Conference PaperPDF Available

Investigating microstructural heterogeneity of white matter hyperintensities in Alzheimer’s disease using single-shell 3-tissue constrained spherical deconvolution

Authors:
Remika Mito at Florey Institute of Neuroscience and Mental Health
Remika Mito
Thijs Dhollander at Murdoch Children's Research Institute
Thijs Dhollander
David Raffelt at Florey Institute of Neuroscience and Mental Health
David Raffelt
Ying Xia at University of Queensland
Ying Xia
Show all 9 authorsHide
Download full-text PDF
Read full-text
Download citation

Abstract

White matter hyperintensities (WMH) observed on FLAIR MRI are highly prevalent in Alzheimer’s disease. Although often associated with cognitive decline, such associations are highly variable, likely due to the underlying pathological heterogeneity within these lesions. Here, we explore this potential heterogeneity in vivo in an Alzheimer’s disease cohort, by investigating relative tissue fractions obtained using single-shell 3-tissue constrained spherical deconvolution (SS3T-CSD). We show distinguishable tissue profiles of lesions based on classification as periventricular or deep, and additionally show heterogeneity within lesions, thus highlighting the pitfalls of binary classification of WMH, and the value of investigating their underlying diffusional properties.
Content uploaded by Thijs Dhollander
Author content
All content in this area was uploaded by Thijs Dhollander on Jun 13, 2018
Content may be subject to copyright.
0135
Investigating microstructural heterogeneity of white matter hyperintensities in Alzheimer’s
disease using single-shell 3-tissue constrained spherical deconvolution
Remika Mito , Thijs Dhollander , David Raelt , Ying Xia , Olivier Salvado , Amy Brodtmann , Christopher Rowe ,
Victor Villemagne , and Alan Connelly
Florey Institute of Neuroscience and Mental Health, Melbourne, Australia, Florey Department of Neuroscience and Mental Health, University of
Melbourne, Melbourne, Australia, The Australian eHealth Research Centre, CSIRO Health and Biosecurity, Brisbane, Australia, Department of Medicine,
Austin Health, University of Melbourne, Melbourne, Australia, Department of Molecular Imaging & Therapy, Centre for PET, Austin Health, University of
Melbourne, Melbourne, Australia
Synopsis
White matter hyperintensities (WMH) observed on FLAIR MRI are highly prevalent in Alzheimer’s disease. Although often associated with cognitive
decline, such associations are highly variable, likely due to the underlying pathological heterogeneity within these lesions. Here, we explore this
potential heterogeneity in vivo in an Alzheimer’s disease cohort, by investigating relative tissue fractions obtained using single-shell 3-tissue
constrained spherical deconvolution (SS3T-CSD). We show distinguishable tissue proles of lesions based on classication as periventricular or
deep, and additionally show heterogeneity within lesions, thus highlighting the pitfalls of binary classication of WMH, and the value of
investigating their underlying diusional properties.
Introduction
White matter hyperintensities (WMH) are commonly observed on T2-weighted FLAIR images in elderly individuals, and thought to be linked to
vascular risk, age, and cognitive decline . They are more prevalent and severe in Alzheimer’s disease (AD) patients ; however, the clinical
signicance of these lesions is insuciently understood, with inconsistent reports of associations with cognitive decline in dementia . Such
inconsistencies may arise from underlying heterogeneity within WMH, which are typically modelled in binary fashion as either present or absent in
FLAIR images. Histologically, these lesions indeed exhibit heterogeneity in their pathological substrates despite appearing homogeneous on FLAIR,
and have been associated with axonal loss, demyelination, gliosis, and arteriolosclerosis, amongst others . Other imaging modalities such as
magnetization transfer imaging have also suggested underlying heterogeneity across lesions .
Advanced diusion imaging could provide greater in vivo insight into heterogeneity both within and across WMHs. Single-shell 3-tissue constrained
spherical deconvolution (SS3T-CSD) enables estimation of white matter (WM) bre orientation distributions (FODs) as well as grey matter (GM) and
CSF compartments. It can be used to quantify the relative WM-GM-CSF-likeness of the signal, which may relate to the underlying tissue
microstructure in pathology .
