A9 IMPLEMENTATION STRATEGIES TO OPTIMIZE DIAGNOSTIC ACCURACY OF COMPUTER-ASSISTED OPTICAL POLYP DIAGNOSIS
Bibliographic record
Abstract
Abstract Background Artificial intelligence (AI) has enabled the development of computer-aided diagnosis (CADx) systems which offer real-time endoscopic pathology prediction of colorectal polyps. However, the clinical benefit of CADx assisted optical diagnosis remains questionable due to lack of diagnostic ability improvement compared to non CADx assisted optical diagnosis. Aims This study aimed to assess diagnostic performance of a novel implementation framework in which optical diagnosis replaces pathology only if CADx and endoscopist agree on polyp diagnosis. We aimed to compare this proposed framework to non CADx assisted optical diagnosis. Methods We performed a secondary analysis of a prospective cohort undergoing optical polyp diagnosis at our center. Polyps measuring ≤5mm (diminutive) with available non CADx assisted and CADx assisted optical polyp diagnosis documentation were included in our analysis. In CADx assisted cases, first the CADx diagnostic output was documented, followed by the endoscopist’s final diagnosis after having seen the CADx diagnostic information. Only cases where the endoscopist agreed with the CADx output were retained for CADx assisted optical diagnosis. Primary outcome was the diagnostic accuracy for cases in which the endoscopist agreed with the CADx diagnostic output versus non CADx assisted OD, using pathology as a reference standard. Secondary outcomes included surveillance interval agreement, prevalence of high-confidence diagnoses, and rectosigmoid negative predictive value (NPV). Results A total of 817 polyps were included in our analysis. Endoscopists agreed with the CADx diagnostic output in 74.3% of cases. Using CADx-assisted optical diagnosis based on CADx and endoscopist diagnostic agreement, 326/439 diminutive polyps could undergo optical polyp diagnosis in the CADx assisted arm and 378/378 in the non CADx assisted OD arm. Using diagnostic agreement between endoscopist and CADx as a framework for CADx assisted optical diagnosis demonstrated superior diagnostic accuracy, 82.8% (95% CI, 78.7-86.9), compared to accuracy of non CADx assisted optical diagnosis, 76.7% (95% CI, 71.3-80.0) (p=0.0256). Conclusions Our study demonstrates that using cases with diagnostic agreement between endoscopist and CADx increases diagnostic accuracy for CADx-assisted OD implementation. Using this framework, CADx assisted OD outperforms non CADx assisted optical diagnosis. Funding Agencies None
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 | 0.016 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.002 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.000 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
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".