Construction Of A Prediction Model For Prolong Air Leakage After Thoracoscopic Anatomic Pulmonary Resection

Objective: prolonged air leak(PAL)after thoracoscopic anatomic pulmonary resection is a common postoperative complication in tho racic surgery,and its early prediction is vital,The purpose of this study was to evaluate the association between the degree o f persistent air leakage after thoracoscopic anatomic pneumonect omy and pleural pressure and their ability to predict PAL risk.Methods: The clinical data of patients undergoing thoracoscopic anatomic pulmonary resectio n in the Depa rtment of Thoracic Surge ry of the First Affiliated Hospital of Xinjiang Medical University from January 2022 to January 2023 were retrospectively analyzed,inc luding 411 patients in group A.All patients were recorded with digital pleural mano meter at 12 h ours after surgery to re cord the maximum and minimum pleural pressure,and were asked to visually measure the degree of air leakage on the first day after surg ery.Independent predictors of PAL were determined by Logistic regression analysis.128 lobectomi es completed in differen t treatment groups were included as group B to test the predicti on model,and receiver operating characteristic(ROC)curves were used to assess the model’s discriminability.The calibration power was evaluated using th e HosmerLemeshow test and calibratio n curve.The clinical application value of the prediction model was evaluated by decision curve analysis(DCA).Results:A total of 539 patien ts were included in this study,of which 98(18.1%)were confirmed to be PAL by logistic r egression screening out 6independent risk factors : Prognostic nutrition index(PNI),PO D1 air leakage,thoracic adherence,mean thoracic pressure difference(ΔP),ROC area under curve(AU C)of the training cohort was 0.8952[95% confidence interval(CI):0.848-0.942].The area und er ROC curve(AUC)of the validation cohort was 0.918(95%CI: 0.864--0.972).The P values of Hosmer-Lemeshow test for the training and valid ation queues were 0.7655 and 0.7494,respectively,and the calibration curves were well fitted.The prediction mod el of DCA decision curve analysis shows that it is useful in clinic.Conclusion: The prediction model is effective in predicting PAL after minimally inv asive pulmonary resection.High risk PAL patients can be identified using this model and preventive meas ures can be taken in advance.The pleural pressure measured at 1 2 hours after operation was correlated with the duration of air leakage to varying degre es.Interpretation of data measured at early points in time by digi tal thoracomanometer can estimate the risk of subsequent PAL development.May be helpful in planning postoperative management for safer and more accurate implementation of thoracic tube man agement strategies.