Computer-assisted optical diagnosis of colorectal polyps up to 10 mm
Bibliographic record
Abstract
Background: Computer-aided diagnosis (CADx) of colorectal polyps during colonoscopy could replace pathology for certain polyps. This study aimed to evaluate CADx-assisted optical diagnosis for polyps of ≤10 mm in the context of established quality benchmarks. Methods: We performed a post-hoc analysis of a randomized controlled trial evaluating assistive versus autonomous computer-aided optical diagnosis. Our primary outcome was achievement of the American Society for Gastrointestinal Endoscopy PIVI1 threshold for resect-and-discard implementation when CADx was used for polyps ≤3 mm. Secondary outcomes included PIVI1 threshold achievement when using CADx with a size cutoff of ≤5 mm and ≤10 mm, as well as diagnostic performance and prevalence of advanced histology across the polyp size groups. Results: We included 313 patients with a total of 463 polyps of ≤10 mm undergoing optical diagnosis with CADx assistance. Compared with pathology-based intervals, surveillance interval agreement was 94.6% (95%CI 91.3%–96.7%), 89.5% (95%CI 85.4%–92.5%), and 85.9% (95%CI 81.5%–89.5%) when CADx was used with size cutoffs ≤3 mm, ≤5 mm, and ≤10 mm, respectively. The diagnostic accuracy of CADx-assisted optical diagnosis was 76.2%, 76.6%, and 72.5% for polyps sized ≤3 mm, >3 to ≤5 mm, and >5 to ≤10 mm, respectively. The negative predictive value for rectosigmoid adenomas was >90% for all size groups (PIVI2). The prevalence of advanced or serrated pathology was higher in polyps >3 mm, which resulted in a higher number of incorrectly assigned surveillances intervals. Conclusions: In our study, CADx-assisted optical diagnosis met the resect-and-discard PIVI1 threshold only with a size cutoff of ≤3 mm, and the diagnose-and-leave PIVI2 threshold for polyps ≤10 mm.
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 | 0.000 | 0.000 |
| 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".