Plasma inflammatory proteome profiles identify MASLD among children with overweight or obesity.

BACKGROUND & AIMS: Pediatric metabolic dysfunction-associated steatotic liver disease (MASLD) is increasingly prevalent among children with overweight or obesity, yet its early diagnosis remains a major clinical challenge. This study aimed to identify circulating inflammatory proteins associated with MASLD and to develop a proteomic risk score (ProScore) to improve diagnostic accuracy. METHODS: In this cross-sectional study of 161 children (median age 8.5 years) with overweight or obesity, MASLD was assessed by vibration-controlled transient elastography, with 42 cases identified. Plasma concentrations of 92 inflammation-related proteins were quantified using a high-throughput proximity extension assay. The ProScore was compared with eleven conventional anthropometric/metabolic indices (WHtR, METS-IR, SPISE, PNFI, VAI, LAP, TyG, TyG-ALT, TyG-WC, TyG-WHtR, and TyG-BMI) and a genetic risk score (GRS). Six machine learning algorithms were employed and diagnostic performance was assessed using area under the curve (AUC) with fivefold cross-validation. RESULTS: Fifteen proteins were significantly associated with MASLD. A six-protein panel (FGF-21, CDCP1, CD244, OPG, Flt3L, MCP-1) achieved the highest diagnostic accuracy (AUC = 0.84), exceeding that of all conventional indices (AUC = 0.65-0.78; all P < 0.05). ProScore performance remained robust in school-based validation (AUC = 0.83), with no substantial improvement when combined with conventional indices. Diagnostic accuracy was higher in children with lower GRS (AUC = 0.92) than in those with higher GRS (AUC = 0.80; P = 0.003). CONCLUSIONS: A proteomic signature of systemic inflammation provides accurate, non-invasive identification of MASLD in at-risk children, outperforming conventional metabolic and genetic tools, and may have utility in clinical and public health settings.