Answers are grounded in this paper and cite their sources.
Importance Early discrimination of major hepatic glycogen storage disease (GSD) subtypes is crucial for effective GSD management. Objective To discriminate between major hepatic GSD subtypes—types I, III, and VI/IX—using routine blood biochemical parameters. Methods A total of 159 pediatric patients with genetically confirmed hepatic GSD between July 2014 and July 2025 were retrospectively collected and were divided into a training set and a validation set. A random forest algorithm was first applied to identify the most predictive biochemical variables. Subsequently, a classification and regression tree model was developed based on the variables. Model performance was evaluated using overall accuracy, sensitivity, specificity, positive predictive value (PPV), and Cohen's kappa. Bootstrap resampling was used to assess model stability. Results The blood levels of triglycerides (TG), lactate (LAC), and creatine kinase (CK) were first identified as the top predictors. The resulting decision tree model classified GSD subtypes: TG ≥3.0 mmol/L and LAC ≥4.4 mmol/L for GSD I; either TG ≥3.0 mmol/L, LAC < 4.4 mmol/L, and CK ≥135.0 U/L, or TG <3.0 mmol/L, and CK ≥198.0 U/L for GSD III; TG <3.0 mmol/L and CK <198.0 U/L for GSD VI/IX. In the validation cohort, the model achieved an overall accuracy of 86.5% (kappa = 0.744), with a PPV of 91.7% for GSD I and 100% for GSD III. Interpretation This study provides an exploratory decision tree based on clinical laboratory parameters that may help discriminate among hepatic GSD subtypes, pending prospective multicenter validation.
No takes yet. Share an insight, caveat, or question.
Zhang et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: