Impact of study design on adenoma detection in the evaluation of artificial intelligence–aided colonoscopy: a systematic review and meta-analysis
Bibliographic record
Abstract
BACKGROUND AND AIMS: Randomized controlled trials (RCTs) have reported that artificial intelligence (AI) improves endoscopic polyp detection. Different methodologies-namely, parallel and tandem designs-have been used to evaluate the efficacy of AI-assisted colonoscopy in RCTs. Systematic reviews and meta-analyses have reported a pooled effect that includes both study designs. However, it is unclear whether there are inconsistencies in the reported results of these 2 designs. Here, we aimed to determine whether study characteristics moderate between-trial differences in outcomes when evaluating the effectiveness of AI-assisted polyp detection. METHODS: A systematic search of Ovid MEDLINE, Embase, Cochrane Central, Web of Science, and IEEE Xplore was performed through March 1, 2023, for RCTs comparing AI-assisted colonoscopy with routine high-definition colonoscopy in polyp detection. The primary outcome of interest was the impact of study type on the adenoma detection rate (ADR). Secondary outcomes included the impact of the study type on adenomas per colonoscopy and withdrawal time, as well as the impact of geographic location, AI system, and endoscopist experience on ADR. Pooled event analysis was performed using a random-effects model. RESULTS: = 62%; P < .001), with no statistical subgroup difference between study design. Both tandem and parallel study designs revealed improvement in adenomas per colonoscopy in AI-aided colonoscopies, but this improvement was more marked among tandem studies (P < .001). AI assistance significantly increased withdrawal times for parallel (P = .002), but not tandem, studies. ADR improvement was more marked among studies conducted in Asia compared to Europe and North America in a subgroup analysis (P = .007). Type of AI system used or endoscopist experience did not affect overall improvement in ADR. CONCLUSIONS: Either parallel or tandem study design can capture the improvement in ADR resulting from the use of AI-assisted polyp detection systems. Tandem studies powered to detect differences in endoscopic performance through paired comparison may be a resource-efficient method of evaluating new AI-assisted technologies.
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 | 0.021 | 0.055 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.018 | 0.038 |
| Bibliometrics | 0.003 | 0.004 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.004 | 0.002 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.004 | 0.003 |
| Insufficient payload (model declined to judge) | 0.003 | 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".