Abstracts

Rationale: Due to the morbidities inherently associated with invasive neurosurgery and intracranially indwelling foreign substance, the extent of intracranial electrode implantation has to be judiciously determined based on clinical data and noninvasive investigative studies. As a result, intracranial electrodes are often implanted exclusively or predominantly in the cerebral hemisphere where the focus of habitual seizure onset is suggested by presurgical data. Given the increased morbidity associated with bilateral craniotomies and extensive implantation of intracranial electrodes, it is important to determine the need of implanting less number of electrodes over the hemisphere with less supporting presurgical localizing data. Methods: We retrospectively reviewed our epilepsy surgery database between 1999 and 2011 to identify patients who had asymmetric number of electrode contacts implanted in both hemispheres for seizure localization. We also proposed a system of scoring each presurgical noninvasive datum for predicting the likelihood of identifying seizures at the hemisphere with less number of intracranial electrode contacts (HLEC) (Table 1). Results: Thirty five patients had asymmetric number of intracranial electrode contacts implanted in both hemispheres. Thirteen patients (37.1%) had recorded seizure involving the HLEC. One patient (2.9%) developed complication (right hemiparesis and aphasia) during the intracranial EEG (iEEG) monitoring. Factors that were statistically significant in predicting HLEC seizures on iEEG are: 1) temporal lobe epilepsy (p = 0.02); 2) interictal scalp EEG discharges present at the HLEC (p = 0.04); 3) both interictal and ictal scalp EEG discharges present at the HLEC (p = 0.01). The median (range) aggregate score was 2 (1,3) for patients with HLEC seizures recorded on iEEG, and 1 (0,3) for patients without HLEC seizures (p = 0.001). Using this model of scoring system, the odds ratio of identifying HLEC seizures on iEEG was about 6 for each one point increment in the aggregate score. The area under the Receiver Operating Characteristic (ROC) curve for this model was 0.836, which suggests an excellent ability of the aggregate score in discriminating between patients with and without HLEC seizures on iEEG. Conclusions: Our study shows there was no increased complication rate associated with asymmetric implantation of intracranial electrodes over both hemispheres. Implanting electrodes over the hemisphere with less supporting presurgical data could be useful in selected patients, especially in patients with temporal lobe epilepsy, scalp interictal discharges involving the HLEC, or both scalp interictal and ictal discharges involving the HLEC. In addition, our proposed scoring system could be helpful in selecting patients with complicated epilepsy for implantation with asymmetric number of intracranial electrodes between the hemispheres.