Artificial Intelligence and colorectal neoplasia detection performances in FIT+patients: a meta-analysis and systematic review
Bibliographic record
Abstract
Aims The combination of fecal immunochemical test (FIT) followed by a colonoscopy has established itself as one of the preferred population-based screening strategies. Optimizing endoscopists’ detection performances is essential for enhancing the effectiveness of Colorectal Cancer (CRC) screening programs in reducing incidence and mortality due to CRC. Despite extensive exploration of various techniques and technologies (ie mucosal exposure devices, chromoendoscopy), their impact on adenoma detection rate (ADR) has shown inconsistency across studies in this specific setting -FIT+population-. The aim of this meta-analysis is pooling data of all the randomized trials focused on this strategic subpopulation in order to address whether the implementation of a CADe system may increase the identification of CR neoplasia precursors within a structured colorectal cancer screening program based on FIT. Methods We searched MEDLINE, EMBASE, and Scopus databases until September 2023 for RCTs reporting diagnostic accuracy of CADe systems in the detection of colorectal neoplasia (PROSPERO: CRD42023462438). The primary outcome was pooled adenoma detection rate (ADR), and secondary outcomes were adenoma per colonoscopy (APC); advanced APC; serrated lesions; and non-neoplastic (i.e. hyperplastic) per colonoscopy. We calculated risk ratios (RRs), and performed meta-regression analysis in case of heterogeneity. Results Ten randomized trials on 5421 patients were included. ADR was higher in the CADe group than in the standard colonoscopy group (62.38% versus 48.35%; RR 1.18, 95% CI 1.08-1.30, I 2 :49,76%). CADe also resulted in higher detection performances of both advanced adenomas (RR 1.31, 95% CI 1.20-1.68, I 2 :46.62%), and serrated lesions (RR 1.20, 95% CI 1.10-1.31, I 2 :0%). On the other hand, more non-neoplastic polyps were removed in the CADe than the standard group (RR 1.17, 95% CI 1.02-1.34, I 2 :50.79%). In multivariable meta-regression, baseline ADR, and withdrawal time were simultaneous significant predictors of the proportion of the CADe effect on both ADR, and Advanced APC. Conclusions The use of CADe during colonoscopy results in an increased detection of adenomas, advanced adenomas, and serrated lesions in a FIT+setting. The expected higher prevalence of advanced adenomas in this subpopulation may have enhanced the risk of lesions overlooking and, thus the potential benefit of CADe systems implementation. The level of heterogeneity found appeared to be associated with variability in colonoscopy quality performances (ie Baseline ADR, and withdrawal time) across the studies, with relevant effect of AI on both ADR, and Advanced APC across those studies with low quality indicators. Higher rates of unnecessary removal of non-neoplastic polyps were also reported. Publication History Article published online: 15 April 2024 © 2024. European Society of Gastrointestinal Endoscopy. All rights reserved. Georg Thieme Verlag KG Rüdigerstraße 14, 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 | 0.008 | 0.025 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.011 | 0.027 |
| Bibliometrics | 0.003 | 0.005 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.003 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.002 | 0.002 |
| 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".