Abstracts

RATIONALE:_There is a group of patients with medically intractable epilepsy that have abnormalities on various imaging modalities, such as MRI, SPECT and PET, with each in its own right being suggestive, but when combined are highly supportive. However, this is usually restricted to "overlaying" the modalities, and visually "fusing" them. There is obvious problem associated with lack of objectivity with this and potentially limitation of the surgical resection. The aim of this study was to use a mathematical algorithm to co-register modalities and integrate them on an image-guided system to facilitate maximum resection of the epileptogenic focus. METHODS: Nine patients with medically intractable epilepsy underwent extensive investigations for their epilepsy. Investigations showed that 3 patients had maximal information on SPECT, 3 on PET and that 3 had very large lesions that required subdural grids. Using techniques of segmentation, surface matching and/or voxal-based matching, were able to co-register these modalities to the 3-D space of the volumetric MRI and use these images on the StealthStation Image-Guided System? to resect the lesion. RESULTS: Two lesions were in the frontal lobe and seven in the temporal lobe. Pathology included a low grade astrocytoma, a dysembryoplastic neuroepithelial tumour, but most cases only showed subplial gliosis and/or minor dysplastic changes. Surgical resection incorporating the epileptogenic focus as indicated by SPECT, PET and/or subdural grids was performed. Seizure outcome was 7 patients were Engel Class 1 (72%), 1 was Class 2 (11%), and 1 was Class 4 (11%). CONCLUSIONS: Mulimodality Image-Guided surgery for epilepsy offers further scope in the surgical management of patients with intractable epilepsy and means that we are likely to offer more patients surgey in the future.