Abstracts

Rationale: Magnetoencephalography (MEG) is a non-invasive diagnostic tool used in the presurgical evaluation of patients with medically refractory epilepsy. Source localization using the Equivalent Current Dipole (ECD) method is the gold standard for localizing interictal spikes on MEG. However, the ECD method has a number of limitations. Synthetic Aperture Magnetometry¹ (SAM) (g2) is a method that has been used with some success in interictal spike localization. In this study we evaluate SAMepi, a modification of SAM(g2), for localization of interictal spikes. Like SAM(g2), SAMepi searches for excess kurtosis in virtual channels created using a beamformer technique. SAMepi was created to be more sensitive to frequent epileptic activity than SAM(g2). Methods: 41 subjects were retrospectively selected for inclusion in this study who 1) received a good quality MEG study at the National Institutes of Health (NIH) Clinical Center; 2) had epilepsy surgery at the NIH or Children's National Medical Center; and 3) had at least one year seizure outcome data available. The single run for each subject with the least artifact and the most MEG spikes was selected for analysis using the SAMepi algorithm. SAMepi results were then compared to the surgical resection location and seizure outcome for each subject. Results: The probability of SAMepi analysis providing a localized result was 67%. SAMepi provided a localizing result in 80% of cases where subjects received extra-temporal resections and 62% of cases where subjects received temporal resections. Results based on SAMepi overlap with surgical resection area and seizure freedom are detailed in Table 1. In this analysis, the sensitivity of SAMepi was 59% and the specificity was 63%. The positive predictive value of a concordant SAMepi result predicting Engel Class I outcome was 65%. The negative predictive value of a null SAMepi result was 43%, while the negative predictive value of a discordant SAMepi result was 86%. Conclusions: Like SAM(g2), SAMepi is a partially automated technique that facilitates the detection of MEG spikes, making analysis more efficient and less dependent on expert readers. This method is designed to be more sensitive to frequent spiking than SAM(g2). In this patient population with medically refractory epilepsy, the probability of SAMepi providing a localized result is similar to that historically reported for the ECD method³. In this study, failure to include the area of SAMepi spike localization in the surgical resection appears to be a predictor of poor surgical outcome. Future work will include a direct comparison with ECD results and SAM(g2). 1. Vrba, J. & Robinson, S. E. Signal Processing in Magnetoencephalography. Methods 25, 249-271 (2001). 2. Robinson, S. E., Nagarajan, S. S., Mantle, M., Gibbons, V. & Kirsch, H. Localization of interictal spikes using SAM(g2) and dipole fit. Neurol. Clin. Neurophysiol. NCN 2004, 74 (2004). 3. Stefan, H. et al. Magnetic brain source imaging of focal epileptic activity: a synopsis of 455 cases. Brain 126, 2396-2405 (2003).