Abstract:

Objective: To analyze the risk factors for restenosis after interventional treatment of tuberculous airway stenosis, construct a risk prediction model, and verify its predictive performance. Methods: A total of 350 patients with tuberculous airway stenosis who received interventional therapy in the Suzhou Fifth People’s Hospital from April 2021 to October 2024 were retrospectively collected as the research subjects. The patients were randomly divided into a modeling group and a validation group at a ratio of 4∶1 using the set-away method. Clinical data of patients was collected and LASSO regression was performed to identify key variables and do multivariable analysis, and a nomogram prediction model was constructed based on independent risk factors. Receiver operating characteristic (ROC) curve, Hosmer-Lemeshow test, and decision curve analysis (DCA) were used to verify performance of the model. Results: The incidence of restenosis after interventional treatment in patients included in this study was 41.71% (146/350). Ten key variables were screened by LASSO regression for multivariable analysis. The results showed that diabetes (OR=3.126, 95%CI:1.017-9.604), microscopic active period (OR=2.646, 95%CI:1.197-5.851), stenosis length >3 cm (OR=5.437, 95%CI:2.682-11.022), and increased systemic immune inflammation index (OR=1.008, 95%CI:1.006-1.011) were all risk factors for restenosis after treatment of tuberculous airway stenosis, while receiving systemic antituberculosis treatment before treatment (OR=0.326, 95%CI:0.150-0.707) was protective factor. The ROC curve analysis showed that the AUCs of the modeling group and the validation group were 0.878 (95%CI:0.832-0.924) and 0.867 (95%CI:0.771-0.962), respectively; The Hosmer Lemeshow test results for the modeling group and validation group were 5.419 (P=0.712) and 6.331 (P=0.610), respectively; The DCA results showed that within the high-risk threshold range of 0.1-0.8, using a nomogram model for preventive intervention in patients could achieve higher standardized net benefit compared to average intervention and no intervention. Conclusion: The restenosis of patients with tuberculous airway stenosis after interventional treatment is related to diabetes, microscopic active period, stenosis length >3 cm, and increased systemic immune inflammation index factors. Based on the relevant factors, a nomogram prediction model can be constructed to conduct risk assessment and early prediction of restenosis, which can provide reference for disease management plans for patients after interventional treatment.

Key words: Tuberculosis, Tracheal stenosis, Bronchoscopy, Prognosis, Nomograms