Abstracts

Rationale: While seizures are a common finding in patients with brain tumors, the spatial relationships of the tumor and the seizures are not well understood. We analyzed spatial relationship of tumor and seizures in patients who underwent 2-stage epilepsy surgery. Methods: Patients with brain tumor were included if they underwent 2-stage epilepsy surgery with implantation of intracranial electrodes (ICE). A total of 11 patients were included in the study. The number of ICE placed ranged from 40 to 108, with total of 779 electrodes (all patients combined). Each patient had a pre-op high resolution MRI and a post-implantation CT scan. First, T1W MRI was imported using BrainSuite, the skull stripped away leaving only the underlying brain. After creating a 3-D surface rendering of the brain, it was imported into a custom software package, where a post-implantation CT showing ICE locations was aligned and coregistered to the MRI. The software was then able to overlay the CT onto the MRI of the patient s brain and extrapolate electrode positions. All electrodes were then manually verified using intraoperative photographs to ensure correct placement. Each ICE was classified as Seizure onset (earliest sustained rhythmic changes on EEG distinct from background), Seizure spread (involved in seizure within 10 sec of onset) or neither. In addition, each ICE was designated as tumor (directly over or within the tumor), peri-tumor (adjacent to tumor) or no tumor. The tumor was also manually outlined on each slice of a pre-operative MRI, and these markings were then turned into three-dimensional renderings by the Smap software. Finally, with the reconstruction of the patient s brain, the electrode locations, and the tumor location, linear distance to tumor was calculated for each electrode in each patient using the Smap software. Results: When all electrodes were included in the analysis seizure onset electrodes were more likely to be outside of the tumor or peri-tumoral regions. When individual electrodes over the seizure onset region were compared, the tumor vs no tumor location were significantly different, with the seizure onset zone more likely to be over the no tumor region than tumor. Out of 779 total electrodes analyzed across all patients, 73 were on seizure onset (9.4%), 98 on seizure spread (12.6%) and 608 were on neither (78.1 %). The average distance of nearest tumor location for seizure onset electrode was 20.37 mm (SD=14.89), seizure spread electrode was 19.88 mm (SD=19.87) and neither electrodes was 28.50 mm (SD=18.77). The distance information was available on 61 of 77 seizure onset electrodes and it was more than 20 mm from the nearest tumor in 26 (42.62%). Conclusions: In most centers, patients with brain tumors and epileptic seizures are operated on with a primary goal of removing the tumor. However, since patients with tumors and seizures often have seizure onset regions far from the tumor location (more than 2 cms away in 43%), as a two-stage approach with electrocorticography is needed to also be certain that the patient s epileptic disorder is fully treated.