Abstracts

Rationale: To be a surgical epilepsy candidate a patient must often obtain a combination of brain MRI, interictal FDG PET, and ictal/interictal HMPAO SPECT imaging to accurately localize an epileptic focus. We hypothesize that arterial spin labeling (ASL) can provide supplemental information in addition to routine MRI that can help localize an epileptic focus. We retrospectively analyzed patients referred for EEG monitoring who also had MRI with ASL querying whether regional interictal hypoperfusion from ASL correlates with focal activity seen on EEG. Methods: We retrospectively reviewed 42 patients who had EEG monitoring and brain MRI with ASL. Brain MRI was performed on a 3.0T clinical scanner using an 8-channel head coil. CBF (ml/100g/min) was estimated using pseudocontinuous ASL (pCASL).The EEGs were recorded using the international 10/20 system of scalp electrode placement. The EEGs were reviewed by a single epileptologist who was blinded to patients' imaging and previous EEG reports. MRI, including ASL data, was reviewed by a single neuroradiologist who was similarly blinded. Evaluation of ASL was performed by visual interpretation of color and greyscale cerebral blood flow (CBF) maps as well as symmetric placement of uniform regions of interest in the frontal, parietal, and temporal lobar cortical regions. Significant lobar hypoperfusion was defined by >10 ml/min/100 g of tissue or 20% relative difference from the contralateral region. This threshold was based on a prior voxel-level reproducibility study at our institution that demonstrated a standard deviation of 9.8 ± 1.9 ml/min/100 g of tissue or 23 ± 4% relative difference in grey matter. Results: We reviewed 42 patients that had both EEG and ASL. Of these, 25/42 had lateralized hypoperfusion on ASL, of which 14/25 had corresponding lateralization on EEG, 4/25 had conflicting lateralization on EEG, and 7/25 had normal or global changes on EEG. In the 17/42 that had normal or global changes on ASL, 7/17 had corresponding normal EEG and 10/17 had lateralized EEG findings. In terms of the EEG findings, 28/42 had lateralized EEG findings, of which 14/28 had corresponding lateralization on ASL, 4/28 had conflicting lateralization on ASL, and 10/28 had normal or global changes on ASL. In the 14/42 that had normal or global changes on EEG, 7/14 had corresponding normal or global changes on ASL and 7/14 had lateralized hypoperfusion on ASL. 10/42 had identifiable structural lesions, of which 8/10 had corresponding lateralization to both ASL and EEG, 1/10 lateralized with ASL but not with EEG, and 1/10 lateralized with EEG but not with ASL. Conclusions: In this study, we showed that epilepsy patients with lateralized findings on both ASL and EEG, 14 (78%) had corresponding lateralization as opposed to 4 (22%) with conflicting lateralization. Only 10 (24%) patients had identifiable structural lesions, which when present, 80% had corresponding lateralization to both ASL and EEG. The addition of ASL to routine brain MRI provides supplementary information that is helpful in the evaluation of epilepsy.