Machine learning-based model assists in differentiating Mycobacterium avium Complex Pulmonary Disease from Pulmonary...

Area of Science:

• Pulmonology
• Infectious Diseases
• Medical Imaging
• Machine Learning

Background:

• Mycobacterium avium-intracellulare complex (MAC) pulmonary disease is rising globally.
• Differentiating MAC pulmonary disease from pulmonary tuberculosis (TB) is challenging due to overlapping clinical and radiological features.
• Accurate diagnosis is crucial for appropriate treatment and patient outcomes.

Purpose of the Study:

• To develop and validate a machine learning (ML) model for distinguishing MAC pulmonary disease from pulmonary TB.
• To integrate clinical data with computed tomography (CT) imaging features for improved diagnostic accuracy.
• To assess the performance of different ML algorithms, including logistic regression, random forest, and support vector machine (SVM).

Main Methods:

• A multi-centered, retrospective study involving 169 patients diagnosed with MAC pulmonary disease or pulmonary TB.
• Analysis of clinical data (age, symptoms) and CT imaging features (cavity morphology, bronchiectasis).
• Development and validation of logistic regression, random forest, and SVM models, evaluated using ROC and precision-recall curves.

Main Results:

• The SVM model demonstrated superior performance in differentiating the two conditions.
• The SVM model achieved an Area Under the Curve (AUC) of 0.960 in the training set and 0.885 in the validation set.
• Distinct differences were observed in patient demographics (MAC patients were older) and radiological findings (cavity and bronchiectasis characteristics) between the groups.

Conclusions:

• The SVM-based ML model integrating clinical and CT features shows excellent diagnostic performance.
• This model can serve as a valuable tool to assist clinicians in differentiating MAC pulmonary disease from pulmonary TB.
• Improved diagnostic capabilities can lead to earlier and more accurate treatment initiation for patients with these lung infections.