The effect of hypointense white matter lesions on automated gray matter segmentation in multiple sclerosis

Rose Gelineau-Morel, Valentina Tomassini, Mark Jenkinson, Heidi Johansen-Berg, Paul M. Matthews, Jacqueline Palace

Research output: Contribution to journalArticle

Abstract

Previous imaging studies assessing the relationship between white matter (WM) damage and matter (GM) atrophy have raised the concern that Multiple Sclerosis (MS) WM lesions may affect measures of GM volume by inducing voxel misclassification during intensity-based tissue segmentation. Here, we quantified this misclassification error in simulated and real MS brains using a lesion-filling method. Using this method, we also corrected GM measures in patients before comparing them with controls in order to assess the impact of this lesion-induced misclassification error in clinical studies. We found that higher WM lesion volumes artificially reduced total GM volumes. In patients, this effect was about 72% of that predicted by simulation. Misclassified voxels were located at the GM/WM border and could be distant from lesions. Volume of individual deep gray matter (DGM) structures generally decreased with higher lesion volumes, consistent with results from total GM. While preserving differences in GM volumes between patients and controls, lesion-filling correction revealed more lateralised DGM shape changes in patients, which were not evident with the original images. Our results confirm that WM lesions can influence MRI measures of GM volume and shape in MS patients through their effect on intensity-based GM segmentation. The greater effect of lesions at increasing levels of damage supports the use of lesion-filling to correct for this problem and improve the interpretability of the results. Volumetric or morphometric imaging studies, where lesion amount and characteristics may vary between groups of patients or change over time, may especially benefit from this correction.

Original languageEnglish
Pages (from-to)2802-2814
Number of pages13
JournalHuman Brain Mapping
Volume33
Issue number12
DOIs
Publication statusPublished - Dec 2012

Fingerprint

Multiple Sclerosis
Atrophy
Gray Matter
White Matter
Brain

Keywords

  • Brain atrophy
  • Gray matter segmentation
  • Lesion-filling
  • MRI
  • Multiple sclerosis
  • White matter lesions

ASJC Scopus subject areas

  • Clinical Neurology
  • Anatomy
  • Neurology
  • Radiology Nuclear Medicine and imaging
  • Radiological and Ultrasound Technology

Cite this

Gelineau-Morel, R., Tomassini, V., Jenkinson, M., Johansen-Berg, H., Matthews, P. M., & Palace, J. (2012). The effect of hypointense white matter lesions on automated gray matter segmentation in multiple sclerosis. Human Brain Mapping, 33(12), 2802-2814. https://doi.org/10.1002/hbm.21402

The effect of hypointense white matter lesions on automated gray matter segmentation in multiple sclerosis. / Gelineau-Morel, Rose; Tomassini, Valentina; Jenkinson, Mark; Johansen-Berg, Heidi; Matthews, Paul M.; Palace, Jacqueline.

In: Human Brain Mapping, Vol. 33, No. 12, 12.2012, p. 2802-2814.

Research output: Contribution to journalArticle

Gelineau-Morel, R, Tomassini, V, Jenkinson, M, Johansen-Berg, H, Matthews, PM & Palace, J 2012, 'The effect of hypointense white matter lesions on automated gray matter segmentation in multiple sclerosis', Human Brain Mapping, vol. 33, no. 12, pp. 2802-2814. https://doi.org/10.1002/hbm.21402
Gelineau-Morel R, Tomassini V, Jenkinson M, Johansen-Berg H, Matthews PM, Palace J. The effect of hypointense white matter lesions on automated gray matter segmentation in multiple sclerosis. Human Brain Mapping. 2012 Dec;33(12):2802-2814. https://doi.org/10.1002/hbm.21402
Gelineau-Morel, Rose ; Tomassini, Valentina ; Jenkinson, Mark ; Johansen-Berg, Heidi ; Matthews, Paul M. ; Palace, Jacqueline. / The effect of hypointense white matter lesions on automated gray matter segmentation in multiple sclerosis. In: Human Brain Mapping. 2012 ; Vol. 33, No. 12. pp. 2802-2814.
@article{62381ef6299844858ceeb60e00be6138,
title = "The effect of hypointense white matter lesions on automated gray matter segmentation in multiple sclerosis",
abstract = "Previous imaging studies assessing the relationship between white matter (WM) damage and matter (GM) atrophy have raised the concern that Multiple Sclerosis (MS) WM lesions may affect measures of GM volume by inducing voxel misclassification during intensity-based tissue segmentation. Here, we quantified this misclassification error in simulated and real MS brains using a lesion-filling method. Using this method, we also corrected GM measures in patients before comparing them with controls in order to assess the impact of this lesion-induced misclassification error in clinical studies. We found that higher WM lesion volumes artificially reduced total GM volumes. In patients, this effect was about 72{\%} of that predicted by simulation. Misclassified voxels were located at the GM/WM border and could be distant from lesions. Volume of individual deep gray matter (DGM) structures generally decreased with higher lesion volumes, consistent with results from total GM. While preserving differences in GM volumes between patients and controls, lesion-filling correction revealed more lateralised DGM shape changes in patients, which were not evident with the original images. Our results confirm that WM lesions can influence MRI measures of GM volume and shape in MS patients through their effect on intensity-based GM segmentation. The greater effect of lesions at increasing levels of damage supports the use of lesion-filling to correct for this problem and improve the interpretability of the results. Volumetric or morphometric imaging studies, where lesion amount and characteristics may vary between groups of patients or change over time, may especially benefit from this correction.",
keywords = "Brain atrophy, Gray matter segmentation, Lesion-filling, MRI, Multiple sclerosis, White matter lesions",
author = "Rose Gelineau-Morel and Valentina Tomassini and Mark Jenkinson and Heidi Johansen-Berg and Matthews, {Paul M.} and Jacqueline Palace",
year = "2012",
month = "12",
doi = "10.1002/hbm.21402",
language = "English",
volume = "33",
pages = "2802--2814",
journal = "Human Brain Mapping",
issn = "1065-9471",
publisher = "John Wiley and Sons Inc.",
number = "12",

