Abstracts

Rationale: Insular/Peri-insular Cortex Epilepsy (IPICE) consists of seizures arising from insula alone or simultaneously from insula and the adjacent structures. Magnetoencephalography (MEG) may fail to detect insular activity due to its deep location and radially oriented currents. Here, we sought to evaluate MEG’s sensitivity for the detection of epileptic sources in the peri-insular region and assess the accuracy of MEG localization of these sources by comparing with the ictal onset zone identified by stereo-electroencephalography (SEEG). Methods: We retrospectively analyzed the data of 33 patients who had insular/peri-insular epilepsy confirmed by SEEG and who underwent successful MEG recording at Cleveland Clinic Epilepsy Center from 2009 to 2017. Patients’ demographics, pre-surgical workup, surgical management and outcome all were reviewed. MEG dipole types were classified into cluster (5 or more within 1 cm or less) or scatter (less than 5 dipoles or dipole sources distributed within more than 1 cm). Stable dipoles orientation (80% of the dipoles had a consistent orientation) are subdivided with reference to the sylvain fissure into perpendicular, parallel or oblique. Dipole sources were considered MEG unique if MEG spikes were not correlated with EEG spikes. MEG/SEEG concordance was defined by a complete or partial overlap between MEG dipoles and SEEG ictal onset zone, while MEG/SEEG discordance if there was no overlap. Results: Mean age at evaluation was 25.8 years (SD± 13.57; range: 3 - 62). 18 patients (54.5%) were female and 15 patients (45.5%) were male. 12 patients (36.3%) had failed prior epilepsy surgery. MRI brain showed relevant lesions in 12 patients (36.3%). SEEG confirmed the ictal onset zone from insula in 7 patients (21.2%) and simultaneously from insula and adjacent structures (16 from frontal, temporal and parietal operculum; 6 from frontal lobe; 5 from temporal lobe; 1 from cingulate) in 26 patients (78.7%). MEG showed abnormal results in 22 patients (66.6%) and was noncontributory (either normal or with only physiological benign variants) in 11 patients (33.3%). The dipole sources were classified as a cluster in 15 patients (68.1%) and scatter in 7 patients (31.8%). The dipoles had stable orientation in half of the patients (8 oblique and 3 perpendiculars) while the remaining half showed variable orientation. 5 patients had MEG unique spikes. 28 patients (84.4%) were treated surgically. The mean follow-up after surgery was 22.9 months (SD ± 17.70 months, range 3—84 months). Concordance analysis demonstrated MEG dipole sources localization were concordant with SEEG ictal onset zone in 19 patients (86.3%) and discordant only in 3 patients (13.6%) but lateralizing to the same hemisphere. Favorable surgical outcome (class I-II) was seen in 14 patients who showed MEG/SEEG concordance compared to 9 patients with discordance or unremarkable MEG.  The concordant MEG positive predictive value for the favorable surgical outcome was 0.87 (95% CI, 0.69–0.95). Conclusions: Our study demonstrates that MEG is a valuable modality for Insular/Peri-insular Cortex Epilepsy (IPICE) pre-surgical workup. MEG detected insular and peri-insular dipoles sources in two-thirds of our sample and had significant concordance with the SEEG ictal onset zone. Funding: No funding