White Matter Microstructural Damage on Diffusion Tensor Imaging in Cerebral Small Vessel Disease

Abstract

HomeStrokeVol. 47, No. 6White Matter Microstructural Damage on Diffusion Tensor Imaging in Cerebral Small Vessel Disease Free AccessResearch ArticlePDF/EPUBAboutView PDFView EPUBSections ToolsAdd to favoritesDownload citationsTrack citationsPermissions ShareShare onFacebookTwitterLinked InMendeleyRedditDiggEmail Jump toSupplementary MaterialsFree AccessResearch ArticlePDF/EPUBWhite Matter Microstructural Damage on Diffusion Tensor Imaging in Cerebral Small Vessel DiseaseClinical Consequences Marco Pasi, MD Inge W.M. van Uden, MD Anil M. Tuladhar, MD Frank-Erik de Leeuw, and MD, PhD Leonardo PantoniMD, PhD Marco PasiMarco Pasi From the NEUROFARBA Department, Neuroscience Section, University of Florence, Florence, Italy (M.P., L.P.); and Department of Neurology, Donders Institute for Brain, Cognition and Behaviour, Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands (I.W.M.v.U., A.M.T., F.-E.d.L.). Search for more papers by this author , Inge W.M. van UdenInge W.M. van Uden From the NEUROFARBA Department, Neuroscience Section, University of Florence, Florence, Italy (M.P., L.P.); and Department of Neurology, Donders Institute for Brain, Cognition and Behaviour, Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands (I.W.M.v.U., A.M.T., F.-E.d.L.). Search for more papers by this author , Anil M. TuladharAnil M. Tuladhar From the NEUROFARBA Department, Neuroscience Section, University of Florence, Florence, Italy (M.P., L.P.); and Department of Neurology, Donders Institute for Brain, Cognition and Behaviour, Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands (I.W.M.v.U., A.M.T., F.-E.d.L.). Search for more papers by this author , Frank-Erik de LeeuwFrank-Erik de Leeuw From the NEUROFARBA Department, Neuroscience Section, University of Florence, Florence, Italy (M.P., L.P.); and Department of Neurology, Donders Institute for Brain, Cognition and Behaviour, Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands (I.W.M.v.U., A.M.T., F.-E.d.L.). Search for more papers by this author and Leonardo PantoniLeonardo Pantoni From the NEUROFARBA Department, Neuroscience Section, University of Florence, Florence, Italy (M.P., L.P.); and Department of Neurology, Donders Institute for Brain, Cognition and Behaviour, Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands (I.W.M.v.U., A.M.T., F.-E.d.L.). Search for more papers by this author Originally published21 Apr 2016https://doi.org/10.1161/STROKEAHA.115.012065Stroke. 2016;47:1679–1684Other version(s) of this articleYou are viewing the most recent version of this article. Previous versions:IntroductionCerebral small vessel disease (SVD) is a major health problem for its contribution to ≈45% of dementias, and about a fifth of all strokes worldwide, representing one of the most important causes of disabilities.1 The term SVD refers to a group of pathological processes with various etiologies that affect the small arteries, arterioles, venules, and capillaries of the brain. The most common forms are age- and hypertension-related SVD and cerebral amyloid angiopathy (CAA).2 Vessel wall changes may lead to both ischemic and hemorrhagic consequences: (1) a state of chronic hypoperfusion or vascular dysfunction responsible for incomplete infarction,3,4 (2) acute focal necrosis (lacunar infarct), or (3) vessel rupture manifesting as hemorrhagic SVD. The clinical consequences of SVD are various and mainly consist of cognitive, mood, and motor dysfunctions leading to functional disability in the late stages of the disease.Magnetic resonance imaging (MRI) has become crucial in the diagnosis of SVD enabling the evaluation of the disease progression both in the clinical and research settings. However, correlations between clinical features of SVD and conventional MRI measures have been partially discordant. Some authors suggested that the cumulative effect of SVD lesions, rather than the individual lesions themselves determines the clinical impact,5 whereas others suggested that the presence and severity of alterations nonvisible on conventional MRI might also be an explanation.6In the past decade, diffusion tensor imaging (DTI) has been increasingly used for the evaluation of SVD patients because it is sensitive to tissue damage and can show abnormalities in both areas of white matter hyperintensities (WMH) and in normal appearing WM (NAWM). Despite the high sensitivity in detecting cerebral damage, DTI has a low specificity in detecting the underlying cause. In fact, we can only infer that DTI changes reflect a loss of WM integrity because of damage to structures that restrict molecular movement along the primary axis of the axons, such as axonal cell membranes, myelin sheaths, and neurofilaments. Growing evidences indicate definite structural