Electrocardiographic Alterations Combined with Hematological, Biochemical, and Metabolic Profiles Predict Prognosis in Kawasaki Disease.
Combining heart and blood test findings with machine learning accurately identifies children with Kawasaki disease at high risk for coronary complications.
A prospective cohort of 255 KD children with random forest ML models demonstrates that ECG parameters add incremental value over CBC/metabolic biomarkers alone, with a comprehensive multimodal model achieving AUC 0.92 for both coronary artery lesion prediction and IVIG resistance. This validates CBC-based ML as a clinically useful risk stratification tool for pediatric inflammatory vasculitis.
What the study was
- Study design
- Prospective cohort study with machine learning random forest models
- Population
- Hospitalized children with Kawasaki disease (West China Second University Hospital, 2022-2024)
- Sample size
- 255
- Category
- Diagnostics
- Maturity
- Validated
- Journal
- Journal of Cardiovascular Development and Disease
Why it surfaced
Prospective cohort with strong AUC (0.92) for clinically impactful outcomes in pediatric KD; CBC + ECG multimodal ML directly applicable to clinical triage; PMC open access.
A plain-language summary of published research — not medical advice. Talk to a clinician about your own care.