A novel prognostic model for high-grade endometrial cancer using combined neutrophillymphocyte ratio (NLR) and platelet-lymphocyte ratio (PLR)

e17622 Background: Endometrial carcinoma (EC) as a common malignant tumor of the female reproductive system, posing a serious threat to the health of women. Compared to low-grade EC, high-grade EC has more rapid progression and a worse prognosis. Among the high-grade EC, grade3 endometrioid carcinoma, serous carcinoma, and clear cell carcinoma were the most common. Neutrophil-lymphocyte ratio (NLR) and platelet-lymphocyte ratio (PLR) are potential biomarkers of the systemic inflammatory response, and the correlation with prognosis of high-grade EC remains unclear. We want to evaluate the predictive significance of NLR and PLR in the prognosis of high-grade endometrial carcinoma and to establish a predictive model combining clinical factors. Methods: This multicenter retrospective cohort study included 910 high-grade epithelial endometrial carcinoma patients from the Chinese Endometrial Carcinoma Consortium database who underwent initial surgical treatment between January 1, 2005, and December 31, 2019. Univariable Cox survival analysis(the 2-sided Log-rank test) was used to assess each factor. Then the least absolute shrinkage and selection operator (LASSO) method was used to select the independent risk factors for DFS (disease-free survival) and OS (overall survival) from the outcomes of the univariate analysis to proceed with model fitting. Cox proportional hazard regression models were developed to predict the risk of recurrence and death at 3, 5, and 10 years, and the models were validated and calibrated. The area under the curve was used to measure the predictive performance of the model. The decision curve analysis was used to explore the clinical net benefit of the nomograms. Results: NLR and PLR were risk factors for recurrence, and NLR was a risk factor for death. We established models for predicting death and recurrence. In the training cohort, the AUC of recurrence model at 3, 5 and 10-year was 0.85, 0.85 and 0.86. And for the death model, the AUC at 3, 5 and 10-year was 0.83, 0.84 and 0.86. In the validation cohort, the AUC of recurrence model at 3,5 and 10-year was 0.72, 0.77 and 0.77. And for death model was 0.72, 0.81 and 0.84. Kaplan–Meier survival analysis showed a significant difference between low-risk and high-risk groups. The calibration curve also revealed excellent predictive accuracy both in the training and validation cohort. Conclusions: NLR and PLR are risk factors affecting the prognosis of high-grade endometrial carcinoma patients. Combining NLR/PLR and clinicopathological factors, we established a novel and more comprehensive prognostic model for patients to achieve risk stratification. This prognosis predictive model has the potential to offer risk scores and precise risk stratification for patients with high-grade endometrial cancer.