Automated detection of white matter hyperintensities of all sizes in cerebral small vessel disease

Identifying sources of bias when testing three available algorithms for quantifying white matter lesions: BIANCA, LPA and LGA

Brain magnetic resonance imaging frequently reveals white matter lesions (WMLs) in older adults. They are often associated with cognitive impairment and risk of dementia. Given the continuous search for the optimal segmentation algorithm, we broke down this question by exploring whether the output of algorithms frequently used might be biased by the presence of different influencing factors. We studied the impact of age, sex, blood glucose levels, diabetes, systolic blood pressure and hypertension on automatic WML segmentation algorithms. We evaluated three widely used algorithms (BIANCA, LPA and LGA) using the population-based 1000BRAINS cohort (N = 1166, aged 18-87, 523 females, 643 males). We analysed two main aspects. Firstly, we examined whether training data (TD) characteristics influenced WML estimations, assessing the impact of relevant factors in the TD. Secondly, algorithm's output and performance within selected subgroups defined by these factors were assessed. Results revealed that BIANCA's WML estimations are influenced by the characteristics present in the TD. LPA and LGA consistently provided lower WML estimations compared to BIANCA's output when tested on participants under 67 years of age without risk cardiovascular factors. Notably, LPA and LGA showed reduced accuracy for these participants. However, LPA and LGA showed better performance for older participants presenting cardiovascular risk factors. Results suggest that incorporating comprehensive cohort factors like diverse age, sex and participants with and without hypertension in the TD could enhance WML-based analyses and mitigate potential sources of bias. LPA and LGA are a fast and valid option for older participants with cardiovascular risk factors.

Automatic segmentation of white matter lesions on multi-parametric MRI: convolutional neural network versus vision transformer

BACKGROUND AND PURPOSE

White matter hyperintensities in brain MRI are key indicators of various neurological conditions, and their accurate segmentation is essential for assessing disease progression. This study aims to evaluate the performance of a 3D convolutional neural network and a 3D Transformer-based model for white matter hyperintensities segmentation, focusing on their efficacy with limited datasets and similar computational resources.

MATERIALS AND METHODS

We implemented a convolution-based model (3D ResNet-50 U-Net with spatial and channel squeeze & excitation) and a Transformer-based model (3D Swin Transformer with a convolutional stem). The models were evaluated on two clinical datasets from Kaohsiung Chang Gung Memorial Hospital and National Center for High-Performance Computing. Four metrics were used for evaluation: Dice similarity coefficient, lesion segmentation, lesion F1-Score, and lesion sensitivity.

RESULTS

The Transformer-based model, with appropriate adjustments, outperformed the well-established convolution-based model in foreground Dice similarity coefficient, lesion F1-Score, and sensitivity, demonstrating robust segmentation accuracy. DRLoc enhanced the Transformer's performance, achieving comparable results on internal and benchmark datasets despite limited data availability.

CONCLUSION

With comparable computational overhead, a Transformer-based model can surpass a well-established convolution-based model in white matter hyperintensities segmentation on small datasets by capturing global context effectively, making them suitable for clinical applications where computational resources are constrained.

Visit-to-visit blood pressure variability and progression of white matter hyperintensities over 14 years

Purpose: There is evidence that blood pressure variability (BPV) is associated with cerebral small vessel disease (SVD) and may therefore increase the risk of stroke and dementia. It remains unclear if BPV is associated with SVD progression over years. We examined whether visit-to-visit BPV is associated with white matter hyperintensity (WMH) progression over 14 years and MRI markers after 14 years.

Materials and methods: We included participants with SVD from the Radboud University Nijmegen Diffusion tensor Magnetic resonance-imaging Cohort (RUNDMC) who underwent baseline assessment in 2006 and follow-up in 2011, 2015 and 2020. BPV was calculated as coefficient of variation (CV) of BP at all visits. Association between WMH progression rates over 14 years and BPV was examined using linear-mixed effects (LME) model. Regression models were used to examine association between BPV and MRI markers at final visit in participants.

Results: A total of 199 participants (60.5 SD 6.6 years) who underwent four MRI scans and BP measurements were included, with mean follow-up of 13.7 (SD 0.5) years. Systolic BPV was associated with higher progression of WMH (β = 0.013, 95% CI 0.005 - 0.022) and higher risk of incident lacunes (OR: 1.10, 95% CI 1.01-1.21). There was no association between systolic BPV and grey and white matter volumes, Peak Skeleton of Mean Diffusivity (PSMD) or microbleed count after 13.7 years.