Purpose
We aimed to explore the underlying diusional properties of WMH in an AD cohort using SS3T-CSD, to determine if the technique could reveal and
characterise underlying heterogeneity across dierent lesion types, and within lesions.
Methods
DWI-data were acquired from 48 AD and 94 healthy elderly control (HC) subjects (demographics in Table 1) from the Australian Imaging, Biomarkers
and Lifestyle (AIBL) study on a 3T Siemens Trio scanner (2.3mm isotropic voxels, 60 directions at b=3000s/mm , 8 b=0 images), along with FLAIR
(0.9x1x1mm ) images. All data were denoised , corrected for motion/eddy-currents and bias elds . WMH segmentations were automatically
performed on FLAIR images using the HyperIntensity Segmentation Tool (HIST) , and classied as periventricular or deep, based on distance of
lesion volumes from ventricles (see Fig. 1).
WM FODs and GM/CSF compartments were computed with SS3T-CSD using average WM/GM/CSF response functions obtained from the data
themselves . Spatial correspondence was achieved by registering each subject’s FOD image to a study-specic FOD template, along with FLAIR and
WMH segmentations. A normal-appearing WM (NAWM) mask was created by subtracting WMH segmentations from a WM mask in template space.
The WM-GM-CSF compartments from SS3T-CSD were normalised to obtain signal fractions (that summed to 1). Mean signal fractions were
computed within periventricular WMH, deep WMH, and NAWM for each subject.
Results
AD patients exhibited signicantly greater volume in periventricular WMH, but not deep WMH, compared to HC (Table 1). In AD, mean relative
tissue fractions derived from SS3T-CSD showed dierent proles of WM-GM-CSF fractions in periventricular and deep WMH, with higher CSF-
likeness in periventricular, and higher GM-likeness in deep WMH (Fig. 2). Periventricular and deep WMH formed separate clusters based on their
relative tissue fraction proles, as did NAWM (Fig. 3). In addition, heterogeneity was consistently observed within segmented lesions (Fig. 4).
Discussion
Accumulating evidence suggests that WMH are heterogeneous in their underlying pathology, which may explain diculties in untangling their
association with clinical and pathological progression of Alzheimer's disease. Using SS3T-CSD, we reveal underlying heterogeneity within these
lesions, and identify “GM-like” compartments within WMH, which have been suggested to represent gliosis, as well as “CSF-like” compartments that
likely indicate increased interstitial uid .
We show that periventricular and deep WMH exhibit distinct WM-GM-CSF proles, and can thus be distinguished by their diusional properties
from one another (and from NAWM), likely due to diering pathological substrates. As shown in Figs. 2 and 3, periventricular WMH exhibited higher
relative CSF-like fraction than deep WMH across AD subjects, suggesting increased interstitial uid within these lesions, likely related to substantial
myelin and axonal loss. Given the higher periventricular WMH volume in AD compared to HC, these lesions could be more deleterious than deep
WMH, and more closely associated with AD as previously suggested . Deep lesions, which were equally extensive in HC as in AD, exhibited greater
1,2 1 1 3 3 1,2 4,5
4,5 1,2
1 2
3 4
5
1,2 3,4
5
3,6
7
8
9
3 2
3 10 11 12
13
8
14
9
15
GM-like fraction, which may reect gliosis in response to white matter damage. However, these classications alone did not capture the
heterogeneity within lesions, and substantial variability was consistently observed in relative WM-GM-CSF-like fractions within regions classed
together as periventricular or deep WMH (Fig. 4). We thus highlight the importance of investigating these lesions as heterogeneous entities when
probing associations with histopathology and clinical progression in AD. To this end, relative tissue fractions from SS3T-CSD will likely reect
histological dierences better than FLAIR, which could have widespread disease-based applications, and could additionally guide investigation of
lesions and their potential association with tract-specic changes.
Acknowledgements
No acknowledgement found.
References
1. Debette, S. & Markus, H. S. The clinical importance of white matter hyperintensities on brain magnetic resonance imaging: systematic review and
meta-analysis. BMJ 341, c3666 (2010).
