Development of a novel predictive model for lymph node metastasis in patients with endometrial endometrioid carcinoma

BACKGROUND: Globally, the burden of endometrial endometrioid carcinoma (EEC) increases annually. However, the histological grade of EEC remains unelucidated. We developed a novel model for predicting lymph node metastasis (LNM) in patients with endometrioid carcinoma (EC), which has not been well established. METHODS: A total of 344 patients with EEC were classified into training (n = 226) and validation (n = 118) cohorts. To develop a nomogram to predict LNM, independent predictors were defined using univariate and multivariate regression analyses. The calibration curve, area under the decision curve analysis (DCA), and receiver operating characteristic curve were used to evaluate the performance of the nomogram. RESULTS: Independent predictors of LNM in EC were identified in the univariate analysis, including mitosis; microcystic, elongated, and fragmented patterns; lymphovascular invasion (LVI); necrosis; and high-grade pattern. Mitosis, LVI, and high-grade pattern remained independent predictors of LNM in multivariate analysis. An LNM nomogram that was constructed by incorporating the five predictors showed reliable discrimination and calibration. DCA showed that the LNM nomogram scoring system had significant clinical application value. In addition, a high nomogram score (score > 150) was a significant prognosticator for survival in both LNM-positive and LNM-negative ECs. CONCLUSIONS: Our novel predictive model for LNM in patients with EC has the potential to assist surgeons in making optimal treatment decisions.