Preoperative risk stratification in endometrial cancer using ESGO/ESTRO/ESP 2021 guidelines: accuracy with and without molecular classification

Abstract

Background: The study aimed to evaluate the impact of integrating molecular classification with imaging-based preoperative staging on risk stratification prediction in endometrial cancer patients in accordance with ESGO/ESTRO/ESP (European Society of Gynaecological Oncology/European Society for Radiotherapy and Oncology/European Society of Pathology) 2021 guidelines. Methods: A retrospective cohort of 143 endometrial cancer patients was analyzed to assess changes in preoperative risk stratification after incorporating molecular classification into clinical evaluation. Preoperative clinical staging was primarily based on transvaginal ultrasound imaging. The overall agreement between preoperative risk group estimates (with/without molecular classification) and final postoperative outcomes was assessed using weighted Cohen’s Kappa, with bootstrap 95% confidence intervals and quadratic weights. Results: The addition of molecular classification significantly improved preoperative risk stratification accuracy (from 59.4 to 73.4%), particularly for patients post-operatively classified as high-risk. Kappa values indicated an improvement in overall agreement between preoperative and postoperative risk stratification following the addition of molecular classification, from 0.551 (95% CI: 0.430–0.671) to 0.767 (95% CI: 0.675–0.849). The non-overlapping confidence intervals indicated statistical significance. Preoperative assessment without molecular input tended to underestimate risk stratification. However, 26.6% of patients remained misclassified due to other factors, mostly within the intermediate and high-intermediate risk groups. Conclusions: Incorporating molecular classification enhances preoperative risk stratification and has the potential to tailor surgical treatment. Further validation through prospective multicentric studies is needed to support our findings. © The Author(s) 2025.