2. Brickman, A. M. Contemplating Alzheimer’s disease and the contribution of white matter hyperintensities. Curr. Neurol. Neurosci. Rep. 13, 1–9
(2013).
3. Scheltens, P. et al. Histopathologic correlates of white matter changes on MRI in Alzheimer’s disease and normal aging. Neurology 45, 883–888
(1995).
4. Yoshita, M. et al. Extent and distribution of white matter hyperintensities in normal aging, MCI, and AD. Neurology 67, 2192–2198 (2006).
5. Schmidt, R. et al. MRI-detected white matter lesions: do they really matter? J. Neural Transm. 118, 673–681 (2011).
6. Gouw, A. A. et al. Heterogeneity of white matter hyperintensities in Alzheimer’s disease: post-mortem quantitative MRI and neuropathology.
Brain 131, 3286–3298 (2008).
7. Spilt, A. et al. Not all age-related white matter hyperintensities are the same: a magnetization transfer imaging study. AJNR Am. J. Neuroradiol. 27,
1964–1968 (2006).
8. Dhollander, T. & Connelly, A. A novel iterative approach to reap the benets of multi-tissue CSD from just single-shell (+ b= 0) diusion MRI data.
Proceedings of the International Society for Magnetic Resonance in Medicine. Singapore, Singapore 24, 3010 (2016).
9. Dhollander, T., Raelt, D. & Connelly, A. Towards interpretation of 3-tissue constrained spherical deconvolution results in pathology. in Proc. Intl.
Soc. Mag. Reson. Med 25, 1815 (2017).
10. Veraart, J., Fieremans, E. & Novikov, D. S. Diusion MRI noise mapping using random matrix theory. Magn. Reson. Med. (2016).
11. Andersson, J. L. R. & Sotiropoulos, S. N. An integrated approach to correction for o-resonance eects and subject movement in diusion MR
imaging. Neuroimage 125, 1063–1078 (2016).
12. Tustison, N. J. et al. N4ITK: improved N3 bias correction. IEEE Trans. Med. Imaging 29, 1310–1320 (2010).
13. Manjón, J. V. et al. HIST: HyperIntensity Segmentation Tool. in Patch-Based Techniques in Medical Imaging 92–99 (Springer, Cham, 2016).
14. Dhollander, T., Raelt, D. & Connelly, A. Unsupervised 3-tissue response function estimation from single-shell or multi-shell diusion MR data
without a co-registered T1 image. Proceedings of the International Society for Magnetic Resonance in Medicine Workshop on Breaking the Barriers
of Diusion MRI 5, (2016).
15. Fazekas, F., Chawluk, J. B., Alavi, A., Hurtig, H. I. & Zimmerman, R. A. MR signal abnormalities at 1.5 T in Alzheimer’s dementia and normal aging.
American journal of roentgenology 149, 351–356 (1987).
16. Hamilton, N. ggtern: An Extension to ‘ggplot2’, for the Creation of Ternary Diagrams. R package version 2.2.2. https://CRAN.R-
project.org/package=ggtern (2017).
Figures
Table 1: Demographic data. Data presented as mean (SD) or number of males (%) for sex. Reported p-values from student’s t-tests for age and
intracranial volume (ICV), chi-square test for independence for sex, and one-way ANCOVA (with age and ICV as covariates) for PVWMH
(periventricular WMH), DWMH (deep WMH) and total WMH volumes. No signicant dierences were observed between groups for age, gender, ICV,
or DWMH volume. PVWMH volume was signicantly greater in AD compared to controls (as was total WMH volume, as expected).
Figure 1: Appearance of WMH on FLAIR, segmentation classications, and tissue compartments from SS3T-CSD. Left: WMH are typically segmented
from FLAIR images, where they appear hyperintense. Middle: Segmentations were classied into “periventricular” and “deep” WMH based on the
minimum and average distance of a lesion from the ventricles (classied as periventricular if the minimum distance < 5.0 mm or average distance <
20.0 mm, and deep otherwise). Right: SS3T-CSD enables modelling of WM FODs and GM/CSF compartments. Heterogeneity with regard to the
underlying WM-GM-CSF-likeness of WMH can be observed with SS3T-CSD.
