Willem H. Bouvy

MRBrainS challenge: online evaluation framework for brain image segmentation in 3T MRI scans

Many methods have been proposed for tissue segmentation in brain MRI scans. The multitude of methods proposed complicates the choice of one method above others. We have therefore established the MRBrainS online evaluation framework for evaluating (semi)automatic algorithms that segment gray matter (GM), white matter (WM), and cerebrospinal fluid (CSF) on 3T brain MRI scans of elderly subjects (65–80 y). Participants apply their algorithms to the provided data, after which their results are evaluated and ranked. Full manual segmentations of GM, WM, and CSF are available for all scans and used as the reference standard. Five datasets are provided for training and fifteen for testing. The evaluated methods are ranked based on their overall performance to segment GM, WM, and CSF and evaluated using three evaluation metrics (Dice, H95, and AVD) and the results are published on the MRBrainS13 website. We present the results of eleven segmentation algorithms that participated in the MRBrainS13 challenge workshop at MICCAI, where the framework was launched, and three commonly used freeware packages: FreeSurfer, FSL, and SPM. The MRBrainS evaluation framework provides an objective and direct comparison of all evaluated algorithms and can aid in selecting the best performing method for the segmentation goal at hand.

Associations between retinal microvascular changes and dementia, cognitive functioning, and brain imaging abnormalities: a systematic review.

Retinal microvascular changes can be visualized noninvasively and have been associated with cognitive decline and brain changes in relation to aging and vascular disease. We systematically reviewed studies, published between 1990 and November 2012, on the association between retinal microvascular changes and dementia, cognitive functioning, and brain imaging abnormalities, in the context of aging and vascular risk factors. In cross-sectional studies (k = 26), retinal microvascular changes were associated with the presence of dementia (range of odds ratios (ORs) 1.17;5.57), with modest decrements in cognitive functioning in nondemented people (effect sizes -0.25;0.03), and with brain imaging abnormalities, including atrophy and vascular lesions (ORs 0.94;2.95). Longitudinal studies were more sparse (k = 9) and showed no consistent associations between retinal microvascular changes and dementia or cognitive dysfunctioning 3 to 15 years later (ORs and hazard ratios 0.77;1.55). However, there were indications of prospective associations with brain imaging abnormalities ((ORs) 0.81;3.19). In conclusion, particularly in cross-sectional studies there is a correlation between retinal microvascular changes and dementia, cognitive impairment, and brain imaging abnormalities. Associations are strongest for more severe retinal microvascular abnormalities. Retinal microvascular abnormalities may offer an important window on the brain for etiological studies.

Visualization of perivascular spaces and perforating arteries with 7 T magnetic resonance imaging.

ObjectivesThe objectives of this study were to explore the possibilities of 7 T brain magnetic resonance imaging to visualize perivascular spaces (PVS) and to depict their related blood vessels. Materials and MethodsFive subjects aged 19 to 27 years and 5 subjects aged 51 to 72 years were scanned. High-resolution 3-dimensional T1-, T2-, as well as T2*-weighted sequences and time-of-flight angiography were used for the visualization of PVS, veins, and perforating arteries. Three extra subjects were scanned with a 2-dimensional time-of-flight sequence tailored to visualize small arteries and veins in the semioval center. The anatomy of PVS and their spatial relation with blood vessels were examined. The number and size of PVS in the semioval center were compared between the 2 groups. ResultsIn the basal ganglia, PVS were connected to the basal cisterns. Prominent dilations were observed in these PVS around the lower end of the putamen. From here, they ran upward and frequently showed caliber changes along their track. In the semioval center, smoothly shaped PVS started a few millimeters below the cortex, converged and tapered toward the ventricles, and ended 1 to 2 cm before the ventricle wall. Perivascular spaces correlated spatially with lenticulostriate arteries and with perforating arteries in the semioval center, but not with veins. The mean (SD) number of PVS was larger in the older subjects (55 [14]) than in the younger subjects (22 [11]), (P = 0.004). No difference in PVS diameter was observed. ConclusionsSeven-tesla magnetic resonance imaging offers detailed 3-dimensional visualization of PVS, their morphological features, and their related perforating arteries. This may offer new opportunities to study the role of PVS in ageing and cerebral small vessel disease.

Assessment of blood flow velocity and pulsatility in cerebral perforating arteries with 7‐T quantitative flow MRI

