift because it integrates molecular features, particularly POLE mutations and p53 abnormalities, into the traditional anatomical staging framework [17]. Several studies have reported that the incorporation of molecular classification improves prognostic stratification compared with purely anatomical staging systems [8,18]. However, the FIGO 2023 staging system remains primarily an anatomical system and is not directly used to guide treatment decisions.
In contrast, ESGO 2025 was specifically designed to guide clinical management, particularly decisions regarding adjuvant therapy. The 2021 guidelines already incorporated molecular classification into risk stratification and the updated 2025 guidelines further refined risk groups by incorporating ER expression and detailed subclassification of the NSMP group. These refinements reflect the increasing recognition that the biological characteristics of tumors contribute substantially to recurrence risk.
In the present study, ESGO 2025 demonstrated lower AIC values and numerically higher concordance indices for recurrence prediction, whereas FIGO 2023 demonstrated lower AIC values and numerically higher concordance indices for disease-specific mortality prediction. These findings suggest that the two systems may provide different prognostic information depending on the clinical endpoint evaluated.
This finding is biologically plausible because recurrence risk is strongly influenced by tumor biology, including molecular subtype, histologic grade, and hormone receptor expression [19,20]. Given that ESGO 2025 incorporates these factors more comprehensively than FIGO 2023, it may better reflect risk stratification for recurrence.
Conversely, anatomical tumor extent, including deep myometrial invasion, cervical involvement, and extrauterine disease, which remain central to FIGO staging, may be more closely associated with disease-specific mortality [21]. Importantly, however, the bootstrap-derived 95% CIs of the concordance indices substantially overlapped between the two systems. Therefore, the observed differences should be interpreted cautiously and should not be regarded as definitive evidence of the statistical superiority of one classification system over the other.
Another notable finding of our study is the heterogeneity observed within FIGO stage IIC disease. Even within this stage, ESGO 2025 stratified patients into intermediate-risk, high-intermediate-risk, and high-risk groups. This observation is consistent with previous reports demonstrating that molecular features may further refine prognosis within the same anatomical stage [22]. Such heterogeneity supports the concept that combined anatomical and molecular stratification may provide the most accurate prognostic assessment.
Interestingly, no disease-specific deaths were observed in either the low-risk or intermediate-risk groups in the present cohort. In addition, no recurrences occurred in the intermediate- risk group. These findings may suggest limited prognostic separation between the low-risk and intermediate-risk groups within our cohort. However, this observation should be interpreted cautiously because the number of patients in the intermediate-risk group was relatively small, and the median follow-up duration may not have been sufficient to capture late recurrences or disease-specific mortality.
Moreover, the ESGO-ESTRO-ESP classification was originally developed to guide clinical decision-making, rather than to maximize survival discrimination between adjacent low-risk categories. Therefore, the absence of events in the intermediate- risk group does not necessarily indicate a lack of clinical relevance of the classification system.
Although the recurrence rate in the low-risk group was low overall, four patients classified as low risk experienced vaginal- cuff recurrence despite not receiving postoperative adjuvant treatment. Notably, all four patients underwent surgery between 2010 and 2012, before several surgical techniques currently used to minimize tumor spillage had become routine practice at our institution. During that period, prophylactic bilateral tubal coagulation before uterine manipulator insertion and protective specimen retrieval maneuvers to reduce vaginal tumor spillage were not routinely performed. Although the precise impact of these factors on recurrence risk remains uncertain, they may have contributed to the occurrence of isolated vaginal-cuff recurrence in these otherwise low-risk patients. In addition, occasional recurrence in low-risk patients may also reflect unrecognized biological heterogeneity, occult microscopic disease, or limitations inherent in retrospective risk stratification.
The present findings may have practical clinical implications for contemporary EC management. ESGO 2025 may provide additional value in recurrence-oriented risk assessment and guiding adjuvant treatment decision-making. In particular, ESGO 2025 was originally developed as a treatment-oriented framework incorporating molecular and clinicopathological risk factors associated with tumor relapse. In contrast, FIGO 2023 may remain useful for mortality risk assessment and prognostic counseling. Therefore, FIGO 2023 may continue to play an important role in prognostic counseling and surveillance planning. Taken together, these observations support the concept that ESGO 2025 and FIGO 2023 should not necessarily be viewed as competing systems but rather as complementary frameworks that provide different types of clinically relevant prognostic information.
