Abstracts

Rationale: Focal Cortical dysplasia (FCD) is a common cause of medically refractory focal epilepsy. Identification of FCD lesion on conventional MRI is challenging as up to 30% of histologically proven FCD cases are negative on conventional MRI. The entire pre-surgical planning can be strongly affected by presence of lesion on the MRI, and surgical outcome can be affected by completeness of lesion resection. In this study we aim to retrospectively test a voxel-based MRI post-processing technique that can help the pre-surgical evaluation by reinforcing the areas of suspicion during patient management conference (PMC). Methods: We retrospectively studied 20 medically refractory focal epilepsy patients (10 frontal, 8 temporal and 1 frontotemporal and 1 temporo-occipital) and 20 age-matched controls with normal MRI. The analysis was blinded by de-identifying and randomizing the patient group and the normal controls. All selected patients (pts) underwent scalp video-EEG and had negative reported MRI and histology proven FCD. During multidisciplinary PMC the neuroradiologist/epileptologist/neurogurgeon re-assessed the MRI, in light of all other pre-surgical data, and noted a probable/questionable lesion that may be indicative of an epileptic focus. MRI post-processing was performed with methodology consistent with Huppertz et al., in a Morphometric Analysis Program (MAP) using Matlab SPM5. In both the patient and the control group, the gray/white junction file generated by MAP was analyzed by two readers for possible abnormalities with z score>4. Final analysis to account a lesion as significant (i.e., MAP+) was completed by a neuroradiologist in a blinded fashion. Results: Mean age was 31.35± 11.16 years and 31.55± 14.64 years for control and patient group respectively. In the control group, no areas of MAP abnormality were detected with z-score >4 and therefore were MAP negative. Of the 20 epilepsy pts; 10 pts had clear MAP+ areas (z>4) with corresponding abnormality included in the area of resection, indicating that MAP detected subtle dysplasia 50% of the time. 14/20 (70%) pts were seizure free (mean duration of follow-up was 25.1 month) of which 8/14 were MAP positive (57%). In the remaining 6/20 (30%) pts seizure recurred (mean time to seizure recurrence 7.21 months). Only 2 of these 6 pts were MAP positive. However, in both pts, resection did not include the MAP+ abnormality in its entirety, and therefore incomplete resection of the epileptogenic FCD lesion may have played a role in seizure recurrence. Conclusions: MAP is a promising and easily applicable MRI post-processing tool to identify cryptogenic FCD lesion in up to 50% of refractory epilepsy pts with presumed normal MRI in this cohort. MAP cannot detect FCD in some pts who attain seizure-freedom following surgical resection. At the same time incomplete resection of an MAP positive area may be associated with postoperative seizure recurrence. This study is part of a large-scale ongoing study on the utility of MAP in MRI negative focal epilepsies.