Abstracts

RATIONALE: Conventional MRI sequences exploit differences in the proton density, T1- and T2-relaxation times to provide tissue contrast and lesion conspicuity. Of these, T2-weighted imaging is the most sensitive in identifying cerebral pathology, which is often characterised by a change in either total tissue water or altered compartmentalisation of water within the tissue. Quantitative assessment of T2 signal intensity (T2-mapping) is more sensitive than qualitative evaluation in, for example, hippocampal sclerosis and multiple sclerosis. T2 mapping in the neocortex may be confounded by partial volume effects of cerebrospinal fluid. This is resolved by using a fast fluid attenuated inversion recovery (FLAIR) sequence.
Our objective was to assess whether T2 mapping, analysed using statistical parametric mapping (SPM) would identify areas of abnormal T2 signal in patients with localisation-related epilepsy.
METHODS: Thirty healthy volunteers and 73 patients with partial seizures were scanned with conventional MRI and T2 mapping. The patient group comprised: acquired lesions (n=11), malformations of cortical development (n=18), normal conventional MRI (n=44).
T2 mapping was performed using FLAIR imaging (TR/TI/TE=5000/1638/15,120ms) with a resolution of 0.94x0.94x5mm. T2 maps were calculated and normalised to Talairach space using SPM99. The patients[ssquote] maps were then compared to the 30 control subjects[ssquote] maps and significant differences were detected at a threshold of p[lt]0.001 (multiple comparisons: p[lt]0.05).
RESULTS: In all 11 patients with acquired lesions, and in 16 out of 18 patients with MCD, SPM detected areas of increased T2. In 5 out of 11 patients with acquired lesions and in 6 out of 18 patients with MCD, areas of abnormal T2 signal were also detected in regions previously reported as normal. Twenty out of 44 MRI-negative patients had areas of increased T2, of which 15 were concordant with the localisation of EEG abnormality.
CONCLUSIONS: T2-mapping was sensitive in patients with acquired lesions and malformations of cortical development. In these patients, abnormalities were also detected in normal appearing brain suggesting occult injury or more widespread MCD than apparent on visual inspection of conventional MRI. A clinically concordant abnormality of T2 signal was identified in 34% of MRI-negative patients. This could be due to either aetiological factors, such as occult dysgenesis, or as a result of chronic seizures, such as atrophy and neuronal loss. This technique, along with other new methods, increases the yield of identifying potentially curable aetiologies of epilepsy.
[Supported by: Action Research and The Gwyneth Forrester Trust]