Abstracts

Rationale: Among patients with pharmacoresistant focal epilepsy, focal cortical dysplasia (FCD) is identified as the most common substrate. Gray-white matter junction blurring occurs most frequently as an imaging characteristic of FCD. Subtle FCD lesions may be completely missed by conventional visual analysis of MRI. It has been estimated that at least 20% of patients undergoing epilepsy surgery have no identifiable MRI abnormality, even when reviewed by dedicated neuroradiologists. MRI-negative patients often have compromised surgical outcome. Voxel-based morphometric MRI post-processing technique facilitates detection of FCD by enhancing the gray-white matter junction, therefore has promise in increasing the diagnostic yield of MRI reading in patients previously interpreted as negative. The relevance and epileptogenicity of post-processing detected lesions in epilepsy surgical candidates are unclear; therefore it is important to confirm their pertinence by performing electrophysiological recordings.Methods: We included in this study all patients who: 1) underwent a preoperative MEG; 2) were considered MRI-negative prior to surgery and 3) had a postoperative follow-up time >6 months. Patients with negative MEG were excluded. A total of 15 patients were identified and retrospectively evaluated using the post-processing MRI analysis proposed by Huppertz et al. Regions with z-score greater than 4 standard deviations on the junction file were considered to be significant. MEG source localization was performed using standard single equivalent current dipole methods. Analyses with both modalities were performed blindly and independently. Statistical significance was assessed using Fisher exact test.Results: The detection rate of the post-processing technique in this cohort is 73% (11/15). Ten of the detected lesions had a concordant MEG correlate. Five patients had these concordant areas completely resected, and all 5 remained seizure free (exemplified by Figure 1). Surgical pathology indicates that all 5 had cortical dysplasia, with 1 type IIb, 4 type Ia. The other 5 MEG/post-processing concordant patients did not have the area of interest completely resected (exemplified by Figure 2), and only 1 remained seizure free. Patients in whom a concordant area was identified by both MEG and MRI post processing, had a significantly higher chance of achieving a seizure-free outcome following resection of this area (p ? 0.05).Conclusions: Improving noninvasive localization is paramount in MRI-negative focal epilepsies. MRI post-processing shows great promise in identifying subtle abnormalities. Concordant MRI post-processing and MEG analysis leads to identification of a structurally and electrically abnormal subtle lesion that should be more confidently localized and targeted. Positive areas detected by MRI post-processing, when confirmed to be epileptic, are likely harboring focal cortical dysplasia. Complete resection of the lesion may be necessary to achieve seizure freedom.