Conclusions: Visit-to-visit systolic BPV is associated with increased progression of WMH volumes and higher risk of incident lacunes over 14 years in participants with SVD. Future studies are needed to examine causality of this association.

Decoding the best automated segmentation tools for vascular white matter hyperintensities in the aging brain: a clinician's guide to precision and purpose

White matter hyperintensities of vascular origin (WMH) are commonly found in individuals over 60 and increase in prevalence with age. The significance of WMH is well-documented, with strong associations with cognitive impairment, risk of stroke, mental health, and brain structure deterioration. Consequently, careful monitoring is crucial for the early identification and management of individuals at risk. Luckily, WMH are detectable and quantifiable on standard MRI through visual assessment scales, but it is time-consuming and has high rater variability. Addressing this issue, the main aim of our study is to decipher the utility of quantitative measures of WMH, assessed with automatic tools, in establishing risk profiles for cerebrovascular deterioration. For this purpose, first, we work to determine the most precise WMH segmentation open access tool compared to clinician manual segmentations (LST-LPA, LST-LGA, SAMSEG, and BIANCA), offering insights into methodology and usability to balance clinical precision with practical application. The results indicated that supervised algorithms (LST-LPA and BIANCA) were superior, particularly in detecting small WMH, and can improve their consistency when used in parallel with unsupervised tools (LST-LGA and SAMSEG). Additionally, to investigate the behavior and real clinical utility of these tools, we tested them in a real-world scenario (N = 300; age > 50 y.o. and MMSE > 26), proposing an imaging biomarker for moderate vascular damage. The results confirmed its capacity to effectively identify individuals at risk comparing the cognitive and brain structural profiles of cognitively healthy adults above and below the resulted threshold.

Baseline and Longitudinal MRI Markers Associated With 16-Year Mortality in Patients With Cerebral Small Vessel Disease

BACKGROUND AND OBJECTIVES

Information on whether small vessel disease (SVD) reduces life expectancy is limited. Moreover, the excess mortality risk attributed specifically to SVD compared with controls from the general population has not been evaluated. This study aimed to investigate the baseline and progression of MRI markers of SVD associated with mortality in a 16-year follow-up cohort study and to determine the excess long-term mortality risk of patients with SVD.

METHODS

Participants with SVD from the Radboud University Nijmegen Diffusion Tensor and Magnetic Resonance Imaging Cohort (RUN DMC) study (with MRI assessments in 2006, 2011, 2015, and 2020) were followed until their death or December 1, 2021. Adjusted Cox regression analyses and linear mixed-effect regression models were used to investigate the association between MRI markers of SVD and mortality. The excess mortality risk of SVD was calculated by comparing mortality data of the RUN DMC study with the general population matched by sex, age, and calendar year.

RESULTS

200 of 503 (39.9%) participants died during a follow-up period of 15.9 years. Cause of death was available for 182 (91%) participants. Baseline white matter hyperintensity volume (HR 1.3 per 1-SD increase [95% CI 1.1-1.5], p = 0.010), presence of lacunes (1.5 [95% CI 1.1-2.0], p = 0.008), mean diffusivity (HR 1.1 per 1-SD increase [95% CI 1.1-1.2], p = 0.001), and total brain volume (HR 1.5 per 1-SD decrease [95% CI 1.3-1.9], p < 0.001) were associated with all-cause mortality after adjusting for age, sex, and vascular risk factors. Total brain volume decrease over time was associated with all-cause mortality after adjusting for age, sex, and vascular risk factors (HR 1.3 per 1-SD decrease [95% CI 1.1-1.7], p = 0.035), and gray matter volume decrease remained significant after additionally adjusting for its baseline volume (1.3 per 1-SD decrease [1.1-1.6], p = 0.019). Participants with a Fazekas score of 3, presence of lacunes, or lower microstructural integrity had an excess long-term mortality risk (21.8, 15.7, 10.1 per 1,000 person-years, respectively) compared with the general population.

DISCUSSION

Excess long-term mortality risk only exists in patients with severe SVD (Fazekas score of 3, presence of lacunes, or lower microstructural integrity). This could help in assisting clinicians to predict the clinical outcomes of patients with SVD by severity.

