Can Artificial Intelligence (AI) Be Used to Accurately Detect Tuberculosis (TB) from Chest X-Rays? An Evaluation of Five AI Products for TB Triaging in a High TB Burden Setting
Bibliographic record
Abstract
Background: Artificial intelligence-powered computer aided detection (CAD) products can be trained to recognize tuberculosis (TB)-related abnormalities on chest radiographs in order to screen and triage for TB. There are a number of these products available commercially and interest in these tools in the field of TB has grown over the past few years. Methods: We evaluated five commercially available AI software products using a large dataset collected in three TB screening centers in Dhaka, Bangladesh. A total of 23,566 individuals whom visited these centers were consecutively enrolled in the study. All individuals received a CXR and an Xpert test. All CXR were read independently by a group of three Bangladeshi board-certified radiologists and five AI products: CAD4TB (v6.3.0), InferRead®DR (v2), Lunit INSIGHT for Chest Radiography (v4.9.0), JF CXR-1 (v2) and qXR (v3). Findings: All five AI products significantly outperformed the human readers. The areas under the receiver operating characteristic curves are qXR: 0·91 (95% CI:0·90-0·91), CAD4TB: 0.90 (95% CI:0·90-0·91), Lunit INSIGHT CXR: 0·89 (95% CI:0·88-0·89), InferRead®DR: 0·85 (95% CI:0·84-0·86) and JF CXR-1: 0·85 (95% CI:0·84-0·85). We also proposed a new analytical framework to evaluate tests used for screening and triaging and to inform threshold selection by consideration of both cost-effectiveness and ability to triage. Further, AI products performed differently across the subgroups of age, use cases, and prior TB history. Interpretation: These AI products can be useful screening and triage tools for active case finding in high TB-burden regions. Funding Statement: Government of Canada. Declaration of Interests: None declared. Ethics Approval Statement: All enrolled participants provided informed written consent. The study protocol was reviewed and approved by the Research Review Committee and the Ethical Review Committee at the International Centre for Diarrheal Disease Research, Bangladesh (icddr,b).
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.010 | 0.025 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".