Computational protocol: Perfusion and Diffusion Abnormalities of Multiple Sclerosis Lesions and Relevance of Classified Lesions to Disease Status

Similar protocols

Protocol publication

[…] To accurately perform quantification in the specific lesions (patients) and WM region (controls), the perfusion and diffusion image distortion induced by eddy currents has to be corrected. Despite the general promising results, mutual information-based matching still has its own limitations [] which could be the reason for failure to rectify the distortion of our data. We therefore used an image registration software known as bUnwarpJ [], an ImageJ plug in which was designed to align pairs of images distorted by both physical and acquisition related distortions, for the task. This approach combines the ideas of elastic and consistent image registration and minimizes the similarity error between the target and source image after imposing a consistency constraint. As a bidirectional registration, the algorithm calculates deformation fields for both direct (from source image to target image) and inverse (from target image to source image) transformation simultaneously.In practice, perfusion-weighted (pre-contrast injection) or diffusion-weighted (b=0) images were used as the source images, while T2-FLAIR or 3D images were the targeted images. The obtained deformation fields were applied for direct transformation of CBV or FA, respectively. The representative registration results are shown in . Each section of original CBV or FA for all patients was aligned onto its corresponding target T2-FLAIR image on which lesion ROIs were created. Similarly, all sections of the original CBV or FA for all controls were registered onto their corresponding target 3D images, where WM region was delineated after segmentation.Since the values of CBV and FA may change after registration procedure, the parameter evaluation was not performed on the registered image. Instead, the lesion and WM ROIs based on target image (T2-FLAIR or 3D image) were inversely transformed to source image (original CBV or FA) where the estimation was conducted for all sections studied.A total of 1135 lesion spots were evaluated for MS patients. Among them, 50, 468 and 617 were identified as active, T1-hypointense and T1-isointense lesion, respectively. The WM regions on all supratentorial sections were assessed for healthy controls. […]

Pipeline specifications

Software tools bUnwarpJ, ImageJ
Application Microscopic phenotype analysis
Organisms Homo sapiens
Diseases Multiple Sclerosis, Mitochondrial Diseases