vascular abnormalities associated with WMH, strengthening the argument that WMH have a vascular pathogenesis.7 DTI is suited to study cortical disconnection because it provides indices of microstructural integrity within interconnected neural networks. Most DTI studies evaluated WM microstructural damage in aging, Alzheimer disease and mild cognitive impairment patients, but recently, studies in SVD patients have documented a significant association between WM microstructural damage and clinical features, providing new insight in the biological basis of this condition.8–10The aim of this review is to analyze the evidence of the role of WM microstructural damage beyond the standard structural MRI sequences, evaluated with DTI, in the clinical consequences of cerebral SVD. Although some definitions have been proposed,11 it is explicitly not our intention to clinically define a SVD patient, as SVD is often accompanied by other processes, such as aging and neurodegeneration, leading to a broad spectrum of clinical manifestations. Our review is based on those studies that enrolled patients with a predominant SVD pathology and does not include those that enrolled non-SVD participants (eg, healthy aging, Alzheimer disease, and mild cognitive impairment) even though they evaluated the association between SVD markers and WM microstructural damage.The first part of this review briefly examines the methodological aspects of DTI, the role of DTI in understanding SVD pathophysiology and the relationship between risk factors and WM microstructural damage. In the second part, we review clinical studies that reported on the association between DTI and different clinical manifestations of SVD, focusing on cognition, mood disorders, and motor dysfunctions. The last part of the review outlines possible future applications of DTI, in particular its role as a sensitive marker to evaluate SVD progression in clinical trials.Article Search StrategyArticles were identified through Pubmed searches using these terms: DTI, diffusion, WM microstructural damage, WM integrity, structural network, brain connectivity AND each of the following: SVD, subcortical small vessel, Binswanger, vascular cognitive impairment, CAA, Cerebral Autosomal Dominant Arteriopathy With Subcortical Infarcts and Leukoencephalopathy (CADASIL), vascular dementia, motor symptoms, gait, falls, balance, parkinsonism, depression, depressive symptoms, mood, smoking, hypertension, blood pressure, diabetes mellitus, blood glucose, body mass index, hypercholesterolemia, physical activity, leukoaraiosis, and lacunes, from any date to November 15, 2015.Methodological AspectsDTI is a quantitative MR technique that measures the movement of water within the tissue microstructure applying a magnetic diffusion gradient in more directions (at least 6) to acquire a diffusion tensor.12 From the tensor, 2 commonly derived quantitative measures that provide information about the in vivo WM microstructure are fractional anisotropy (FA) and mean diffusivity (MD). FA is a measure of anisotropic water diffusion, which reflects the degree of directionality of cellular structures within the WM tracts and ranges from 0 (diffusivity equal in all directions) to 1 (entirely unidirectional). MD is the average rate of diffusion in the noncollinear directions and an increasing value represents an increase in water diffusion. A lower FA and corresponding higher MD are generally believed to reflect lower microstructural connectivity. Other tensor indices have been proposed as markers of neuronal damage, such as axial diffusivity and radial diffusivity. Once FA and MD maps are generated, postprocessing procedures start with mainly 3 approaches: region of interest (Figure [A]), tract-based spatial statistics (TBSS; Figure [B]), and voxel-based analysis (Figure [C]). Each of these techniques has pro and cons. Technical aspects related to brain network analysis are briefly described in the figure (Figure [D]).Download figureDownload PowerPointFigure. Diffusion tensor imaging (DTI) post processing techniques with pros and cons, construction of structural network based on diffusion tensor tractography and network measures. A, Region-of-interest-based analysis, (B) tract-based spatial statistics, (C) voxel-based analysis. D, Construction of structural network based on DTI followed by tractography. Each individual network is represented as a graph, a mathematical model of a network with nodes (brain regions) linked by edges (white matter connections), which can then be explored using graph theory. There is a wide variety of network measures, only some of which are depicted in the figure. Basic network measures, including degree and strength of a node, are depicted in the figure. Degree of a node represents the number of connections linked to a node in a network. Strength of a node is defined