Article Figures & Data
Figures
- Fig 1.
Illustration of the EPVS segmentation process. A, A coronal MR brain image of a patient with aMCI. The original image fitted with a brain mask (B) and the original image fitted with a mask generated with first-order k-means clustering to remove the CSF (C). D, Contrast-enhanced image of the brain fitted with the mask generated with second-order k-means clustering to exclude the cortical gray matter. E, A mask for pixels with high and positive spatial gradient values obtained from a pixel-wise spatial gradient map. The brain stem was manually removed. F, EPVS selected on the basis of object size after removing the edges of the brain structures. G, Color-coded EPVS overlaid on the original brain image.
- Fig 2.
A side-by-side comparison of EPVS in a cognitively healthy control versus a patient with aMCI A, A coronal MR brain image of a cognitively healthy control. B, Segmented EPVS color-overlaid on the MR brain image of a cognitively healthy control. C, A coronal MR brain image of a patient with aMCI. D, Color-coded EPVS overlaid on the MR brain image of a patient with aMCI.
- Fig 3.
Representative multislice MIP projections of EPVS in the subcortical brain structures and the basal ganglia of a control and a subject with aMCI. A, MIP image of subcortical structures within the middle of the brain of a control subject. B, The MIP image of subcortical structures within the middle of the brain of a subject with aMCI. C, MIP image of the basal ganglia of a control subject. D, MIP image of the basal ganglia of a subject with aMCI.
- Fig 4.
Summarized EPVS densities within the middle part of the subcortical brain (SB) and the basal ganglia (BG) of healthy controls and subjects with aMCI. The asterisk indicates P < .001.
- Fig 5.
The receiver operating characteristic curve for using EPVS as a differentiating biomarker between aMCI and cognitively healthy individuals. The area under the curve is 0.96, the 95% confidence interval is 0.89–1.00, and the standard error is 0.04. The cutoff point that gives the best sensitivity (92.86%) and the specificity (93.33%) is 3.35 v/v%.
Jump to section
Related Articles
Cited By...
- A Systematic Review and Meta-Analysis of Automated Methods for Quantifying Enlarged Perivascular Spaces in the Brain
- Lesion Volume in Relapsing Multiple Sclerosis is Associated with Perivascular Space Enlargement at the Level of the Basal Ganglia
- Increasing age is independently associated with higher free water in non-active MS brain - A multi-compartment analysis using FAST-T2