Analysis of scalp EEG and quantitative MRI in cases of temporal lobe epilepsy requiring intracranial electrographic monitoring.

Abstract Purpose. Contradictory scalp electroencephalographic (sEEG) and discordant imaging features are common in temporal lobe epilepsy (TLE). We assessed the relative importance of sEEG features and their relation to quantitative magnetic resonance (MR) imaging measures in regard to surgical outcome. Methods and materials. Patients with a putative TLE underwent extraoperative electrocorticography (eECoG) with bitemporal subdural electrodes for clarification of the ictogenic source. Patients were categorized by sEEG ictal patterns (IPs) as showing unilateral or bilateral onset. Concordance with the side of resection, as determined by eECoG, to that suggested by the predominant sEEG IP was further analysed as: (a) entirely ipsilateral eECoG IPs with discordant nonelectrographic data; (b) ipsilateral preponderant (> 80%) eECoG IPs; and c) contralateral preponderant (> 80%) eECoG IPs. Quantitative hippocampal volumes and signal characteristics were applied for comparison. Results. Of 26 patients, eECoG confirmed a unilateral IP on sEEG in 19 (73%). Of these 19, exclusively ipsilateral sEEG interical epileptiform discharges (IEDs) were identified in 9 (47%). When bilateral, generalized, absent or contralateral IEDs were found, 6 cases (60%) still showed a preponderantly ipsilateral IP identifying the epileptogenic side. In patients with sEEG bilateral IPs, 5 (71%) also had bilateral IEDs. Of the 16 patients who underwent resection, 14 (87.5%) achieved favourable outcomes and 9 (56%), seizure cessation. Hippocampal volumetry in 23 patients correctly lateralized 7 (30%) whereas fluid-attenuated inversion recovery (FLAIR) signal measures applied in 23 patients lateralized 9 (39%). Conclusions. Favourable surgical outcomes are attainable following eECoG in cases where ambiguity exists regarding the laterality of TLE on sEEG, even for those in whom bilateral IPs and either bilateral or no IEDs are demonstrated on sEEG. Neither volumetric nor FLAIR signal ratios were sufficiently reliable for lateralizing TLE in the majority of cases.