Dynamic multimodal brain function monitoring enables quantitative severity grading and prognostic prediction in pediatric neurocritical care.

OBJECTIVE: Multimodal Brain Function Monitoring (MBFM) enables simultaneous assessment of cerebral blood flow, intracranial pressure (ICP), and brain tissue oxygenation, providing dynamic evaluation for pediatric neurocritical care. This study evaluated MBFM-derived parameters, including perturbation factor (PF), edema factor (EF), ICP, and regional cerebral oxygen saturation (rSO₂), for disease severity stratification and prognostic prediction. METHODS: We prospectively enrolled 120 pediatric patients (6 months-12 years) in the PICU of the Children's Hospital, the Weifang People's Hospital (January 2022- November 2025). Continuous MBFM monitoring was performed for 72 h. Patients were stratified into mild, moderate, and severe groups using K-means and hierarchical clustering. Key predictive thresholds were identified via CART decision tree. Predictive performance for adverse outcomes, including Glasgow Outcome Scale scores, complications, and survival, was assessed by ROC analysis, logistic regression, Kaplan-Meier survival analysis, and decision curve analysis. RESULTS: PF, ICP, EF, and rSO₂ significantly differed among severity groups. PF predicted adverse outcomes with an AUC 0.86 (cutoff 0.34), while the combined model (PF + EF + ICP + rSO₂) improved the AUC to 0.91. Patients with PF > 0.34 and severe disease showed reduced 6-month survival (Log-rank χ² = 7.76, P = 0.005). CONCLUSIONS: MBFM provides a quantitative framework for grading pediatric neurocritical illness. PF is a core predictor, and integrating multimodal metrics enhances prognostic accuracy, guiding PICU interventions and outcome prediction.