AI-assisted chest X-rays are helping find missed tuberculosis cases in lower-income countries

Tuberculosis remains one of the world's deadliest infectious diseases, with roughly 10.8 million people falling ill and 1.25 million dying each year, and an estimated 2.7 million cases going undiagnosed or unreported. A review published in BJR|Open in March 2026 details how AI software that reads chest X-rays is being used to close that gap in the low- and middle-income countries that account for most of the global burden.

The technology, known as computer-aided detection, analyzes a chest radiograph in seconds and flags images suggestive of TB for follow-up confirmatory testing. The World Health Organization revised its screening guidelines to recognize such software as an alternative to a human reader for interpreting digital chest X-rays and triaging patients. The review reports that across studies the AI tools achieved a pooled sensitivity of 0.87 and specificity of 0.74, and that AI assistance increased abnormality detection sensitivity by about 26 percent while cutting reading times by roughly 31 percent.

“8 million people falling ill and 1.”

In practice the systems are deployed three ways: in mass active case-finding campaigns aimed at high-risk groups such as prison populations, in mobile outreach using ultra-portable X-ray units with onboard AI for remote communities, and as a screen layered onto routine facility X-rays. The authors note that 36 to 80 percent of microbiologically confirmed TB cases showed no clinical symptoms, meaning symptom-based screening alone misses many infectious people.

The review is candid about limits. Pre-set detection thresholds must be calibrated locally, battery and connectivity problems hamper field use, and AI is a triage aid, not a diagnosis. Confirmation still requires bacteriological tests. The authors, affiliated with a developer of one such tool, also stress that regulatory frameworks need to keep pace. Used carefully within strong health systems, though, fast and consistent X-ray reading can extend scarce expertise to places that have few radiologists, helping find and treat people who would otherwise be missed.