Abstracts

Intracranial EEG monitoring using subdural electrodes is often required to delineate the epileptogenic zone and eloquent cortex during presurgical evaluation of patients with medically resistant epilepsy. Furthermore, in order to tailor the resection, it is essential to localize the implanted electrodes precisely. Among the available techniques for this purpose, co-registration of a 3D CT and MRI from the same patient has the best potential for realistic display all implanted electrodes relative to cortical anatomy. Our goal was to determine the accuracy of co-registering 3D CT and MRI to localize subdural electrodes in planning epilepsy surgery. Data from 18 consecutive surgical candidates who underwent intracranial EEG monitoring were used for the study. Intra-operatively, we obtained digital photos of subdural electrodes that were visible in the craniotomy and identified their location relative to cerebral anatomical landmarks. The location of subdural electrodes, obtained from a post-operative volumetric CT, was co-registered onto the surface of the patient[apos]s cortex, that was reconstructed from a pre-implant volumetric MRI using [italic]Curry[/italic] software and a minimum of three external fiducial markers (nasion, inion, left and right pre-auricular points). The location of subdural electrodes derived from the co-registered CT/MRI was compared to the digital operative photographs by using gyral/sulcal patterns and cortical vasculature. Subdural electrodes were accurately localized with co-registered CT/MRI in 15 of the 18 patients studied. Mean positional error compared to photography was approximately 3-4 mm. In the other 3 patients, two or more of the standard external fiducials were not imaged in either the MRI or CT, and co- registration was preformed with combination of external and internal markers. The mean co-registration error in these patients was approximately 8-10mm. Localization of subdural electrodes by 3D CT/MRI co-registration can be reliably and accurately performed in the routine clinical setting. Accuracy requires that both MRI and CT include the standard external fiducials. This technique is particularly useful when electrode contacts are not visible in the craniotomy, as is commonly the case when recording from non-convexity cortex. This method provides a foundation for multimodal imaging that integrates anatomical, electrical, pathological and functional data to improve preoperative planning.