Two-centers machine learning analysis for predicting acid-fast bacilli results in tuberculosis sputum tests
Background: Tuberculosis (TB) is a chronic respiratory infectious disease caused by Mycobacterium tuberculosis, typically diagnosed through sputum smear microscopy for acid-fast bacilli (AFB) to assess the infectivity of TB. Methods: This study enrolled 769 patients, including 641 patients from the...
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-02-01
|
| Series: | Journal of Clinical Tuberculosis and Other Mycobacterial Diseases |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2405579425000026 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Background: Tuberculosis (TB) is a chronic respiratory infectious disease caused by Mycobacterium tuberculosis, typically diagnosed through sputum smear microscopy for acid-fast bacilli (AFB) to assess the infectivity of TB. Methods: This study enrolled 769 patients, including 641 patients from the First Affiliated Hospital of Guangxi Medical University as the training group, and 128 patients from Guangxi Hospital of the First Affiliated Hospital of Sun Yat-sen University as the validation group. Among the training cohort, 107 patients were AFB-positive, and 534 were AFB-negative. In the validation cohort, 24 were AFB-positive, and 104 were AFB-negative. Blood samples were collected and analyzed using machine learning (ML) methods to identify key factors for TB diagnosis. Results: Several ML methods were compared, and support vector machine recursive feature elimination (SVM-RFE) was selected to construct a nomogram diagnostic model. The area under the curve (AUC) of the diagnostic model was 0.721 in the training cohort and 0.758 in the validation cohort. The model demonstrated clinical utility when the threshold was between 38% and 94%, with the NONE line above the ALL line in the decision curve analysis. Conclusion: We developed a diagnostic model using multiple ML methods to predict AFB results, achieving satisfactory diagnostic performance. |
|---|---|
| ISSN: | 2405-5794 |