AI-powered TB screening project aims to transform diagnosis
Rather than relying solely on sputum samples healthcare workers could use the technology to rapidly identify people.
A new international research project led by Stellenbosch University (SU) aims to transform tuberculosis (TB) diagnosis by combining artificial intelligence (AI) with a simple finger-prick blood test to improve detection in areas with limited access to healthcare.
The project, known as AddiCAD, will combine AI-powered chest X-ray analysis with a biomarker blood test that measures the body’s immune response to TB, offering a faster and potentially more accurate way to identify people with the disease.
ALSO READ | Tuberculosis cases hit record high: WHO
Funded with R46 million (€2.5 million) by the Global Health European and Developing Countries Clinical Trials Partnership 3 (Global Health EDCTP3), the initiative officially launched in May and brings together researchers and organisations from South Africa, Namibia, The Gambia and Europe.
Despite being preventable and curable, tuberculosis continues to place a heavy burden on healthcare systems, particularly in low-resource settings.
Of the estimated 10.7 million new TB cases recorded globally each year, around 2.5 million people remain undiagnosed, largely because current testing methods can be expensive, laboratory-dependent and difficult to access.
The AddiCAD project seeks to address those challenges by combining CAD4TB, an AI system that analyses digital chest X-rays for signs of tuberculosis, with a fingerstick blood test that detects biomarkers linked to the disease.
Researchers say preliminary findings show the combined approach achieved a 20% improvement in specificity compared with CAD4TB alone, without reducing its ability to detect positive cases.
ALSO READ | TB still killing millions amid poverty and weak healthcare
This could reduce false-positive diagnoses while ensuring patients with TB are identified and treated more quickly.
Rather than relying solely on sputum samples, which are often difficult to obtain and process, healthcare workers could use the technology to rapidly identify people most likely to have TB and refer them for treatment sooner.
Over the course of the project, researchers will develop the biosensor and companion mobile application before validating the technology in a clinical study involving about 1,000 adults with suspected TB in South Africa, Namibia and The Gambia.
The consortium also plans to work closely with healthcare providers, patient representatives, regulators and commercial partners to support the future rollout of the technology if the trial proves successful.
According to SU associate professor in immunology and AddiCAD project co-ordinator, Professor Stephanus Malherbe, a timely diagnosis can prevent negative impacts.
ALSO READ | Old TB hospital now a haven for people of Eastwood in Pietermaritzburg
“For many people, a timely TB diagnosis can prevent negative consequences like transmission, lung damage, or death. Yet far too many diagnoses are delayed or missed,” said Malherbe.
What excites us about AddiCAD is its potential to bring together cutting-edge science and real-world usability in a way that could make accurate diagnosis more accessible where it is needed most.
The AddiCAD consortium includes Stellenbosch University, Delft Imaging Systems, Life SADX, LINQ Management GmbH, the London School of Hygiene & Tropical Medicine and the University of Namibia.
If successful, researchers believe the technology could significantly improve access to life-saving TB diagnosis and treatment in resource-limited communities.
