RESUMO
Objective To construct and appraise a new model for predicting the prognosis of rectal cancer patients using the Lasso-Cox strategy.Methods The clinical pathological data of 599 rectal cancer patients who underwent radical resection were analyzed.Comparison between groups,Lasso and Cox regression were used to select varia-bles and construct a model,and its discrimination,consistency,and clinical benefits were appraised by the receiv-er operating characteristic(ROC),calibration curve,and decision curve analysis.Results Comparison between groups showed that age,body mass index(BMI),preoperational nutrition status,carbohydrate antigen199(CA199),preoperative chemotherapy,intraoperative blood transfusion,vascular or nerve invasion,cancer nod-ules,pathologic T,N,and TNM stages,tumor recurrence or metastasis,radiotherapy and postoperative survival time were associated with grouping of death or survival in rectal cancer patients.Among them,8 variables were se-lected by lasso and contained into the Cox regression model.Age(HR=1.04,P<0.05),BMI(HR=0.89,P<0.05),blood transfusion(HR=2.29,P<0.05),postoperative chemotherapy(HR=0.16,P<0.01),recur-rence(HR=43.67,P<0.01),and metastasis(HR=2.75,P<0.05)were identified as independent prognostic factors,which were used to construct a nomogram model.The area under the curve(AUC)and the 95%confi-dence interval of the receiver operating characteristic(ROC)curve of the predictive model was 0.95(0.91-0.99),P<0.01.The predicted probability of 1-year and 3-year survival was close to the actual probability.The DCA curve of the model was far away from a decision line parallel to the X-axis and another line with a negative slope.Conclusion The newly established nomogram has good discrimination,consistency and clinical benefits,which help predict the prognosis of rectal cancer after surgery.