Expanding molecular diagnostic coverage for tuberculosis by combining computer-aided chest radiography and sputum specimen pooling: a modeling study from four high-burden countries

This cost-effectiveness modeling study assessed the impact of combining AI-assisted chest X-ray (CXR) interpretation and sputum specimen pooling on reducing tuberculosis (TB) diagnostic costs and expanding access to molecular testing in Bangladesh, Nigeria, Vietnam, and Zambia. The study used Xpert testing and AI probability scores from CAD4TB (Zambia) and qXR (all other countries) to evaluate four screening and testing approaches, ensuring a false negative rate of 5%. The optimal strategy involved using AI to rule out low-risk cases and guide pooled vs. individual testing in moderate- and high-risk individuals, leading to Xpert test savings of 50.8%–61.5% across countries. These savings could expand diagnostic coverage by 34%–160%, enabling more people to access molecular testing. The study concludes that AI-assisted screening and pooled testing strategies can optimize TB diagnostic resources, though customized approaches are needed for different populations and settings.