Screening
Comparative Performance of Metabolic Risk Indices for Predicting MASLD in Adolescents: Evidence from the Ewha Birth and Growth Study Hyeseung Jang* Hyeseung Jang Jang Jang Jang Jang Department of Preventive Medicine, college of Medicine, Ewha Womans University
Background
With the increasing prevalence of obesity, insulin resistance, and dyslipidemia among adolescents, metabolic dysfunction–associated steatotic liver disease (MASLD) has emerged as an important public health concern. Because pediatric MASLD is often asymptomatic and imaging-based diagnosis is limited in large-scale settings, noninvasive metabolic indices are widely used. Although most indices were developed in adults, recent studies highlight the need for pediatric-specific criteria, including cutoff values based on the fatty liver index (FLI) and hepatic steatosis index (HSI).
Methods
This study included 242 adolescents aged 13–15 years from the Ewha Birth and Growth Cohort and compared the predictive performance of the cholesterol–glucose index (CHG), triglyceride–glucose index (TyG), and single-point insulin sensitivity estimator (SPISE) for MASLD. MASLD was defined using pediatric-specific cutoff values (FLI ≥20 or HSI ≥30). Associations were evaluated using logistic regression. Predictive performance was assessed using receiver operating characteristic (ROC) curves and area under the curve (AUC), with between-index differences tested using ROC contrast tests. Optimal cutoffs for the metabolic indices were determined by maximizing the Youden index.
Results
The prevalence of MASLD defined by pediatric criteria was 24.4%. All three indices were significantly associated with MASLD. In adjusted analyses, CHG and TyG showed positive associations with MASLD risk, whereas SPISE showed an inverse association. In ROC analysis, SPISE demonstrated the highest predictive performance (AUC = 0.928, 95% CI 0.893–0.962), significantly outperforming TyG (AUC = 0.686) and CHG (AUC = 0.666). At the optimal cutoff, SPISE achieved high sensitivity (82.8%) and specificity (88.9%). CHG and TyG showed lower sensitivities (55.2% and 51.7%) and specificities (80.7% and 84.2%), respectively.
Conclusion
Using pediatric MASLD criteria proposed in prior studies, this study compared the predictive performance of CHG, TyG, and SPISE in adolescents. Among the candidate indices, SPISE showed superior predictive accuracy, supporting its utility as a simple tool for identifying MASLD risk in adolescents when imaging-based diagnosis is not feasible.
