Machine Learning Study to Predict Acute Exacerbation Risk and Support Clinical Decisions in Adults with Allergic Bronchopulmonary Aspergillosis

What this Trial Is About

Researchers are conducting a multicenter observational study to develop and validate a machine learning model that predicts the risk of acute exacerbation within one year in adults with allergic bronchopulmonary aspergillosis ABPA during its stable phase. The study aims to improve risk stratification and support clinical decision-making using this model, which is based on the 2024 ISHAM diagnostic criteria for ABPA. Participants are grouped by whether they experience an acute exacerbation within one year. Participants diagnosed with ABPA will be divided into an acute exacerbation group or a non-exacerbation group based on their condition over one year. The study uses multiple machine learning algorithms and expert consensus to select features and develop 12 independent models. The models performance is evaluated through measures like ROC curves, sensitivity, specificity, and decision curve analysis. An interactive nomogram and SHAP method are used to enhance clinical interpretability. All participants will be followed for 12 months with regular clinical and laboratory evaluations. Outcome measures include occurrence and timing of exacerbations, lung function tests, chest CT changes, and various immunological markers. The study monitors participants health status and collects data on lung function and immune responses to assess the models predictive value and utility in managing ABPA.

ML-Based ABPA Recurrence Prediction and Clinical Utility