S655 AI-Enhanced Colonoscopy Improves Adenoma Detection Without Delaying Procedure: Meta-Analysis of 9,500 Patients
Bibliographic record
Abstract
Introduction: Colorectal cancer (CRC) is a major global health challenge, with nearly 2 million new cases reported in 2020. Colonoscopy is the cornerstone of CRC prevention, and adenoma detection rate (ADR) is a validated performance metric inversely associated with interval CRC. Yet, ADR varies significantly among endoscopists. Computer-aided detection (CADe) systems using artificial intelligence (AI) offer real-time support that may improve detection consistency. While single studies and trials suggest benefit, a robust synthesis across diverse settings is lacking. Methods: Following PRISMA guidelines, we searched PubMed, Embase, Cochrane Library, and ClinicalTrials.gov (2000–2024) using terms such as “AI-assisted colonoscopy” and “adenoma detection.” Eligible studies included randomized controlled trials and cohort studies comparing AI-assisted to standard colonoscopy and reporting ADR, polyp detection rate (PDR), and missed adenoma rate (MAR). Two independent reviewers assessed study quality using Cochrane Risk of Bias and Newcastle-Ottawa tools. Pooled odds ratios (ORs) with 95% confidence intervals (CIs) were calculated using a random-effects model. Publication bias was evaluated using Egger’s test and funnel plots. Results: Thirty-eight studies (22 RCTs, 16 cohort studies) including 9,500 patients were analyzed. AI-assisted colonoscopy significantly increased ADR (31.5% vs 25.3%; odds ratio [OR] 1.48, 95% CI 1.35–1.62; P < 0.01) and PDR (48.2% vs 39.4%; OR 1.42, 95% CI 1.30–1.53; P < 0.01). MAR from tandem studies was significantly lower in the AI group (16.9% vs 24.5%; P = 0.01). Detection of diminutive adenomas (<5 mm; P = 0.02) and sessile serrated lesions (P = 0.03) also improved. No significant difference in withdrawal time was found (P = 0.07). Findings align with key multicenter trials including COLO-DETECT (OR 1.47, 95% CI 1.21–1.78; P < 0.0001). No significant publication bias was observed. Conclusion: AI-enhanced colonoscopy significantly improves ADR and reduces missed lesions, including subtle and flat polyps, without extending procedure time. These findings support routine AI integration into colorectal cancer screening to enhance diagnostic consistency and potentially reduce interval cancers.
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 | 0.015 | 0.031 |
| Meta-epidemiology (narrow) | 0.003 | 0.001 |
| Meta-epidemiology (broad) | 0.012 | 0.045 |
| Bibliometrics | 0.004 | 0.004 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.003 | 0.002 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.002 | 0.002 |
| Insufficient payload (model declined to judge) | 0.006 | 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".