Esophageal cancer, Pulmonary infection, Economic burden, Survival analysis, Risk factors, Predictive model, Surgery, RD1-811, Neoplasms. Tumors. Oncology. Including cancer and carcinogens, RC254-282

Abstract Background As a major complication after esophageal cancer (EC) surgery, postoperative pulmonary infection (PPI) is speculated to be associated with quality of life and survival after surgery. This study is aimed to explore the influence of PPI on the reality and establish a nomogram to predict PPI. Methods Data of patients undergoing esophagectomy was collected between January, 2016 and December, 2020 and divided into PPI and without PPI groups. Hospital costs and overall survival (OS) were compared between two groups. Univariate-multivariate analysis and LASSO-multivariate logistic regression were carried out to identify risk factors, and then two models were established based on them. To choose the better one, the receiver operating characteristic (ROC), the area under curve (AUC) and K-fold cross validation were compared between the models. Results The incidence of PPI in 633 esophageal cancer patients was 30.2% (191/633). PPI caused a total economic burden of RMB11,872.31 yuan on each patient and a poorer overall survival (60.5% vs. 54.0%, P = 0.002). The final nomogram was established by Univariate-multivariate logistic regression, including four independent risk factors of BMI