This study has several limitations. First, it was a retrospective analysis conducted at a single institution, which may limit the generalizability of the findings. Second, although our cohort size was comparable to that of previous molecular classification studies, larger multicenter cohorts are needed to validate these findings. Adjuvant treatment allocation was not randomized and was likely influenced by clinicopathological risk factors incorporated into the 2025 ESGO-ESTRO-ESP classification. Therefore, confounding by indication may have affected survival outcomes, particularly in higher-risk groups that more frequently received postoperative adjuvant therapy. To improve transparency, the distribution of adjuvant treatment according to risk group has been provided in Supplementary Table 2. Nevertheless, residual treatment-related confounding cannot be completely excluded.
Despite these limitations, our study provides real-world evidence through a direct comparison of two contemporary classification systems in EC. Our findings suggest complementary prognostic roles for molecularly integrated risk stratification and anatomical staging in EC management.
Footnotes
Conflict of interest
The authors declare that they have no conflicts of interest.
Ethical approval
The study was approved by the Institutional Review Board of Korea University Guro Hospital (IRB number: 2024GR0367).
Patient consent
Informed consent was waived for this retrospective study, as approved by the Institutional Review Board.
Funding information
This study did not receive any external funding.
Acknowledgement
The biospecimens and data used in this study were provided by the Biobank of Korea University Guro Hospital.
Supplementary Information
References
- 1.Agarwal S, Melgandi W, Sonkar DR, Ansari FA, Arora S, Rathi AK, et al. Epidemiological characteristics of endometrial cancer patients treated at a tertiary health center in National Capital Territory of India. J Cancer Res Ther. 2023;19:452–6. doi: 10.4103/jcrt.jcrt_2029_21. [DOI] [PubMed] [Google Scholar]
- 2.Bourou MZ, Matsas A, Vrekoussis T, Mastorakos G, Valsamakis G, Panoskaltsis T. Conservative treatment of endometrial cancer in women of reproductive age (Review) Mol Clin Oncol. 2023;19:55. doi: 10.3892/mco.2023.2651. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Bassette E, Ducie JA. Endometrial cancer in reproductive- aged females: etiology and pathogenesis. Biomedicines. 2024;12:886. doi: 10.3390/biomedicines12040886. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Drab A, Kanadys W, Malm M, Wdowiak K, Dolar-Szczasny J, Barczyński B. Association of endometrial cancer risk with hypertension- an updated meta-analysis of observational studies. Sci Rep. 2024;14:24884. doi: 10.1038/s41598-024-76896-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Park B. Associations between obesity, metabolic syndrome, and endometrial cancer risk in East Asian women. J Gynecol Oncol. 2022;33:e35. doi: 10.3802/jgo.2022.33.e35. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Chen W, Zhou W, Liu S. The key role of natural products in the fight against endometrial cancer. Int Immunopharmacol. 2025;151:114344. doi: 10.1016/j.intimp.2025.114344. [DOI] [PubMed] [Google Scholar]
- 7.Zhou X, Zeng Y, Zheng R, Wang Y, Li T, Song S, et al. Natural products modulate cell apoptosis: a promising way for treating endometrial cancer. Front Pharmacol. 2023;14:1209412. doi: 10.3389/fphar.2023.1209412. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Berek JS, Matias-Guiu X, Creutzberg C, Fotopoulou C, Gaffney D, Kehoe S, et al. FIGO staging of endometrial cancer: 2023. Int J Gynaecol Obstet. 2023;162:383–94. doi: 10.1002/ijgo.14923. [DOI] [PubMed] [Google Scholar]
- 9.Concin N, Matias-Guiu X, Cibula D, Colombo N, Creutzberg CL, Ledermann J, et al. ESGO-ESTRO-ESP guidelines for the management of patients with endometrial carcinoma: update 2025. Lancet Oncol. 2025;26:e423–35. doi: 10.1016/S1470-2045(25)00167-6. [DOI] [PubMed] [Google Scholar]
