Screening high-risk groups, such as people living with HIV (PLHIV), for tuberculosis (TB) is crucial for TB elimination. However, choosing the best approach for TB screening requires careful consideration of cost-effectiveness. Automated nucleic acid testing has increased diagnostic precision in Tanzanian regions with a high prevalence of tuberculosis. However, because of its high cost, the screening initiative has been ineffective in finding many TB cases.

To address this challenge, MUHAS has established a research and development laboratory for Emerging Technologies for Healthcare (mETH) in Tanzania. The laboratory’s biomedical engineering unit is developing an AI algorithm using deep learning, specifically the Convolutional Neural Network (CNN), for detecting TB on chest X-rays.

The project’s principal investigator is Dr. Deogratias Mzurikwao, a lecturer at MUHAS and the head of the mETH laboratory. He has a Ph.D. in the application of AI in healthcare and has published several machine learning papers in peer-reviewed journals.