Abstracts

18-flourodeoxyglucose positron emission tomography (¹⁸FDG-PET) is an imaging modality commonly used to detect a seizure focus in patients with temporal lobe epilepsy in conjunction with other localizing modalities such as magnetic resonance imaging (MRI) and video EEG telemetry. The localizing value of PET imaging in the presurgical evaluation for temporal lobe resection remains unclear. We wished to evaluate the value of PET imaging in relation to presurgical MRI findings.
We investigated retrospectively the relationship between brain PET and MRI in 27 refractory epilepsy patients evaluated at the UC Davis Medical Center for anterior temporal lobe resection. This cohort of patients was selected on the basis of having received a PET imaging study as part of their pre-surgical evaluation. To ensure accurate localization, only patients with ictal video EEG recordings according to a previously established criteria were included in the study. MRI studies were evaluated for evidence of mesial temporal sclerosis (MTS) according to a previously established criteria. MRI study results were used to classify the patients into those with either MTS or non-MTS. Patients were also classified based on PET imaging reports showing lateralizing temporal lobe metabolic changes or non-lateralizing changes. Of the patients studied, 9 patients proceeded to temporal lobotomy surgery. The surgical outcomes were classified into either seizure free (Engel Class I) or not seizure free (Engel Class II-IV).
Out of the 27 patients studied, 16 patients had normal MRI studies. Of these 16 patients, 7 had normal PET studies (44%). In the remaining patients with abnormal MRI studies, 10 had evidence of unilateral MTS with concordant PET findings in 8. Two of the unilateral MTS MRI patients had symmetric PET studies. In the 7 patients with normal PET studies, all 7 of them (100%) also demonstrated normal MRI studies. There were 17 patients with lateralizing PET study findings. Of these patients, 7 of them demonstrated concordant MRI findings (41%), 8 had normal MRI scans. Review of the data suggests that patients with normal MRI are more likely to have normal PET scans compared to patients that have evidence of MTS on MRI scan (p[lt]0.001). Five of the 9 surgical patients became seizure free, of which 4 patients had MRI findings concordant to the resected lobe; 4 patients also had PET findings that were concordant to the surgical site. One seizure free patient had a normal MRI, while another had symmetric PET findings. Of the 4 non-seizure free patients, 3 had concordant MRI and PET findings.
The preliminary results of this study suggest that a normal MRI is more strongly correlated with a normal PET study. With regard to surgical outcomes, it appears that there is no significant difference in surgical free outcome between patients with lateralizing and non-lateralizing PET findings.