Abstracts

RATIONALE:To investigate the histopathologic correlates of Hippocampal Volumetry and 1H Magnetic resonance spectroscopic imaging (MRSI)in temporal lobe epilepsy METHODS: We prospectively acquired hippocampal volumes and metabolic ratios in patients undergoing temporal lobe resections for intractable epilepsy. Hippocampal volumes were acquired using 3D-brain T1-Weighted 1.5 mm images in the coronal plane. The hippocampal volumes in mm3 were normalized using previous methods. MRSI studies were obtained at high field (4.1T) using a PRESS sequence with short TE in the hippocampal plane. The Cr/NA ratio was calculated from the the mean of the voxels spanning the hippocampus. Representatives samples from cornu ammonis (CA) and Facia dentata (FD)were obtained and the neuronal to glial ratio were calculated for each section by quantitative cell counts. Control data from 14 age-matched autopsy specimens were used for comparison. RESULTS:Thirty-six patients had HV and pathology. Three had poor MRSI data. Complete correlations were obtained in 33 patients. Cell counts revealed a significant difference between autopsy controls and patients in CA1-FD regions. HV correlated with CA ratios (P<0.01) but not with FD ratios (P<0.09). No correlation was found between MRSI and Volumetry (0.7)and a negative correlation was found between neuronal/glial ratios and the Cr/NA ratio (-0.42 Pearson). CONCLUSIONS: Structural changes measured with MRI correlate with hippocampal cell loss as previously reported. Similar to temporal lobe hypometabolism measured with FDG-PET however,no correlation was found between hippocampal MRSI ratios and neuronal/glial cell ratios. These findings support the concept that in vivo measurements of reduced hippocampal NAA reflect functional changes underscoring neuronal dysfunction rather than neuronal cell loss. Future studies using quantitative NAA concentrations may yield further information about the relationship between cell dysfunction and NAA.