Abstracts

Rationale: To elucidate the utility and limitation of short-time MEG study for presurgical evaluation in epilepsy patients, supplementing the standard long-term in-patient scalp EEG (VEEG) monitoring. Methods: We reviewed the medical records of consecutive epilepsy patients who were considered candidates for epilepsy surgery, underwent VEEG and, within 6 months, also had a routine1-hour MEG, between February 2008 and January 2012. The detectability, localization and concordance of epileptiform discharges between VEEG and MEG were analyzed. Chi-square test with Bonferroni correction was used for statistics. Results: A total of 300 consecutive patients (160 male, 140 female) were included. Nine patients underwent 2 MEGs, and 5 out of 9 also underwent a second VEEG. Median age was 20.3 years (range 9 months - 68 years) and duration of epilepsy was 9.0 years (0 - 52 years). The mean duration of VEEG was 4.8 days (1 - 13 days). Interictal epileptiform discharges were identified in 245 out of 305 (80.3%) VEEGs and in 228 out of 309 (73.8%) MEGs. Ictal discharges were identified in 274 (89.8%) VEEGs and in 39 (12.6%) MEGs. Taken together, epileptic discharges were identified more often in the 5-day VEEG than in the 1-hour MEG: 294 (96.4%) VEEG vs 237 (76.7%) MEG (p<0.0001). However, results which localized the discharges were more similar: 241 (79.0%) 5-day VEEG vs 231 (74.8%) 1-hour MEG, (p=0.21). The locations of these discharges totaled 249 (28.1%) in frontal lobe, 280 (31.6%) in temporal, 118 (13.3%) in central, 157 (17.7%) in parietal, and 83 (9.4%) in occipital. The concordance of the discharges between VEEG and MEG was 22.9% in frontal, 38.6% in temporal, 18.6% in central, 21.7% in parietal, and 27.7% in occipital. Furthermore, the proportion of the localizable discharges only detected by MEG was 28.9% in frontal, 23.2% in temporal, 19.5% in central, 40.8% in parietal, and 49.4% in occipital. The concordance between VEEG and MEG was significant in the temporal lobe (p<0.05), and MEG's ability to detect and localize discharges not localized on VEEG was significantly higher in the occipital lobe (p<0.05). Of 111 patients who underwent lobar or multilobar resection, the resection lobe(s) was indicated by both VEEG and MEG in 35 (31.5%), by the union of VEEG and MEG in 7 (3.3%), by only VEEG in 27 (24.3%), and by only MEG in 13 (11.7%); and the resection was partially indicated by VEEG and MEG in 8 (7.2%), by only VEEG in 1 (0.9%) and by only MEG in 2 (1.8%). Thusfar, 6-month surgical outcome is available in 92 patients, and 52 (56.5%) were seizure free. The patients whose resected lobe(s) was indicated by both VEEG and MEG are more likely to be seizure free (19 of 27 patients, 70.4%) (p=0.08). For those patients who underwent surgery, the MEG provided unique localizing information in an additional 22 (19.8%). Conclusions: Compared to a 5-day VEEG in-patient monitoring stay, a one-hour MEG study localizes the majority of the epileptiform discharges in all lobes, providing unique additional localization especially in the occipital lobes.