Artificial Intelligence in Colonoscopy: A Systematic Review of Adenoma Versus Polyp Detection Rates
Bibliographic record
Abstract
Colonoscopy is the gold standard in the prevention of colorectal cancer, but the miss rates of adenoma are high, which restricts its efficacy. To improve lesion recognition, artificial intelligence (AI), especially computer-aided detection (CADe) systems, has been introduced. The aim of this systematic review was to compare AI-assisted colonoscopy in terms of its ability to improve adenoma detection rate (ADR) and polyp detection rate (PDR). An extensive search was performed on PubMed, Embase, and Cochrane Library from 2015 to 2025. There were 17 randomized controlled trials (RCTs) comparing the use of AI-assisted colonoscopy with normal colonoscopy. The methodological quality measure of the included RCTs was Cochrane Risk of Bias 2.0 (RoB 2.0), which subdivided the studies based on low risk, some concerns, or high risk of bias based on whether they were biased in this or that domain. The robVis tool was used to produce the visual summaries. AI-aided colonoscopy effectively enhanced both adenoma detection rate (ADR) and polyp detection rate (PDR) in all of the included studies over conventional colonoscopy. In adenoma detection, accuracy was more than 85%, and in polyp detection, more than 90%. The advantage was also found especially in the detection of small and flat adenomas, which are very often missed in routine practice. The use of AI in colonoscopy is strongly associated with an increase in the detection rate of adenoma and polyps, minimizing the risk of underdiagnosis. The results highlight the clinical promise of AI in the form of a decision-support tool across gastroenterologists and suggest that AI can be applied to enhance the outcomes of preventive and screening colorectal cancer. Future research should be cost-efficient and practical, and combined with some clinical activities.
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 | 0.013 | 0.060 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.011 | 0.012 |
| Bibliometrics | 0.009 | 0.009 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.003 | 0.002 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.002 | 0.001 |
| 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".