Abstracts

Rationale: Patients with temporal lobe epilepsy (TLE) most commonly present with unilateral hippocampal sclerosis (HS) on MRI. Neuropathology studies have demonstrated variable subregional involvement of the hippocampal head, body, and tail in individual subjects with HS. In addition, autopsy studies have demonstrated subtle neuronal loss contralateral to the seizure focus in subjects with TLE. In the present study, we therefore used in vivo MRI to characterize subregional atrophy patterns in subjects with TLE and unilateral HS.  Methods: Eleven patients (age: 22-58; nine males) with TLE and unilateral HS were compared with ten healthy controls (age: 21-41; five males). MR Images were acquired with a 4.7T MRI system using a T2-weighted fast-spin echo sequence (spatial resolution: 0.26 x 0.34 x 1 mm3). Hippocampi were manually segmented into head, body and tail subregions using anatomical landmarks. Intra-rater reliability was evaluated by measuring spatial overlap with Dice Similarity Coefficients. Control volumes were then used to construct a normal range (10-90th percentile) for each subregion. Volume measurements from patients with TLE (both ipsilateral and contralateral to known HS) which fell within the normal range were classified as normal whereas measurements below this range were defined as abnormal. Results: Reliability of the segmentation protocol in healthy controls and subjects with TLE is shown in Table 1. Volumetric analysis of the ipsilateral hippocampus revealed that while all subjects had atrophy of both the head and body, only seven subjects had atrophy of the hippocampal tail (Figure 1). Volumetry of the contralateral hippocampus revealed that six subjects had normal volumes of all subregions (head, body, and tail). However, the remaining five subjects were found to have atrophy of at least one subregion of the contralateral hippocampus.  Conclusions: Our results demonstrate variable involvement of hippocampal subregions between subjects with TLE, both ipsilateral and contralateral to known HS. Further studies with a larger sample size are required to investigate the neuropathological, clinical, and prognostic correlates of these findings.  Funding: This work was supported by an operating grant held by DWG from the Canadian Institutes of Health Research (funding reference number 81083) and by the E. (Ben) & Mary Hochhausen Fund for Epilepsy & Cerebral Palsy Research from the University Hospital Foundation.