MRI-based Morphometric Analysis Of Focal Cortical Dysplasia

Focal cortical dysplasia(FCD) is one kind of the typical developmental malformation of the cerebral cortex, and it is also a common causative factor of refractory epilepsy. Since FCD was first reported by Taylor in 1971, recent studies have revealed a higher prevalence of FCD than previously estimated, and more and more attention has been paid to FCD.Because FCDs may be very subtle in appearance and might escape conventional visual analysis, detection of epileptogenic lesions on MRI is still challenging. Many advanced image processing strategies have been investigated. Among them, voxel-based morphometric analysis is an important method, which has been applied to structural magnetic resonance images by many researchers to evaluate whole brain gray matter and white matter lesions objectively.In this thesis, we carried out VBM-DARTEL method and morphometric analysis program method(MAP) based on the voxel-based morphometric analysis. In terms of FCD patients group, we established a VBM-DARTEL based approach to detect gray matter volume abnormalities in FCD and reveal the main brain regions where FCD often occurs. For FCD type I patient individual diagnosis, MAP was performed on structural MR images to visualize the abnormal but subtle blurring of gray-white matter junctions to improve the detection of subtle cortical abnormalities and presurgical evaluation in the epileptogenic zone.We first established a VBM-DARTEL based approach to detect gray matter volume abnormalities in FCD. This method includes image segmentation, mask making with DARTEL, image standardization and so on. Two sample t tests were applied to compute the gray matter volume difference of FCD group and control group. Thus, the whole-brain abnormal changes in FCD were detected objectively and comprehensively. Compared to normal controls, we found that FCD patients have diffuse gray matter volume changes in bilateral cerebral hemisphere. We also found that focal gray matter decrease and increase both exist, particularly in frontal and temporal lobe areas.We also implemented MAP to T2 FLAIR images of type I FCD patients, and improved MAP of threshold calculation resulting in two volumetric statistical maps, namely the gray-white junction and extension images. Comparing preoperative images with junction image and extension image revealed that the two feature images could be more sensitive to show the lesions than original MRI. In addition, junction image is more sensitive than extension image. In short, MAP can detect I FCD with a high degree of sensitivity. This analysis proves to be useful for clinician evaluating and treating affected patients.