Related Experiment Video
Updated: Mar 10, 2026
09:06
Whole-brain Segmentation and Change-point Analysis of Anatomical Brain MRI—Application in Premanifest Huntington's Disease
Published on: June 9, 2018
12.7K
Reproducible segmentation of white matter hyperintensities using a new statistical definition
Soheil Damangir1, Eric Westman2, Andrew Simmons2,3
1Department of Neurobiology, Care Sciences and Society, Karolinska Institutet, Hälsovägen 7, Huddinge, 14157, Stockholm, Sweden. soheil.damangir@ki.se.
Magma (New York, N.Y.)
|December 13, 2016
Summary
A new statistical method accurately identifies white matter hyperintensities (WMH) using various MRI sequences. This approach offers high reproducibility, surpassing manual segmentation for dementia research.
Area of Science:
- Neuroimaging
- Biomedical image analysis
- Neurology
Background:
- White matter hyperintensities (WMH) are crucial biomarkers in neurodegenerative diseases.
- Accurate segmentation of WMH is essential for clinical diagnosis and research.
- Current methods often rely on manual delineation or specific MRI sequences.
Purpose of the Study:
- To introduce a novel statistical definition for segmenting white matter hyperintensities (WMH).
- To develop a method adaptable to any combination of conventional magnetic resonance (MR) sequences.
- To evaluate the accuracy and reproducibility of the proposed WMH segmentation technique.
Main Methods:
- Utilized T1-weighted, T2-weighted, FLAIR, and PD MRI sequences from 119 subjects.
- Applied a statistical definition based on the one-tailed Kolmogorov-Smirnov test for WMH segmentation.
- Compared the proposed method against manual segmentations and existing automated techniques.
Main Results:
- Achieved high similarity (Dice 0.85-0.91) between proposed method and manual WMH segmentations.
- Demonstrated high similarity (Dice 0.83-0.94) across different combinations of input MRI sequences.
- Outperformed other available methods in accuracy and reproducibility on the test dataset.
Conclusions:
- The proposed statistical definition provides accurate WMH segmentation.
- The method exhibits superior reproducibility compared to manual delineation.
- This approach serves as a viable alternative to existing manual and automatic WMH segmentation techniques.