Figure 2: Boxplots showing relative signal fractions within WMH and NAWM. The mean relative signal fractions for WM-GM-CSF for all AD subjects
(n=48) are summarised into boxplots. Boxplots display median, rst and third quartiles, and 95% condence interval of the median across subjects.
NAWM exhibits high WM-like fraction as expected, with relatively low GM/CSF-like signal fractions. In contrast, WMHs exhibited higher GM/CSF-like
signal fractions. Periventricular WMH exhibited higher CSF-likeness compared to deep WMH, which exhibited higher GM-likeness. These lesions
could be distinguished based on their relative signal fraction proles, as similarly shown in Fig. 3.
Figure 3: Ternary plot exhibiting relative signal fractions within lesions and NAWM. For each subject, the periventricular WMH (red circles), deep
WMH (green triangles), and NAWM (blue squares) are displayed on a ternary plot (created using ggtern package in R) , with the location
corresponding to the relative WM-GM-CSF fraction of the lesions (or NAWM). The relative tissue fraction (as a percentage) is shown along the left
(WM-like), right (GM-like), and bottom (CSF-like) axes. Remarkably, the periventricular WMH, deep WMH, and NAWM appear in distinct clusters,
exhibiting their dierent proles with regard to relative tissue fractions obtained from SS3T-CSD diusion data.
Figure 4: Heterogeneity within WMH. WMH appear as a homogeneous lesion on FLAIR images from which they are most commonly segmented
(segmentation outline shown). Tissue maps derived from SS3T-CSD show that there is heterogeneity within a single lesion with regard to the
relative tissue components. The insets on the left display the heterogeneity of the relative tissue compartments within the same lesion
segmentation. The underlying pathological changes within these lesions is also likely to be heterogeneous, of which the diusional changes are
likely reective.
16
Proc. Intl. Soc. Mag. Reson. Med. 26 (2018)
0135
... However, the interpretation of those tissue types remains challenging in diseased brains. Leveraging previous works [9,10] on tissue heterogeneity in white matter hyper intensities, we characterise tissue heterogeneity with brain metastases. ...
... Multi-shell multi-tissue constrained spherical deconvolution [8] (MSMT-CSD) was applied to the diffusion data using a set of 3-tissue group-averaged response functions, followed by image intensity inhomogeneity correction of the resulting tissue maps (WMfrac, GMfrac, CSFfrac). The relative tissue fractions were then normalized [10] so that WMfrac + GMfrac + CSFfrac = 1. The T1w images of each patient were registered to their respective total Apparent Fiber Density maps [17] using ANTs [18] . ...
... Discussion Extrapolating from the recommendations given by Dhollander et al. [9] and Miko et al. [10] regarding the interpretation of signal fractions obtained from MSMT-CSD, we observed: ...
Characterising tissue heterogeneity in cerebral metastases using multi-shell multi-tissue constrained spherical deconvolution
Conference Paper
Full-text available
  • May 2019
Synopsis Considerable attention has focused on characterizing brain tumours using diffusion tensor imaging, and only more recently using advanced modelling techniques. Building on the observation that metastatic tumors exhibit different signal intensities depending on their histological/cellular composition, we investigate how multi-shell multi-tissue constrained spherical deconvolution can characterise tissue heterogeneity within brain metastases.
View
... WM, GM, and CSF tissue response functions were generated using the Dhollander algorithm [65], and single-shell 3-tissue constrained spherical deconvolution was used to generate the WM fiber orientation distribution (FODs) and GM and CSF representations. 3-Tissue Constrained Spherical Deconvolution [66][67][68][69] was used to calculate the voxel-wise maps of the fraction of signal arising from each of 3 compartments: an intracellular anisotropic, intracellular isotropic, and extracellular isotropic freely diffusing water compartment by setting the sum of all FOD coefficients equal to unity. WM-FODs were then used to create a cohort-specific template with a subset of 40 individuals counterbalanced between sex and diagnosis [70]. ...
