Abstracts

Rationale: Approximately 30-40% of people with epilepsy will continue to experience seizures following anti-epileptic drug (AED) treatment. Current neuroimaging methods are not capable of differentiating patients who will, or will not, achieve seizure freedom with AED treatment. Recently developed diffusion methods offer the ability to probe more deeply into tissue characteristics, notably fiber ball imaging (FBI; Jensen, J. H., et al. Neuroimage, 2016;124:824-33). FBI uses highly diffusion weighted MRI (dMRI) to model the fiber orientation distribution function (fODF) of white matter (WM) fiber bundles. FBI allows the estimation of microstructural properties associated with axonal fiber bundles including the fractional anisotropy of the intra-axonal space (FAA). Fiber ball white matter (FBWM; McKinnon, E. T., et al. Neuroimage, 2018;176:11-21) modelling uses both high and low dMRI to estimate additional microstructural parameters including the intrinsic intra-axonal diffusivity (Da), the mean extra-axonal diffusivity (De-mean), and the axonal water fraction (AWF). These measures provide insights into the microstructure of WM. Therefore, we hypothesised that FBI could be sensitive to different epilepsy phenotypes regarding medication response. If this hypothesis is confirmed, this preliminary study would provide the bases for a promising line of investigation for prognostic markers in epilepsy. Methods: Fourteen patients with medically refractory focal epilepsy, seven patients with well-controlled focal epilepsy, and five healthy controls have been recruited into this study to date. All participants underwent diffusion kurtosis imaging (DKI; b values of 1000 and 2000 s/mm2, 30 directions) and FBI (b value of 5000 s/mm2, 128 directions) on a 3T Siemens Prisma MRI scanner. DKI maps of mean diffusivity (MD), fractional anisotropy (FA), and mean kurtosis (MK) were generated using Diffusional Kurtosis Estimator (https://www.nitric.org/projects.dke/). FBI and FBWM (https://github.com/m-ama/FBWM) specific maps were generated to include Da, De-mean, AWF, FAA, radial De (De-rad), axial De (De-ax), and zeta, and assessment of global FBWM was performed as previously described in McKinnon et al. Results: Using a two-way ANOVA with Bonferroni correction, patients who had refractory epilepsy had significantly increased global De-ax relative to controls (p=0.01). A concomitant difference between patients with well-controlled seizures and healthy controls was not observed. There were no other significant differences in mean WM values for AWF, Da, De, De-rad, FA, FAA, MK, MD, or zeta between any of the groups. Conclusions: In this preliminary study, only patients with medically refractory epilepsy had higher global WM values for axial extra-axonal diffusivity compared to the control group. This measure of the WM microstructural environment is not amenable to investigation using conventional DTI and therefore may provide new insights into the nature of WM involvement in focal epilepsy, however, as this study is not longitudinal, it remains unclear if changes seen are attributable to the cause or consequence of epilepsy. Nonetheless, this can resolved by subsequent longitudinal studies and it is particularly promising that FBI may present a sensitive marker of outcome. In further work, we will seek to determine whether FBI can provide useful biomarkers of pharmacoresistance as recruitment into our study grows. Funding: PhD project funded by the Medical Research Council