Journal Article

Advanced neuroimaging in pediatric epilepsy surgery: state of the art and future perspectives

Consales, Alessandro, Parodi, Costanza, Resaz, Martina, Pacetti, Mattia, Couto, Rosa, Panzeri, Sofia, Tortora, Domenico, Nobile, Giulia, Francione, Stefano, Severino, Mariasavina, Rossi, Andrea, Ramaglia, Antonia

2026

Overview

PurposeTo review recent advances in structural MRI post-processing for pediatric drug-resistant epilepsy, with emphasis on artificial intelligence–driven and quantitative techniques, including MELD-Graph, MAP18, FLAT1, and SUPR-FLAIR, and to evaluate their impact on lesion detection, epileptogenic zone localization, and presurgical planning.MethodsNovel post-processing approaches were examined with respect to their computational foundations, imaging requirements, and diagnostic performance. Techniques employing machine learning, deep learning, voxel-based morphometry, cortical surface projection, and FLAIR/T1 ratio mapping were assessed for their applicability in children and their integration into multimodal evaluation pathways alongside electrophysiology and functional imaging.ResultsAdvanced post-processing tools substantially increase sensitivity for detecting subtle cortical abnormalities, particularly in MRI-negative pediatric epilepsy. MELD-Graph identify features of focal cortical dysplasia through automated surface-based analysis and deep neural network classification, achieving notable lesion detection even when conventional MRI findings are normal. MAP18 provides complementary voxel-wise morphometric assessment, improving specificity and benefiting from optimized structural sequences. FLAT1 enhances lesion conspicuity by quantifying FLAIR/T1 signal relationships, while SUPR-FLAIR improves visualization of cortical signal abnormalities through normalized FLAIR intensity projection onto the cortical surface. When incorporated into multimodal diagnostic workflows, these methods refine epileptogenic zone localization, inform individualized surgical strategies, and can reduce reliance on invasive testing.ConclusionAdvanced structural MRI post-processing is transforming the neuroradiological evaluation of pediatric drug-resistant epilepsy. By revealing subtle cortical abnormalities not visible on conventional imaging, these tools support more precise lesion characterization and surgical planning. Ongoing efforts toward standardization, clinical validation, and workflow integration will be essential to ensure widespread adoption and maximize clinical impact within precision-medicine approaches to pediatric epilepsy.Advanced structural MRI post-processing tools significantly improve the detection of subtle epileptogenic lesions like focal cortical dysplasia in pediatric epilepsy.Accurate localization requires multimodal integration of structural, metabolic, and functional data, with electrical imaging.Imaging informs personalized surgical planning by mapping eloquent cortex and predicting post-surgical seizure and cognitive outcomes using virtual models

• Publisher: Springer Nature B.V
• Subject: Abnormalities / Algorithms / Artificial intelligence / Artificial neural networks / Automation / Clinical medicine / Conspicuity / Convulsions & seizures / Deep learning / Drug resistance / Dysplasia / Electrophysiology / Epilepsy / Lesions / Localization / Machine learning / Magnetic resonance imaging / Mapping / Medical imaging / Metabolism / Morphometry / Neural networks / Neuroimaging / Pediatrics / Seizures / Sensory integration / Standard scores / Structure-function relationships / Surgery