An interpretable machine learning model based on routine clinical indicators for early identification of obesity-related non-alcoholic fatty liver disease in children: a single-center retrospective study.
Source check: Source checked. Not independent scientific confirmation.
Finding: A single-center retrospective study of 439 children with obesity found that a seven-variable LightGBM model predicted ultrasound-defined fatty liver disease with test AUC 0.870; external validation remains needed.
Technology: Six machine-learning models were trained on seven routine clinical variables from 439 children with obesity; LightGBM performed best and SHAP ranked feature contributions.
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- Six machine-learning models were trained on seven routine clinical variables from 439 children with obesity; LightGBM performed best and SHAP ranked feature contributions.
- Test and result
- In a single-center retrospective cohort, LightGBM predicted ultrasound-defined fatty liver disease with test AUC 0.870, accuracy 0.823, sensitivity 0.787, specificity 0.855 and F1 0.807.
- Limitation / still unknown
- Only 439 children from one hospital were studied. External validation and prospective clinical utility before confirmatory ultrasound are not established.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://tp.amegroups.org/article/view/155877/html
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 2/5 · 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.