Abstracts

Rationale: One-third of drug resistant epilepsy patients are potential candidates for epilepsy surgery. Finding a lesion is crucial for surgery outcome: seizure freedom rates may be up to 70% with vs. 46% without a lesion. Several imaging modalities are used to address this question but often expose patients to radiation, contrast agents, and lack spatial resolution. We studied Arterial Spin Labeling (ASL) MR imaging, a non-contrast, radiation-free technique, for presurgical seizure focus localization. Methods: Nine healthy controls (five females, ages 20-63) and thirteen patients (seven females, ages 19-57) with drug-resistant focal epilepsy were studied. Patients underwent presurgical testing with long-term video EEG, neuropsychology, magnetic resonance imaging (MRI), positron emission tomography (PET), functional MRI (fMRI), and magnetoencephalography (MEG) if indicated. 3D pseudocontinuous Arterial Spin Labeling (pCASL) images (3T GE Medical Systems) and 3D volumetric MPRAGE images (3T Philips Achieva) were acquired for all subjects. The pCASL image was coregistered to T1 MRI, and cerebral blood flow (CBF) weighted images were calculated from aligned ASL images after subtraction of labeled from control images. Volumetric segmentation was carried out using FreeSurfer (v6.0.0). Segmentations used to define regions of interest (ROIs) were applied to CBF-weighted images. Lesional ROIs were drawn in patients with MRI identified focal cortical dysplasias (FCD) and amygdala/hippocampus in patients with mesial temporal sclerosis (MTS). For patients with MTS, the regional relative CBF values within hippocampus and amygdala were measured and compared to both contralateral hemisphere and healthy controls. For patients with FCD, lesions masks were manually drawn by an experienced neurologist, and relative blood flow values within the subject’s lesion mask were compared to respective values within the same region in healthy controls. CBF weighted images were analyzed qualitatively by visual analysis and quantitatively by using an asymmetry index (AI) comparing ROIs to similar masks on healthy controls. Results: Three patients had extratemporal focal cortical dysplasia, six temporal lobe epilepsy (5 MTS), one bitemporal epilepsy, three temporal plus epilepsy. Visual analysis of pCASL MR images identified abnormalities within presumed seizure foci (from video EEG, MRI, PET) in 10/13 patients. Preliminary quantitative analysis in FCD regions showed statistically significant AI in 1/3 patients. AI values for amygdala and hippocampus ROI comparison between healthy controls (HC) and patients with MTS did not show clear statistical significance. HC mean AI value -0.179 (+/- 0.114). MTS patients mean AI value in the presumed seizure focus was -0.07 (+/-0.157). Conclusions: Our preliminary results suggest that, within a multimodal imaging approach, pCASL MRI may be used to provide additional evidence regarding seizure focus localization. Our preliminary data are limited by variability in healthy control relative regional blood flow. This can potentially be addressed by the use of a larger control cohort. Funding: No funding