Abstracts

Rationale: The extent of the minimally effective resection remains a controversy in the surgical treatment of medically intractable mesial temporal lobe epilepsy (MTLE). In the ROSE Trial cohort, we compare postsurgical volume changes by seizure remission status to link anatomical changes with outcomes.  Methods: Patients with MTLE were randomized to stereotactic radiosurgery (SRS) or anterior temporal lobectomy (ATL). Seizure remission was defined as absence of all but simple partial seizures between months 25-36 after intervention. The mesial structure volumes were obtained from semi-automated segmentation techniques on baseline and postoperative MRI (SRS 36m, ATL 3m). Fractional volume changes (FVC) for regions of interest ipsilateral to surgery were calculated as (baseline-pre)/baseline and corrected for contralateral changes. Median differences between seizure-free and not-seizure-free groups within treatment arms were tested with Mann-Whitney nonparametric tests. Overlay threshold brain maps were created for post-treatment images to highlight common and disparate regions of changes for each arm/outcome.   Results: 31 patients underwent SRS and 27 underwent ATL. Drop-out and suitable neuroimaging left 22 (71%) SRS and 16 (59%) ATL patients for study. Of remaining SRS patients, 11 (50%) were seizure-free and 11 (50%) not-seizure-free. Except for the perimesial temporal white matter, median FVC of seizure-free SRS subjects were significantly larger (16±11%) than not-seizure-free subjects, indicating that seizure-free SRS subjects had overall greater postoperative atrophy than not-seizure-free subjects. The regions with worse relative atrophy in SRS seizure-free patients were the entorhinal cortex, hippocampus, and temporal pole. Of ATL patients, 18 (81%) were seizure-free and 3 (19%) were not. FVC did not differ between seizure-free and not-seizure-free ATL patients (seizure-free FVC tending to be larger by 12±7%, indicating a tendency for smaller residual tissue left behind in not-seizure-free patients). Resective volume changes ranged widely despite a specific surgical protocol of resection. For example, excised volume changes of hippocampi ranged from nearly complete (~1.0) to 0.4, indicating that hippocampal changes could have a greater than two-fold difference, but patients remained seizure-free. Threshold maps showed that for either arm, the posterior extent of post-treatment lesions differed the most. After SRS, seizure-free patients had more posterior hippocampal atrophy. After ATL, non-seizure-free patients had more of the posterior subtemporal region resected. Conclusions: In patients with MTLE, MRI volumetry demonstrated that seizure remission after SRS follows the severity of postradiation encephalomalacia. Seizure remission after open surgery can occur across a wide range of resective volumes. Future study will evaluate the effect of postsurgical changes on neuropsychological data to help calculate the risks and benefits of minimally invasive surgery and extent of surgical resection after open surgery. Funding: NIH-NINDS R01 NS 058634-01A2; Elekta AB, Stockholm, Sweden