as the sum of the weights of all edges connected to a node. Characteristic path length is defined as the shortest path or the minimum number of connections between 2 given nodes. Efficiency is inversely related to path length, which is easier to use in a disconnected graph. ROI indicates region-of-interest; and TBSS, tract-based spatial statistics.DTI and Cerebral SVDAmong DTI studies in patients with sporadic SVD, mainly defined as presence of moderate/severe WMH and lacunar infarcts not related to a monogenic disease, the predominant findings are that FA is decreased and MD is increased both in NAWM and WMH suggesting decline in the composition and integrity of the WM.13 Similar results have been reported also in CADASIL patients (Table I in the online-only Data Supplement).A substantial loss of anisotropy in the regions of WMH compared with NAWM has been reported,14 whereas NAWM showed lower FA values close to WMH compared with more distant areas15 For this reason, the term WM penumbra has been proposed to describe the area just surrounding the WMH which is still composed of NAWM based on conventional MRI, but with already lower structural integrity as compared with more remote areas of WM.15A positive correlation has been reported between the total WMH load and diffuse WM injury in NAWM, suggesting WM damage to be more widespread rather than region specific. Normal WM, WMH penumbra, and WMH all show a similar decline in WM integrity over time.15,16 Accordingly, NAWM regions that ultimately converted into WMH had already significant lower FA and higher MD at baseline in both growing WMH (defined as WMH expanding from already present WMH at baseline) and de novo WMH (defined as a new WMH not adhering to an already present WMH at baseline) compared with persistent NAWM.17 These results highlight that WMH develop gradually, and that WMH are only the tip of the iceberg of WM pathology.DTI tractography can be used to spatially characterize WM diffusion abnormalities along the pathway of a specific tract. Using a reconstructed WM tract containing a lacunar infarct, Reijmer et al18 showed that WM microstructural damage attenuates with increasing distance from the primary lesion. This finding was replicated also in CAA patients.19 Duering et al20 applied serial cortical thickness measurements and tractography in CADASIL patients and showed focal cortical thinning in cortical regions with high probability of connectivity with the incident infarct. This result provided evidence for cortical neurodegeneration after subcortical ischemia as one mechanism for brain atrophy in cerebrovascular disease. The same group has replicated this finding in a non-CADASIL cohort.21Recently, new advances in network analysis have been used to study the whole brain connectivity using graph theory. This can be applied after structural networks have been reconstructed from diffusion tensor tractography (Figure [D]). SVD patients have been reported to have networks less densely connected, and reductions in both global and local efficiency, compared with controls, especially in interhemispheric and prefrontal tracts.22 A similar approach showed network disturbances, most pronounced in the occipital, parietal, and posterior temporal lobes, in CAA patients.5Risk Factors for Microstructural Damage Within the WM in SVDTo date, the Radboud University Nijmegen Diffusion Tensor and Magnetic resonance Cohort (RUN DMC) is the only study investigating cross-sectionally the role of vascular risk factors on microstructural changes in SVD (Table II in the online-only Data Supplement).23,24 Increased blood pressure (average of 3 measurements of systolic and diastolic blood pressures) and hypertension (defined as blood pressure >140/80 mm Hg and use of blood pressure–lowering agents) were associated with loss of WM integrity in both the NAWM and WMH.23 In particular, hypertension was associated with lower FA in the splenium of the corpus callosum and higher MD in both the anterior body and the splenium of the corpus callosum. These associations disappeared after adjustment for other SVD markers, such as WMH volume, lacunes, and gray matter atrophy, evoking the possible role of mediator of SVD between hypertension and low microstructural integrity.24In one SVD cohort, both history and duration of smoking were associated with a low WM microstructural integrity, and diffusion values were comparable between those who had quit smoking more than 20 years and those who had never smoked. This may suggest a beneficial role of quitting smoking on WM structural integrity.25One cross-sectional study, using TBSS, investigated the relationship between physical activity and WM microstructural integrity showing that poor physical activity was associated with lower microstructural integrity in almost all voxels of the TBSS skeleton.26No study has to