A groundbreaking study presented at the ESCMID Global 2025 conference in Vienna has unveiled a new artificial intelligence (AI)-driven lung ultrasound tool, ULTR-AI, which surpasses human experts in diagnosing pulmonary tuberculosis (TB).
This advancement is promising for enhancing TB detection, particularly in resource-limited African regions.
Conducted at a tertiary hospital in Benin, the study involved 504 patients, with 192 confirmed cases of pulmonary TB through molecular testing.
Notably, 15 percent of participants were HIV-positive, and 13 percent had a history of TB, demographics often challenging for conventional diagnostics.
ULTR-AI demonstrated a sensitivity of 93 percent and specificity of 81 percent, outperforming human radiologists by 9 percent and exceeding the World Health Organization’s benchmarks for TB triage tools.
ALSO READ: African Stock Market Performance In The Week Ending April 11th
The ULTR-AI suite comprises three deep learning models: ULTR-AI, which predicts TB directly from lung ultrasound images; ULTR-AI (signs), detecting ultrasound patterns as interpreted by human experts; and ULTR-AI (max), which utilizes the highest risk score from both models to optimize accuracy.
By analyzing images from portable, smartphone-connected ultrasound devices, ULTR-AI offers a rapid, non-invasive, and scalable alternative for TB detection, eliminating the need for sputum samples or extensive laboratory infrastructure.
This innovation addresses longstanding challenges in TB diagnosis across Africa, where reliance on sputum samples, costly GeneXpert tests, and limited radiology services hinder timely detection. By enabling real-time interpretation and reducing operator dependency, ULTR-AI has the potential to significantly improve early detection and reduce patient drop-out rates, particularly in rural areas lacking specialized medical personnel.
Health officials and researchers are optimistic that ULTR-AI could transform Africa’s fight against TB by facilitating earlier, scalable detection, even in facilities lacking radiologists or lab infrastructure.
