Rationale:
Magnetoencephalography (MEG) source localization of interictal epileptiform discharges (IEDs), grouped into clusters by equivalent current dipole (ECD), is the current standard of analysis for patients with medically refractory focal epilepsy. There is growing evidence that the propagation of IEDs can represent the pathway of the epileptogenic network. Moreover, treating both the onset zone and the epileptogenic network provides a greater chance for seizure freedom.1 We report the results of a retrospective study on ten patients with refractory epilepsy, where we show the spatiotemporal propagation of spikes has a strong predictive power for surgical outcomes.
Methods:
For this study, we selected ten refractory epilepsy patients who, as part of their surgical evaluation, had a MEG study, a resection surgery, a postop MRI scan, and a follow-up at least one year following the surgery. Five have been completely seizure-free, and the other five had no appreciable change (1a and 4b on the Engel Epilepsy Surgery Outcome Scale, respectively). The MEG source localization was performed using dipoles clustering and source distribution analysis (dSPM) over 100ms from the spikes onset. We used a novel algorithm to detect the resection region and detected whether the dipoles’ clusters and the temporal-spatial source distributions were localized inside or outside this region.
Results:
The spatiotemporal activity was localized in the region of the resection for all the patients with good outcomes. For the patients with no appreciable change, the spatiotemporal distribution had wide propagation and wasn’t localized to the resection zone. In Figure 1, we present two representative patients (1a and 4 Engel scores). For the first patient, clusters of dipoles were mapped on the border of the resection zone (A and B). The spatiotemporal activity was localized inside the resection zone (C and D). In E, we present the localizing of the resection zone. In F, presenting the second patient, two clusters of dipoles were localized, only one on the border of the resection zone. The spatiotemporal activity shows wide propagation, originating in the contralateral hemisphere.
Conclusions:
In this study, we show the MEG spatiotemporal propagation of IEDs has a strong predictive power for surgical outcomes. These results suggest that this kind of analysis should be considered to become part of the surgical planning procedure for refractory epilepsy patients. We also demonstrate a novel approach for analyzing the spatiotemporal propagation and identifying whether it localizes inside or outside the resection zone. More work is needed in order to show that method is significantly more accurate than dipoles clustering, which is the standard of care.
Funding: NCRR (S10RR014978) and NIH (S10RR031599, R01-NS069696, 5R01-NS060918,
U01 MH093765)