Common Optimal Network Between Laser Interstitial Thermal Therapy and Thalamic DBS for Temporal Lobe

1736 

Abstract Submission 

Min Jae Kim¹, Frederic Schaper², Rob Rouhl³, Yasin Temel³, Michael Fox², Kathryn Davis¹, H. Isaac Chen¹, Ramya Raghupathi¹, Joon-Yi Kang⁴, William Anderson⁴

¹Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA, ²Brigham and Women’s Hospital, Harvard Medical School, Boston, MA, ³Maastricht University Medical Center, Maastricht, Maastricht, ⁴Johns Hopkins School of Medicine, Baltimore, MD

Min Jae Kim  
Perelman School of Medicine, University of Pennsylvania
Philadelphia, PA

Frederic Schaper  
Brigham and Women’s Hospital, Harvard Medical School
Boston, MA

Rob Rouhl  
Maastricht University Medical Center
Maastricht, Maastricht

Yasin Temel  
Maastricht University Medical Center
Maastricht, Maastricht

Michael Fox  
Brigham and Women’s Hospital, Harvard Medical School
Boston, MA

Kathryn Davis, MD  
Perelman School of Medicine, University of Pennsylvania
Philadelphia, PA

H. Isaac Chen, MD  
Perelman School of Medicine, University of Pennsylvania
Philadelphia, PA

Ramya Raghupathi, MD  
Perelman School of Medicine, University of Pennsylvania
Philadelphia, PA

Joon-Yi Kang  
Johns Hopkins School of Medicine
Baltimore, MD

William Anderson  
Johns Hopkins School of Medicine
Baltimore, MD

Laser interstitial thermal therapy (LiTT) and thalamic deep brain stimulation (DBS) are both widely accepted surgical therapies for patients with temporal lobe epilepsy (TLE). LiTT is used to ablate the seizure onset zone in the temporal lobe, while DBS is used to modulate the anterior thalamus (ANT) and its connected network to reduce seizures. While both therapies have shown significant clinical promise for TLE, postoperative seizure outcomes are variable across patients, presumably due to a lack of understanding of their mechanisms of action. Identifying which brain circuits are involved in improved clinical outcome after both therapies may increase our understanding of their seizure suppressive mechanisms. In this study, we first explore the network-level effect of LiTT and ANT-DBS by studying functional connectivity patterns associated with improved seizure outcome, and second by identifying a common brain network that may predict seizure reduction.

27 TLE patients who underwent unilateral amygdalohippocampal LiTT were classified into two cohorts (14 patients with left and 13 patients with right TLE) based on one-year postoperative seizure outcome. Seizure outcome was measured as either seizure-free (ILAE Class I) or seizures-present (ILAE Class II – VI). Individual ablation volumes were segmented from the intraoperative MRI sequence. From another cohort of TLE patients who underwent bilateral ANT-DBS, their volumes of tissue activated (VTA) were generated based on active electrode contact. Segmented ablation volumes and VTA were then seeded into a normative functional connectome (Brain Genomics Superstruct Project, n=1000) to extract functional connectivity (FC) profiles associated with respective ablation volumes or VTA. For LiTT patients, functional connections associated with seizure freedom were identified by comparing FC patterns between the two cohorts, resulting a voxel-wise "T-map". For ANT-DBS patients, FC patterns were correlated with a percentage degree of reduction (%) relative to preoperative seizure frequency to create a voxel-wise "R-map".

For the LiTT dataset, 10 TLE patients achieved complete seizure freedom. Seizure free patients had right LiTT ablation volumes that were more functionally connected to bilateral calcarine area, cuneus, and thalamic anterior and medial pulvinar nucleus compared to non-seizure free patients. (Figure 1B). The spatial similarity between the "T-map" produced from right-side ablation and each patients' connectivity profiles was highly associated with seizure outcomes by fitting a logistical regression model (log-likelihood ratio = 10.25, p = 0.0014**) as shown in Figure 1C. In the "R-map" produced from ANT-DBS, DBS sites associated with increased seizure control were more functionally connected to right cuneus and thalamic medial dorsal nucleus and medial pulvinar nucleus (Figure 2B). Finally, "T-map" produced from right-side LiTT and "R-map" from ANT-DBS was positively correlated (r = 0.29, p < 0.0001), suggesting that the predictive functional maps from the two procedures for TLE are highly similar.

In this study, we identified a common brain network connected to laser ablation and thalamic DBS sites associated with improved seizure outcome. This common network included functional connections to the cuneus and medial pulvinar nucleus, which may potentially improve surgical targeting and DBS programming to control seizures in TLE. Specifically, with more evidence of medial pulvinar nucleus playing a larger role in seizure propagation in TLE [1,2], our study further proposes medial pulvinar nucleus as an alternative target for controlling seizures through its networks engaged from neuromodulatory or ablative procedures.

Deep Brain Stimulation ²

Neurodegenerative/ Late Life (eg. Parkinson’s, Alzheimer’s)

Connectivity (eg. functional, effective, structural)

fMRI Connectivity and Network Modeling ¹

Epilepsy

FUNCTIONAL MRI

Thalamus

Treatment