Decoding the Best Automated Segmentation Tools for Vascular White Matter Hyperintensities in the Aging Brain: A Clinician's Guide to Precision and Purpose

Cerebrovascular damage from small vessel disease (SVD) occurs in healthy and pathological aging. SVD markers, such as white matter hyperintensities (WMH), are commonly found in individuals over 60 and increase in prevalence with age. WMHs are detectable on standard MRI by adhering to the STRIVE criteria. Currently, visual assessment scales are used in clinical and research scenarios but is time-consuming and has rater variability, limiting its practicality. Addressing this issue, our study aimed to determine the most precise WMH segmentation software, offering insights into methodology and usability to balance clinical precision with practical application. This study employed a dataset comprising T1, FLAIR, and DWI images from 300 cognitively healthy older adults. WMHs in this cohort were evaluated using four automated neuroimaging tools: Lesion Prediction Algorithm (LPA) and Lesion Growth Algorithm (LGA) from Lesion Segmentation Tool (LST), Sequence Adaptive Multimodal Segmentation (SAMSEG), and Brain Intensity Abnormalities Classification Algorithm (BIANCA). Additionally, clinicians manually segmented WMHs in a subsample of 45 participants to establish a gold standard. The study assessed correlations with the Fazekas scale, algorithm performance, and the influence of WMH volume on reliability. Results indicated that supervised algorithms were superior, particularly in detecting small WMHs, and can improve their consistency when used in parallel with unsupervised tools. The research also proposed a biomarker for moderate vascular damage, derived from the top 95th percentile of WMH volume in healthy individuals aged 50 to 60. This biomarker effectively differentiated subgroups within the cohort, correlating with variations in brain structure and behavior.

Perivascular Spaces, Diffusivity Along Perivascular Spaces, and Free Water in Cerebral Small Vessel Disease

BACKGROUND AND OBJECTIVES

Previous studies have linked the MRI measures of perivascular spaces (PVSs), diffusivity along the perivascular spaces (DTI-ALPS), and free water (FW) to cerebral small vessel disease (SVD) and SVD-related cognitive impairments. However, studies on the longitudinal associations between the three MRI measures, SVD progression, and cognitive decline are lacking. This study aimed to explore how PVS, DTI-ALPS, and FW contribute to SVD progression and cognitive decline.

METHODS

This is a cohort study that included participants with SVD who underwent neuroimaging and cognitive assessment, specifically measuring Mini-Mental State Examination (MMSE), cognitive index, and processing speed, at 2 time points. Three MRI measures were quantified: PVS in basal ganglia (BG-PVS) volumes, FW fraction, and DTI-ALPS. We performed a latent change score model to test inter-relations between the 3 MRI measures and linear regression mixed models to test their longitudinal associations with the changes of other SVD MRI markers and cognitive performances.

RESULTS

In baseline assessment, we included 289 participants with SVD, characterized by a median age of 67.0 years and 42.9% women. Of which, 220 participants underwent the follow-up assessment, with a median follow-up time of 3.4 years. Baseline DTI-ALPS was associated with changes in BG-PVS volumes (β = -0.09, p = 0.030), but not vice versa (β = -0.08, p = 0.110). Baseline BG-PVS volumes were associated with changes in white matter hyperintensity (WMH) volumes (β = 0.33, p-corrected < 0.001) and lacune numbers (β = 0.28, p-corrected < 0.001); FW fraction was associated with changes in WMH volumes (β = 0.30, p-corrected < 0.001), lacune numbers (β = 0.28, p-corrected < 0.001), and brain volumes (β = -0.45, p-corrected < 0.001); DTI-ALPS was associated with changes in WMH volumes (β = -0.20, p-corrected = 0.002) and brain volumes (β = 0.23, p-corrected < 0.001). Furthermore, baseline FW fraction was associated with decline in MMSE score (β = -0.17, p-corrected = 0.006); baseline FW fraction and DTI-ALPS were associated with changes in cognitive index (FW fraction: β = -0.25, p-corrected < 0.001; DTI-ALPS: β = 0.20, p-corrected = 0.001) and processing speed over time (FW fraction: β = -0.29, p-corrected < 0.001; DTI-ALPS: β = 0.21, p-corrected < 0.001).

DISCUSSION

Our results showed that increased BG-PVS volumes, increased FW fraction, and decreased DTI-ALPS are related to progression of MRI markers of SVD, along with SVD-related cognitive decline over time. These findings may suggest that the glymphatic dysfunction is related to SVD progression, but further studies are needed.

