Diagnostic accuracy of a commercially available, deep learning-based chest X-ray interpretation software for detecting culture-confirmed pulmonary tuberculosis
Bibliographic record
Abstract
BACKGROUND: Few evaluations of computer-aided detection (CAD) software for analyzing chest radiographs for tuberculosis have used mycobacterial culture as the reference standard. METHODS: Using data from a prospective study of symptomatic adults and household contacts of persons with tuberculosis who were seeking care in Karachi, we evaluated the accuracy of LUNIT INSIGHT version 3.1.0.0 (LUNIT, South Korea) for detecting pulmonary tuberculosis in the triage use case. The reference standard was liquid culture. We estimated the diagnostic accuracy using three developer-recommended threshold scores for tuberculosis: 15, 30, and 45. RESULTS: A total 269 of 2190 (12%) participants had culture-confirmed pulmonary tuberculosis. LUNIT-reported abnormalities of nodule, consolidation, fibrosis, and pleural effusion were more common with culture-confirmed tuberculosis. At the tuberculosis threshold score of 30, sensitivity and specificity were, respectively, 87.7% [95% CI: 83.2-91.4%] and 64.3% [62.1-66.4%]. Sensitivity was similar at scores of 15, 88.1% [95% CI: 83.6-91.7%] and 45, 86.6% [82.0 - 90.5%]; and specificity was 57.9% [55.7-60.2%] and 69.9% [67.8-71.9%], respectively. Sensitivity was lower for smear-negative disease, and specificity was lower with increasing age, previous tuberculosis, and decreasing body mass index. Diabetes and tobacco smoking did not modify accuracy. CONCLUSION: In a population where most tuberculosis was smear-positive, LUNIT-reported radiographic abnormalities were associated with culture-confirmed disease. Manufacturer-recommended threshold scores had limited sensitivity.
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 | 0.003 | 0.013 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 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 | 0.000 |
| 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".