Abstract:

Objective: To investigate the independent risk factors for post-treatment bronchial stenosis in patients with lymph node fistula-type tracheobronchial tuberculosis (TBTB), construct and validate a risk prediction nomogram model, and provide evidence for early clinical identification of high-risk patients and formulation of individualized intervention strategies. Methods: A retrospective study was conducted, and 914 patients with lymph node fistula-type TBTB diagnosed at Hunan Chest Hospital from January 2019 to December 2023 were enrolled as study subjects. Data including demographic characteristics, clinical symptoms, laboratory indicators, imaging features, etiological test results and bronchoscopic findings were collected. The subjects were divided into the stenosis group and the non-stenosis group according to the presence of post-treatment bronchial stenosis. Post-treatment bronchial stenosis was defined as airway lumen stenosis >25% observed by follow-up bronchoscopy, accompanied by dyspnea or obstructive changes on imaging. Predictive variables were sequentially screened using LASSO regression, variance inflation factor test and recursive feature elimination (RFE). A multivariate logistic regression prediction model was constructed and visualized as a nomogram. The model performance was evaluated using the area under the receiver operating characteristic curve (AUC), calibration curve and decision curve analysis. Results: The incidence of post-treatment bronchial stenosis among the included patients was 24.95% (228/914). Five key predictive variables were screened out by LASSO regression and RFE. Multivariate analysis showed that each 1-year increase in age (OR=1.06, 95%CI: 1.04-1.08), endoscopic ulceration (OR=2.42, 95%CI: 1.41-4.17), absence of hemoptysis (OR=3.48, 95%CI: 1.28-9.48) and absence of lymph node tuberculosis at other sites (OR=2.67, 95%CI: 1.52-4.70) were independent risk factors for bronchial stenosis, while male sex (OR=0.56, 95%CI: 0.37-0.86) was a protective factor. The AUC of the nomogram model was 0.806 in the training cohort and 0.720 in the validation cohort. The calibration curve demonstrated high consistency between the predicted probability of the nomogram and actual observed outcomes. Decision curve analysis confirmed that the model had favorable clinical utility across a wide range of threshold probabilities. Conclusion: The nomogram model constructed with 5 predictors including age, sex, hemoptysis, endoscopic ulceration and lymph node involvement at other sites has favorable discriminative ability and clinical practical value, which can provide quantitative evidence for early risk stratification and individualized follow-up of post-treatment bronchial stenosis in patients with lymph node fistula-type TBTB.

Key words: Tuberculosis, Bronchial diseases, Tracheal stenosis, Fistula, Lymph nodes