FV 4. U-Fiber Density Imaging identifies specific microstructural alterations in the epileptogenic zone in individual patients with cryptogenic focal epileps

Abstract

Patients with cryptogenic focal epilepsy (cFE), where conventional MRI does not show any structural lesion represent the fastest growing number of epilepsy surgery candidates in most tertiary referral centers. The absence of a visible surgical target challenges presurgical evaluation and requires complementary diagnostic methods to guide further investigations, especially intracranial EEG recordings. Here we present a new diagnostic approach, based on the quantification of regional u-fiber density to identify specific microstructural abnormalities in the epileptogenic zone of patients with cFE. 55 patients with cFE and 60 healthy controls were investigated. Diffusion Tensor Imaging (DTI) data was acquired on a GE Signa HDx 3T MRI Scanner, using an acquisition scheme with 64 diffusion weighted directions, a b-value of 1000 s/mm[b] and 2.4 mm slice thickness. Whole brain tractography was performed, creating an average of 1.05 million fiber tracts per subject. From this dataset, we specifically selected u-fibers, based on length, curvature and shape criteria. Density images were created, specifying the number of u-fibers within any 1 mm[c] voxel of the brain. These u-fiber density images (UFDI) were spatially normalized to a template in MNI-space and smoothed with an 8 mm kernel. For every single patient the UDFI was statistically compared against the control population, using SPM software. Clusters of significant reductions in UFDI were determined, thresholding at a minimum size of 5 cm[c] and a p-value <0.001, correcting for multiple comparisons using the family wise error (FWE) method. Localization of UFDI reductions was compared with clinical data from the presurgical evaluation. All patients had normal conventional MRI, using a specific epilepsy protocol on a 3T scanner. In 43 patients, a conclusive hypothesis on the epileptogenic zone could be derived, based on ictal and interictal EEG, seizure semiology, FDG-PET and ictal SPECT imaging and neuropsychological assessment. Only those 43 patients were included in the final analysis. In the remaining 12, clinical data was inconclusive, preventing further analysis of their UDFI alterations. Clusters of significant UFDI reductions were found in 91% of the patients and 78% of these reductions were consistent with the clinically suspected epileptogenic zone. Twelve patients underwent intracranial EEG recording with depth electrodes, confirming seizure onset in 74% of the detected UFDI reductions. Eight patients had resective surgery and the most frequent histopathological finding was a mild architectural disturbance with blurred gray-white matter border and ectopic neurons in the white matter. Quantitative UFDI has proven sensitive and specific to detect microstructural alterations in patients with cFE, superior to other complementary diagnostic methods. Histopathology has shown mild malformations of cortical development (MCD), a well-known cause of cFE. MCD typically affect the gray-white matter border, and this is exactly where u-fibers are located, spanning between neighboring gyri, directly underneath the cortex. We hypothesize the ectopic neurons disrupt the microstructural order of subcortical white matter, thereby artificially reducing the number of u-fibers we could reconstruct in this region. UFDI has become a routine diagnostic method in our department for all patients with cFE and is now routinely used to guide implantation of intracranial electrodes.