A high-risk prediction model for endometrial cancer: exploring the synergistic interaction between polycystic ovary syndrome and metabolic syndrome.

Abstract

Objective: To investigate the synergistic interaction between polycystic ovary syndrome (PCOS) and metabolic syndrome (MetS) in relation to the risk of endometrial cancer (EC). Additionally, we aimed to develop a clinically applicable, high-risk early-warning model that incorporates these interactive factors, enhancing the precision and clinical utility of EC screening.

Methods: We conducted a retrospective case-control study involving 445 newly diagnosed EC patients and 299 healthy female controls from the First People's Hospital of Changde City, between January 2018 and January 2025. Multivariate logistic regression was used to assess the independent and combined effects of PCOS and MetS on EC risk. A nomogram-based predictive model was developed and validated rigorously using training, internal validation, and external validation cohorts. The model's performance was evaluated based on discrimination (area under the curve [AUC]), calibration (Hosmer-Lemeshow test), and clinical utility (decision curve analysis). The diagnostic performance of our comprehensive model was compared to traditional tumor markers (cancer antigen 125/199, human epididymis protein 4).

Results: LASSO regression identified 14 clinically significant predictors. Logistic regression revealed that HE4 levels, endometrial thickness, and fasting blood glucose were independent risk factors for EC, while high-density lipoprotein was an independent protective factor. The nomogram based on these variables demonstrated excellent discrimination, with AUCs of 0.984 in the training set, 0.987 in the internal validation set, and 0.964 in the external validation set. The integrated risk model significantly outperformed individual markers in diagnostic accuracy across all datasets (P<0.001).

Conclusion: Our PCOS-MetS interaction-based EC risk prediction model showed robust and consistent performance across multiple validation cohorts. This tool significantly improves early detection accuracy and holds substantial clinical promise, laying the foundation for personalized EC risk management strategies.