Computer-aided diagnosis for colorectal polyp in comparison with endoscopists: A systematic review and meta-analysis
Bibliographic record
Abstract
Aims Computer-aided diagnosis (CADx) is anticipated to enhance the prediction of colorectal polyp histology. This study aims to clarify the diagnostic accuracy of CADx in surface pattern diagnosis of colorectal polyps compared with experienced and inexperienced endoscopists. Methods Registered in the International Prospective Register of Systematic Review (PROSPERO) (ID: CRD42019136918) and published this protocol in Open Science Framework (OSF) (https://osf.io/2bajy/), this systematic review included studies assessing the diagnostic accuracy of CADx versus colonoscopists. We conducted a comprehensive bibliographic search of the following databases: MEDLINE (Ovid), Embase (Ovid), and the Cochrane Central Register of Controlled Trials (CENTRAL) (Ovid). A bivariate random effects model was employed. The primary outcome was the comparison of sensitivity and specificity between CADx and experienced endoscopists; the secondary outcome was the comparison between CADx and inexperienced endoscopists. To mitigate the influence of variability and explore potential sources of heterogeneity, we performed a subgroup analysis. This was categorized into real-time imaging, which represents the direct, in situ analysis during colonoscopy, and still imaging, which involves post-procedure analysis of static images. To assess whether results were robust enough for the conclusions drawn in the review, we performed the sensitivity analysis including only studies with clear definitions of experienced endoscopists. Results Twenty-one studies involving 5,477 polyps were included. The prevalence of adenoma ranged from 13% to 83%. The pooled sensitivities of CADx and experienced endoscopists were 0.874 (95% confidence interval [CI] 0.822-0.912) and 0.876 (95% CI 0.826-0.914), respectively (p=0.932). The pooled specificities were 0.850 (95% CI 0.784-0.898) for CADx and 0.873 (95% CI 0.815-0.915) for experienced endoscopists (p=0.534). In nine studies comparing CADx with inexperienced endoscopists, the pooled sensitivities were 0.879 (95% CI 0.818-0.921) for CADx and 0.849 (95% CI 0.778-0.900) for inexperienced endoscopists (p=0.460). The pooled specificities were 0.838 (95% CI 0.775-0.883) for CADx and 0.774 (95% CI 0.701-0.833) for inexperienced endoscopists (p=0.161). A subgroup analysis was performed to investigate the impact of sequential endoscopic imaging of CADx data. We compared real-time imaging (n=12) with still imaging (n=9), and the sensitivity and specificity between the two groups showed no significant differences. We performed sensitivity analysis, excluding one study without clear definitions of experienced endoscopists. Even with this exclusion, there were no significant differences between CADx and experienced endoscopist groups, consistent with the primary results. Conclusions CADx does not demonstrate superior diagnostic accuracy in surface pattern diagnosis of colorectal polyps compared to endoscopists, regardless of their experience level. Publication History Article published online: 27 March 2025 © 2025. European Society of Gastrointestinal Endoscopy. All rights reserved. Georg Thieme Verlag KG Oswald-Hesse-Straße 50, 70469 Stuttgart, Germany
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 | 0.005 | 0.017 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.015 | 0.028 |
| Bibliometrics | 0.003 | 0.005 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.003 | 0.002 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.003 | 0.002 |
| Insufficient payload (model declined to judge) | 0.004 | 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".