Abstracts

RATIONALE: Proton MR Sectroscopic Imaging is a non-invasive and sensitive technique in lateralizing the seizure focus in Temporal Lobe Epilepsy (TLE). It also allows assessment of the distribution of the biochemical disturbances in the temporal seizure focus. Dipole Modeling of scalp-recorded interictal and ictal EEG has been demonstrated to have high predictability in localizing the seizure onset into a sublobar cortex in TLE. We compared the localization value of both techiques in patients with TLE.
METHODS: Eleven patients who underwent presurgical evaluation for intractable TLE, were evaluated by multivoxel 1H spectroscopic imaging of both temporal lobes performed on a 1.5 T Siemens Vision Scanner. The imaging parameters were TR = 1.5 sec, TE = 270 ms, resulting in an imaging time of 6 minutes 31 seconds. We measured the spectra of N-Acetyl Aspartate (NAA), Choline, Creatine, Myoinositol and Lactate from a 2 X 3 voxel grid with the long axis parallel to the hippocampus. The region of interest sampled the amygdala/head of the hippocampus, the body and the tail, parahippocampal region and the temporal neocortex. Decreased NAA peak and NAA/Choline and Creatine ratio were mainly considered in the lateralization and localization of the temporal seizure focus and in predicting neuronal loss. We also performed dipole source analysis of scalp EEG of the temporal lobe spikes and seizures of all eleven patients. We used spherical head modeling (BESA 2000) to localize the EEG generators into temporal sublobar sources (temporal mesial, anterior, inferolateral and lateral sources). We compared the predicted localization of both techniques in all patients.
RESULTS: Nine patients demonstrated significant NAA depletion and Choline and Creatine elevation in the temporal lobe ipsilateral to the seizure focus. The changes were more profound in the amygdala/head of the hippocampus with less profoud changes in the hippocampal body/tail and lateral cortex. EEG source modeling predicted temporal mesial and/or anterior sources in eight patients. Four of these patients underwent standard temporal lobectomy and had successful outcome after surgery. Neuropathological examination revealed neuronal loss and gliotic tissue in the resected temporal lobes of all four patients. The other two patients demonstrated significant bilateral changes on spectroscopic imaging and had bilateral independent temporal lobe seizures. EEG source modeling of spikes and seizures arising from both sides revealed mesial or inferolateral sources.
CONCLUSIONS: Both 1H MR spectroscopic imaging and EEG source modeling of temporal spikes and seizures demonstrate high correlation in lateralizing and localizing the temporal lobe seizure origins into a temporal sublobar level. Both techniques may add further benefit in the non-invasive surgical evaluation.