Abstracts

Rationale: Temporal lobe epilepsy is the commonest form of refractory focal epilepsy in adults. About 30% of temporal lobe epilepsy is due to visible structural lesions on MRI scans, which provide useful information regarding location of epilepsy surgery. Majority of these lesions have neuropathological findings consistent with hippocampal sclerosis. On the other hand, despite recent advent of neuroimaging technology, approximately 60-70% of temporal lobe seizure patients still have normal MRI findings, which are often referred as non-lesional temporal lobe epilepsy (NLTLE). Previous studies indicated that some NLTLE patients might have hippocampal sclerosis when neuropathology examination was performed. However, it is not clear whether there are clear association between FDG-PET findings and neuropathology findings of hippocampal sclerosis. The aim of this study is to compare the incidence of hippocampal sclerosis among FDG PET-positive and FDG PET-negative epilepsy patients whose MRI scans showed normal findings. Methods: We collected clinical information retrospectively on patients with intractable temporal lobe epilepsy who had negative high-resolution MRI (3T) and FDG-PET scans, and subsequently underwent epilepsy surgery between 2007 and 2009. The data set included patient demographics, MRI findings, FDG-PET findings, postoperative pathology reports, and their seizure outcomes. The study was performed at a Level 4 Comprehensive Epilepsy Center (Barrow Neurological Institute). Results: 42 patients who underwent surgical resection had normal 3T MRI brain scans). Among them, 15 (36%) had abnormal FDG-PET scan. Neuropathological findings in this MRI-negative but PET-positive subgroup includes: hippocampal sclerosis (15/15;100%) and cortical dysplasia/microdysgenesis (one patient who had dual pathology). Neuropathological findings in the remaining 27 patients (64%) who are MRI-negative PET-negative includes: hippocampal sclerosis ( 12/27; 44%); nodular heterotopia ( 2/27; 7%) and focal cortical dysplasia ( 1/27; 3%). In 12 out of 27 patients (44%) histopathological diagnosis was indeterminate. A 6 months epilepsy surgery outcome was correlated with histopathological diagnosis and PET scan findings ( Table 1). PET scan when positive has a predictive value of 100% in identifying hippocampal sclerosis and 94% in predicting an Engel Class I surgical outcome at 6 months. Conclusions: Our study suggests that MRI-negative but PET-positive patients have high-percentage of hippocampal sclerosis on neuropathological findings. In MRI negative and PET-negative patients, hippocampal sclerosis was far less common and large percentage of them had no clear pathological findings. In addition, PET-positive patients had significantly better postsurgical seizure outcomes compare to PET-negative patients.