Abstracts

Rationale: A significant number of children with intractable focal epilepsy (IFE) have normal MR imaging. With no obvious structural target, few proceed to surgery. However, there is evidence that pathology is present even when it is not seen on MRI. Our aim was to determine if voxel based morphometry (VBM) of structural MRI of children with IFE could detect both visible focal cortical lesions and in those with normal MRI, abnormality concordant with the seizure focus. Methods: Consecutive children with IFE and focal cortical dysplasia or normal MRI were recruited from the Epilepsy Clinics at Great Ormond Street Hospital, London. A control group of similar age was also recruited. MR imaging was carried out on a 1.5T Avanto scanner (Siemens, Erlangen, Germany). 3D FLASH (TR 11ms, TE 4.94ms, FOV 256mm, voxel size 1mm³) and 3D FLAIR (TR 6000ms, TE 353ms, TI 2200ms, FOV 256mm, voxel size 1mm³) datasets were acquired. Images were reviewed by an experienced Neuroradiologist, blind to clinical data. VBM was performed using SPM5 (Wellcome Institute of Cognitive Neuroscience, London) running in MATLAB 7.1 (The MathWorks, Natick, Massachusetts, USA). Two contrasts were used to detect whether voxels had higher or lower probabilities of being grey/white matter compared to controls.Results: Ten patients were enrolled with focal cortical dysplasia (age 7.9-17.3y, median 14.5y). Lesions were frontal (2), parietal (2), temporal (2), temporoparietal (1) and opercular (1). The lesion was visible on FLAIR in all, but visible on FLASH in only 4/10. VBM of FLASH datasets detected 3/10 dysplasias.On FLASH images dysplasias were detected as areas with increased probability that voxels were grey matter. VBM of FLAIR datasets detected 7/10 dysplasias. Of the undetected dysplasias, 2 were subtle on visual inspection and 1 was confined to the superior frontal gyrus. On FLAIR images dysplasias were detected as areas with reduced probability that voxels were white matter. 14 patients with IFE and normal MRI (age 7.8-16.8y, median 13y) were enrolled. The seizure focus was frontal (10), frontotemporal (3) and centrotemporal (1). VBM of FLASH datasets did not detect any cortical abnormality. VBM of FLAIR datasets detected abnormality in 4/14 children. In 2/4 the abnormality was singular and concordant with the seizure focus. In 2/4 multiple abnormalities were detected, one area being concordant with the seizure focus in one patient. Conclusions: Many of our children with IFE had dysplasias visible only on FLAIR. VBM of FLAIR was superior to FLASH in detecting visible dysplasias and importantly could detect abnormality in 4/14 children with IFE and normal MRI. In 3/4 the VBM abnormality was concordant with the seizure focus suggesting that FLAIR VBM could potentially identify a subgroup of children with IFE and normal conventional MRI for whom neurosurgery might be considered. This work, funded by SPARKS, was undertaken at GOSH/ICH(UCL) who receive a proportion of their funding from the Department of Health's NIHR Biomedical Research Centre funding scheme.