Electroclinical correlates of flumazenil and fluorodeoxyglucose PET abnormalities in lesional epilepsy

Objective: To analyze the clinical utility of [C-11]flumazenil (FMZ) PET to detect perilesional and remote cortical areas of abnormal benzodiazepine r eceptor binding in relation to MRI, 2-deoxy-2-[F-18] fluoro-D-glucose (FDG) PET, and electrocorticographic (ECoG) findings as well as clinical charact eristics of the epilepsy in epileptic patients with brain lesion. Backgroun d: The success of resective surgery in patients with medically intractable epilepsy and brain lesion depends not only on removal of the lesion itself but also on the reliable presurgical delineation of the epileptic cortex th at commonly extends beyond it. PET could provide a noninvasive identificati on of such epileptogenic areas. Methods: Seventeen patients underwent high resolution MRI, FDG and FMZ PET, and presurgical EEG evaluation, including chronic intracranial ECoG monitoring or intraoperative ECoG. Regional corti cal FDG/FMZ PET abnormalities were defined on partial volume-corrected PET images using an objective method based on a semiautomated definition of are as with abnormal asymmetry. Structural lesions were defined on coregistered MRI. The marked PET abnormalities visualized on three-dimensional cortical surface were compared with each other, to the extent of MRI-defined lesion , as well as to ECoG findings, Results: The mean surface extent of FMZ PET abnormalities was significantly larger than the corresponding structural le sions, but it was significantly smaller than areas of glucose hypometabolis m. The size of perilesional FDG PET abnormalities showed a correlation with the lifetime number of seizures (r = 0.93, p = 0.001). The extent of peril esional FMZ PET abnormalities was independent of the seizure number and sho wed an excellent correspondence with spiking cortex, the resection of which resulted in seizure-free outcome in all but one operated patient. Remote F MZ PET abnormalities (n = 6) were associated with early age at seizure onse t (p = 0.048) and appeared in ipsilateral synaptically connected regions fr om the lesion area. Conclusions: Three-dimensional surface-rendered FMZ PET is able to delineate perilesional epileptic cortex, and it may be especial ly useful to localize such areas in patients with extensive perilesional gl ucose hypometabolism associated with a large number of seizures. Remote FMZ PET abnormalities in patients with early onset and long duration of epilep sy might represent secondary epileptogenesis, but this requires further stu dy.