MétaCan
Menu
Back to cohort
Record W4323350964 · doi:10.1093/jcag/gwac036.125

A125 TANDEM STUDY DESIGN IS LESS LIKELY TO DEMONSTRATE IMPROVED ADENOMA DETECTION RATE THAN PARALLEL STUDY DESIGN IN THE ASSESSMENT OF ARTIFICIAL INTELLIGENCE-ASSISTED COLONOSCOPY

2023· article· en· W4323350964 on OpenAlexaff
Michelle Chae Min Lee, Thurarshen Jeyalingam, Colleen H. Parker, L W Liu

Bibliographic record

VenueJournal of the Canadian Association of Gastroenterology · 2023
Typearticle
Languageen
FieldMedicine
TopicRadiomics and Machine Learning in Medical Imaging
Canadian institutionsUniversity of TorontoUniversity Health Network
Fundersnot available
KeywordsColonoscopyRandomized controlled trialMedicineMEDLINESystematic reviewMeta-analysisAdenomaInternal medicineColorectal cancerCancer

Abstract

fetched live from OpenAlex

Abstract Background Randomized controlled trials have reported that artificial intelligence (AI) improves adenoma detection rate (ADR). Different methodologies, namely parallel and tandem study designs, have been employed to evaluate the efficacy of AI-assisted colonoscopy in randomized controlled trials. In systematic reviews and meta-analyses, a pooled effect that includes both study designs have been reported. However, it is unclear whether there are inconsistencies in the reported results of these two designs. Purpose To determine if there are differences in ADR using AI-aided technologies during colonoscopy between parallel and tandem study designs Method A systematic search of Ovid MEDLINE (1946 to October 2022) and EMBASE (1947 to October 2022) for randomized controlled trials comparing AI-assisted colonoscopy with routine high-definition colonoscopy in polyp detection was conducted. Reference lists of systematic reviews were searched for additional studies. The publications were divided based on trial design: parallel vs. tandem. Analysis was conducted using Review Manager 5.4.1 using a random effects model. Result(s) The search identified 540 articles. After screening the title and abstract for relevance, 19 randomized controlled trials involving a total of 14 657 patients were included for full-text review. Fourteen were parallel studies (14 136 patients) and 5 were tandem studies (521 patients). ADR was reported in 17 studies, and there was overall improvement in ADR with AI-assisted colonoscopy (risk ratio [RR] 1.33, 95% CI 1.22-1.44; p<.0001). Based on a separate pooled analyses of 13 parallel studies and 4 tandem studies, ADR significantly improved with AI assistance compared to routine colonoscopy, regardless of study design (RR 1.35, 95% CI 1.24-1.47 and p<.0001; RR 1.15, 95% CI 1.03-1.28; p=0.02, respectively). A significant increase in ADR with AI assistance were found in 84.6% (11/13) of parallel design studies, but in only 25% (1/4) of tandem studies. Image Conclusion(s) AI assistance during colonoscopy significantly increased ADR in both parallel and tandem studies; however, improvement in ADR was less likely to be seen in tandem compared to parallel studies. It remains unclear why this difference exists, but some hypotheses include smaller sample sizes in the tandem studies, significant heterogeneity in the tandem design, and differences in operator bias depending on study design. Better understanding the differences in these study designs will inform future studies of new endoscopic technologies. Disclosure of Interest None Declared

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 imitation

Not 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.

metaresearch head score (Codex)0.176
metaresearch head score (Gemma)0.302
Version: metacan-v3-hybrid-931329e0061cValidation status: machine_predicted_unvalidated
Candidate categoriesMetaresearch
Consensus categoriesnone
DomainCandidate signal: Methods · Consensus signal: none
Study designCandidate signal: Simulation or modeling · Consensus signal: none
GenreCandidate signal: Empirical · Consensus signal: none
Teacher disagreement score0.824
Threshold uncertainty score0.929

Distilled classifier scores by category (both heads)

CategoryCodexGemma
Metaresearch0.1760.302
Meta-epidemiology (narrow)0.0030.002
Meta-epidemiology (broad)0.0140.031
Bibliometrics0.0050.006
Science and technology studies0.0020.004
Scholarly communication0.0050.005
Open science0.0030.002
Research integrity0.0070.003
Insufficient payload (model declined to judge)0.0100.001

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.

Opus teacher head0.043
GPT teacher head0.327
Teacher spread0.284 · how far apart the two teachers sit on this one work
Validation statusscore_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it

Classification

machine, unvalidated

Machine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.

Study designSimulation or modeling
DomainMethods
GenreEmpirical

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".

Quick stats

Citations0
Published2023
Admission routes1
Has abstractyes

Explore more

Same venueJournal of the Canadian Association of GastroenterologySame topicRadiomics and Machine Learning in Medical ImagingFrench-language works237,207