S611 Diagnostic Accuracy of Artificial Intelligence in the Diagnosis of Intestinal Metaplasia and Dysplasia in Patients With Barrett’s Esophagus: A Diagnostic Test Accuracy Meta-Analysis
Bibliographic record
Abstract
Introduction: Diagnosing dysplasia in patients with Barrett’s esophagus is crucial in preventing esophageal cancer but challenging in clinical practice. Artificial intelligence (AI) could potentially provide better diagnostic accuracy. Methods: A comprehensive electronic search was conducted of cross-sectional studies examining the accuracy of AI in diagnosing intestinal metaplasia or dysplasia. Study selection, data extraction and quality assessment were completed by 2 authors independently. When a study used several models, the model with the highest sensitivity was used in meta-analysis. QUADAS-2 was used to assess risk of bias and applicability. Meta-analysis was performed using a bivariate model to obtain summary estimates of sensitivity, specificity, and diagnostic odds ratio. Results: Twenty-five out of 1479 articles were included. The diagnosis of intestinal metaplasia by AI algorithms had a sensitivity, specificity, and diagnostic odds ratio of 0.95 (CI 0.76-0.99), 0.96 (CI 0.60-0.99), and 394.05 (CI -102.36-890.45) respectively. The diagnosis of dysplasia compared to intestinal metaplasia by AI algorithms had a sensitivity, specificity, and diagnostic odds ratio of 0.91 (CI 0.88-0.94), 0.88 (CI 0.81-0.92), and 74.24 (CI -28.10-120.37) respectively. A sensitivity analysis by removing the largest studies did not change the overall accuracy of AI. Conclusion: AI algorithms seem to be accurate at detecting the presence of intestinal metaplasia and dysplasia. Future studies should evaluate the use of AI in combination with endoscopist opinion as this technology could be utilized as a clinical decision tool to better target biopsies for dysplasia.
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 | 0.044 | 0.122 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.010 | 0.059 |
| Bibliometrics | 0.009 | 0.009 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.005 | 0.003 |
| Open science | 0.003 | 0.002 |
| Research integrity | 0.003 | 0.003 |
| Insufficient payload (model declined to judge) | 0.005 | 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".