Platform Session – Electroencephalography/Epilepsy: Magnetoencephalography imaging of surgically resected inter-ictal high frequency oscillation (80–200 Hz) was associated with seizure freedom

Abstract

Introduction Accurate delineation of epileptogenic zone(EZ) is crucial in patients with medically refractory epilepsy (MRE) warranting surgery and at least 33% will have seizure recurrence post-surgery. Intracranial EEG studies had showed that high frequency oscillations (HFO; 80–200 Hz) can accurately localize the EZ and patients had better post-surgical seizure freedom, if brain regions generating HFOs were resected out. However, none of the noninvasive studies evaluated the clinical utility of HFOs. So, we investigated epileptic HFOs source imaging to delineate the presurgical EZ and predict the seizure freedom. Methods Fifty-two patients (M: F = 37:15; age = 23 ± 8.9 years; disease duration 11.4 ± 7.7 years) with MRE were longitudinally recruited and studied with MEG. Each interictal epileptic discharge were marked by assessing spatial topography. Subject specific semi-realistic head model and a 3D grid voxels of 10 mm resolution was constructed. Source reconstruction was performed with frequency beamformer, by attenuating the source power at a specified location, subject to unit-gain constraint. Gain matrix was determined and cross-spectral density matrix on a Fourier transformed data was calculated at 7 distinct frequency bands; 8–14 Hz, 14–30 Hz, 1–30 Hz, 30–54 Hz, 55–80 Hz, 30–80 Hz, and 80–200 Hz. From the available clinical data, EZ was defined, respective surgery was performed and surgical outcome was assessed. The congruency of source construction with the EZ and surgically resected cortex was established at four levels, for each band & strength was quantified by kappa statistic. Sensitivity, positive predictive value (PPV) and accuracy was evaluated. Source localization error (SLE) was determined on the basis of minimum euclidean distance between the margin of reconstructed sources with the volume of the resected cortex. Results With surgery, 44 (84.6%) patients had good (Engel-I) and 8 (15.4%) patients had poor (Engel-II-IV) clinical outcome at 22.1 ± 10 months follow up. Compared to other frequency bands, interictal HFO imaging accurately delineated the epileptogenic cortex in 80.3% (k = 0.44 ± 0.12) of the patients, with 80.8% sensitivity, 100% PPV, and 80.32% accuracy. The highest possible concordance rate of 78.8% (k = 0.46 ± 0.1) with the surgically resected cortical volume, was observed for the reconstructed HFO sources. HFO source reconstruction did correlate with the seizure freedom (p Conclusion To date, this is the first ever largest prospective & noninvasive (MEG) study to investigate the clinical role of the HFOs, where HFOs precisely delineated the EZ & predicted the patients who might achieve seizure freedom with surgery, emerging as a key noninvasive epilepsy biomarker.