Thus far, blood flow velocity measurements with MRI have only been feasible in large cerebral blood vessels. High‐field‐strength MRI may now permit velocity measurements in much smaller arteries. The aim of this proof of principle study was to measure the blood flow velocity and pulsatility of cerebral perforating arteries with 7‐T MRI. A two‐dimensional (2D), single‐slice quantitative flow (Qflow) sequence was used to measure blood flow velocities during the cardiac cycle in perforating arteries in the basal ganglia (BG) and semioval centre (CSO), from which a mean normalised pulsatility index (PI) per region was calculated (n = 6 human subjects, aged 23–29 years). The precision of the measurements was determined by repeated imaging and performance of a Bland–Altman analysis, and confounding effects of partial volume and noise on the measurements were simulated. The median number of arteries included was 14 in CSO and 19 in BG. In CSO, the average velocity per volunteer was in the range 0.5–1.0 cm/s and PI was 0.24–0.39. In BG, the average velocity was in the range 3.9–5.1 cm/s and PI was 0.51–0.62. Between repeated scans, the precision of the average, maximum and minimum velocity per vessel decreased with the size of the arteries, and was relatively low in CSO and BG compared with the M1 segment of the middle cerebral artery. The precision of PI per region was comparable with that of M1. The simulations proved that velocities can be measured in vessels with a diameter of more than 80 µm, but are underestimated as a result of partial volume effects, whilst pulsatility is overestimated. Blood flow velocity and pulsatility in cerebral perforating arteries have been measured directly in vivo for the first time, with moderate to good precision. This may be an interesting metric for the study of haemodynamic changes in aging and cerebral small vessel disease.

Bayesian Model Selection for Pathological Neuroimaging Data Applied to White Matter Lesion Segmentation

In neuroimaging studies, pathologies can present themselves as abnormal intensity patterns. Thus, solutions for detecting abnormal intensities are currently under investigation. As each patient is unique, an unbiased and biologically plausible model of pathological data would have to be able to adapt to the subjects individual presentation. Such a model would provide the means for a better understanding of the underlying biological processes and improve ones ability to define pathologically meaningful imaging biomarkers. With this aim in mind, this work proposes a hierarchical fully unsupervised model selection framework for neuroimaging data which enables the distinction between different types of abnormal image patterns without pathological a priori knowledge. Its application on simulated and clinical data demonstrated the ability to detect abnormal intensity clusters, resulting in a competitive to improved behavior in white matter lesion segmentation when compared to three other freely-available automated methods.

Microbleeds as a predictor of intracerebral haemorrhage and ischaemic stroke after a TIA or minor ischaemic stroke: a cohort study

Objectives We examined whether patients with cerebral microbleeds on MRI, who started and continued antithrombotic medication for years, have an increased risk of symptomatic intracerebral haemorrhage (ICH). Design Prospective cohort study. Settings Multicentre outpatient clinics in the Netherlands. Participants We followed 397 patients with newly diagnosed transient ischaemic attack (TIA) or minor ischaemic stroke receiving anticoagulants or antiplatelet drugs. 58% were men. The mean age was 65.3 years. 395 (99%) patients were white Europeans. MRI including a T2*-weighted gradient echo was performed within 3 months after start of medication. 48 (12%) patients had one or more microbleeds. They were followed every 6 months by telephone for a mean of 3.8 years. Primary and secondary outcome measures Primary outcome was a symptomatic ICH. Secondary outcome were all strokes, ischaemic stroke, myocardial infarct, death from all vascular causes, death from non-vascular causes and death from all causes. Results Five patients (1%) suffered from a symptomatic ICH. One ICH occurred in a patient with microbleeds at baseline (adjusted HR 2.6, 95% CI 0.3 to 27). The incidence of all strokes during follow-up was higher in patients with than without microbleeds (adjusted HR 2.3, 95% CI 1.0 to 5.3), with a dose–response relationship. The incidences of ischaemic stroke, vascular death, non-vascular death and death of all causes were higher in patients with microbleeds, but not statistically significant. Conclusions In our cohort of patients using antithrombotic drugs after a TIA or minor ischaemic stroke, we found that microbleeds on MRI are associated with an increased risk of future stroke in general, but we did not find an increased risk of symptomatic ICH.

Robustness of automated methods for brain volume measurements across different MRI field strengths

Introduction Pooling of multicenter brain imaging data is a trend in studies on ageing related brain diseases. This poses challenges to MR-based brain segmentation. The performance across different field strengths of three widely used automated methods for brain volume measurements was assessed in the present study. Methods Ten subjects (mean age: 64 years) were scanned on 1.5T and 3T MRI on the same day. We determined robustness across field strength (i.e., whether measured volumes between 3T and 1.5T scans in the same subjects were similar) for SPM12, Freesurfer 5.3.0 and FSL 5.0.7. As a frame of reference, 3T MRI scans from 20 additional subjects (mean age: 71 years) were segmented manually to determine accuracy of the methods (i.e., whether measured volumes corresponded with expert-defined volumes). Results Total brain volume (TBV) measurements were robust across field strength for Freesurfer and FSL (mean absolute difference as % of mean volume ≤ 1%), but less so for SPM (4%). Gray matter (GM) and white matter (WM) volume measurements were robust for Freesurfer (1%; 2%) and FSL (2%; 3%) but less so for SPM (5%; 4%). For intracranial volume (ICV), SPM was more robust (2%) than FSL (3%) and Freesurfer (9%). TBV measurements were accurate for SPM and FSL, but less so for Freesurfer. For GM volume, SPM was accurate, but accuracy was lower for Freesurfer and FSL. For WM volume, Freesurfer was accurate, but SPM and FSL were less accurate. For ICV, FSL was accurate, while SPM and Freesurfer were less accurate. Conclusion Brain volumes and ICV could be measured quite robustly in scans acquired at different field strengths, but performance of the methods varied depending on the assessed compartment (e.g., TBV or ICV). Selection of an appropriate method in multicenter brain imaging studies therefore depends on the compartment of interest.

