FLUMAZENIL PET VERSUS INTRACRANIAL EEG FOR LOCALIZATION OF TEMPORAL AND EXTRATEMPORAL SEIZURE ONSET ZONE IN ADULTS
Abstract number :
1.231
Submission category :
Year :
2003
Submission ID :
2246
Source :
www.aesnet.org
Presentation date :
12/6/2003 12:00:00 AM
Published date :
Dec 1, 2003, 06:00 AM
Authors :
Christian Vollmar, Stephan Arnold, Rainer Linke, Alexander Drzezga, Peter Bartenstein, Soheyl Noachtar Department of Neurology, University of Munich, Munich, Germany; Department of Nuclear Medicine, University of Munich, Munich, Germany; Department of Nuc
[11C]-Flumazenil-PET (FMZ-PET) has been reported to localize the seizure onset zone, particularly in pediatric epilepsy patients. Aim of the study was to correlate localization of reduced GABA-receptor binding by FMZ-PET with the results from invasive EEG recording, using 3D image registation of FMZ-PET and MRI in adult temporal and extratemporal epilepsy patients.
We investigated 18 patients (13 male, 5 female, aged 15 to 52 years, mean age 36 years) with temporal (n=5) and extratemporal (n=13) epilepsy. All patients underwent cranial MRI, CT and FMZ-PET scans and and invasive monitoring with subdural grid and strip electrodes. MRI showed lesions in 15 of the 18 patients. In all patients the epileptogenic zone was identified by subdural EEG recording. So far 15 patients underwent resective epilepsy surgery. Analysis of FMZ-PET images was performed by automated quantitative analysis based on a set of 63 regions of interest and visual inspection by four different observers. PET findings were rated abnormal with a left-right asymmetry of more than 10% for automated quantification or after consensus of the observers for visual inspection.
The exact anatomic relationship between subdural electrodes and FMZ-PET findings was determined threedimensionally by image coregistration and volume rendering of MRI, CT and FMZ-PET scans.
Automated quantification of FMZ-PET localized the epileptogenic zone correctly in 6 of 18 (33%) patients and revealed false positive findings in 13 patients (72%). Consensus of visual evaluation detected the correct seizure onset zone in 9 patients (50%), false positive findings were reported in 8 cases (44%). Colorencoded 3D rendering of coregistered MRI and FMZ-PET data helped to identify the correct epileptogenic zone in 4 additional cases (22%) and ruled out false positive findings in 5 patients (28%). The results were similar in temporal and extratemporal epilepsies.
3D rendering of coregistered images improved interpretation of FMZ-PET studies in 50%, primarily because it allows discrimination of FMZ-PET hypointensities into enlarged CSF-space and true reductions of benzodiazepin-receptor density in morphologically normal brain.