Abstracts

Rationale: Infantile Epileptic Spasm Syndrome (IESS) is a severe multifocal epilepsy affecting infants aged 3-12 months, with a prevalence of 4 in 10,000. Approximately 35% of IESS cases are of unknown etiology. This study aims to identify neuroimaging biomarkers using diffusion MRI (dMRI) predicting treatment response for IESS patients (those who achieve spasm freedom vs. patients who continue to have spasms).

Methods: This retrospective study analyzed dMRI parameters of patients diagnosed with IESS, who had non-lesional MRI brain and were followed for at least six months at Children’s Wisconsin (CW), Milwaukee, WI, USA, from January 2017 to December 2023. The study focused on clinical features of IESS, including response to medication used to control spasms, and collected imaging data. MRIs were obtained using Siemens or GE 3T scanners, employing more than 36 directions for obtaining diffusion imaging. dMRI data was processed using an MPCA-based denoising algorithm and corrected for head movement, eddy currents, and EPI distortions using FSL tools. Bias field inhomogeneities were corrected using B0 images. Mean diffusivity (MD) and fractional anisotropy (FA) maps were generated using Mrtrix3, while T1-weighted images, segmented with infant-freesurfer recon-all, were registered to skull-stripped B0 images and aligned to the MNI152 space using age-specific templates from the UNC-BCP 4D atlas. Regions of interest (ROIs) in the brain and corpus callosum were identified, binarized, and utilized to extract FA values from the FA maps.


Results: This study involved 26 patients (13 males) with IESS with non-lesional MRI. The average age at spasm onset was 6.23 months (SD 2.31), with treatment starting approximately 1.53 months post-diagnosis (SD 2.05). Genetic abnormalities were identified in 42.3% (11/26) of patients, and developmental delays was noted in 53.8% (14/26). Initial treatment for spasms included ACTH (11/26), oral steroids (10/26), and vigabatrin (5/26), which resulted in 42.3% (11/26) of patients with resolution of spasms. Among those who relapsed, 53.3% (8/15) responded to subsequent treatments (Table 1).

Differences in FA and MD maps were analyzed to explore the association between responders vs. non-responders to initial and second IESS specific medications. White matter integrity in the corpus callosum, thalamus, and overall brain white matter were adjusted for age-related changes in myelination (using Pearson partial correlation). The results, detailed in Table 2, showed no statistically significant differences in the mean FA and MD values between responders and non-responders to the first and second spasm medications.


Conclusions: This study is one of the few that have assessed dMRI parameters to predict treatment outcomes in IESS with non-lesional MRI. Results indicated no significant disparities between treatment responders and non-responders in FA and MD parameters. Further evaluation is needed using a multicenter study approach and incorporating tractography to identify neuroimaging markers for predicting treatment responses in IESS.


Funding: This study was supported by Wisconsin Children's Neuroscience Executive Team Grant.