Abstracts

Single Photon Emission Computed Tomography (SPECT) has been used to assist in localizing the ictal onset zone in patients with intractable epilepsy. Several methods including visual analysis, difference analysis, and subtraction ictal SPECT co-registered with MRI (SISCOM), have been used to analyze SPECT images in an effort to identify the ictal onset zone. However, visual and difference analyses are subject to interobserver variability in interpretation of results. We used Analysis of Functional Neuroimaging (AFNI), a powerful software tool, to analyze ictal and interictal SPECT studies, to lessen subjectivity in the analysis, and to compare results with ANALYZE. We analyzed pre-operative ictal and interictal SPECT studies done on nine patients who underwent epilepsy surgery and were subsequently seizure free for six or more months. Ictal and interictal scans were obtained by injecting Tc-99m ECD (ethyl cysteinate dimer, Neurolite). ANALYZE analysis included anatomic registration of ictal and interictal SPECT studies, and normalization of the scans to the mean intensity. Regions more than two standard deviations over the mean in the difference image between normalized ictal and interictal scans were calculated and identified. AFNI analysis involved registration of the interictal and template anatomy with the ictal SPECT scans using FLIRT (Functional Magnetic Resonance Imaging of the Brain[apos]s Linear Image Registration Tool), calculation of the mean and normalization of the scans to mean intensity, converting the anatomy image to Talairach coordinates, calculation and identification of brain voxels more than 2 standard deviation above the difference mean, and finally calculation of the voxel volume for both the ictal CBF increases and decreases and identification of the Talairach coordinates for the mean voxel volume location. We were able to visually compare the results from Analyze and AFNI. The regions of ictal CBF relative hyperperfusion and hypoperfusion evident on difference images were similar in ANALYZE and AFNI in all nine subjects. We were also able to obtain the voxel volumes and the Talairach coordinates for both hyperfused and hypoperfused regions using AFNI. Using AFNI, we obtained results nearly identical to those generated by ANALYZE. The voxel volume and Talairach coordinates of the largest regions of CBF increases and decreases were readily obtained. Simultaneous visualization of both the ictally hyperperfused and hypoperfused regions could be easily accomplished using AFNI as compared to ANALYZE, where this required a separate session of image processing. Using AFNI, active thresholding of the difference scans assisted in further evaluating regions of blood flow from 1-4SD above or below the mean, which is an advantage compared to ANALYZE. SPECT image analysis using AFNI can potentially assist in further localizing and characterizing seizure-induced perfusion changes in patients with intractable epilepsy.