Abstracts

Rationale: Multimodal neuroimaging refers to the coregistration of one or more imaging data sets, often including both anatomic and functional imaging modalities. Multimodal imaging is commonly employed in the presurgical evaluation of medically-intractable epilepsy and is generally accepted to improve localization of the epileptogenic zone, therefore leading to superior seizure free outcomes following epilepsy surgery. However, few published studies support this hypothesis. We compare a cohort of patients evaluated at our institution before and after the use of coregistered multimodal neuroimaging as part of the presurgical evaluation to determine impact on surgical planning and postoperative outcome. Methods: We retrospectively reviewed charts of patients who underwent presurgical evaluation and resective epilepsy surgery with at least 1 year follow up for treatment of intractable epilepsy between 1/2006-12/2012 using the Cook Children's Comprehensive Epilepsy Surgery Database. Included patients had at least one anatomic and one functional neuroimaging study completed as part of their evaluation. Patients that did not grant consent, those with preoperatively designated palliative procedures, and those without adequate data to determine surgical variables were excluded. Multiple pre-, peri-, and postoperative variables were abstracted and compared between patients evaluated before and after the institution of coregistered multimodal neuroimaging at our facility. Results: One hundred and fifty seven patients were included in the cohort; 58 in the pre-coregistration group and 99 who underwent coregistered imaging. All patients had localization related epilepsy. Patients in the pre-coregisteration group had more AED exposures prior to surgery (M₀=3.78, M₁=2.71, p =.002) and more often had multifocal onset on preoperative video eeg (p₀=17%, p ₁=2%, p =.028), though other presurgical variables were similar. Controlling for onset observed on preoperative video eeg, patients undergoing multimodal coregistration less often required invasive subdural monitoring for localization of the epileptogenic zone (F_(1,154) = 7.08, p=.009). When subdural monitoring was employed in this group, fewer electrodes were required (M₀=60 , M ₁=51), though this did not meet statistical significance (p=.07). Patients with coregistered imaging more often experienced favorable outcome (Engel's class 1 or 2) at all follow up time periods (1 month [82%], χ²=10.30, p =.006; 6 months [72%], χ² = 6.14, p =.046; 1 year [69%], χ² = 10.36, p =.016; and 2 years follow up [60%], χ² = 11.45,p =.01). Conclusions: Coregistration of multimodal neuroimaging improves localization of the epileptogenic zone reducing need for invasive subdural monitoring. Patients evaluated with coregistered multimodal imaging are more likely to experience favorable post-operative seizure frequency outcomes. These findings support the assertion that coregistration of multimodal neuroimaging positively impacts surgical planning and outcome and should be included as part of the presurgical evaluation of patients with intractable epilepsy.