Long-Term Longitudinal Course of Cognitive and Motor Symptoms in Patients With Cerebral Small Vessel Disease

BACKGROUND AND OBJECTIVES

Patients with cerebral small vessel disease (SVD) show a heterogenous clinical course. The aim of the current study was to investigate the longitudinal course of cognitive and motor function in patients who developed parkinsonism, dementia, both, or none.

METHODS

Participants were from the Radboud University Nijmegen Diffusion Tensor and Magnetic Resonance Cohort study, a prospective cohort of patients with SVD. Parkinsonism and dementia were, respectively, diagnosed according to the UK Parkinson's Disease Society brain bank criteria and the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition, criteria for major neurocognitive disorder. Linear and generalized linear mixed-effect analyses were used to study the longitudinal course of motor and cognitive tasks.

RESULTS

After a median follow-up of 12.8 years (interquartile range 10.2-15.3), 132 of 501 (26.3%) participants developed parkinsonism, dementia, or both. Years before diagnosis of these disorders, participants showed distinct clinical trajectories from those who developed none: Participant who developed parkinsonism had an annual percentage of 22% (95% CI 18%-27%) increase in motor part of the Unified Parkinson's Disease Rating Scale score. This was significantly higher than the 16% (95% CI 14%-18%) of controls, mainly because of a steep increase in bradykinesia and posture and gait disturbances. When they developed dementia as well, the increase in Timed Up and Go Test time of 0.73 seconds per year (95% CI 0.58-0.87) was significantly higher than the 0.20 seconds per year increase (95% CI 0.16-0.23) of controls. All groups, including the participants who developed parkinsonism without dementia, showed a faster decline in executive function compared with controls: Annual decline in Z-score was -0.07 (95% CI -0.10 to -0.05), -0.09 (95% CI -0.11 to -0.08), and -0.11 (95% CI -0.14 to -0.08) for participants who developed, respectively, parkinsonism, dementia, and both parkinsonism and dementia. These declines were all significantly faster than the annual decline in Z-score of 0.07 (95% CI -0.10 to -0.05) of controls.

DISCUSSION

A distinct pattern in deterioration of clinical markers is visible in patients with SVD, years before the diagnosis of parkinsonism and dementia. This knowledge aids early identification of patients with a high risk of developing these disorders.

Structural Network Efficiency Predicts Conversion to Incident Parkinsonism in Patients With Cerebral Small Vessel Disease

BACKGROUND

To investigate whether structural network disconnectivity is associated with parkinsonian signs and their progression, as well as with an increased risk of incident parkinsonism.

METHODS

In a prospective cohort (Radboud University Nijmegen Diffusion Tensor and Magnetic Resonance Cohort study) consisting of 293 participants with small vessel disease (SVD), we assessed parkinsonian signs and incident parkinsonism over an 8-year follow-up. In addition, we reconstructed the white matter network followed by graph-theoretical analyses to compute the network metrics. Conventional magnetic resonance imaging markers for SVD were assessed.

RESULTS

We included 293 patients free of parkinsonism at baseline (2011), with a mean age 68.8 (standard deviation [SD] 8.4) years, and 130 (44.4%) were men. Nineteen participants (6.5%) developed parkinsonism during a median (SD) follow-up time of 8.3 years. Compared with participants without parkinsonism, those with all-cause parkinsonism had higher Unified Parkinson's Disease Rating scale (UPDRS) scores and lower global efficiency at baseline. Baseline global efficiency was associated with UPDRS motor scores in 2011 (β = -0.047, p < .001) and 2015 (β = -0.84, p < .001), as well as with the changes in UPDRS scores during the 4-year follow-up (β = -0.63, p = .004). In addition, at the regional level, we identified an inter-hemispheric disconnected network associated with an increased UPDRS motor score. Besides, lower global efficiency was associated with an increased risk of all-cause and vascular parkinsonism independent of SVD markers.

CONCLUSIONS

Our findings suggest that global network efficiency is associated with a gradual decline in motor performance, ultimately leading to incident parkinsonism in the elderly with SVD. Global network efficiency may have the added value to serve as a useful marker to capture changes in motor signs.

Cerebral Small Vessel Disease Progression Increases Risk of Incident Parkinsonism

OBJECTIVE

Cerebral small vessel disease (SVD) is associated with motor impairments and parkinsonian signs cross-sectionally, however, there are little longitudinal data on whether SVD increases risk of incident parkinsonism itself. We investigated the relation between baseline SVD severity as well as SVD progression, and incident parkinsonism over a follow-up of 14 years.