... It should also be noted that extensive work has shown that extracellular water increases in the axonal WM skeleton with age and cerebrovascular injury [68,[107][108][109], but that none of the axonal ROIs tested was significant for ECI signal fraction differences, suggesting that there Fig 7, including standard error. ECI signal fraction increased in the composite cortical ROI, aggregate g-ratio decreased in the composite WM ROI, and aggregate conduction velocity increased in the composite WM ROI. ...
Conduction velocity, G-ratio, and extracellular water as microstructural characteristics of autism spectrum disorder
Article
Full-text available
The neuronal differences contributing to the etiology of autism spectrum disorder (ASD) are still not well defined. Previous studies have suggested that myelin and axons are disrupted during development in ASD. By combining structural and diffusion MRI techniques, myelin and axons can be assessed using extracellular water, aggregate g-ratio, and a new approach to calculating axonal conduction velocity termed aggregate conduction velocity, which is related to the capacity of the axon to carry information. In this study, several innovative cellular microstructural methods, as measured from magnetic resonance imaging (MRI), are combined to characterize differences between ASD and typically developing adolescent participants in a large cohort. We first examine the relationship between each metric, including microstructural measurements of axonal and intracellular diffusion and the T1w/T2w ratio. We then demonstrate the sensitivity of these metrics by characterizing differences between ASD and neurotypical participants, finding widespread increases in extracellular water in the cortex and decreases in aggregate g-ratio and aggregate conduction velocity throughout the cortex, subcortex, and white matter skeleton. We finally provide evidence that these microstructural differences are associated with higher scores on the Social Communication Questionnaire (SCQ) a commonly used diagnostic tool to assess ASD. This study is the first to reveal that ASD involves MRI-measurable in vivo differences of myelin and axonal development with implications for neuronal and behavioral function. We also introduce a novel formulation for calculating aggregate conduction velocity, that is highly sensitive to these changes. We conclude that ASD may be characterized by otherwise intact structural connectivity but that functional connectivity may be attenuated by network properties affecting neural transmission speed. This effect may explain the putative reliance on local connectivity in contrast to more distal connectivity observed in ASD.
View
... Resolving these different compartments using either the three-tissue CSD method (ie, MSMT-CSD or SS3T-CSD) holds value beyond improving WM tractography: It can also serve as a proxy for the evaluation of brain microstructure and tissue composition. [24][25][26] By interrogating brain voxels for diffusion signal patterns that look like compositions of the diffusion signals represented by the WM/GM/CSF response functions, it might be possible to gain quantifiable information about microstructure ( Figure 1). Using these basic compartments as a diffusion signal model focuses | 2163 NEWMAN Et Al. more on coarse properties of brain-tissue microstructure rather than separating similar cell types (e.g., different populations of glial cells) or separating different types of pathology (e.g., edema, CSF infiltration in neurodegeneration, and damage from ischemic stroke). ...
... They provide information on a voxel-wise or region-wise basis for analysis of subcortical structures, lesions, or developing brains. [24][25][26]46 Related microstructural analysis of free-water signal fractions has been performed in the context of Parkinson's disease, 57,58 Schizophrenia, 59,60 and concussion. 61 Three-tissue CSD techniques may therefore have the potential to be applied to a variety of these and other neurological conditions. ...
