A nomogram integrating YTHDC2 expression and clinicopathological parameters for predicting endometrial cancer recurrence: a multicenter retrospective study

Endometrial cancer (EC) is one of the most common gynecological malignancies worldwide, with persistently rising incidence and mortality. Reliable molecular biomarkers for predicting EC recurrence remain lacking. This study aimed to develop a nomogram integrating YTHDC2 expression with clinicopathological parameters to predict postoperative recurrence-free survival (RFS) in EC patients.

A total of 520 stage I-III EC patients undergoing initial standard surgical treatment were enrolled as the training cohort, and 243 patients from three independent hospitals constituted the validation cohort. Univariate and multivariate Cox regression analyses were performed to identify independent prognostic factors for RFS in the training cohort, and a nomogram was constructed accordingly. Internal validation was conducted using the concordance index (C-index) and calibration curves in the training cohort, and external validation was performed in the validation cohort. Risk stratification was based on the Nomogram-predicted 3-year RFS probability, and Kaplan-Meier curves were used to evaluate the clinical utility of the model. The nomogram's discriminative performance was further compared with three published EC prediction models.

Univariate and multivariate Cox regression analyses identified age (HR = 2.02, P = 0.006), International Federation of Gynecology and Obstetrics (FIGO) stage (HR = 4.20, P < 0.001), lymphovascular space invasion (LVSI; HR = 1.81, P = 0.022), CA125 level (HR = 1.66, P = 0.048), depth of myometrial invasion (HR = 1.68, P = 0.039), histological type (HR = 3.42, P < 0.001), p53 protein status (HR = 2.61, P < 0.001), and YTHDC2 high expression level (HR = 0.26, P < 0.001) as independent prognostic factors for RFS. A nomogram incorporating these eight factors was constructed to predict RFS in EC. Internal and external calibration curves demonstrated good agreement between predicted and observed outcomes. The model achieved a C-index of 0.892 in the training cohort and 0.859 in the validation cohort, significantly outperforming single-indicator models. It also surpassed three published EC prediction models in both cohorts. The nomogram consistently outperformed single-indicator models in predicting 1-, 3-, and 5-year RFS.

The nomogram integrating YTHDC2 expression with clinicopathological parameters demonstrated favorable accuracy in predicting RFS for patients with EC and may effectively guide long-term prognostic management.