date investigated the association between diabetes mellitus and microstructural damage specifically in a SVD population. A recent review, however, reported on 5 cross-sectional studies examining the relation between DTI parameters and diabetes mellitus and all found lower microstructural integrity in patients with type 2 diabetes mellitus compared with controls, adjusted for different confounders.27Taken together, these studies may suggest that vascular risk factors could damage WM integrity in elderly patients with SVD and that their control might be associated with better DTI parameters. This may not apply to normal aging because recently in a population-based cohort cardiovascular risk factors were not associated with longitudinal changes in white matter microstructure.28Clinical Expressions of SVD and WM Microstructural DamageCognitionSVD patients are prone to develop cognitive impairment and their neuropsychological profile is generally characterized by a predominant impairment of executive functions, attention, and psychomotor speed. One of the most accepted mechanisms of cognitive impairment in SVD is based on the disconnection theory by which it is hypothesized that impairment in attention, processing speed, and executive function is related to the disruption of fronto-subcortical circuits. Indeed, it has been demonstrated in both sporadic SVD and CADASIL patients that the forceps minor and the thalamic radiation are strategic WM tracts for processing speed.29,30O’Sullivan et al13 demonstrated that in SVD patients DTI indices, especially in NAWM, correlated more strongly with cognitive function than T2-lesion volume, after controlling for conventional MRI parameters. Similarly, diffusion changes predict faster decline in psychomotor speed, executive functions, and working memory regardless of conventional MRI findings.31 Other groups have confirmed the strong association between WM microstructural damage and cognitive impairment in sporadic SVD patients, especially in terms of executive functions, attention, and psychomotor speed (Table III in the online-only Data Supplement). In CADASIL patients, executive performances were reported to be correlated with MD in the frontal WM and through the major antero-posterior fasciculus of the cingulum bundle.32A further contribution to the understanding of the relationship between WM microstructural damage and cognition comes from the RUN DMC study in which more than 500 independently living, nondemented patients with cerebral SVD, aged between 50 and 85 years, were enrolled. In this large cohort, the microstructural integrity of both WMH and NAWM was related to global cognitive function, memory, and executive function.33 Moreover, TBSS postprocessing analyses were performed and corpus callosum especially in the genu and splenium showed the highest significant relation with global cognitive index. Analyses for each cognitive domain showed the strongest relationship between (1) cingulum bundle microstructural integrity and verbal memory performance and (2) frontal WM and psychomotor speed.8 However, in the same cohort, the main predictors for the development of incident dementia at 5 years were WM and hippocampal volumes,34 whereas baseline WM integrity was not associated with decline in cognitive performances35In the Vascular Mild Cognitive Impairment Tuscany study, WM microstructural damage was more strongly reflected in Montreal cognitive assessment than mini mental status examination performances,36 possibly for the presence in Montreal cognitive assessment of items reflecting executive functions and psychomotor speed.Interesting insights in the development of cognitive impairment related to SVD come from the evaluation of network connectivity. In both SVD and CAA cohorts, the importance of network disruption as a mediating mechanism between SVD MRI burden and cognitive dysfunction, especially in executive functions, has been demonstrated.5,22,37 Moreover, it has been shown that structural network efficiency is a predictor of conversion to dementia.38Depressive SymptomsPrevious cross-sectional studies showed a positive association between conventional SVD characteristics and depressive symptoms in older age, both at a cross-sectional level39 and prospectively40 DTI studies performed in patients with late life depression consistently showed lower microstructural integrity in the fronto-striatal and limbic networks.41To date, 4 studies investigated the role of the WM microstructure in SVD in relation to depressive symptoms (Table IV in the online-only Data Supplement).9,42–44 The first study found that microstructural WM damage, measured by median FA, at least partially mediated the association between SVD and depression.42 The second study, using TBSS, showed that low WM microstructural integrity in the genu and body of the corpus callosum, bilateral