Abstracts

Magnetic Resonance (MR) techniques can assess brain tissue characteristics. Tissue signal can be quantified by measurement of the T2 relaxometry time (T2-R), and brain metabolites by MR Spectroscopy (MRS). Both show characteristic changes in epilepsy, but it is not known whether they measure the same pathology. Here we analyse the relationship between T2-R and MRS findings.
We assessed 29 patients, recruited from a tertiary epilepsy program, and 28 controls on a 3T GE LX Horizon scanner. For T2-R, ten coronal slices with eight images per location (CPMG, TE 28ms - 231ms, TR 5000ms) were acquired. T2 maps were generated using a GE software package (Functool[reg]). Three bilateral temporal regions of interest (ROI) were measured (hippocampus, anterior temporal lobe, amygdala). Values were expressed as percentage normal (100/(mean control T2-R) x (individual patient T2- R)), and an average value of the three ROI was calculated. For MRS, bilateral temporal lobe single voxel spectra, centred over the hippocampus, were acquired (1.5 cm3, TR, 3.0 s; TE 30 ms; data points, 2048; spectral width 5000Hz). Concentrations of the following metabolites were obtained using LCModel: N-acetylaspartate (NA), creatine (Cr), phosphocholine (Cho) and myoinositol (mI). The level of significance was set at 5%.
Patients showed bilateral increase of T2-R values in all ROI. The average signal increase in the three ipsilateral ROI was 6% [plusmn] 7 (p[lt]0.0001). The patients[rsquo] temporal lobe MRS values showed a reduction of the ipsilateral NA (patients 5.0[plusmn]1.2, controls 5.6[plusmn]0.8, p=0.03) and Cho (patients 1.4[plusmn]0.3, controls 1.5[plusmn]0.2, p=0.03), whereas the other metabolites or contralateral metabolite concentrations did not differ from controls.
The ipsilateral T2-R (average of three ROI) was correlated with the ipsilateral Cho (r=0.4, p=0.003), and Cr (r=0.3, p=0.01), but not with NA or mI. Contralateral T2-R times did not correlate with contralateral metabolite concentrations. Seizure frequency and age of onset were also not related to the ipsilateral T2 signal, and neither were metabolite concentrations.
In epilepsy, reduced neuron cell counts and gliosis have been described. Increased hippocampal T2 relaxation times have been attributed mainly to gliosis, whereas NA reduction to reduced neuron cell density. Both values were abnormal in our patient group, but were not correlated. This suggests that they indeed measure different aspects of the epileptic pathology. Interestingly, T2- R values were correlated with Cho and Cr, which may reflect increased cell turnover, including gliotic reaction, and may support that alteredT2-R time reflects changes in the glial cell population.
[Supported by: National Health and Medical Research Council, Neuroscience Victoria and Brain Imaging Research Foundation, Australia ]