METHODS

This study included 503 participants with SVD, and without parkinsonism at baseline, from the RUN DMC prospective cohort study. Baseline inclusion was performed in 2006 and follow-up took place in 2011, 2015, and 2020, including magnetic resonance imaging (MRI) and motor assessments. Parkinsonism was diagnosed according to the UK Brain Bank criteria, and stratified into vascular parkinsonism (VaP) and idiopathic Parkinson's disease (IPD). Linear mixed-effect models were constructed to estimate individual rate changes of MRI-characteristics.

RESULTS

Follow-up for incident parkinsonism was near-complete (99%). In total, 51 (10.2%) participants developed parkinsonism (33 VaP, 17 IPD, and 1 progressive supranuclear palsy). Patients with incident VaP had higher SVD burden compared with patients with IPD. Higher baseline white matter hyperintensities (hazard ratio [HR] = 1.46 per 1-SD increase, 95% confidence interval [CI] = 1.21-1.78), peak width of skeletonized mean diffusivity (HR = 1.66 per 1-SD increase, 95% CI = 1.34-2.05), and presence of lacunes (HR = 1.84, 95% CI = 0.99-3.42) were associated with increased risk of all-cause parkinsonism. Incident lacunes were associated with incident VaP (HR = 4.64, 95% CI = 1.32-16.32).

INTERPRETATION

Both baseline SVD severity and SVD progression are independently associated with long-term parkinsonism. Our findings indicate a causal role of SVD in parkinsonism. Future studies are needed to examine the underlying pathophysiology of this relation. ANN NEUROL 2023.

Role of small acute hyperintense lesions in long-term progression of cerebral small vessel disease and clinical outcome: a 14-year follow-up study

BACKGROUND

Small hyperintense lesions are found on diffusion-weighted imaging (DWI) in patients with sporadic small vessel disease (SVD). Their exact role in SVD progression remains unclear due to their asymptomatic and transient nature. The main objective is to investigate the role of DWI+lesions in the radiological progression of SVD and their relationship with clinical outcomes.

METHODS

Participants with SVD were included from the Radboud University Nijmegen Diffusion tensor MRI Cohort. DWI+lesions were assessed on four time points over 14 years. Outcome measures included neuroimaging markers of SVD, cognitive performance and clinical outcomes, including stroke, all-cause dementia and all-cause mortality. Linear mixed-effect models and Cox regression models were used to examine the outcome measures in participants with a DWI+lesion (DWI+) and those without a DWI+lesion (DWI-).

RESULTS

DWI+lesions were present in 45 out of 503 (8.9%) participants (mean age: 66.7 years (SD=8.3)). Participants with DWI+lesions and at least one follow-up (n=33) had higher white matter hyperintensity progression rates (β=0.36, 95% CI=0.05 to 0.68, p=0.023), more incident lacunes (incidence rate ratio=2.88, 95% CI=1.80 to 4.67, p<0.001) and greater cognitive decline (β=-0.03, 95% CI=-0.05 to -0.01, p=0.006) during a median follow-up of 13.2 (IQR: 8.8-13.8) years compared with DWI- participants. No differences were found in risk of all-cause mortality, stroke or dementia.

CONCLUSION

Presence of a DWI+lesion in patients with SVD is associated with greater radiological progression of SVD and cognitive decline compared with patients without DWI+lesions.

FLAIR MRI biomarkers of the normal appearing brain matter are related to cognition

•

Normal appearing brain matter (NABM) biomarkers in FLAIR MRI are related to cognition.

•

NABM texture in FLAIR MRI is correlated to mean diffusivity (MD) in dMRI.

•

Analysis conducted on large multicentre FLAIR MRI dataset: 1400 subjects, 87 centers.

•

NABM biomarkers vary differently across age and MoCA categories.

•

Biomarkers showed differences in patients with AD dementia and vascular disease.

