Analysis of influencing factors of perioperative complications of laparoscopic right hemicolectomy and constructing a risk model based upon nomogram
Objective To explore the influencing factors for perioperative complications of right laparoscopic hemicolectomy and construct a risk model of column graph.Methods From October 2019 to June 2023,the relevant clinical data were retrospectively reviewed for 223 patients undergoing right laparoscopic hemicolectomy.They were divided into two groups of complication(n=42)and non-complication(n=181)according to whether or not complications occurred in perioperative period.Multivariate Logistic regression was utilized for examining the influencing factors for perioperative complications.R4.3.2 software was utilized for constructing a prediction model of perioperative complications.Receiver operating characteristic(ROC)curve was employed for measuring the predictive value of nomogram and Hosmer-Lemeshow goodness of fit calibration curve for evaluating the fitting degree of nomogram.Decision curve analysis(DCA)was employed for examining the clinical utility of predictive model.Results Among them,42 patients had complications during perioperative period.Multivariate Logistic regression analysis indicated that diabetes mellitus(DM),operative duration>3 h,low tumor differentiation,external abdominal anastomosis and APACHE-Ⅱ score≥14 were independent risk factors for perioperative complications(P<0.05).The prediction probability of perioperative complications could be calculated by constructing the visualization risk prediction diagram of risk factors.DCA curve of nomogram prediction model was plotted.When prediction probability threshold was between 0 and 0.6,net return rate of nomogram model for perioperative complications was higher.The area under ROC curve was 0.941(95%CI:0.893-0.988)with a sensitivity of 87.46%and a specificity of 81.48%.It suggested that the nomogram prediction model had a decent discriminant capability.Hosmer-Lemeshow goodness of fit test χ2=5.236,P=0.732,actual curve of calibration curve approximated an ideal curve.Predicted probability accorded well with actual probability.Conclusion Concurrent DM,operative duration>3 h,low tumor differentiation,external abdominal anastomosis and APACHE-Ⅱ score≥14 are independent risk factors for perioperative complications of laparoscopic right hemicolectomy.The prediction model based upon the above nomogram is both accurate and well-differentiated.Clinical practicability is excellent.