Abstracts

Rationale: Subtraction ictal SPECT co-registered to MRI (SISCOM) can be an essential tool in the presurgical evaluation and planning of patients with non-lesional, medically intractable focal epilepsy. Cerebral blood flow changes (both hyper and hypoperfusion) can be observed via this study which may demonstrate areas of brain activated and hence may reflect seizure onset and pathway of seizure propagation. Although some studies have previously described the perfusion patterns in subcortical structures such as thalamus, basal ganglia or cerebellum, many of subcortical regions and their perfusion patterns during seizures have not been well described. The purpose of our study is to describe subcortical ictal perfusion changes of various subcortical regions during temporal lobe epilepsy and its usefulness in seizure lateralization and localization. Methods: We studied a series of 41 consecutive temporal lobe epilepsy patients who underwent interictal and ictal SPECT perfusion studies, video EEG monitoring, volumetric MRI as well as stereo-encephalopagraphy (SEEG) for localization of the epileptogenic zone. Temporal lobe epilepsies were further subdivided into neocortical and mesial and into bi-temporal, right or left temporal in origin. The subcortical regions of interest were subdivided into basal ganglia (putamen, globus pallidus, caudate), thalamus, cerebellum (vermis or either of the cerebellar hemispheres), corpus callosum, brainstem (medulla, pons or midbrain), internal capsule and external capsule. Results: Median age of our patient population was 31 (range: 8-60) years. 39 % (16) of the patients were female. 27 (66%) patients had neocortical, 12 (29%) had mesial temporal and 2 (5%) had both mesial and neocortical temporal lobe epilepsy. 22 (54%) patients had left temporal, 10 (24%) had right temporal and 9 (22%) had bi-temporal lobe epilepsy. Basal ganglia (88% with hyperperfusion), thalamus (80%) and cerebellum (95%) regions were the most frequently hyperperfused subcortical regions during seizures on SISCOM. Basal ganglia and thalamus were more likely to be activate ipsilateral whereas cerebellum was more likely to be activated contralateral to the side of seizure onset. When basal ganglia, thalamus and cerebellum were co-activated in this fashion (basal ganglia, thalamus with ONLY ipsilateral activation along with contralateral cerebellar activation), this had a positive predictive value of 100 % in correctly lateralizing the side of seizure onset. When external capsule was activated only on one side, it was always the side ipsilateral to the side of seizure onset. When midbrain was activated only on one side, the positive predictive value of that side being ipsilateral to seizure focus was found to be 86%. Conclusions: Subcortical SISCOM hyperperfusion could offer additional information in terms of lateralization of temporal lobe epilepsy. It may serve as a useful confirmatory additive tool in lateralization of epilepsy. However, larger, multi-center studies are needed to expand on these initial findings. Funding: No funding