Test–retest reliability and long‐term stability of three‐tissue constrained spherical deconvolution methods for analyzing diffusion MRI data
Article
Full-text available
Purpose Several recent studies have used a three‐tissue constrained spherical deconvolution pipeline to obtain quantitative metrics of brain tissue microstructure from diffusion‐weighted MRI data. The three tissue compartments, consisting of white matter, gray matter, and CSF‐like (free water) signals, are potentially useful in the evaluation of brain microstructure in a range of pathologies. However, the reliability and long‐term stability of these metrics have not yet been evaluated. Methods This study examined estimates of whole‐brain microstructure for the three tissue compartments, in three separate test–retest cohorts. Each cohort had different lengths of time between baseline and retest, ranging from within the same scanning session in the shortest interval to 3 months in the longest interval. Each cohort was also collected with different acquisition parameters. Results The CSF‐like compartment displayed the greatest reliability across all cohorts, with intraclass correlation coefficient (ICC) values being above 0.95 in each cohort. White matter–like and gray matter–like compartments both demonstrated very high reliability in the immediate cohort (both ICC > 0.90); however, this declined in the 3‐month interval cohort to both compartments having ICC > 0.80. Regional CSF‐like signal fraction was examined in bilateral hippocampus and had an ICC > 0.80 in each cohort. Conclusion The three‐tissue constrained spherical deconvolution techniques provide reliable and stable estimates of tissue‐microstructure composition, up to 3 months longitudinally in a control population. This forms an important basis for further investigations using three‐tissue constrained spherical deconvolution techniques to track changes in microstructure across a variety of brain pathologies.
View
Single-shell derived tissue signal fraction maps show increased contrast between hippocampal subfields compared to multi-shell analysis
Conference Paper
Full-text available
  • Aug 2020
Recent advances in the analysis of diffusion MRI have allowed for the estimation of 3 tissue compartments in the brain from data with only a single non b=0 shell. There is currently no published quantitative comparison between signal fractions derived from either single- or multi-shell methods. Applying both single-shell analysis and multi-shell analysis to the same dataset shows high b-value single-shell analysis may increase contrast between different hippocampal subfields. While this effect may occur due to differences in microstructure between ROIs it should be a noted factor when applying either model and deserving of further study.
View
View
HIST: HyperIntensity Segmentation Tool
Conference Paper
Full-text available
  • Oct 2016
  • Lect Notes Comput Sci
Accurate quantification of white matter hyperintensities (WMH) from MRI is a valuable tool for studies on ageing and neurodegeneration. Reliable automatic extraction of WMH biomarkers is challenging, primarily due to their heterogeneous spatial occurrence, their small size and their diffuse nature. In this paper, we present an automatic and accurate method to segment these lesions that is based on the use of neural networks and an overcomplete strategy. The proposed method was compared to other related methods showing competitive and reliable results in two different neurodegenerative datasets.
View
Unsupervised 3-tissue response function estimation from single-shell or multi-shell diffusion MR data without a co-registered T1 image
Conference Paper
Full-text available
  • Sep 2016
To overcome the fact that the fibre orientation distribution (FOD) from constrained spherical deconvolution (CSD) assumes a single-fibre white matter (WM) response function—and is thus inappropriate and distorted in voxels containing grey matter (GM) or cerebrospinal fluid (CSF)—multi-shell multi-tissue CSD (MSMT-CSD) was proposed. MSMT-CSD can resolve WM, GM and CSF signal contributions, but requires multi-shell data. Very recently, we proposed a novel method that can achieve the same results using just single-shell data. We refer to this method as "single-shell 3-tissue CSD" (SS3T-CSD). Both MSMT-CSD and SS3T-CSD require WM, GM and CSF response functions. These can be obtained from manually selected exemplary voxels of the tissue classes, or via the procedure described initially in the MSMT-CSD paper, which relies on a highly accurately co-registered T1 image. We propose an unsupervised procedure that does not depend on a T1 image, nor registration, and works for both single-shell and multi-shell data.
View
A novel iterative approach to reap the benefits of multi-tissue CSD from just single-shell (+b=0) diffusion MRI data
Conference Paper
Full-text available
  • May 2016
Constrained spherical deconvolution (CSD) is a robust approach to resolve the fibre orientation distribution (FOD) from diffusion MRI data. However, the FOD from CSD only aims to represent "pure" white matter (WM) and is inappropriate/distorted in regions of (partial voluming with) grey matter (GM) or cerebrospinal fluid (CSF). Multi-shell multi-tissue CSD was proposed to solve this issue by estimating WM/GM/CSF components, but requires multi-shell data to do so. In this work, we provide the first proof that similar results can also be obtained from only simple single-shell (+b=0) data, and propose a novel specialised optimiser that achieves this goal.
