Abstracts

Rationale:
Single photon emission computed tomography (SPECT) and magnetoencephalography (MEG) are diagnostic tools used to assist localization of the seizure onset zone (SOZ) during presurgical workup for patients with pharmacoresistant focal epilepsy, and are most often used in cases where there is no clear causative lesion identified on magnetic resonance imaging (MRI). MEG helps determine potential SOZ by measuring cortically generated magnetic fields and the resulting dipoles which are then overlaid onto MRI images. Ictal SPECT scans incorporate injection of a radioactive isotope tracer at the seizure onset  followed by CT perfusion scan. Subsequent interictal-ictal SPECT scans are then subtracted to produce an image highlighting the region(s) of maximal cerebral perfusion during a seizure that serves as a surrogate for the SOZ. Both modalities present specific barriers that make them challenging to obtain, due to cost, scarcity, and in the case of SPECT, difficulty coordinating timing of tracer with an electrographic seizure. Our aim was to compare the predictive accuracy of MEG vs. ictal SPECT in localizing the SOZ by comparing to subsequent intracranial EEG (iEEG) localization, and to determine if temporal or extratemporal seizure onset is better localized by either modality.
Method:
This was a single center retrospective cohort study. Thirty-seven adult patients with medically refractory focal epilepsy who were treated at our facility underwent either ictal SPECT, MEG, or both prior to proceeding to iEEG as part of a presurgical workup between 2015 to present, and were included in this analysis. MEG and subtraction ictal-interictal SPECT localization were compared to the SOZ as determined by iEEG.
Results:
Thirty-seven patients were included in this study. Thirty-two patients had subtraction ictal-interictal SPECT. Eight of these were non-diagnostic due to improper timing of SPECT isotope or seizures outside of the SPECT window. An additional nine patients did not complete iEEG. Fifteen were included in the study in the SPECT arm. Seven patients were included who underwent MEG scan and went on to complete iEEG. Two of these patients also had a prior SPECT scan. Subtraction SPECT successfully predicted the SOZ in 11/15 (73.3%). MEG successfully predicted the SOZ in 5/7 (71.4%). There was no statistical significance between these groups (p = 0.67). SPECT successfully predicted 6/8 (75%) of temporal onset seizures and 5/7 (71.4%) extratemporal seizures (p=0.31). MEG successfully predicted 2/3 (66.7%) temporal onset seizures and 3/4 (75%) extratemporal onset seizures (p=0.54).
Conclusion:
There was no statistically significant difference between SPECT and MEG in effectiveness of predicting the SOZ as compared to iEEG. Both modalities are useful non-invasive tools in predicting the SOZ that can aid in pre-surgical localization in both temporal and extra-temporal epilepsy. Further research is warranted to delineate comparisons of these difficult-to-obtain diagnostic tools in terms of effectiveness, practicality, and outcomes in the context of a pre-surgical evaluation.
Funding:
:N/A