Authors :
Presenting Author: Essi Tsatsou, BS – University of Kansas
Xavier Scherschligt, BS – University of Kansas
Christopher Park, BS – University of Kansas
Namie Fotion, BS – University of Kansas
Hannah Judd, BS – University of Kansas
McKenzie Rynard, MS – University of Kansas
Patrick Landazuri, MD – University of Kansas
Caleb Pearson, PhD – University of Kansas
Carol Ulloa, MD – University of Kansas
Adam Rouse, MD, PhD – University of Kansas
Rationale:
Drug resistant epilepsy (DRE), characterised by persistent seizures despite trials of at least two appropriate antiseizure medications, affects roughly one-third of epilepsy patients. Identifying the epileptogenic zones, regions where seizures originate, is crucial for potential surgical intervention. This study evaluates the efficacy of Diffusion Tensor Imaging (DTI) in presurgical assessment and localization of epileptogenic zones in such cases.Methods:
We assessed the effectiveness of DTI for lateralizing seizure onset zone in DRE patients using fractional anisotropy (FA) values from DSI Studio. The research was conducted in two phases: the first in summer 2023 focused on mesial temporal lobe epilepsy, and the second in summer 2024 expanded to various types of focal epilepsy, University of Kansas Comprehensive Epilepsy Center. We analyzed patients’ DTI imaging from individuals referred to a comprehensive epilepsy center who had subsequently undergone surgical evaluation. For analysis, the laterality of seizure onset was retrospectively identified following surgical intervention.
Results:
Summer 2023 Findings: For mesial temporal lobe epilepsy patients, FA values were lower on the side of epilepsy in 76.8% of cases. Localization accuracy improved to 86.7% when FA differences exceeded 2.5 standard deviations from the mean. Summer 2024 Findings: In the replication phase, involving a broader range of epilepsy types, FA values were lower on the epileptogenic side in 66.2% of cases. To further refine our findings, linear mixed-effects modeling was employed to identify patients that we could most confidently lateralize the seizure onset zone. This advanced analysis identified a small subset of 8 out of the 75 patients with localization accuracy improved to 87.5%. Suggesting that while the overall localization accuracy varied, it remains a potentially valuable tool under specific conditions.
Conclusions:
DTI shows potential for presurgical assessment and localization of epileptogenic zones in DRE. The initial findings from summer 2023 demonstrated strong predictive value with reduced FA correlating with seizure foci. However, the 2024 replication phase, with a more diverse cohort, showed FA was lower on the epileptogenic side in 66.2% of cases. Advanced analysis highlighted that while overall accuracy varies, DTI remains a valuable tool under certain conditions. Future research will focus on optimizing tract identification, enhancing clinical data presentation, and developing advanced predictive models for better surgical planning and outcomes.
Funding:
This study was funded by the AES Summer Internship Program, whose support was crucial for the research’s execution and progression.