- 10.Kommoss S, McConechy MK, Kommoss F, Leung S, Bunz A, Magrill J, et al. Final validation of the ProMisE molecular classifier for endometrial carcinoma in a large-population based case series. Ann Oncol. 2018;29:1180–8. doi: 10.1093/annonc/mdy058. [DOI] [PubMed] [Google Scholar]
- 11.Kim G, Lee SK, Suh DH, Kim K, No JH, Kim YB, et al. Clinical evaluation of a droplet digital PCR assay for detecting POLE mutations and molecular classification of endometrial cancer. J Gynecol Oncol. 2022;33:e15. doi: 10.3802/jgo.2022.33.e15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Joe S, Lee M, Kang J, Kim J, Hong SH, Lee SJ, et al. Enhanced risk stratification in early-stage endometrial cancer: integrating POLE through droplet digital PCR and L1CAM. Cancers (Basel) 2023;15:4899. doi: 10.3390/cancers15194899. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Ouh YT, Cho HW, Min KJ, Lee JK, Chun Y, Hong JH. Diagnostic accuracy of the droplet digital PCR POLE mutation test in endometrial cancer: comparison with Sanger sequencing and NGS. J Gynecol Oncol. 2026;37:e83. doi: 10.3802/jgo.2026.37.e83. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Peters EEM, León-Castillo A, Smit VTHBM, Boennelycke M, Hogdall E, Hogdall C, et al. Defining substantial lymphovascular space invasion in endometrial cancer. Int J Gynecol Pathol. 2022;41:220–6. doi: 10.1097/PGP.0000000000000806. [DOI] [PubMed] [Google Scholar]
- 15.Kandoth C, Schultz N, Cherniack AD, Akbani R, Liu Y, Shen H, et al. Integrated genomic characterization of endometrial carcinoma. Nature. 2013;497:67–73. doi: 10.1038/nature12113. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Talhouk A, McConechy MK, Leung S, Li-Chang HH, Kwon JS, Melnyk N, et al. A clinically applicable molecular- based classification for endometrial cancers. Br J Cancer. 2015;113:299–310. doi: 10.1038/bjc.2015.190. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Talhouk A, McAlpine JN. New classification of endometrial cancers: the development and potential applications of genomic-based classification in research and clinical care. Gynecol Oncol Res Pract. 2016;3:14. doi: 10.1186/s40661-016-0035-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Concin N, Matias-Guiu X, Vergote I, Cibula D, Mirza MR, Marnitz S, et al. ESGO/ESTRO/ESP guidelines for the management of patients with endometrial carcinoma. Int J Gynecol Cancer. 2021;31:12–39. doi: 10.1136/ijgc-2020-002230. [DOI] [PubMed] [Google Scholar]
- 19.León-Castillo A, de Boer SM, Powell ME, Mileshkin LR, Mackay HJ, Leary A, et al. Molecular classification of the PORTEC-3 trial for high-risk endometrial cancer: impact on prognosis and benefit from adjuvant therapy. J Clin Oncol. 2020;38:3388–97. doi: 10.1200/JCO.20.00549. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Stelloo E, Nout RA, Osse EM, Jürgenliemk-Schulz IJ, Jobsen JJ, Lutgens LC, et al. Improved risk assessment by integrating molecular and clinicopathological factors in early-stage endometrial cancer-combined analysis of the PORTEC cohorts. Clin Cancer Res. 2016;22:4215–24. doi: 10.1158/1078-0432.CCR-15-2878. [DOI] [PubMed] [Google Scholar]
- 21.Soslow RA, Tornos C, Park KJ, Malpica A, Matias-Guiu X, Oliva E, et al. Endometrial carcinoma diagnosis: use of FIGO grading and genomic subcategories in clinical practice: recommendations of the International Society of Gynecological Pathologists. Int J Gynecol Pathol. 2019;38(Suppl 1):S64–74. doi: 10.1097/PGP.0000000000000518. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Imboden S, Nastic D, Ghaderi M, Rydberg F, Siegenthaler F, Mueller MD, et al. Implementation of the 2021 molecular ESGO/ESTRO/ESP risk groups in endometrial cancer. Gynecol Oncol. 2021;162:394–400. doi: 10.1016/j.ygyno.2021.05.026. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.