The abstract is read at the publisher; the summary is JClub's.
AI-based computer-aided detection of pulmonary tuberculosis on chest X-rays in clinical cohorts shows pooled sensitivity 0.896 and specificity 0.689, with WHO triage thresholds largely within uncertainty.
- Why it matters: Screening and triage in settings with few radiologists depend on reliable CXR accuracy, but clinical-cohort performance and evidence certainty remain unclear.
- What they did: A systematic review and diagnostic accuracy meta-analysis searched five databases on 15 March 2026 for 2020–March 2026 studies, pooled 2×2 data from clinical cohorts using a bivariate random-effects model with 95% prediction intervals and QUADAS-2/GRADE.
- The result: Only nine evaluations from three cohorts were pooled; most evidence had high risk of bias, specificity was near but not clearly above WHO targets, and local threshold calibration plus prospective external validation are needed.
The findingWhy it mattersWhat they didThe result
AI summary of the abstract · may contain errors · not medical advice · read the paper.