Alzheimer's is a progressive neurodegenerative disease and the most common cause of dementia. The most well-known characteristic features of Alzheimer's disease (AD) are the extracellular accumulation of amyloid beta, the intracellular deposition of neurofibrillary tangles, and the shrinkage of the brain (atrophy) which occurs due to the loss of synapses. One of the overlooked aspects of AD is the vascular alterations that commonly occur in association with the disease.
These alterations include the morphological changes of the micro vessels (becoming tortuous, kinking, looping, twisting), the quantitative changes in the micro vessels (e.g., the reduction of the micro vessel's density), atherosclerosis in the circle of Willis, arterial stiffness, etc. Figure 1 displays the vascular alterations in AD.
Fig. 1: Vascular alterations in AD
The quantitative changes of the micro vessels (arterioles and capillaries) have been widely investigated through post-mortem studies. The quantitative changes of the larger vessels, however, have been rarely investigated. With the high-resolution MRI, these vessels can be visualized invivo providing the opportunity to evaluate the quantitative vascular alterations associated with different diseases including AD. The vascular changes can be captured either by using quantification metrics including vessel density, vessel volume, vessel diameter, vessel length, etc or by using brain vessel Atlases developed by several studies. Each of these approaches has limitations. Quantification metrics compress vascular patterns into a single number and don't provide any information regarding the probable changes in the vascular patterns and the Atlases are not reliable since they register all the subjects into one common space. Developing brain vessel atlases is very challenging due to high inter-subject variability in the vascular pattern. Even though the big vessels pattern like internal carotid arteries, middle carotid arteries, etc are quite similar between subjects, the variability increases as the size of the vessels become smaller. To address these challenges, a new quantification metric called vessel distance mapping (VDM) is introduced. In this approach, the Euclidean distance transform is used and the distance of each voxel to its closest vessel voxel is calculated. This means that each voxel (even non-vessel voxels) includes information about the closest vessel voxel. With this approach, a cloud-shaped probability map of the vascular distribution is generated which includes both information about the vascular pattern and the distance values. This probability map is visualized in figure2. The left figure shows the segmented vessels. The left and right hippocampus regions are marked in blue. The right figure visualized VDM. It can be seen that as we get closer to the vessels the map gets darker (distance values get closer to zero) and as we get farther, the map becomes darker. These figures were generated using Fiji software.
(a) segmented vessels (b) VDM
Fig.2: (a) shows the segmented vessels in one slice of a TOF-MRA dataset and blue contours show left and right hippocampus. (b) Shows the VDM probability map for the same slice
To evaluate the functionality of this new method, I have applied this technique to the vascular pattern of the people suffering from AD in my thesis.
Comments