Abstracts

RATIONALE: Focal cortical dysplasia (FCD) is the commonest neurodevelopmental lesion in refractory epilepsy. Feautures of FCD on optimal MRI include focal cortical thickening, grey-white interface blurring and changes in signal extending into the white matter. Surgical series show 40-50% of FCD patients have a poor postoperative outcome with continuing seizures despite complete lesion resection seen on postoperative MRI. This suggests the presence, undetected on preoperative MRI, of additional epileptogenic abnormalities in about 40-50% of FCD patients. We have previously published a method quantifying regional distribution of grey (GM) and white matter (WM) in patients with epilepsy. The presence of abnormalities correlated with statistically worse outcome following lobectomy for hippocampal sclerosis¹. Our previous method involved manual GM and WM segmentation and was time consuming and subjective. We now present a fully automated, operator independent protocol and apply it to an FCD cohort.
METHODS: 30 subjects with FCD (15 men) and 100 age-matched controls (44 men) had optimal high resolution MRI. FCD was diagnosed by visual assessment by two expert neuroradiologists. No abnormalities were seen in control subjects. Postprocessing was performed on the T1 volumetric data set, giving meaures of regional GM and WM distribution, cortical thickness, fractal dimension and volume/surface area relationships. The controls defined normal ranges (mean[plusminus]3sd) for each measure.
RESULTS: 10(33%) patients had quantitative abnormalities. In all 10 there were additional extralesional abnormalities, in visually normal areas, ranging from abnormalities in contiguous areas(4); ipsilateral distant areas(6) and in the contralateral hemisphere(8).
CONCLUSIONS: We present a quick automated method for quantitative MR analysis applicable easily and objectively to large numbers of subjects. These quantitative techniques have detected extralesional abnormalities not seen on visual inspection in a large group of patients with refractory epilepsy and FCD which could be subtle additional malformation. The proportion of subjects with additional nonvisualised quantitative abnormalities in our group is comparable to that with a poor outcome following focal FCD resection.
Other novel MRI techniques have tried to identify single areas of FCD. None have yet proven useful as a prognostic indicator in epilepsy surgery. Our technique has potential application as a presurgical screening tool to identify those who may have a poor surgical result. This would prevent patients unlikely to achieve seizure freedom from being needlessly exposed to the risks of surgery. We are now validating our findings in a group of surgical patients.
1.Sisodiya SM et al. Ann Neurol 1997;41:490-496.
Support: MRC(UK).