A novel biomarker panel was proposed to quantify macro and microstructural biomarkers from the normal-appearing brain matter (NABM) in multicentre fluid-attenuation inversion recovery (FLAIR) MRI. The NABM is composed of the white and gray matter regions of the brain, with the lesions and cerebrospinal fluid removed. The primary hypothesis was that NABM biomarkers from FLAIR MRI are related to cognitive outcome as determined by MoCA score. There were three groups of features designed for this task based on 1) texture: microstructural integrity (MII), macrostructural damage (MAD), microstructural damage (MID), 2) intensity: median, skewness, kurtosis and 3) volume: NABM to ICV volume ratio. Biomarkers were extracted from over 1400 imaging volumes from more than 87 centres and unadjusted ANOVA analysis revealed significant differences in means of the MII, MAD, and NABM volume biomarkers across all cognitive groups. In an adjusted ANCOVA model, a significant relationship between MoCA categories was found that was dependent on subject age for MII, MAD, intensity, kurtosis and NABM volume biomarkers. These results demonstrate that structural brain changes in the NABM are related to cognitive outcome (with different relationships depending on the age of the subjects). Therefore these biomarkers have high potential for clinical translation. As a secondary hypothesis, we investigated whether texture features from FLAIR MRI can quantify microstructural changes related to how “structured” or “damaged” the tissue is. Based on correlation analysis with diffusion weighted MRI (dMRI), it was shown that FLAIR MRI texture biomarkers (MII and MAD) had strong correlations to mean diffusivity (MD) which is related to tissue degeneration in the GM and WM regions. As FLAIR MRI is routinely collected for clinical neurological examinations, novel biomarkers from FLAIR MRI could be used to supplement current clinical biomarkers and for monitoring disease progression. Biomarkers could also be used to stratify patients into homogeneous disease subgroups for clinical trials, or to learn more about mechanistic development of dementia disease.

The Hyperintense study: Assessing the effects of induced blood pressure increase and decrease on MRI markers of cerebral small vessel disease: Study rationale and protocol

Background:

Neuroimaging markers of cerebral small vessel disease (SVD) are common in older individuals, but the pathophysiological mechanisms causing these lesions remain poorly understood. Although hypertension is a major risk factor for SVD, the direct causal effects of increased blood pressure are unknown. The Hyperintense study is designed to examine cerebrovascular and structural abnormalities, possibly preceding SVD, in young adults with hypertension. These patients undergo a diagnostic work-up that requires patients to temporarily discontinue their antihypertensive agents, often leading to an increase in blood pressure followed by a decrease once effective medication is restarted. This allows examination of the effects of blood pressure increase and decrease on the cerebral small vessels.

Methods:

Hyperintense is a prospective observational cohort study in 50 hypertensive adults (18–55 years) who will temporarily discontinue antihypertensive medication for diagnostic purposes. MRI and clinical data is collected at four timepoints: before medication withdrawal (baseline), once antihypertensives are largely or completely withdrawn (T = 1), when patients have restarted medication (T = 2) and reached target blood pressure and 1 year later (T = 3). The 3T MRI protocol includes conventional structural sequences and advanced techniques to assess various aspects of microvascular integrity, including blood-brain barrier function using Dynamic Contrast Enhanced MRI, white matter integrity, and microperfusion. Clinical assessments include motor and cognitive examinations and blood sampling.

Discussion:

The Hyperintense study will improve the understanding of the pathophysiological mechanisms following hypertension that may cause SVD. This knowledge can ultimately help to identify new targets for treatment of SVD, aimed at prevention or limiting disease progression.

Determinants and Temporal Dynamics of Cerebral Small Vessel Disease: 14-Year Follow-Up

BACKGROUND

The aim of this study is to investigate the temporal dynamics of small vessel disease (SVD) and the effect of vascular risk factors and baseline SVD burden on progression of SVD with 4 neuroimaging assessments over 14 years in patients with SVD.

METHODS

Five hundred three patients with sporadic SVD (50-85 years) from the ongoing prospective cohort study (RUN DMC [Radboud University Nijmegen Diffusion Tensor and Magnetic Resonance Cohort]) underwent baseline assessment in 2006 and follow-up in 2011, 2015, and 2020. Vascular risk factors and magnetic resonance imaging markers of SVD were evaluated. Linear mixed-effects model and negative binomial regression model were used to examine the determinants of temporal dynamics of SVD markers.

RESULTS

A total of 382 SVD patients (mean [SD] 64.1 [8.4]; 219 men and 163 women) who underwent at least 2 serial brain magnetic resonance imaging scans were included, with mean (SD) follow-up of 11.15 (3.32) years. We found a highly variable temporal course of SVD. Mean (SD) WMH progression rate was 0.6 (0.74) mL/y (range, 0.02-4.73 mL/y) and 13.6% of patients had incident lacunes (1.03%/y) over the 14-year follow-up. About 4% showed net WMH regression over 14 years, whereas 38 out of 361 (10.5%), 5 out of 296 (2%), and 61 out of 231 (26%) patients showed WMH regression for the intervals 2006 to 2011, 2011 to 2015, and 2015 to 2020, respectively. Of these, 29 (76%), 5 (100%), and 57 (93%) showed overall progression across the 14-year follow-up, and the net overall WMH change between first and last scan considering all participants was a net average WMH progression over the 14-year period. Older age was a strong predictor for faster WMH progression and incident lacunes. Patients with mild baseline WMH rarely progressed to severe WMH. In addition, both baseline burden of SVD lesions and vascular risk factors independently and synergistically predicted WMH progression, whereas only baseline SVD burden predicted incident lacunes over the 14-year follow-up.

