Intraoperative Metabolomic-Guided Precision Surgery for Pediatric Brain Tumors: A Systematic Review of Multi-Modal Molecular Imaging Platforms and Artificial Intelligence Integration

BackgroundPediatric brain tumors are the leading cause of cancer death in children, with surgical resection critical for survival and neurodevelopment. Intraoperative molecular imaging has advanced in adults but remains limited in pediatrics. This review examines the availability of intraoperative metabolomic and molecular imaging including fluorescence-guided surgery, magnetic resonance imaging, and mass spectrometry, AI integration, and multi-modal imaging in pediatric brain tumor surgery. MethodsLiterature search was done in PubMed, Scopus, Web of Science, and Embase from 2010-2025. Included studies addressed intraoperative molecular imaging in pediatrics, metabolomic neurosurgery approaches, fluorescence-guided surgical techniques, or AI application in pediatric brain tumor care. ResultsOf 2,856 articles, 84 met criteria. Pediatric intraoperative imaging predominantly relies on magnetic resonance imaging (21 studies), with more limited metabolomic approaches (16 studies) and emerging fluorescence-guided surgery applications (9 studies). Intraoperative MRI increased gross total resection rates from approximately 67% with conventional surgery to 84-89% with iMRI guidance, while maintaining similar rates of new neurological deficits around 8%. Mass spectrometry shows promise for real-time tissue characterization but remains largely confined to adult neurosurgical populations. Fluorescence-guided surgery using 5-aminolevulinic acid (5-ALA) and sodium fluorescein has demonstrated safety in over 249 pediatric cases, with fluorescence utility correlating with tumor grade and proving most effective in glioblastoma (85% fluorescence rate) and anaplastic ependymoma (77%), but limited in pilocytic astrocytoma (26%) and medulloblastoma (39%). Artificial intelligence in pediatric neuroimaging improved tumor segmentation and outcome prediction across 15 studies, while two multimodal imaging studies integrating MRI with diffusion and PET demonstrated modified surgical plans in most cases involving eloquent brain regions and improved progression-free survival. Key gaps include: (1) limited pediatric metabolomic databases, (2) absence of real-time metabolomic platforms optimized for developing brains, (3) age-dependent variability in fluorescence-guided surgery efficacy, (4) insufficient integration of neurodevelopmental considerations into surgical planning, and (5) lack of standardized protocols for multi-modal imaging integration. ConclusionThe review highlights opportunities to advance intraoperative molecular imaging in pediatric neurosurgery via metabolomic-guided, fluorescence-guided, and AI-integrated approaches. Future research should develop pediatric-specific metabolomic platforms, optimize fluorescence imaging protocols for younger children, establish age-specific biomarker libraries, and create integrated decision-support systems considering oncological and neurodevelopment outcomes.