Related Experiment Video
Updated: Feb 3, 2026
Whole-brain Segmentation and Change-point Analysis of Anatomical Brain MRI—Application in Premanifest Huntington's Disease
Published on: June 9, 2018
Enlarged perivascular spaces in brain MRI: Automated quantification in four regions
Florian Dubost1, Pinar Yilmaz2, Hieab Adams2
1Biomedical Imaging Group Rotterdam, Department of Radiology, Department of Medical Informatics, Erasmus MC - University Medical Center Rotterdam, the Netherlands.
Abstract:
Enlarged perivascular spaces (PVS) are structural brain changes visible in MRI, are common in aging, and are considered a reflection of cerebral small vessel disease. As such, assessing the burden of PVS has promise as a brain imaging marker. Visual and manual scoring of PVS is a tedious and observer-dependent task. Automated methods would advance research into the etiology of PVS, could aid to assess what a "normal" burden is in aging, and could evaluate the potential of PVS as a biomarker of cerebral small vessel disease. In this work, we propose and evaluate an automated method to quantify PVS in the midbrain, hippocampi, basal ganglia and centrum semiovale. We also compare associations between (earlier established) determinants of PVS and visual PVS scores versus the automated PVS scores, to verify whether automated PVS scores could replace visual scoring of PVS in epidemiological and clinical studies. Our approach is a deep learning algorithm based on convolutional neural network regression, and is contingent on successful brain structure segmentation. In our work we used FreeSurfer segmentations. We trained and validated our method on T2-contrast MR images acquired from 2115 subjects participating in a population-based study. These scans were visually scored by an expert rater, who counted the number of PVS in each brain region. Agreement between visual and automated scores was found to be excellent for all four regions, with intraclass correlation coefficients (ICCs) between 0.75 and 0.88. These values were higher than the inter-observer agreement of visual scoring (ICCs between 0.62 and 0.80). Scan-rescan reproducibility was high (ICCs between 0.82 and 0.93). The association between 20 determinants of PVS, including aging, and the automated scores were similar to those between the same 20 determinants of PVS and visual scores. We conclude that this method may replace visual scoring and facilitate large epidemiological and clinical studies of PVS.
Insights
An automated deep learning method accurately quantifies enlarged perivascular spaces (PVS) in brain MRIs, outperforming manual scoring and offering a reliable tool for cerebral small vessel disease research.
Area of Science:
- Neuroimaging
- Radiology
- Medical Image Analysis
Background:
- Enlarged perivascular spaces (PVS) are MRI-visible brain changes associated with aging and cerebral small vessel disease.
- Assessing PVS burden is a promising neuroimaging biomarker for neurological conditions.
- Current visual PVS scoring is time-consuming and subjective, limiting large-scale studies.
Purpose of the Study:
- To develop and validate an automated deep learning method for quantifying PVS in key brain regions.
- To compare the performance of automated PVS scoring against traditional visual scoring.
- To assess the utility of automated PVS scores in epidemiological and clinical research.
Main Methods:
- A convolutional neural network regression model was employed for automated PVS quantification.
- The method utilized FreeSurfer segmentations on T2-contrast MRI scans from 2115 participants.
- Deep learning model was trained and validated using expert-rated visual PVS scores.
Main Results:
- Excellent agreement was observed between automated and visual PVS scores (ICCs 0.75–0.88), surpassing inter-observer reliability (ICCs 0.62–0.80).
- High scan-rescan reproducibility (ICCs 0.82–0.93) confirmed method stability.
- Associations between PVS determinants and automated scores mirrored those with visual scores.
Conclusions:
- The automated PVS quantification method demonstrates high accuracy and reliability.
- This deep learning approach can effectively replace manual PVS scoring in large-scale studies.
- The validated method facilitates research into PVS etiology and its role as a biomarker for cerebral small vessel disease.
Related Concept Videos
01:22Enlargement of the Plasma Membrane
01:20Anatomy of the Brain: Major Regions
The cerebrum is the largest section of the brain and divides into left and right hemispheres, separated by a deep fissure. The cerebral outer layer of grey matter — the cerebral cortex — comprises elevations called gyri and shallow groves called sulci. The inner portion of white matter includes long nerve fibers known as axons, which connect...
01:25Space Trusses
At the core of a space truss lies the fundamental unit known as the tetrahedron. This structure is composed of six members that form a three-dimensional shape...
01:27State Space Representation
Consider an RLC circuit, a...
01:03IR Frequency Region: Fingerprint Region
01:29Space Trusses: Problem Solving
Consider a tripod consisting of a tetrahedral space truss with a ball-and-socket joint at C. Suppose the height and lengths of the horizontal and vertical...