CONCLUSIONS

SVD shows pronounced progression over time, but mild WMH rarely progresses to clinically severe WMH. WMH regression is noteworthy during some magnetic resonance imaging intervals, although it could be overall compensated by progression over the long follow-up.

Prediction of dementia using diffusion tensor MRI measures: the OPTIMAL collaboration

OBJECTIVES

It has been suggested that diffusion tensor imaging (DTI) measures sensitive to white matter (WM) damage may predict future dementia risk not only in cerebral small vessel disease (SVD), but also in mild cognitive impairment. To determine whether DTI measures were associated with cognition cross-sectionally and predicted future dementia risk across the full range of SVD severity, we established the International OPtimising mulTImodal MRI markers for use as surrogate markers in trials of Vascular Cognitive Impairment due to cerebrAl small vesseL disease collaboration which included six cohorts.

METHODS

Among the six cohorts, prospective data with dementia incidences were available for three cohorts. The associations between six different DTI measures and cognition or dementia conversion were tested. The additional contribution to prediction of other MRI markers of SVD was also determined.

RESULTS

The DTI measure mean diffusivity (MD) median correlated with cognition in all cohorts, demonstrating the contribution of WM damage to cognition. Adding MD median significantly improved the model fit compared to the clinical risk model alone and further increased in all single-centre SVD cohorts when adding conventional MRI measures. Baseline MD median predicted dementia conversion. In a study with severe SVD (SCANS) change in MD median also predicted dementia conversion. The area under the curve was best when employing a multimodal MRI model using both DTI measures and other MRI measures.

CONCLUSIONS

Our results support a central role for WM alterations in dementia pathogenesis in all cohorts. DTI measures such as MD median may be a useful clinical risk predictor. The contribution of other MRI markers varied according to disease severity.

Comparison of domain adaptation techniques for white matter hyperintensity segmentation in brain MR images

Robust automated segmentation of white matter hyperintensities (WMHs) in different datasets (domains) is highly challenging due to differences in acquisition (scanner, sequence), population (WMH amount and location) and limited availability of manual segmentations to train supervised algorithms. In this work we explore various domain adaptation techniques such as transfer learning and domain adversarial learning methods, including domain adversarial neural networks and domain unlearning, to improve the generalisability of our recently proposed triplanar ensemble network, which is our baseline model. We used datasets with variations in intensity profile, lesion characteristics and acquired using different scanners. For the source domain, we considered a dataset consisting of data acquired from 3 different scanners, while the target domain consisted of 2 datasets. We evaluated the domain adaptation techniques on the target domain datasets, and additionally evaluated the performance on the source domain test dataset for the adversarial techniques. For transfer learning, we also studied various training options such as minimal number of unfrozen layers and subjects required for fine-tuning in the target domain. On comparing the performance of different techniques on the target dataset, domain adversarial training of neural network gave the best performance, making the technique promising for robust WMH segmentation.

Triplanar ensemble U-Net model for white matter hyperintensities segmentation on MR images

White matter hyperintensities (WMHs) have been associated with various cerebrovascular and neurodegenerative diseases. Reliable quantification of WMHs is essential for understanding their clinical impact in normal and pathological populations. Automated segmentation of WMHs is highly challenging due to heterogeneity in WMH characteristics between deep and periventricular white matter, presence of artefacts and differences in the pathology and demographics of populations. In this work, we propose an ensemble triplanar network that combines the predictions from three different planes of brain MR images to provide an accurate WMH segmentation. In the loss functions the network uses anatomical information regarding WMH spatial distribution in loss functions, to improve the efficiency of segmentation and to overcome the contrast variations between deep and periventricular WMHs. We evaluated our method on 5 datasets, of which 3 are part of a publicly available dataset (training data for MICCAI WMH Segmentation Challenge 2017 - MWSC 2017) consisting of subjects from three different cohorts, and we also submitted our method to MWSC 2017 to be evaluated on the unseen test datasets. On evaluating our method separately in deep and periventricular regions, we observed robust and comparable performance in both regions. Our method performed better than most of the existing methods, including FSL BIANCA, and on par with the top ranking deep learning methods of MWSC 2017.

