Abstracts

RATIONALE: For temporal lobe epilepsy (TLE) patients, the association of lateralized [¹⁸F]fluorodeoxyglucose positron emission tomography (PET) hypometabolism with successful temporal lobectomy has been well established. The preponderance of patients in these previous analyses had lateralized and concordant interictal encephalograms (EEG) and magnetic resonance imaging (MRI) scans. In this study we determined the association of nonlateralized and nonconcordant EEG and MRI with lateralized hypometabolism on PET. This is a necessary first step in evaluating if PET can identify epileptic foci in TLE patients who have nonconcordant EEG and MRI.
METHODS: We searched our epilepsy center[ssquote]s electronic patient database and identified patients who had clinical as well as EEG or MRI evidence of TLE and who did not have any structural epileptogenic foci besides mesial temporal sclerosis (MTS). Patients were included in the study if they had video EEG, MRI, and PET studies and if the extent of lateralization on these studies was specifically reported. EEG was determined to be lateralized if [gte]80% of the interictal spikes arose from one temporal lobe. MRI and PET scans were interpreted visually. MTS was identified as T2 hyperintensity and/or atrophy in the hippocampi. The clinical radiographic reports categorized the severity of the MTS and hypometabolism as absent, mild, moderate, and severe. For each patient, we scored the MRI and PET studies as lateralizing if the two hippocampi were classified in different severity levels.
RESULTS: 78 of the 111 patients with nonlesional TLE met the inclusion criteria. 66 (85%) of these 78 patients had lateralized hypometabolism on PET. 11 of the 12 patients with nonlateralized interictal EEG had lateralized PET; statistically the same fraction as those with lateralized EEG (p=0.53). 16 of the 23 patients (70%) with nonlateralized MRI (93% for lateralized MRI, p[lt]0.01) and 27 of the 35 (77%) patients with nonconcordant EEG and MRI (93% concordant, p[lt]0.05) had lateralized hypometabolism on PET.
CONCLUSIONS: Fewer patients with nonconcordant EEG and MRI studies have lateralized PET scans than those with concordant studies. However a clinically significant fraction of the nonconcordant patients ([gte]70%) do have lateralized PET scans. Further analysis of these patients will help determine if a lateralized PET scan in the setting of nonconcordant EEG and MRI studies will predict a good surgical outcome.