Loading
Life sciences · Journal article
BMC Medical Informatics and Decision Making · October 21, 2026
No summary has been generated for this record yet. What follows is drawn from its source metadata only.
Journal article.
No findings were extractable from the material analysed.
Safety was not reported in the material analysed. Check the source before drawing any conclusion about harm.
The source did not state who this applies to in practice.
Graded across the dimensions that decide whether you should act, each from what the source actually supports. There is no single score, and where a dimension was not assessed it says so.
This record has not been graded across any dimension yet. Treat the label above as provisional and read the source.
What is missing. This record has no bottom line, key findings, reported figures, evidence dimensions. That is a gap in the analysis, not a judgement about the study.
Abstract Background Childhood obesity is increasingly accompanied by metabolic dysfunction-associated steatotic liver disease (MASLD), creating a need for accessible screening strategies in routine clinical care. We aimed to develop and externally validate a machine learning (ML) model based on routinely available clinical data to identify ultrasound-defined MASLD in children with overweight or obesity. Methods In this retrospective multicenter study, overweight or obese children aged 5–16 years were identified from outpatient records at two hospitals. Anthropometric measurements, routine blood tests, and biochemical indicators were collected, and MASLD status was determined by abdominal ultrasound. A stacking ensemble model integrating ten ML algorithms was developed using routinely available clinical variables, and its performance was assessed by 10-fold cross-validation and external validation. Complementary feature-contribution analysis was performed using SHAP for the AdaBoost base classifier. Results A total of 7,454 children (3,475 events and 3,979 non-events) from Nanjing and 353 children (129 events and 224 non-events) from Suzhou were included. Using 46 clinical variables, the stacking ensemble model achieved an area under the receiver operating characteristic curve (AUC) of 0.877 from patient-level out-of-fold predictions in 10-fold cross-validation within the development set and 0.865 in the independent external validation set, indicating good discrimination. However, external validation showed overprediction of absolute risk, highlighting the need for further recalibration and validation in larger and more diverse populations. SHAP analysis of the AdaBoost base classifier identified liver enzymes and metabolic markers as major contributors to its predictions. A web-based application was further developed to demonstrate individualized model output. Conclusion This multicenter machine-learning model demonstrated good discrimination for identifying ultrasound-defined MASLD in children with overweight or obesity. Built on routinely available clinical variables, this model may support future identification of ultrasound-defined MASLD after further validation and recalibration.