View
An integrated approach to correction for off-resonance effects and subject movement in diffusion MR imaging
Article
Full-text available
  • Oct 2015
  • NEUROIMAGE
In this paper we describe a method for retrospective estimation and correction of eddy current (EC)-induced distortions and subject movement in diffusion imaging. In addition a susceptibility-induced field can be supplied and will be incorporated into the calculations in a way that accurately reflects that the two fields (susceptibility- and EC-induced) behave differently in the presence of subject movement. The method is based on registering the individual volumes to a model free prediction of what each volume should look like, thereby enabling its use on high b-value data where the contrast is vastly different in different volumes. In addition we show that the linear EC-model commonly used is insufficient for the data used in the present paper (high spatial and angular resolution data acquired with Stejskal–Tanner gradients on a 3 T Siemens Verio, a 3 T Siemens Connectome Skyra or a 7 T Siemens Magnetome scanner) and that a higher order model performs significantly better. The method is already in extensive practical use and is used by four major projects (the WU-UMinn HCP, the MGH HCP, the UK Biobank and the Whitehall studies) to correct for distortions and subject movement.
View
The clinical importance of white matter hyperintensities on brain magnetic resonance imaging: systematic review and meta-analysis
Article
Full-text available
  • Jul 2010
  • Br Med J
To review the evidence for an association of white matter hyperintensities with risk of stroke, cognitive decline, dementia, and death. Systematic review and meta-analysis. PubMed from 1966 to 23 November 2009. Prospective longitudinal studies that used magnetic resonance imaging and assessed the impact of white matter hyperintensities on risk of incident stroke, cognitive decline, dementia, and death, and, for the meta-analysis, studies that provided risk estimates for a categorical measure of white matter hyperintensities, assessing the impact of these lesions on risk of stroke, dementia, and death. Population studied, duration of follow-up, method used to measure white matter hyperintensities, definition of the outcome, and measure of the association of white matter hyperintensities with the outcome. 46 longitudinal studies evaluated the association of white matter hyperintensities with risk of stroke (n=12), cognitive decline (n=19), dementia (n=17), and death (n=10). 22 studies could be included in a meta-analysis (nine of stroke, nine of dementia, eight of death). White matter hyperintensities were associated with an increased risk of stroke (hazard ratio 3.3, 95% confidence interval 2.6 to 4.4), dementia (1.9, 1.3 to 2.8), and death (2.0, 1.6 to 2.7). An association of white matter hyperintensities with a faster decline in global cognitive performance, executive function, and processing speed was also suggested. White matter hyperintensities predict an increased risk of stroke, dementia, and death. Therefore white matter hyperintensities indicate an increased risk of cerebrovascular events when identified as part of diagnostic investigations, and support their use as an intermediate marker in a research setting. Their discovery should prompt detailed screening for risk factors of stroke and dementia.
View
Diffusion MRI noise mapping using random matrix theory
Article
Purpose: To estimate the spatially varying noise map using a redundant series of magnitude MR images. Methods: We exploit redundancy in non-Gaussian distributed multidirectional diffusion MRI data by identifying its noise-only principal components, based on the theory of noisy covariance matrices. The bulk of principal component analysis eigenvalues, arising due to noise, is described by the universal Marchenko-Pastur distribution, parameterized by the noise level. This allows us to estimate noise level in a local neighborhood based on the singular value decomposition of a matrix combining neighborhood voxels and diffusion directions. Results: We present a model-independent local noise mapping method capable of estimating the noise level down to about 1% error. In contrast to current state-of-the-art techniques, the resultant noise maps do not show artifactual anatomical features that often reflect physiological noise, the presence of sharp edges, or a lack of adequate a priori knowledge of the expected form of MR signal. Conclusions: Simulations and experiments show that typical diffusion MRI data exhibit sufficient redundancy that enables accurate, precise, and robust estimation of the local noise level by interpreting the principal component analysis eigenspectrum in terms of the Marchenko-Pastur distribution. Magn Reson Med, 2015. © 2015 Wiley Periodicals, Inc.
