Abstracts

Rationale: High resolution Diffusion Tensor Imaging (DTI) of the hippocampus has demonstrated regional abnormalities of elevated mean diffusivity (MD) in temporal lobe epilepsy (TLE) with hippocampal sclerosis (HS)¹. Using high b value data, Diffusion Kurtosis Imaging (DKI) can provide a more comprehensive characterization of diffusion properties in tissue that may better reflect the complex underlying microstructure². DKI has shown white matter alterations in TLE patients, namely reduced mean kurtosis (MK) in the ipsilateral side of TLE patients with several regions not identified on DTI measurements³. However, DKI has not yet been applied for detecting subtle alterations in the hippocampus of TLE patients, which is the aim of this work.


Methods: This study enrolled 122 healthy controls (26-54 years) and 9 TLE patients (7 unilateral HS, 1 bilateral and 1 non-lesional; 26-53 years). Multi-shell diffusion-weighted images aligned to the hippocampus length were acquired on a 3T Siemens Prisma with 1 mm isotropic resolution over 20 slices for 20 b=0, 64 b=500, and 64 b=2000 s/mm² in 7:50 min. MD, MK, fractional anisotropy - FA, kurtosis FA – KFA were averaged over each hippocampus. The absolute volume of diffusion abnormalities was calculated using thresholds based on the control distribution, below 10^th percentile for MK and above 90^th percentile for FA, MD and KFA). These parameters were compared between groups using unpaired T-test or between measurements using paired T-test (p < 0.05).


Results: Different change patterns were observed for DTI and DKI metrics in non-HS and HS hippocampi. In Figure 1, MK and KFA from non-HS showed striking differences compared to controls that were not apparent with FA, and only minimally abnormal on MD. Furthermore, KFA of non-HS was greater than the HS group. In Figure 2, low MK regions correspond to widespread MD elevations in the HS side of TLE patients. In the non-HS hippocampi, MD is slightly elevated compared to controls, whereas MK is drastically lower. FA did not show visual differences between controls and TLE patients, while KFA is elevated for both hippocampi of TLE patients. Higher volume of diffusion abnormalities was found for Kurtosis metrics in the non-HS side.


Conclusions: DKI provided complementary information to DTI. Significant findings for KFA, but not FA, along with greater volume changes for KFA and MK, suggest DKI is more sensitive to hippocampal abnormalities. Notably, DKI findings were most dramatic in non-HS hippocampi, indicating that subjectively normal appearing hippocampi also have substantial structural abnormalities.

References
1.Treit S, et al. (2019) Regional hippocampal diffusion abnormalities associated with subfield-specific
pathology in temporal lobe epilepsy. Epilepsia Open 4:544
2.Jensen J, et al. (2005) Diffusional kurtosis imaging: the quantification of non-gaussian water
diffusion by means of magnetic resonance imaging. Magn Reson Med 53:1432
3.Glenn GR, et al. (2016) Epilepsy-related cytoarchitectonic abnormalities along white matter
pathways. J Neurol Neurosurg Psychiatry 87:930


Funding: CIHR
Conahcyt