A Simple, Interpretable Machine Learning Model Based on Clinical Factors Accurately Predicts Incident Dysplasia or Malignancy in Barrett's Esophagus.
Source check: Source checked. Not independent scientific confirmation.
Finding: A decision tree trained on a retrospective Barrett esophagus cohort identified segment length alone as a useful risk-stratification predictor for observed progression. In 598 patients (61 progressors), the held-out test set achieved AUROC .79; lowering the threshold raised sensitivity to 93.3% with specificity 58.5%. This is single-center retrospective evidence and is not an externally validated surveillance recommendation.
Technology: Supervised decision tree among six ML classifiers uses five candidate predictors; the selected tree relies only on Barrett segment length.
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- Supervised decision tree among six ML classifiers uses five candidate predictors; the selected tree relies only on Barrett segment length.
- Test and result
- In 598 patients with follow-up through Jan 2024, 61 progressed within five years. The tree’s held-out test AUROC was .79; default sensitivity/specificity were 73.3%/70.4%, and a lower threshold yielded 93.3% sensitivity/58.5% specificity; segment-length thresholds define four risk groups.
- Limitation / still unknown
- The model was developed on 75% of a single-center veteran cohort and tested on a 25% held-out split (n=150); progressor pathology was independently verified by a second expert pathologist. No external cohort was used, calibration was suboptimal, and the sensitivity threshold trades specificity for fewer false negatives.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC12587488/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.