Abstracts

Rationale: The typical pattern of vigabatrin (VGB) associated MRI signal changes include reversible diffusion restriction in deep gray matter and brainstem. Here we report a case series demonstrating restricted diffusion in additional brain regions and some atypical patterns in infants exposed to VGB therapy. Methods: We identified all patients with video-EEG confirmed infantile spasms (IS) between January 2014-December 2018 that received VGB at our center. Imaging studies were systematically reviewed by an experienced neuroradiologist blinded to clinical data of all patients reported to have any signal abnormality. Alternate causes for such appearance, including hypoxic-ischemic injury or metabolic disorders, were ruled out. Abnormalities were specified by region and imaging sequence; all MRIs had DWI including apparent diffusion coefficient (ADC) maps in addition to the usual sequences. For each case, the duration of VGB exposure, timing of MRI in relation to VGB exposure, EEG, and clinical history, especially seizure history, were noted.  Results: We identified 17 patients treated with VGB for infantile spasms with associated MRI changes. Among these patients, the basal ganglia was affected in 16 patients, thalamus was involved in 10, brainstem in 14, and the cerebellar dentate nuclei in 12. In 4 of the 17 patients with typical VGB related changes, there was either restricted diffusion or T2 hyperintensity in the bilateral mesial temporal regions (n=3) or unilaterally in the right hippocampal region (n=1). The mesial temporal lobe signal changes were noted at the same time as the VGB related changes (n=3) (Fig. 1). In the 4th patient, the bilateral hippocampal signal change was not seen on the initial MRI (that showed VGB changes) but was seen on a subsequent MRI done 2 years later. The latter MRI on this patient continued to show VGB related changes. During the time when the hippocampal signal changes were noted, 2 patients were reported to be seizure free while the other 2 had either brief seizures or epileptic spasms.Three of the 4 patients had follow up MRIs (1-3.5 years after the initial MRI showing VGB related changes) which demonstrated resolution of changes including the typical VGB changes as well as the mesial temporal signal changes. At the time of the follow-up MRI, one of 3 patients discontinued VGB while the other 2 remained on VGB. The patient with unilateral mesial temporal change did not have a follow up MRI.Additionally, we noted one patient who received VGB briefly (4 weeks) had asymmetric (R>L) changes in the basal ganglia, brainstem, and right thalami while receiving VGB for 2 weeks (Fig. 2). Conclusions: Typical VGB-associated MRI changes may be associated with restricted diffusion with or without T2 hyperintensities in cortical areas, such as the mesial temporal regions. The temporally synchronous occurrence of these mesial temporal MRI signal changes along with the typical VGB related MRI changes (deep gray regions) and the disappearance of the mesial temporal signal abnormalities along with typical VGB changes support the notion that these cortical signal changes are VGB related. Further study into this novel finding is necessary to establish the association, and treating clinicians and radiologists should be aware of these atypical patterns to avoid unnecessary investigations, if the patient is otherwise asymptomatic and remains on VGB. Funding: No funding