Abstracts

Rationale: Previous work in patients with non-lesional focal epilepsy has shown focal asymmetries in the distribution of fiber tracts reconstructed from Diffusion Tensor Images (DTI) on visual inspection. Here we investigate the contribution of automatic quantitative analysis of fiber density maps to identify structural abnormalities in individual patients with cryptogenic focal epilepsies. Methods: Thirty-one patients with cryptogenic focal epilepsy were examined and compared against 18 healthy controls. All patients had presurgical video EEG monitoring and normal structural MRI at 3T using a dedicated epilepsy protocol. DTI data was acquired on a GE Signa HDx 3T scanner, using an acquisition scheme with 64 diffusion-weighted directions, b-value of 1000 s/m<+>2<+>, 2.4 mm slice thickness and 2 mm in-plane resolution. Images were pre-processed using FSL software, whole brain tractography was performed and fiber density maps were generated for every subject. Statistical analysis was performed with SPM software, comparing every single patient against the control population. Clusters of significantly reduced fiber density were identified using a threshold of p < 0.05 after family-wise error correction for multiple comparisons and cluster size threshold of > 3 cm³. Results: An average of 1,1 million fibers were reconstructed for every subject. Forty-two clusters of fiber density reduction were found in 22 patients (71%). Eleven patients showed a single cluster of reduction and four patients had extensive bilateral changes in fiber density. Ten patients had follow-up with intracranial electrodes and four had resective epilepsy surgery. Thirty-three of the 42 clusters (79%) were consistent with the clinical data and presumed seizure onset zone as defined by scalp and/or intracranial EEG. Two patients with bilateral changes had intracranial EEG recording and this confirmed bilateral seizure onset zones as indicated by DTI in both. Histopathology of the resected specimen showed focal cortical dysplasia with heterotopic neurons in two patients, chronic inflammatory changes in one and one patient showed no histological changes Conclusions: Our data strengthen the role of quantitative DTI analyses as complementary imaging modality in presurgical evaluation of patients with non-lesional focal epilepsy. The statistical analysis of fiber density maps is a user independent method and provides new localizing information in more than two thirds of patients and helps in guiding further invasive procedures. It also provided evidence for bilateral pathological changes in some patients, where surgical intervention might not be an option. The histopathological findings of focal cortical dysplasia and chronic inflammation in the first operated patients provide a plausible biological explanation for microstructural disruption of the white matter underlying the seizure onset zone, which leads to impaired fiber tracking and reduced fiber density.