inferior fronto-occipital fasciculus, uncinate fasciculus, and corona radiata was associated with depressive symptoms. These associations almost fully disappeared after adjustment for WMH and lacunes, suggesting that the visible SVD drives the association.9 The third study reported an association between WM microstructural damage and depressive symptoms in mild cognitive impairment patients with SVD independently of disability or cognitive or motor impairment.43 The last study evaluated the relationship between FA and both apathy and depression, finding that only apathy was related to damage of cortical–subcortical networks.44The majority of these studies suggest that the association between WM microstructural damage and depressive symptoms might be mediated by the underlying SVD and to a lower extent by other factors, such as disability.Motor ProblemsOnly a small number of studies have investigated the relation between WM integrity and motor impairment (gait, parkinsonism, falls, and balance) in SVD using DTI (Table V in the online-only Data Supplement).5,10,45–51 Loss of WM integrity, most pronounced in the corpus callosum, especially the genu,46,47 was associated with lower gait velocity at a cross-sectional level.5,45–47 This association with gait was seen for both NAWM and WMH. Network efficiency was also related to gait velocity in CAA patients, suggesting a role of network disruption in this relation.5 Other studies investigated the cross-sectional associations between microstructural integrity and a clinical scale measuring extrapyramidal motor deficits,48 extrapyramidal movement disorders, such as freezing of gait,49 and mild Parkinsonian signs.10 Three studies found an association between extrapyramidal motor symptoms and low microstructural integrity in both supratentorial (frontal lobes)10,50,51 and infratentorial (pedunculopontine nucleus) regions. A prospective study showed a low baseline microstructural integrity of several bifrontal WM tracts involved in movement control in participants with incident vascular parkinsonism in comparison to those without51 also after adjustment for SVD characteristics.These studies uniformly support the notion that, in SVD, disturbances of frontal WM microstructure, especially the genu of the corpus callosum, are associated with motor deficits, and related to incident vascular parkinsonism.Future Directions and ConclusionsThe evaluation of WM microstructural damage has gained attention during the past 15 years in the study of SVD because it provides in vivo an understanding of the pathogenesis of important clinical and neuroimaging consequences of SVD. The majority of the studies that have used DTI demonstrated a good correlation between WM microstructural damage and several clinical measures linked to SVD such as cognition, mood disorders and motor performances. In the studies where a multimodal approach was used, DTI indices were generally strongly associated with clinical outcome measures also after correction for multiple conventional neuroimaging markers of SVD. Furthermore, longitudinal studies showed that changes in DTI parameters could be detected during a period of 1 or 2 years.52,53 In CADASIL patients, for example, Molko et al54 found important changes in DTI parameters during a period of 20 months, whereas no changes were detected in the control group. These findings suggest that DTI might be considered a sensitive biomarker to monitor the progression of WM damage in patients with SVD. This may be particularly relevant because DTI indices were shown to be predictors of clinical progression in both sporadic SVD and CADASIL.52–55 Therefore, the measurement of diffusion will possibly become one important surrogate marker in future preventive trials in SVD.There are at least 3 possible ways in which DTI can be of aid in a better understanding of the pathogenesis and clinical consequences of SVD: (1) It may provide new insights in the understanding of the mechanisms of the main clinical consequences of SVD, particularly by evaluating the structural integrity of the cerebral WM architecture; (2) it may furnish a reliable surrogate marker, especially in clinical trials, of SVD progression over time to appreciate the effects of beneficial therapeutic interventions; (3) it may help to better appreciate the real SVD burden and its progression.Sources of FundingDr de Leeuw was supported by a VIDI innovational grant from the Netherlands Organization for Scientific Research (Nederlandse Organisatie voor Wetenschappelijk Onderzoeck [NWO], Grant 016-126-351).DisclosuresNone.FootnotesThe online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.115.012065/-/DC1.Correspondence to Leonardo Pantoni, MD, PhD, NEUROFARBA Department, Neuroscience Section, University of Florence, Largo Brambilla 3, 50134 Florence, Italy. E-mail [email protected]