AI Based Cough Analysis for Pulmonary Tuberculosis Triage and Diagnosis: A Technical Review
Artificial intelligence-based cough analysis has emerged as a potential non-sputum approach for pulmonary tuberculosis triage and diagnostic support. This technical review synthesizes evidence from cough-based TB studies, highlighting current datasets, AI methods, performance benchmarks, implementation requirements, and translational barriers. We identify key limitations, including dataset bias, recording variability, limited prospective validation, and insufficient evidence in early-stage or asymptomatic TB populations.
Authors
- Shams Nafisa Ali (ORCID: https://orcid.org/0000-0002-5213-5463)
- Varunya Sakpuntoon
- Lisa Y. Armitige (ORCID: https://orcid.org/0000-0003-0678-9682)
- Nuttada Panpradist (ORCID: https://orcid.org/0000-0002-2733-4110)
- Blanca I. Restrepo (ORCID: https://orcid.org/0000-0002-3743-9098)
- Stephen J. Pont (ORCID: https://orcid.org/0000-0003-0200-0453)
- Umberto Villa (ORCID: https://orcid.org/0000-0002-5142-2559)
- Gutta J. Chowdary
- Adrienne E. Shapiro
- Leonard Kingwara
Institutions
- Austin College (US)
- Texas Department of State Health Services (US)
- Johns Hopkins University (US)
- The University of Texas at San Antonio Health Science Center (US)
- Bangladesh University of Engineering and Technology (BD)
- University of Washington (US)
- National Heart Hospital (BG)
- Ministry of Health (KE)
- Brownsville Public Library (US)
- Texas Center for Infectious Disease (US)
- The University of Texas at Austin (US)
Publication Details
- Journal
- npj Biomedical Innovations.
- Published
- 2026-09-21
- DOI
- https://doi.org/10.1038/s44385-026-00110-9
- Primary Topic
- COVID-19 diagnosis using AI
- Type
- article
- Field-Weighted Citation Impact
- 0.00