A dynamic machine learning model for prediction of NAFLD in a health checkup population: A longitudinal study.
Source check: Source checked. Not independent scientific confirmation.
Finding: An LSTM updated with successive health-check records predicted later NAFLD in retrospective cohorts, with AUROC .818 internally and .792 externally after five checkups; this does not show prospective clinical benefit. The reviewed source supports this bounded result; external generalization, causal effects, and sustained operational impact remain unestablished where the source reports retrospective or internal evaluation.
Technology: LSTM consumes successively updated records; L1-penalized logistic regression selected features and SHAP explained predictions.
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- LSTM consumes successively updated records; L1-penalized logistic regression selected features and SHAP explained predictions.
- Test and result
- In 10,493 retrospective participants, internal AUROC rose .729 to .818; external cohort of 1,728 had .700 to .792.
- Limitation / still unknown
- Training-set five-fold tuning and separate internal/external validation assessed discrimination and calibration; no prospective deployment or clinical benefit was tested.
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/PMC10412833/
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.