View
Contemplating Alzheimer’s Disease and the Contribution of White Matter Hyperintensities
Article
  • Dec 2013
  • CURR NEUROL NEUROSCI
As the older adult segment of the population increases, Alzheimer's disease (AD) has emerged as a significant public health epidemic. Over the past 3 decades, advances in the understanding of the biology of AD have led to a somewhat unified hypothesis of disease pathogenesis that emphasizes the precipitating role of beta amyloid protein. However, several lines of evidence suggest that multiple pathologies are necessary for clinical manifestation of the disease. Our focus over the past several years has been on the contribution of small vessel cerebrovascular disease, visualized as white matter hyperintensities (WMH) on magnetic resonance imaging, to AD. White matter hyperintensity volume, particularly in parietal regions, is elevated among individuals with and at risk for AD, predicts future diagnosis of AD, predicts the rate of progression of cognitive symptoms among individuals with AD, and increases over time among individuals destined to develop AD. White matter hyperintensities may represent an independent source of impairment and/or may interact more fundamentally with "primary" AD pathology. Future work should focus on more inclusive models of that better define "normal" vs "pathological" aging.
View
N4ITK: improved N3 bias correction
Article
  • Jul 2010
A variant of the popular nonparametric nonuniform intensity normalization (N3) algorithm is proposed for bias field correction. Given the superb performance of N3 and its public availability, it has been the subject of several evaluation studies. These studies have demonstrated the importance of certain parameters associated with the B -spline least-squares fitting. We propose the substitution of a recently developed fast and robust B-spline approximation routine and a modified hierarchical optimization scheme for improved bias field correction over the original N3 algorithm. Similar to the N3 algorithm, we also make the source code, testing, and technical documentation of our contribution, which we denote as ??N4ITK,?? available to the public through the Insight Toolkit of the National Institutes of Health. Performance assessment is demonstrated using simulated data from the publicly available Brainweb database, hyperpolarized 3He lung image data, and 9.4T postmortem hippocampus data.
View
MR Signal Abnormalities at 1.5 T in Alzheimer's Dementia and Normal Aging
Article
  • Aug 1987
The type, frequency, and extent of MR signal abnormalities in Alzheimer's disease and normal aging are a subject of controversy. With a 1.5-MR unit we studied 12 Alzheimer patients, four subjects suffering from multiinfarct dementia and nine age-matched controls. Punctate or early confluent high-signal abnormalities in the deep white matter, noted in 60% of both Alzheimer patients and controls, were unrelated to the presence of hypertension or other vascular risk factors. A significant number of Alzheimer patients exhibited a more extensive smooth "halo" of periventricular hyperintensity when compared with controls (p = .024). Widespread deep white-matter hyperintensity (two patients) and extensive, irregular periventricular hyperintensity (three patients) were seen in multiinfarct dementia. Areas of high signal intensity affecting hippocampal and sylvian cortex were also present in five Alzheimer and two multiinfarct dementia patients, but absent in controls. Discrete, small foci of deep white-matter hyperintensity are not characteristic of Alzheimer's disease nor do they appear to imply a vascular cause for the dementing illness. The frequently observed "halo" of periventricular hyperintensity in Alzheimer's disease may be of diagnostic importance. High-signal abnormalities in specific cortical regions are likely to reflect disease processes localized to those structures.
View
MRI-detected white matter lesions: Do they really matter?
Article
  • Feb 2011
  • J NEURAL TRANSM
Despite extensive research over the last decades the clinical significance of white matter lesions (WMLs) is still a matter of debate. Here, we review current knowledge of the correlation between WMLs and cognitive functioning as well as their predictive value for future stroke, dementia, and functional decline in activities of daily living. There is clear evidence that age-related WMLs relate to all of these outcomes on a group level, but the inter-individual variability is high. The association between WMLs and clinical phenotypes exists particularly for early confluent to confluent changes, which are ischaemic in aetiology and progress quickly over time. One reason for the variability of the relationship between WMLs and clinic on an individual level is probably the complexity of the association. Numerous factors such as cognitive reserve, concomitant loss of brain volume, and ultrastructural changes have been identified as mediators between white matter damage and clinical findings, and need to be incorporated in the consideration of WMLs as visible markers of these detrimental processes.
View