Development and validation of a novel interpretable machine learning model integrating immune-inflammatory indicators for intravenous immunoglobulin resistance in Kawasaki disease.
Source check: Source checked. Not independent scientific confirmation.
Finding: Interpretable models using immune-inflammatory indicators predicted IVIG resistance in Kawasaki disease and were evaluated in development/validation cohorts; the retrospective prediction result does not establish treatment benefit.
Technology: LightGBM, random forest, XGBoost, neural network, SVM and ElasticNet logistic regression models predict IVIG resistance in children with Kawasaki disease; SHAP explains feature contributions.
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- LightGBM, random forest, XGBoost, neural network, SVM and ElasticNet logistic regression models predict IVIG resistance in children with Kawasaki disease; SHAP explains feature contributions.
- Test and result
- Among 1,584 patients, LightGBM achieved test-set AUC 0.832 (95% CI 0.766–0.898), sensitivity 0.860 and specificity 0.639; leading features included fever duration, NLR, IL-1β, albumin and AST.
- Limitation / still unknown
- Retrospective single-hospital data were randomly divided into training and test sets; external prospective performance is unknown.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://doi.org/10.21037/tp-2025-1-907 · https://europepmc.org/article/MED/41982958 · https://europepmc.org/articles/PMC13071721 · https://www.ncbi.nlm.nih.gov/research/bionlp/RESTful/pmcoa.cgi/BioC_json/PMC13071721/unicode · https://pubmed.ncbi.nlm.nih.gov/41982958/
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.