Abstract
1841
Objectives To evaluate and compare the performance of both visual and voxel-based analysis (SPM) of the 18F-FDG PET scan (lobar localization and lateralization value of epileptic zone) and to study their role in presurgical investigation of partial epilepsy.
Methods 141 patients with intractable focal epilepsy (age range 10-59 years) underwent interictal 18F-FDG PET scan. PET images were analyzed by both visual assessment and SPM. On SPM analysis, a cut-off threshold p <0.01 was considered statistically significant to find epileptogenic zone (EZ). Lateralization value and performance of lobar localization were compared with the EZ determined by scalp video electroencephalographic monitoring (video-EEG) and MRI results.
Results MRI showed structural lesions in 98/141 (70%) patients and normal findings in 43/141 (30%). Visual analysis was positive in 89% of patients, while SPM in 87%. Although, the sensitivity was 88% for PET visual and 86% for SPM in patients without MRI structural lesions. The sensitivity of both PET visual and SPM was 85% in the correct localization of EZ, but SPM analysis performed higher specificity (99%vs96%) and accuracy (92%vs90%) than visual assessment. Moreover, SPM was better than visual assessment in the correct lateralization (98%vs96%). Also, in more than 50% of cases, visual and SPM analysis have been useful to guide or to avoid the placement of intracranial electrodes and to greatly improved the surgical resection.
Conclusions In our institution 18F-FDG PET with SPM have been routinely performed in the presurgical evaluation of patients with drug-resistant partial epilepsy in combination with clinical, electrophysiological and neuroimaging, demonstrating high values of sensitivity, specificity and diagnostic accuracy in the correct localization of epileptogenic foci. PET-SPM findings were also useful to better guide the placement of intracranial electrodes and to decide the best surgical approach.