Prognostic application of lesion network mapping to epilepsy surgery outcomes in pediatric tuberous sclerosis complex

Objective: Patients with tuberous sclerosis complex (TSC)-related epilepsy often have drug-refractory epilepsy and numerous potential epileptogenic tubers. Current clinical methods target tubers for resection, but prediction of resulting seizure relief is difficult. This study describes implementation of lesion network mapping in TSC patients undergoing epilepsy surgery to associate resection zone with seizure outcomes. Methods: Thirty-nine consecutive patients with TSC who underwent invasive electroencephalography (EEG) and resection or ablation of the seizure onset zone were included. After preprocessing, each resection zone was delineated as the region of interest. Lesion network mapping was performed to determine the association between cortical networks connected to the resection zone and postoperative outcome using a multiple regression, iterative model that included demographic and other variables obtained from analysis of invasive EEG. Results: Of 39 patients, 20 (51%) had a good International League Against Epilepsy outcome (1-3). Resection regions connected to the default mode network and motor network were associated with better seizure outcome. Regions connected to the bilateral insula, visual associative regions, and putamen were associated with poor seizure outcome. Significance: This study provides methodology for lesion network mapping in TSC-related epilepsy. The results suggest a tendency for better outcomes when the resection zone is connected to certain ne