1248 Optimal Connectivity Map Predicts Seizure Freedom Rates After Laser Interstitial Thermal Therapy in Mesial Temporal Lobe Epilepsy

Abstract

INTRODUCTION: Laser Interstitial Thermal Therapy (LiTT) is increasingly becoming a first-line surgical therapy for mesial temporal lobe epilepsy (mTLE). While promising, its seizure outcomes have also been highly variable, presumably due to an incomplete understanding of mechanisms behind LiTT. Epilepsy is a network disorder, and therefore it is crucial to account for both the effect on local destruction of epileptogenic foci, as well as global intervention of limbic circuitry, to elucidate the circuit-level mechanisms of LiTT and devise improved ablation strategies. In this study, we explore LiTT's network-level effect by investigating functional connectivity patterns and identifying optimal connectivity patterns predictive of seizure freedom. METHODS: 27 mTLE patients who underwent unilateral LiTT were classified as seizure-free (ILAE Class I) or seizure-present (ILAE Class II–VI) cohort based on one-year postoperative outcomes. Individual ablation volumes were segmented from intraoperative MRI sequences. Ablation volumes were seeded into a normative functional connectome (Brain Genomics Superstruct Project) to extract patient-specific connectivity profiles associated with ablation volumes. FC patterns were compared between the two cohorts to produce voxel-wise “t-maps” to identify local and global differences in connectivity strengths predictive of seizure freedom. RESULTS: Ten mTLE patients achieved complete seizure freedom. In seizure-free patients, right ablated volumes shared higher FC with regions such as calcarine, cuneus, and thalamic anterior pulvinar nucleus than patients with seizure presence. CONCLUSIONS: For the right-sided ablations, seizure freedom was associated with a distinctive whole-brain connectivity pattern beyond the local surgical target. Further investigation is required to explore if this pattern is specific between the two surgical sides in mTLE population.