Abstract
OBJECTIVE: Lymph node metastasis (LNM) is an important reference indicator for the prognosis of endometrial cancer (EC). Even in patients with early low-risk EC, many people still have LNM. The purpose of this study was to investigate the related factors influencing LNM in early-stage EC and determine the optimal positive threshold of immunohistochemical parameter Ki67 for predicting LNM, providing auxiliary reference indicators for clinical diagnosis and treatment. METHODS: The clinicopathological data of 651 patients with "apparent" early-stage EC who underwent standard surgical treatment were included. Univariate and multivariate logistics regression were used to analyze the correlation between each clinicopathological factor and LNM. Receiver operating characteristic curve (ROC curve) and Youden index were used to determine the optimal positive threshold of Ki67 for predicting LNM. Finally, correlation between Ki67 and various clinicopathological factors was analyzed, and the predictive value of each prognostic factor was compared. RESULTS: Multivariate analysis found that histologic grade (P=0.023), lymphatic vessel space invasion (LVSI) (P < 0.001), serological index Ca125 (P=0.002), immunohistochemical parameter Ki67 (P < 0.001), ER (P < 0.001) and P53 (P=0.001) were independent prognostic factors of LNM. ROC curve and Youden index showed that the optimal positive thresholds of Ki67 to predict LNM were 40%. Based on this, ROC curve showed that the area under the curve (AUC) of Ki67 (AUC=0.714) was larger than other single predictors, and Ki67 combined with other predictors can significantly increase the AUC value (AUC= 0.847 and 0.868, respectively). CONCLUSION: Ki67 was an important predictor for predicting the LNM in early-stage EC and taking a positive percentage of about 40% can be used as the positive threshold of the immunohistochemical parameter Ki67. On this basis, Ki67 combined with other predictive indicators can significantly improve prediction performance and can be used for segmentally predicting LNM of early-stage EC.