Abstracts

Rationale: Functional neuroimaging with Positron Emission Tomography (PET) has been a useful tool for the estimation of seizure onset zone, as part of presurgical workup. However, since not all PET studies can precisely determine the seizure onset zone, we developed a quantitative data analysis algorithm to isolate the cortical elements of PET, in an effort to increase the sensitivity of PET studies. Methods: A new software algorithm developed in our institution with the principal of gray matter segmentation as a means to extract the cortical portions of the co-registered PET, utilizing reproducible cerebral parcels from a modified brain atlas to perform symmetry analysis of the cortical PET, has enabled us to concentrate solely on the abnormally asymmetric cortical metabolism, thereby identifying the putative seizure onset zone. An important point for successful analysis and accuracy of this method is determination of the electrophysiological state of cerebral cortex within the first 30 minutes (uptake phase) of radiotracer injection, to delineate whether the radiotracer uptake took place during a true inter-ictal, ictal, or inter-ictal/ictal continuum. Results: A good example of the utility of our method can be demonstrated in the analysis of PET-MRI co-registration of a surgical patient, a 5-year-old ambidextrous boy with intractable focal epilepsy with potential left parietal, temporal onset on scalp EEG, and significant sleep potentiation of epileptiform discharges in slow wave sleep, with a normal brain MRI and non-localizing PET on conventional analysis.However, after a PET study that was obtained during sleep potentiated spiking, mimicking an ictal study, and using the aforementioned computer algorithm, we were actually able to localize the relative hypermetabolism to the left post central sulcus. A MEG study, which was also performed during sleep-potentiated spiking, revealed concordant data with maximum negativity in the same region. Moreover, another PET with radiotracer uptake during ictal state provided corroborating evidence. Conclusions: This case clearly demonstrates that our novel computer algorithm for quantitative analysis can identify subtle hypermetabolic features in an interictal PET with frequent inter-ictal spikes, at the same localization as ictal PET, and other modalities like MEG. Funding: No funding