O133 Objective detection of cortical glucose metabolic abnormalities in children with epilepsy surgery: Validation with intracranial EEG

Abstract

Objective To evaluate the accuracy of objectively identified hypo- and hypermetabolic cortex to detect the lobe(s) of seizure onset in young children with focal epilepsy. Methods Presurgical 2-deoxy-2[18F]fluoro-D-glucose positron emission tomography (FDG-PET) scans of 37 children (mean age: 9.7 years) with drug-resistant focal epilepsy, who underwent 2-stage epilepsy surgery guided by subdural EEG, were compared to pseudo-normal pediatric control PET data ( N = 64; age 1–18 years, mean age: 9.3 years) by statistical parametric mapping (SPM). The control PET data were from non-surgical epileptic children with normal MRI and FDG-PET on expert visual assessment and were also used to create age-specific pediatric PET templates. Accuracy of intracranial seizure onset zone (SOZ) detection by SPM-defined FDG-PET abnormalities was determined on the lobar level. Results At least one cluster of cortical metabolic abnormality was detected by SPM in 36 children. In the whole group, hypometabolic clusters had 75% sensitivity and 64% specificity to detect the SOZ lobe(s). Hypermetabolic clusters had a slighly higher specificity (69%) but low sensitivity (34%). A similar accuracy of SOZ detection was found in young children ( Discussion Objective detection of metabolic abnormalities by PET is difficult in young children due to lack of age-matched healthy controls. The findings validate the use of SPM with an age-matched pediatric pseudo-normal control group. Conclusion SPM with pediatric pseudo-normal control PET data can identify the lobe(s) of seizure onset with around 70% lobar accuracy even in young children. Significance This approach extends the use of FDG-PET for the objective detection of metabolic abnormalities in the seizure onset lobe(s) to young children above 1 year of age.