Abstracts

Rationale: In the non-invasive workup of patients for epilepsy surgery, we routinely obtain a variety of anatomical MRIs, and one or more ictal SPECT studies leading to intracranial electrode placement. We use these, along with non-invasive electrophysiological data, to target stereotactic implantation to brain areas that are suspected to be involved in seizure onset and propagation. The effective use of these data requires that they be co-registered and fused for the planning of implants, the analysis of intracranial implant EEG and cortical stimulation data, as well as surgical planning and image guidance during resections. Methods: Anatomical MRI studies typically include a variety of T1- and T2-weighted sequences which document the patient’s individual anatomy, as well as any clear or suspected lesions that could be implicated in their seizures. Medically refractory patients who have stereotypical seizures and inadequate non-invasive evidence of seizure localization will have one or more ictal SPECT injections, which are combined with interictal SPECT studies to produce subtraction SPECT images. In our epilepsy center, SPECT images are given strong weight in the planning of intracranial implants, so that SPECT foci in both lesional and non-lesional brain locations are typically included among those sites targeted for implant. We assess our implant localization and intracranial data using co-registered and fused images obtained from pre-implant MRI, post-implant CT (with electrodes in place) and pre-implant subtraction SPECT(s). Pre-implant fMRI and PET studies may also be used in this way. We use the Mayo Analyze™ and SPM programs to process images and can format the resulting composite image data in a variety of ways for visualization and use in image guidance systems. Results: We have several years experience using these techniques and find them invaluable for the correlation and interpretation of anatomical (MRI, CT), metabolic (SPECT, PET) and electrophysiological data (ictal and interictal EEG, cortical stimulation findings). The techniques are relatively easy to implement using standard proprietary and shareware software packages, and the production of images appropriate for importation into image guidance systems is relatively straightforward. The necessary processing steps will be illustrated and examples of the resulting multimodal images and their use in epilepsy surgery cases will be shown. Conclusions: Precise multimodal co-registration and fusion of anatomical, metabolic and implant images, along with electrophysiological and stimulation data in a common coordinate system, is straightforward, and enhances the utility and interpretability of these data for localization purposes.