ABSTRACT
Objective:The purpose of this study is to construct and validate a simple and accurate clinical nomogram for predicting preoperative lymph node metastasis(LNM)in individualized patients with colorectal cancer.Methods:The clinical data of patients came from SEER database, and Logistic regression model was used to build a prediction model.The accuracy of nomograph was evaluated and tested by receiver operating characteristic curve(ROC)curve and calibration curve.Results:78584 eligible colorectal cancer(CRC)patients in SEER database were enrolled in this study.The risk factors related to lymph node metastasis include diagnostic age, T stage, M stage, tumor grade and carcinoembryonic antigen(CEA)level.The model showed good accuracy, with an AUC value of 0.734(95% CI, 0.730-0.741). In the validation queue, the model can still reach a high AUC value of 0.728(95% CI, 0.722-0.734). Conclusions:Age, T stage, M stage, tumor grade and CEA level are independent risk factors for lymph node metastasis of colorectal cancer.The nomogram established based on these factors can predict the probability of lymph node metastasis of colorectal cancer.