Abstracts

MR-based high dimensional mapping (HDM) in patients with mesial temporal sclerosis (MTS) demonstrates regional hippocampal surface anatomy changes, which are likely influenced by normal right-left hippocampal asymmetries and the underlying histopathology of MTS. Using HDM, we compare MTS subjects with controls to investigate the influence of hemispheric asymmetries in the hippocampal surface anatomy changes of MTS, and compare surface anatomy changes with the histopathology of MTS.
The epilepsy surgery series at Saint Louis University was consecutively reviewed for subjects with histopathological confirmation of MTS, and MR-based hippocampal volume (HV) loss (with greater than 10% asymmetry) concordant with the side of MTS. A group of control subjects, with no history of CNS disease, also had MR studies. All MRs were performed on the same MR scanner. MRs from epilepsy subjects were grouped according to the side of MTS. The right and left MTS groups (each considered independently) were matched for both average age and intracranial size with a group of 10 normal control subjects. Hippocampal deformation segmentations and comparisons were performed as previously described (Wang et al. [italic]Neuroimage[/italic] 2001;14:531-45), calculating right and left sided differences in MTS and control groups, then comparing asymmetries of groups. This enables discrimination of effects related to normal right/left asymmetries from those of MTS. For consistency in measuring eigenvector functions predictive of three-dimensional surface structure changes, right to left asymmetry was measured from both RMTS and LMTS groups. Final deformation results were projected on a representative hippocampal surface for each group.
There were 15 subjects in both RMTS and LMTS groups, and 10 subjects in each of the matching control groups. In analysis of HV and hippocampal shape (using the first 9 eigenvectors), using repeated measures ANOVA, results showed p[lt]0.0001 for both LMTS vs. controls and RMTS vs. controls. Analysis for best subset of eigenvectors to discriminate between groups, using logistic regression, yielded eigenvectors 1, 2 for LMTS vs. control, and 1,3 for the RMTS vs. control, suggestive that there are subtle differences in shape changes in LMTS subjects as compared to RMTS subjects. Visual inspection of surface anatomy changes showed similar patterns of regional volume loss in both the LMTS and RMTS groups as compared to controls, with the most accentuated involvement of the region correlating with Sommer sector in the hippocampal body.
HDM shows similar patterns of involvement in hippocampal surface anatomy in both RMTS and LMTS. Accentuated involvement in surface anatomy correlates with the underlying pathology of MTS.[table1]