Objective lateralization and localization of epileptic foci by FDG PET using voxel based analysis in children with intractable epilepsy undergoing surgery

Objectives We previously demonstrated that statistical parametric mapping (SPM) can reliably detect hypometabolic areas in epileptic children >5 yrs of age by using healthy young adults as control (Muzik et al., Neuroimage 2000). However, the role of this objective analysis has not yet been established in the presurgical lateralization and localization of epileptic foci in children with intractable epilepsy. We assessed the performance of SPM analysis of FDG PET in this group.

Methods 31 children with intractable focal epilepsy (age: 11±4 [5-18] yrs) underwent interictal FDG PET, followed by 2-stage epilepsy surgery with chronic intracranial EEG. PET images were analyzed using SPM2 by comparing individual patient’s FDG image with those of 13 young controls (age: 26±3 yrs). Seizure onset localization in intracranial EEG was taken as gold standard.

Results SPM increased the lateralization of seizure foci from 80%, with visual analysis, to 97% (p=0.04). Although visual or SPM analysis (with p<0.001 or 0.0001, uncorrected) had similar performance in the localization of seizure foci, use of higher SPM thresholds helped in identifying the seizure foci more accurately.

Results Visual analysis missed seizure onset in the medial cortex (n=7), whereas SPM missed in the lateral cortex (n=5). While visual analysis performed better in left than right hemispheric seizure foci (sensitivity of 83% vs 54%; p=0.009), SPM performed equally on both sides.

Conclusions SPM can complement the visual analysis of FDG PET in objective lateralization and localization of seizure foci in epileptic children >5 years of age.