Chest radiography has been used to evaluate individuals for tuberculosis for over a century . As is happening with other radiologic modalities, advances in artificial intelligence are transforming chest X-ray (CXR) interpretation by offering the potential to replace human readers with automated computer analysis, often called computer aided detection (CAD).
Although CAD software that analyze CXR for pulmonary tuberculosis are commercially available, there remains uncertainty surrounding their diagnostic accuracy. A systematic review found the majority of published studies focused on development of CAD algorithms rather than their evaluation in clinical contexts, and that common risks of overestimating diagnostic accuracy included selection bias and use of human CXRreading rather than microbiologic testing as reference standards . Meta-analysis could not be performed due to methodologic differences . High quality evidence was particularly scarce for deep learning-based CAD—an artificial intelligence method that is highly effective for image recognition . Given the small evidence base, and that most users will have no field experience with this novel technology, it is important to know if accuracy varies between populations and by patient characteristics. There are no published data on whether human immunodeficiency virus (HIV) infection affects accuracy of deep learning based CXR analysis, and effects of other patient characteristics on sensitivity and specificity were reported in only 1 study .
We performed an individual patient data (IPD) meta-analysis to address gaps in the evidence base on the diagnostic accuracy of CXR analysis with CAD for detecting tuberculosis. We focused on the use of CXR to evaluate individuals self-referring for symptoms of tuberculosis. In this context, chest radiography functions as a triage test: when the CXR is abnormal, sputum microbiologic tests are required to diagnose active pulmonary tuberculosis, whereas a normal CXR is sufficient to rule out active disease .
