A preoperative and intraoperative scoring system to predict nodal metastasis in endometrial cancer

Objective: To develop a scoring system that guides surgical decision-making regarding the need to perform lymphadenectomy. Methods: A retrospective study was performed of patients who underwent complete surgical staging of endometrial cancer between 2003 and 2014 at three centers in Brazil. Preoperative and intraoperative risk factors were used to develop a scoring system to predict lymph node metastasis. Results: Among 329 patients included, 71 (21.6%) had positive lymph nodes and 259 (78.4%) had negative lymph nodes. The characteristics associated with nodal metastasis in univariate analysis included the level of cancer antigen 125 (P<0.001), preoperative histological grade (P<0.001), endometrial thickness (P=0.012), and pathologic features including tumor size (P<0.001), tumor extension (P<0.001), and lower uterine segment involvement (P<0.001). On multivariate logistic regression analysis, tumor grade, tumor extension, and lower uterine segment involvement remained significantly associated. The resulting scoring system showed good accuracy as demonstrated by an area under the receiver operating characteristic curve of 0.858 (95% confidence interval 0.804–0.913). Conclusion: A highly accurate scoring system for the prediction of lymph node metastasis was developed on the basis of three preoperative and intraoperative risk factors. After validation, this model could greatly aid clinicians in the surgical management of endometrial cancer.