Search

ObjectiveMalformations of cortical development (MCD), including focal cortical dysplasia (FCD), are the most common cause of drug-resistant focal epilepsy in children. Histopathological lesion characterisation demonstrates abnormal cell types and lamination, alterations in myelin (typically co-localised with iron), and sometimes calcification. Quantitative susceptibility mapping (QSM) is an emerging MRI technique that measures tissue magnetic susceptibility (χ) reflecting it’s mineral composition.In a retrospective observational study, QSM was investigated abnormal tissue composition group of children with focal epilepsy with comparison to effective transverse relaxation rate (R2*) and Synchrotron radiation X-ray fluorescence (SRXRF) elemental maps. Our primary hypothesis was that reductions in χ would be found in FCD lesions, resulting from alterations in their iron and calcium content. We also evaluated deep grey matter nuclei for changes in χ with age.MethodsQSM and R2* maps were calculated for 40 paediatric patients with suspected FCD (18 histologically confirmed) and 17 age-matched controls.Patients sub-groups were defined based on concordant electro-clinical or histopathology data. Quantitative investigation of QSM and R2* were performed within lesions, using a surface-based approach with comparison to homologous regions, and globally within deep brain regions using a voxel-based approach with regional values modelled with age and epilepsy as covariates.Synchrotron radiation X-ray fluorescence (SRXRF) was performed on brain tissue resected from 4 patients to map changes in iron, calcium and zinc and relate them to MRI parameters.ResultsCompared to fluid-attenuated inversion recovery (FLAIR) or T1Lweighted imaging, QSM improved lesion conspicuity in 5% of patients.In patients with well-localised and confirmed FCDIIb lesions, quantitative profiling demonstrated decreased χ, but not R2*, across cortical depth with respect to the homologous regions. Contra-lateral homologous regions additionally exhibited increased χ at 2-3mm cortical depth that was absent in lesions. The iron decrease measured by the SRXRF in FCDIIb lesions was in agreement with myelin reduction observed by Luxol Fast Blue histochemical staining.SRXRF analysis in two FCDIIb tissue samples showed increased zinc and calcium, and decreased iron in the brain region exhibiting low χ and high R2*. QSM revealed expected age-related changes in the striatum nuclei, substantia nigra, sub-thalamic and red nucleus, but these changes were not altered in epilepsy.ConclusionQSM non-invasively revealed cortical/sub-cortical tissue alterations in MCD lesions and in particular that χ changes in FCDIIb lesions were consistent with reduced iron, co-localised with low myelin and increased calcium and zinc content. Theses findings suggests that the measurements of cortical χ measurements could be used to detect and delineate epilepsy lesions.

ObjectiveA retrospective, cross-sectional study to evaluate the feasibility and potential benefits of incorporating deep-learning on structural MRI into planning stereoelectroencephalography (sEEG) implantation in paediatric patients with diagnostically complex drug-resistant epilepsy. This study aims to assess the degree of co-localisation between automated lesion detection and the seizure onset zone (SOZ) as assessed by sEEG.MethodsA neural network classifier was applied to cortical features from MRI data from three cohorts. 1) The network was trained and cross-validated using 34 patients with visible focal cortical dysplasias (FCDs). 2) Specificity was assessed in 20 paediatric healthy controls. 3) Feasibility for incorporation into sEEG implantation plans was evaluated in 38 sEEG patients. Coordinates of sEEG contacts were coregistered with classifier-predicted lesions. sEEG contacts in seizure onset and irritative tissue were identified by clinical neurophysiologists. A distance of ResultsIn patients with radiologically-defined lesions, classifier sensitivity was 74% (25/34 lesions detected). No clusters were detected in the controls (specificity 100%). Of 34 sEEG patients, 21 patients had a focal cortical SOZ. Of these there was co-localisation between classifier output and SOZ contacts in 62%. The algorithm detected 7/8 histopathologically-confirmed FCDs (86%).ConclusionsThere was a high degree of co-localisation between automated lesion detection and sEEG. We have created a framework for incorporation of deep-learning based MRI lesion detection into sEEG implantation planning. Our findings demonstrate that automated MRI analysis could be used to plan optimal electrode trajectories.