Abstracts

Rationale: Magnetic resonance imaging (MRI) has become a powerful tool in the diagnosis and treatment of epilepsy. Despite technical advances, routine clinical visual inspections may lack sensitivity to establish a diagnosis with a sufficient degree of confidence. To overcome such limitations, previous quantitative MRI studies have established important imaging markers of epilepsy-related pathology, including features sensitive to abnormal cortical morphology, cell loss, myelin damage, and reactive astrogliosis.

Here, we present an analytical pipeline for high-resolution multimodal MRI data of pediatric epilepsy patients. Our quantitative approach provides individualized multivariate asymmetry maps to identify subtle cortical pathology that may be initially overlooked by conventional neuroradiological examination.

Methods: Participants. Fourteen pediatric patients with drug-resistant epilepsy (five males, mean age±SD=13.21±3.58 years, range=8–18 years) were recruited from the British Columbia Children’s Hospital in Vancouver. Nine age- and sex-matched healthy children (six males, mean age±SD=11.67±3.39 years, range=6–17 years) also underwent identical imaging.

MRI acquisition. Participants underwent multimodal research 3T MRI scans with prospective motion correction, including a (i) 3D T1-weighted anatomical scan, (ii) 3D T2-FLAIR, (iii) 3D quantitative T1 mapping, and (iv) 2D diffusion MRI.

Multimodal asymmetry. Patient-specific asymmetry maps comparing left vs. right hemispheres [(L–R)/((L+R)/2)] were derived from multivariate cortical thickness, mean diffusivity, as well as quantitative T1 and T2-FLAIR intensities, and were z-scored relative to controls. Findings were mapped to brain surface models using the ENIGMA Toolbox.

Results: As proofs of concept, we present the cases of two children with drug-resistant epilepsy. Visual inspection of the multimodal MRI scans showed a right frontal FCD (case #1; Fig. 1a) as well as a right parahippocampal ganglioglioma (case #2; Fig. 2a). Quantitative analyses of these multimodal imaging data localized profound changes (i.e., increased atrophy, mean diffusivity, T2-FLAIR signal, qT1 intensity) in areas of high lesion conspicuity (Fig. 2b, Fig. 2c). Seizure onset zones identified by our quantitative analytical pipeline were furthermore cross-referenced with, and confirmed by, postsurgical histological findings. Overall, multimodal asymmetry maps alone could identify a putative surgical target in approximately 71% (10/14) of cases.

Conclusions: While these analyses were confirmatory in the current case studies, the use of multimodal postprocessing methods has shown promise in identifying subtle pathology that may be initially overlooked by conventional neuroradiological examination. Close coordination among epileptologists, radiologists, and epilepsy research scientists can facilitate the integration of advanced neuroimaging techniques into the preoperative clinical workup and ultimately improve personalized diagnosis and prognosis of drug-resistant patients with epilepsy.

Funding: Please list any funding that was received in support of this abstract.: CIHR, FRQS.