Authors :
Presenting Author: Caitlin Neher, MS – Department of Mechanical Engineering, University of Washington
Oleksandr Khegai, PhD – BioMedical Engineering and Imaging Institute, Icahn School of Medicine at Mount Sinai
Lara Marcuse, MD – Department of Neurology, Icahn School of Medicine at Mount Sinai
Madeline Fields, MD – Department of Neurology, Icahn School of Medicine at Mount Sinai
Mehmet Kurt, PhD – Department of Mechanical Engineering, University of Washington
Priti Balchandani, PhD – BioMedical Engineering and Imaging Institute, Icahn School of Medicine at Mount Sinai
Rationale:
When used to identify epileptic foci, clinical MRI at standard field strength fails in approximately 20-30% of drug-resistant patients, emphasizing the need for improved imaging or additional imaging contrasts to provide pre-surgical evaluation of epilepsy (Arch Neurol 1993; 50; 65-9). This study aims to evaluate the feasibility of stiffness mapping and perfusion MRI to localize epilepsy in non-lesional mesial temporal lobe epilepsy (mTLE) patients. Neurosurgeons report using qualitative tactile feedback to differentiate the boundaries of surgical resection, and previous studies report lesional epileptic tissue to be stiffer than normal (Sci Rep 2020; 10; 20978), likely reflecting underlying reactive gliosis. Our central hypothesis is that advanced MRI including stiffness mapping and perfusion will enhance localization and characterization of the epileptic zone.
Methods:
The multi-modal protocol for this study included T1-weighted contrast, magnetic resonance elastography (MRE) and arterial spin labeling (ASL). Seven (2M/5F, age 31-54) unilateral mTLE patients were scanned on a 7 Tesla MRI scanner (Siemens Healthineers) at the Icahn School of Medicine at Mount Sinai. The MRE acquisition used a custom multi-slice 2D echo-planar imaging pulse sequence with 3D motion-encoding gradients, and a vibration frequency of 50 Hz at 1.1 mm3 isotropic resolution (Phys Med Biol 2024; 69(20); 1361-6560). Two subjects’ MRE data were excluded due to low octahedral shear-strain signal-to-noise ratio. The MRE phase data were unwrapped, denoised, and maps of stiffness were reconstructed by algebraic inversion of the Helmholtz equation. The 3.5 mm3 isotropic resolution ASL data (Interface Focus 2025; 15; 20240051) were acquired to generate quantitative perfusion maps using fMRI Software Library FSL (Fig. 1). T1 images (0.7 mm3) were segmented in FreeSurfer for regional analyses of the left and right temporal lobes (TL), mesial temporal lobes (MTL), and hippocampi (HC) in each mTLE patient. A one-tailed t-test was used to evaluate the hypothesis that ipsilateral regions were hypoperfused and stiffer than contralateral regions in each subject.
Results:
Both perfusion and stiffness showed statistically significant alignment with the study hypothesis and each patient’s diagnosis (Fig. 2). Mean TL stiffness was 19.2% higher across the group (n = 5, p = 0.004). Mean MTL and HC perfusion was reduced by 21.8 and 19.9% (n = 7, p = 0.003 and 0.006), respectively. Algebraic inversion of the MRE data limited analysis of finer structures such as the MTL and HC.
Conclusions:
This study presents the first multi-modal analysis of brain stiffness and perfusion in patients with unilateral mTLE to evaluate the techniques’ sensitivity at ultra-high field MRI. Future work will expand the technique to other common epilepsy types and investigate the perfusion-mechanics relationship in epileptic foci. The significance of these preliminary findings is a novel non-invasive imaging scheme which shows sensitivity to epilepsy lateralization and could serve as a complementary diagnostic method for unilateral mTLE patients.
Funding:
The study was supported by the AES Seed Grant. The work of Caitlin Neher was funded by NSF CMMI 1953323.