Abstracts

RATIONALE: The utility of invasive recording in patients with normal MRI is not known.Surgical treatment for the control of medically refractory partial seizures relies on the precise localization of epileptic tissue.MRI and / or PET images and non-invasive monitoring are often adequate for ictal localization. When data from these modalities are negative, uninterpretable, or discordant, invasive recording with subdural or depth electrodes may be helpful. METHODS: We retrospectively reviewed our 12-year experience with 41 patients undergoing invasive chronic intracranial monitoring. Of these 41 patients, 14 had Mesial Temporal Sclerosis (MTS), 2 had other abnormalities, and 25 had MRI scans interpreted as normal. RESULTS: 6/25 patients with normal scans were monitored by depth electrodes alone or in conjunction with subdural electrodes (SDE) when scalp recordings were inadequate for localization. Depth electrodes in 2/6 patients verified localization determined by subdural recordings. In 3/6 patients depth electrodes localized seizures when SDE could not.In the remaining patient who was ultimately found to have neocortical epilepsy, depth electrodes did not demonstrate any seizure activity. 2/6 patients were seizure-free and 1 had a significant reduction in seizure frequency. In comparison, 10/19 patients who underwent surgical resection based on subdural electrodes alone remained seizure-free, while 2/19 had 50% or fewer seizures. Astrogliosis, neuronal loss and cortical dysplasia were the most common abnormalities found in specimens from patients with MTS. In patients with normal scans, perivascular lymphocytic infiltrates and macrophage aggregations consistent with chronic inflammation predominated. CONCLUSIONS: These data suggest surgical outcome in patients with a normal MRI and clinical features of temporal lobe epilepsy may not be improved by invasive monitoring despite adequate localization. Furthermore, the pathologic findings in patients with normal neuroimaging indicate that the underlying clinical entity differs significantly from MTS and may explain differences in outcome. Further studies are merited.