A groundbreaking €10-million grant from the European Union will fund the development of an AI-powered ultrasound tool designed to transform tuberculosis (TB) diagnosis in Africa. The initiative, part of the Global Health EDCTP3 Joint Undertaking, seeks to replace slow, costly diagnostic methods with a portable, smartphone-compatible device capable of delivering rapid, accurate results in low-resource settings.
Dubbed "CAD LUS4TB", the five-year project will roll out initially in Benin, Mali, and South Africa, testing the device on over 3,000 patients. The tool, named "ULTR-AI", uses artificial intelligence to analyse lung ultrasound images in real time, providing frontline healthcare workers with instant diagnostic insights—even in areas lacking radiologists or advanced lab facilities.
AI needs to be more than just powerful; it has to be usable in the places that need it most, said Dr. Mary-Anne Hartley, Director of EPFL’s Laboratory for Intelligent Global Health Technologies. "ULTR-AI was built for real-world challenges, not just lab conditions."
TB remains one of the world’s deadliest infectious diseases, particularly in sub-Saharan Africa, where limited access to chest X-rays and sputum tests delays detection and treatment. ULTR-AI meets World Health Organisation (WHO) triage standards while offering a more practical alternative for remote clinics.
"Ultrasound interpretation is highly skill-dependent," explained Prof. Noémie Boillat-Blanco of Lausanne University Hospital (CHUV), a co-lead on the project. "This tool puts that capability into the hands of frontline workers, closing a critical gap in TB care."
The project is co-led by EPFL and CHUV, with contributions from a 10-member consortium of African and European institutions, including:
- Stellenbosch University (South Africa)
- Carnegie Mellon Africa (Rwanda)
- FIND(Foundation for Innovative New Diagnostics)
-Swiss Tropical and Public Health Institute
This collaboration ensures the technology is both scientifically robust and tailored to Africa’s healthcare realities. The AI algorithms powering ULTR-AI will be open-access, continuously refined using anonymised patient data from trial sites. This approach not only improves accuracy but also allows for future adaptations to diagnose pneumonia, heart failure, and other conditions, making it a versatile tool for primary care in underserved regions.
With TB claiming 1.5 million lives annually, innovations like ULTR-AI could be transformative. By bringing cutting-edge diagnostics to rural and resource-limited areas, the project aligns with global efforts to end TB by 2030, while setting a precedent for AI-driven healthcare solutions in low-income settings.
Article by Nyokabi Wanjiku
Photo/Google
Comment