Development And Validation Of A Nomogram Model For Postoperative Pulmonary Complications In Patients With Esophageal Cancer

ObjectivesTo explore the status of PPC(postoperative pulmonary complications)in esophageal cancer patients,analyze the influencing factors of esophageal cancer patients with PPC,and build a nomogram model for PPC in esophageal cancer patients to assist healthcare practitioners identify high-risk patients with PPC early,adjust treatment plans timely,and reduce the risk of PPC in patients with esophageal cancer.MethodsThe cluster sampling method was adopted,and participants were patients who received radical esophagectomy in a tertiary hospital in Jinan City,Shandong Province,1 January 2019 to 31 December 2020.The perioperative information of patients was collected through the selfdesigned "Questionnaire on the Status and Influencing Factors of Pulmonary Complications in Patients with Esophageal Cancer Surgery",and a total of 780 patients with esophagectomy were randomly assigned into a training group and a validation group.A nomogram for predicting PPC in patients with esophageal cancer was developed based on LASSO regression selecting predictors and multiple Logistic regression analyzing predictors.Internal and external validation of the nomogram model using training group and validation group,area under the ROC(receiver operating characteristic)curve,calibration curve and decision curve analysis(DCA)were used to evaluate the predictive model of PPC in esophageal cancer patients.Results1.This study included 780 patients with esophageal cancer,mainly male(637 cases),accounting for 81.7%.There were 546 patients in the training group,with an average age of 59.79±8.36 years,a total length of stay was 24.23±15.47 days,and the mean BMI was 23.21 ±2.99;the validation group of 234 patients,with an average age of 59.16±8.39 years,and a total length of stay was 23.10 ±12.00 days,the mean BMI was 22.92 ± 2.95.66.7%(520 patients)had a history of smoking,63.8%(498 patients)had a history of drinking;Preoperatively,176(22.6%)patients had hypertension,65(0.08%)patients had diabetes mellitus,44(0.06%)patients had coronary heart disease,and 50(0.06%)patients had pulmonary tuberculosis,4(0.01%)patients had chronic obstructive pulmonary disease(COPD).2.Among 780 patients with esophageal cancer,299 patients had PPC,and the incidence of PPC was 38.3%.The incidence of PPC was 39.7%in the training group(217 patients)and 35.0%in the validation group(82 patients).3.The higher incidence of PPC were pulmonary infection(22.0%),pleural effusion(17.4%),empyema(5.2%)and pneumothorax(3.7%).4.Based on LASSO regression and multivariate logistic regression analysis,four predictors including length of hospital stay,albumin,intraoperative blood loss,and hemoglobin were determined to develop a nomogram model for PPC of esophageal cancer.AUC(area under the ROC curve)of the training group and the validation group were 0.885 and 0.845,respectively,indicating that the nomogram model had better discrimination.Calibration curves for both groups showed agreement between the predicted and actual probabilities of PPC.The clinical decision curves of the two groups showed that the predicted probabilities were 1%-90%and 4%-81%,respectively,indicating that the nomogram model has clinical value.ConclusionsThe length of stay,albumin level,intraoperative blood loss and hemoglobin level are high risk factors for esophageal cancer PPC.The nomogram suggests that clinical staff should pay more attention to esophageal cancer patients with PPC,to dynamically monitor the perioperative characteristics of patients,control the risk factors of PPC,promote the decisionmaking of doctors and patients,and carry out individualized intervention to reduce the risk of PPC in patients with esophageal cancer,and optimize the allocation of health care resources.