A124 A SYSTEMATIC REVIEW OF THE EFFICACY OF ARTIFICIAL INTELLIGENCE IN IDENTIFYING BARRETT'S ESOPHAGUS NEOPLASIA
Bibliographic record
Abstract
Abstract Background Barrett’s Esophagus (BE) is a known precursor for esophageal adenocarcinoma and requires frequent surveillance with esophagogastroduodenoscopy. Due to tumour heterogeneity and logistical demands on endoscopists, identification of BE dysplasia is difficult and random biopsies are riddled with sampling error. Artificial Intelligence (AI), in the form of computer-aided detection, has entered the endoscopic realm to improve BE dysplasia detection. This systematic review aims to evaluate its efficacy in BE screening. Aims To survey the literature regarding the efficacy of machine learning tools in identifying BE dysplasia. The primary outcome was recognition of BE from a database of endoscopic images with histopathologic correlation utilizing a machine learning algorithm, with sensitivity and specificity reported. Methods Using the PRISMA framework, MEDLINE, EMBASE and Compendex databases were searched from inception to Sept 1, 2023. Of 1915 articles identified, 35 were selected for full-text review. Two reviewers, JB and JK, independently completed the literature review and discrepancies were reviewed by the PI. After applying inclusion and exclusion criteria, 20 studies were included in the systematic review. The quality of studies was assessed using the Newcastle-Ottawa Scale. Inclusion Criteria: -Must include Barrett's esophagus and / or EAC (esophageal adenocarcinoma) -Must include novel research (not a systematic review or MA) -Requires use of endoscopic images -English only, full text manuscripts Exclusion Criteria: -No prior surgery (esophagectomy) -Does not include esophageal squamous cell carcinoma Results Of the 20 articles selected, seven were published after 2021. All studies utilized a machine learning algorithm to aid in identification of BE dysplasia. Various imaging modalities were used, including white light imaging, narrow-band imaging, or volumetric laser endomicroscopy. A total of 3,886 patients were included with 5,605 images. Sensitivity ranged from 72 – 100%, specificity ranged from 64 – 94%. The overall quality of studies included was low. Conclusions Surveillance of BE dysplasia remains a difficult and time-intensive task. Computer-aided detection of BE dysplasia demonstrates strong performance independent of the machine learning algorithm or imaging modality. Meta-analyses are required to illustrate heterogeneity and power to bolster these preliminary findings. Funding Agencies None
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 | 0.016 | 0.071 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.012 | 0.010 |
| Bibliometrics | 0.018 | 0.016 |
| Science and technology studies | 0.001 | 0.002 |
| Scholarly communication | 0.004 | 0.004 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.002 | 0.001 |
| Insufficient payload (model declined to judge) | 0.007 | 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".