New trial compares tech-driven TB screening in African communities

Tuberculosis remains a major public health problem in many African communities, where people often face barriers to care and diagnostic tools are limited. Missed or delayed diagnoses are especially harmful in hard-to-reach areas with high rates of TB/HIV co-infection, because they allow disease to spread, raise mortality and increase social and economic costs. To tackle these challenges, researchers are testing new, practical ways to find TB cases earlier and link people to treatment. Led by Aita Signorell, the study described here is a prospective, community-based, diagnostic, pragmatic trial that compares two population screening strategies. The trial takes place in the Butha Buthe District in Lesotho and the Greater Edendale area of Msunduzi Municipality, KwaZulu-Natal in South Africa. It asks whether using Computer-Aided Detection (CAD) software alone, specifically CAD4TBv7, or using CAD together with a point-of-care C-reactive protein (CRP) test (the CAD4TBv7-CRP approach) is better at finding active pulmonary TB in community settings and whether either approach offers superior cost-effectiveness for real-world programs.

The trial enrolls 20,000 adults who receive a posterior anterior digital chest X-ray analyzed by CAD software, which produces an abnormality score indicating the likelihood of TB. Thresholds that guide who receives further testing were chosen using the World Health Organization’s target product profile for a TB screening test. Under the CAD4TBv7 approach, a CAD score above a pre-specified threshold triggers confirmatory testing with Xpert MTB/RIF Ultra. Under the CAD4TBv7-CRP approach, a CAD score in a defined middle window prompts a second screening test, CRP; a CAD score above the upper threshold goes straight to Xpert MTB/RIF Ultra, and a CRP result above the selected cut-off likewise requires Xpert MTB/RIF Ultra. Participants with CAD scores below the lower threshold, and those with CRP levels below the cut-off, are considered screen-negative. The study uses a paired screen-positive design to compare the yield of detected TB cases and the cost-effectiveness of the two strategies. The protocol is registered (ClinicalTrials.gov NCT05526885; SANCTR DOH-27-092022-8096) and described in protocol version 4.0, 19 January 2024.

This pragmatic trial is the first direct comparison of CAD4TBv7 alone versus CAD4TBv7 combined with point-of-care CRP for community-based pulmonary TB screening. Because the inclusion criteria and delivery mirror real-world conditions, the findings should help health programs decide which screening approach is most suitable where resources are limited and access to care is challenging. If one approach detects more TB cases for the same or lower cost, it could be adopted by community outreach teams to improve early diagnosis, speed treatment initiation and reduce onward transmission. Results will be shared through peer-reviewed publications, conference presentations and the project’s website (https://tbtriage.com/), and the trial received ethics approvals from committees in Lesotho, South Africa and Switzerland, ensuring oversight as the team evaluates these promising screening strategies.