Abstracts

Rationale: In addition to electroencephalographic (EEG) data, neuropsychological testing and neurological exam, presurgical epilepsy evaluations typically include acquisition of imaging data that assess anatomy, metabolism and functional measures. Synthesizing a large amount of imaging data into a coherent set of images to correlate with clinical and electrophysiological data can be a challenge. Our data integration technique involves utilizing three dimensional (3D) volume renderings to enhance visualization of anatomy, metabolism and function. Methods: We retrospectively examined the imaging performed as part of a standard epilepsy protocol for pediatric epilepsy patients. Inclusion criteria required that the patient had three imaging modalities: an epilepsy protocol MRI, PET or SPECT abnormality, and a tasked-based fMRI. Some subjects also had functional connectivity MRI (fcMRI) and postoperative MRI. Data from each patient were processed according to modality. In BioImage Suite (1), 3D renderings of the linear co-registration of PET and structural T1 MRI were overlaid to identify the location of any abnormalities in metabolism. Functional MRI scans were post-processed using FSL, and a 3D rendering was linearly co-registered to the structural T1 and PET-T1 volumes. Data from SPECT scans were normalized and processed using BioImage Suite, and were linearly co-registered to the structural T1. Utilizing the metabolic and functional 3D renderings, the brain anatomy was assessed for abnormalities that are often more clearly observable in a 3D volume rendering than on two dimensional MRI views. These images also provide a more clear relationship to other anatomical landmarks. When available, postoperative MRI imaging was non-linearly co-registered with BioImage Suite to the preoperative MRI such that the metabolic and functional preoperative results could be compared to specific resection margins. Results: For a two year period, data from 16 patients (11M/5F; range 1-19 yr) were evaluated using 3D volume renderings of imaging data. Of the 16 patients, metabolic abnormalities occurred in the temporal lobe (4), parietal lobe (6), frontal lobe (4) and a mix of parietal and occipital lobe (2). Six of the patients also had postoperative MRI data. These retrospective reviews of patient data allowed neurologists and surgeons to visualize metabolic and functional data in relation to structural MRI data. Conclusions: Using 3D volume renderings allows the care team to incorporate results from different study modalities in an intuitive fashion. In addition, EEG data can be used to determine the relationship of the ictal and interictal epileptiform abnormalities to the location of the abnormal metabolic and functional data. The imaging is used to guide intracranial electrode studies and tailor resective surgery to increase success and decrease morbidity. Including 3D renderings as part of a standard presurgical epilepsy evaluation may enhance diagnoses from an imaging perspective.