AI-powered X-ray screening proves to be an effective and economical tool for tuberculosis screening in the community

Tuberculosis remains the world's deadliest infectious disease. Yet one in four TB cases are either undiagnosed or unreported, allowing the disease to spread and delaying treatment.

The study evaluated computer-aided detection (CAD), an artificial intelligence-powered tool that analyses chest X-rays and is recommended by World Health Organization (WHO) for TB screening. Researchers compared CAD-supported chest X-ray screening alone with a strategy that added a point-of-care blood test measuring C-reactive protein (CRP), which WHO recommends for TB screening among people living with HIV.

The researchers wanted to find out whether the additional blood test could improve community screening programmes and reduce the need for more expensive confirmatory TB tests without missing cases. 

The trial was implemented by SolidarMed in Lesotho and the Human Sciences Research Council in South Africa. It enrolled more than 20,000 people in communities with a high TB burden and found that the simpler approach – CAD-supported chest X-ray screening alone – detected more TB cases and was less costly. In other words, combining CAD-supported chest X-ray screening with CRP, as done in this trial, offered no advantages in terms of either case detection or cost. Simulations carried out by the team also showed that the cost advantage of CAD-supported chest X-ray screening alone remained consistent across different TB prevalence scenarios and population sizes.

“Countries such as Lesotho, which have a high TB burden but relatively limited resources for controlling the disease, can draw important insights from our findings regarding community-based TB screening approaches,” said Irene Ayakaka, Technical Director SolidarMed Lesotho and senior author of the study.

A key finding was that nearly half of people with confirmed TB did not report classic symptoms such as cough, fever, night sweats or weight loss. This underlines the importance of active screening approaches that do not rely only on people seeking care once symptoms appear.

“If we want to interrupt TB transmission, we need to find people earlier in their communities,” said Aita Signorell, Project Leader at Swiss TPH and co-author of the study. “Well-implemented, community-based screening strategies play a crucial role here, even though they are complex and costly.”

The trial used an innovative design called paired screen-positive design, which enabled the team to examine the two screening approaches in the same person. 

“Compared with the commonly used cluster-randomised trial design, the paired screen-positive design is more efficient, achieving the desired power with a smaller number of participants,” said Fiona Vanobberghen, Senior Statistician at Swiss TPH and co-author of the study. 

The findings will help improve TB detection beyond the passive case-detection strategies that still prevail in many settings. They also provide important evidence for national TB programmes and global health organisations designing future community screening campaigns. 

“TB often remains undetected for too long, especially when people do not have symptoms,” said Klaus Reither, Head of Clinical Tuberculosis Research at Swiss TPH, lead author and PI of the project. “Our findings show that CAD-supported chest X-ray screening alone can be a powerful tool to find TB in communities.”

This work was supported by the European and Developing Countries Clinical Trials Partnership (EDCTP).

For media enquiries, please contact communications@swisstph.ch.