Cerebral microbleeds are not associated with long-term cognitive outcome in patients with transient ischemic attack or minor stroke.

Background: Cerebral microbleeds have been related to cerebrovascular disease and dementia. They occur more frequently in patients with ischemic stroke than in the general population, but their relation to cognition in these patients is uncertain, particularly in the long run. We examined the relationship between microbleeds in patients with a transient ischemic attack (TIA) or minor ischemic stroke, and cognitive performance 4 years later. Methods: Participants were recruited from a prospective multicenter cohort of patients with a TIA or minor ischemic stroke (n = 397). They underwent magnetic resonance imaging (MRI), including a T2*-weighted sequence, within 3 months after their ischemic event. Microbleeds, atrophy, lacunae and white matter hyperintensities (WMH) were rated visually. Cognitive status was examined in 94% of all patients who were still alive after a mean interval of 3.8 years by the Dutch version of the Telephone Interview for Cognitive Status (TICS; n = 280) or by an Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE) obtained from a close relative if a TICS could not be obtained (n = 48). The relationship between presence of microbleeds and TICS or IQCODE score was assessed with linear regression analyses adjusted for age, sex, educational level and time interval between MRI and cognitive evaluation. Results: The mean age was 65 ± 12 years at inclusion. The vascular event at inclusion was a TIA in 170 patients (52%) and a minor ischemic stroke in 155 patients (47%). Microbleeds were present in 11.6% of the patients. Patients with microbleeds were significantly older than patients without microbleeds (70 ± 9 vs. 64 ± 12 years), more often had hypertension, and had more cerebral atrophy, WMH and lacunae on MRI (all p < 0.05). The mean TICS score was 35.3 ± 5.9 for patients with microbleeds (n = 29) and 34.6 ± 5.2 for patients without microbleeds (n = 251); the adjusted mean difference (95% CI) was 1.69 (-0.01 to 3.38). The total IQCODE score was 66.0 ± 10.8 for patients with microbleeds (n = 9) and 63.1 ± 12.9 for patients without microbleeds (n = 39); the adjusted mean difference was 2.43 (-7.55 to 12.41). The relative risk (adjusted for age) for abnormal cognitive performance when having microbleeds was 1.19 (95% CI: 0.63-2.26). Subcortical atrophy was associated with lower TICS score [standardized regression coefficient β: -0.12 (-0.23 to 0.00); p = 0.04] and with lower IQCODE score [0.51 (0.19-0.83); p = 0.00]. The adjusted mean difference of IQCODE scores between patients with and those without a lacunar infarct was 0.39 (0.12-0.65; p = 0.01). Conclusions: In this sample of patients with a recent TIA or minor ischemic stroke, microbleeds were not associated with cognitive performance 4 years later. Apparently, this association is different from other markers of small vessel disease.

Perivascular spaces on 7 Tesla brain MRI are related to markers of small vessel disease but not to age or cardiovascular risk factors

Cerebral perivascular spaces (PVS) are small physiological structures around blood vessels in the brain. MRI visible PVS are associated with ageing and cerebral small vessel disease (SVD). 7 Tesla (7T) MRI improves PVS detection. We investigated the association of age, vascular risk factors, and imaging markers of SVD with PVS counts on 7 T MRI, in 50 persons aged ≥ 40. The average PVS count ± SD in the right hemisphere was 17 ± 6 in the basal ganglia and 71 ± 28 in the semioval centre. We observed no relation between age or vascular risk factors and PVS counts. The presence of microbleeds was related to more PVS in the basal ganglia (standardized beta 0.32; p = 0.04) and semioval centre (standardized beta 0.39; p = 0.01), and white matter hyperintensity volume to more PVS in the basal ganglia (standardized beta 0.41; p = 0.02). We conclude that PVS counts on 7T MRI are high and are related SVD markers, but not to age and vascular risk factors. This latter finding may indicate that due to the high sensitivity of 7T MRI, the correlation of PVS counts with age or vascular risk factors may be attenuated by the detection of “normal”, non-pathological PVS.

Abnormalities of Cerebral Deep Medullary Veins on 7 Tesla MRI in Amnestic Mild Cognitive Impairment and Early Alzheimer’s Disease: A Pilot Study

Cerebral small vessel disease (SVD) contributes to cognitive impairment and dementia. SVD may affect veins, but veins are difficult to detect with 1.5 and 3T MRI. We compared deep medullary veins (DMVs) visualized on 7T-MRI between patients with early Alzheimers disease (eAD; n = 17) or amnestic MCI (aMCI; n = 12) and controls (n = 40). The number and density of DMVs was similar in patients and controls, but tortuosity was higher in eAD (Cohens d = 0.7, 95% CI: 0.1-1.2, p = 0.02) and aMCI (Cohens d = 0.8, 95% CI: 0.2-1.5, p = 0.01), independent of brain atrophy. Venous changes provide a new perspective on vascular involvement in dementia.