}

TY - JOUR

T1 - The effect of hypointense white matter lesions on automated gray matter segmentation in multiple sclerosis

AU - Gelineau-Morel, Rose

AU - Tomassini, Valentina

AU - Jenkinson, Mark

AU - Johansen-Berg, Heidi

AU - Matthews, Paul M.

AU - Palace, Jacqueline

PY - 2012/12

Y1 - 2012/12

N2 - Previous imaging studies assessing the relationship between white matter (WM) damage and matter (GM) atrophy have raised the concern that Multiple Sclerosis (MS) WM lesions may affect measures of GM volume by inducing voxel misclassification during intensity-based tissue segmentation. Here, we quantified this misclassification error in simulated and real MS brains using a lesion-filling method. Using this method, we also corrected GM measures in patients before comparing them with controls in order to assess the impact of this lesion-induced misclassification error in clinical studies. We found that higher WM lesion volumes artificially reduced total GM volumes. In patients, this effect was about 72% of that predicted by simulation. Misclassified voxels were located at the GM/WM border and could be distant from lesions. Volume of individual deep gray matter (DGM) structures generally decreased with higher lesion volumes, consistent with results from total GM. While preserving differences in GM volumes between patients and controls, lesion-filling correction revealed more lateralised DGM shape changes in patients, which were not evident with the original images. Our results confirm that WM lesions can influence MRI measures of GM volume and shape in MS patients through their effect on intensity-based GM segmentation. The greater effect of lesions at increasing levels of damage supports the use of lesion-filling to correct for this problem and improve the interpretability of the results. Volumetric or morphometric imaging studies, where lesion amount and characteristics may vary between groups of patients or change over time, may especially benefit from this correction.

AB - Previous imaging studies assessing the relationship between white matter (WM) damage and matter (GM) atrophy have raised the concern that Multiple Sclerosis (MS) WM lesions may affect measures of GM volume by inducing voxel misclassification during intensity-based tissue segmentation. Here, we quantified this misclassification error in simulated and real MS brains using a lesion-filling method. Using this method, we also corrected GM measures in patients before comparing them with controls in order to assess the impact of this lesion-induced misclassification error in clinical studies. We found that higher WM lesion volumes artificially reduced total GM volumes. In patients, this effect was about 72% of that predicted by simulation. Misclassified voxels were located at the GM/WM border and could be distant from lesions. Volume of individual deep gray matter (DGM) structures generally decreased with higher lesion volumes, consistent with results from total GM. While preserving differences in GM volumes between patients and controls, lesion-filling correction revealed more lateralised DGM shape changes in patients, which were not evident with the original images. Our results confirm that WM lesions can influence MRI measures of GM volume and shape in MS patients through their effect on intensity-based GM segmentation. The greater effect of lesions at increasing levels of damage supports the use of lesion-filling to correct for this problem and improve the interpretability of the results. Volumetric or morphometric imaging studies, where lesion amount and characteristics may vary between groups of patients or change over time, may especially benefit from this correction.

KW - Brain atrophy

KW - Gray matter segmentation

KW - Lesion-filling

KW - MRI

KW - Multiple sclerosis

KW - White matter lesions

UR - http://www.scopus.com/inward/record.url?scp=84868654177&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84868654177&partnerID=8YFLogxK

U2 - 10.1002/hbm.21402

DO - 10.1002/hbm.21402

M3 - Article

C2 - 21976406

AN - SCOPUS:84868654177

VL - 33

SP - 2802

EP - 2814

JO - Human Brain Mapping

JF - Human Brain Mapping

SN - 1065-9471

IS - 12

ER -