Abstract
Objectives
To develop and validate interferon-γ-based early prediction models for MAS in pediatric SJIA patients.Methods
We conducted a retrospective multicentre study of hospitalized SJIA patients complicated with MAS across seven tertiary paediatric centres in China between January 2020 and December 2024. Demographic, clinical, laboratory, and cytokine data were extracted from electronic medical records. Univariate and multivariate logistic regression analyses identified independent risk factors. A nomogram and cytokine-based predictive models were constructed and externally validated.Results
In the training cohort, independent predictors of MAS included thrombocytopenia (≤ 181 × 10⁹/L; OR 31.5, P = 0.002), elevated AST (> 50 U/L; OR 24.5, P 1.76 mmol/L; OR 12.0, P 684 μg/L; OR 33.0, P 5.0 mg/L; OR 26.3, P 15 pg/mL; OR 29.4, P Conclusion
We developed and validated early prediction tools for MAS in paediatric SJIA by integrating serum IFN-γ with clinical parameters. These models may support timely diagnosis and treatment. Key points • Serum IFN-γ is a strong early biomarker for MAS in SJIA. • A cytokine-enhanced model improves MAS risk prediction in paediatric SJIA. • Ferritin, AST, and fibrinogen remain core predictors of MAS.