Abstracts

Rationale:
Up to one third of children with epilepsy develop drug resistance and require neurosurgery to control their seizures. Optimized MRI detects an anatomical lesion related to epilepsy in almost two third of all cases; in these cases, the MRI lesion is considered a reliable guide for the definition of the epileptogenic zone since its resection is associated with a good seizure control. The remaining third are less likely to be offered definitive epilepsy surgery as the epileptogenic zone is difficult to estimate. Detailed electrophysiological methods can identify pathological areas considered non-lesional on conventional MRI. Magnetoencephalography (MEG) and high-density electroencephalography (HD-EEG) using Magnetic and Electric Source Imaging (MSI and ESI, respectively) are increasingly used in this cohort. In this study, we aim to assess the clinical utility of interictal MSI/ESI using dipole clustering in MRI-negative children with medically refractory epilepsy (MRE).
Method:
We localized interictal spikes (using dipoles) on simultaneous MEG and HD-EEG data from 11 MRI-negative children with MRE. For each dipole, we computed level of clustering and and used it to discriminate between clustered and scattered dipoles (Fig. 1a). We computed the distance from the seizure onset zone (SOZ) (Fig. 1b) and irritative zone (IZ) (Fig. 1c) defined by intracranial EEG, and assessed whether those distances were dependent on clustering. Finally, we assessed whether dipole proximity to the resection was predictive of outcome, which was dichotomized into optimal (Engel 1) and suboptimal (Engel 2-3). (Fig. 1d, Fig. 1e) No Engel 4 was observed in our cohort.
Results:
For both MSI and ESI, clustered dipoles had higher precision than scattered dipoles to the SOZ [MSI: 46% vs. 23%; ESI: 49% vs. 21%; p< 0.001] (Fig. 2a, left)and IZ [MSI: 67.2 % vs. 36.4 % ; ESI: 67.9% vs. 29.6%, p< 0.001] (Fig. 2b, left).  Clustered dipoles were closer to the SOZ (MSI: 15.5 mm vs. 28.9mm, p< 0.001; ESI: 15.8 mm vs.33.9 mm, p< 0.001) (Fig. 2a, right) (Fig. 2d, Fig. 2f)and IZ (MSI: 10.9 mm vs. 20.4 mm; ESI: 12.2 cm vs.25.4 mm; p< 0.001) (Fig. 2b, right)compared to scattered.In optimal outcome patients, clustered dipoles were closer to resection than scattered for MSI (distance from resection: 10.3mm vs 35.4 mm, p< 0.001)andESI (distance from resection: 7.1 mm vs 33.4mm, p< 0.001) (Fig. 2c, left); while this was not found in suboptimal outcomes. The proximity to resection was predictive of outcome for clustered dipoles of both MSI (odds-ratio: 0.934; p< 0.001) and ESI (odds-ratio: 0.892; p< 0.001), but not for scattered dipoles.(Fig. 2c, right) (Fig. 2e, Fig. 2g)
Conclusion:
Interictal MSI and ESI using dipole clustering helps localize the SOZ and IZ andfacilitates the prognostic assessment of MRI-negative children with MRE. The level of MSI or ESI clustering appears to allow recognition of interictal activity generated in the epileptogenic zone. Resection of areas indicated by clustered dipoles is associated with good surgical outcomes, in contrast to areas that generate scattered dipoles.
Funding:
:No
Funding:
was used to support this abstract.