Comparing the accuracy of computer-aided detection (CAD) software and radiologists from multiple countries for tuberculosis detection in chest X-Rays
Bibliographic record
Abstract
Nearly a third of TB cases go undetected annually. WHO recommends computer-aided detection (CAD) to enhance TB screening, with studies showing comparable performance to local radiologists. Using 774 chest X-rays from the South African National TB Prevalence Survey, we compared 12 CAD software with 11 radiologists from Nigeria, India, the UK, and the US, against a composite microbiological reference standard. Sensitivity, specificity and Cohen's kappa were calculated. Receiver-operating characteristic curves were developed for CAD and Euclidean distance assessed radiologists' alignment with the best-performing software. Binomial regression tested the impact of radiologists' characteristics on accuracy. Radiologist performance varied. On the restricted read, British radiologists had the highest sensitivity (78.7% [73.2-83.5%]) and Indian radiologists the lowest (67.1% [61.0-72.8%]). Specificity ranged from 75.8% (71.8-79.4%, Nigeria) to 84.3% (80.9-87.3%, the US). Radiologist performance was significantly impacted by HIV, prior TB, and age. The top CAD outperformed all except Indian radiologists when matching specificity. CAD with Conformité Européenne generally matched or surpassed radiologists. British radiologists' sensitivity was closest to the top CAD, while American radiologists were closest in specificity and overall. Experience, TB reads, and country had no significant impact on accuracy. CAD performed well against radiologists globally, highlighting potential to enhance access to care.
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 distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.003 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 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 teacher head, 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".