White Matter Hyperintensities Are No Major Confounder for Alzheimer's Disease Cerebrospinal Fluid Biomarkers

Background:

The cerebrospinal fluid (CSF) biomarkers amyloid-β 1–42 (Aβ₄₂), total and phosphorylated tau (t-tau, p-tau) are increasingly used to assist in the clinical diagnosis of Alzheimer’s disease (AD). However, CSF biomarker levels can be affected by confounding factors.

Objective:

To investigate the association of white matter hyperintensities (WMHs) present in the brain with AD CSF biomarker levels.

Methods:

We included CSF biomarker and magnetic resonance imaging (MRI) data of 172 subjects (52 controls, 72 mild cognitive impairment (MCI), and 48 AD patients) from 9 European Memory Clinics. A computer aided detection system for standardized automated segmentation of WMHs was used on MRI scans to determine WMH volumes. Association of WMH volume with AD CSF biomarkers was determined using linear regression analysis.

Results:

A small, negative association of CSF Aβ₄₂, but not p-tau and t-tau, levels with WMH volume was observed in the AD (r² = 0.084, p = 0.046), but not the MCI and control groups, which was slightly increased when including the distance of WMHs to the ventricles in the analysis (r² = 0.105, p = 0.025). Three global patterns of WMH distribution, either with 1) a low, 2) a peak close to the ventricles, or 3) a high, broadly-distributed WMH volume could be observed in brains of subjects in each diagnostic group.

Conclusion:

Despite an association of WMH volume with CSF Aβ₄₂ levels in AD patients, the occurrence of WMHs is not accompanied by excess release of cellular proteins in the CSF, suggesting that WMHs are no major confounder for AD CSF biomarker assessment.

Longitudinal relation between structural network efficiency, cognition, and gait in cerebral small vessel disease

Background

To investigate changes in gait performance over time and how these changes are associated with the decline in structural network efficiency and cognition in older patients with cerebral small vessel disease (SVD).

Methods

In a prospective, single-center cohort with 217 older participants with SVD, we performed 1.5T MRI scans, cognitive tests, and gait assessments evaluated by Timed UP and Go (TUG) test twice over 4 years. We reconstructed the white matter network for each subject based on diffusion tensor imaging tractography, followed by graph-theoretical analyses to compute the global efficiency. Conventional MRI markers for SVD, that is, white matter hyperintensity (WMH) volume, number of lacunes, and microbleeds, were assessed.

Results

Baseline global efficiency was not related to changes in gait performance, while decline in global efficiency over time was significantly associated with gait decline (ie, increase in TUG time), independent of conventional MRI markers for SVD. Neither baseline cognitive performance nor cognitive decline was associated with gait decline.

Conclusions

We found that disruption of the white matter structural network was associated with gait decline over time, while the effect of cognitive decline was not. This suggests that structural network disruption has an important role in explaining the pathophysiology of gait decline in older patients with SVD, independent of cognitive decline.

Cognition mediates the relation between structural network efficiency and gait in small vessel disease

Cerebral small vessel disease (SVD), including white matter hyperintensities (WMH), microbleeds, lacunes, was related to gait disturbances, while the underlying mechanism is unclear. Here, we investigated the relation between structural network efficiency, cognition and gait performance in 272 elderly subjects with SVD. All participants underwent 1.5 T MRI, gait and neuropsychological assessment. Conventional MRI markers for SVD, i.e. WMH volume, number of lacunes and microbleeds, were assessed. Diffusion tensor imaging-based tractography was used to reconstruct the brain network for each individual, followed by graph-theoretical analyses to compute the well-established network measure, global efficiency. We found that lower global efficiency was associated with worse gait performance, including slower gait speed and shorter stride length, independent of conventional MRI markers for SVD. This association was partly mediated via cognitive function. We identified subnetworks of white matter connections associated with gait and cognition, characterized by dominant involvement of frontal tracts. Our findings suggest that network disruption is associated with gait disturbances through cognitive dysfunction in elderly with SVD. Gait is a highly cognitive process and the crucial role of cognition should be considered when investigating gait disturbances in the elderly with SVD.

An anomaly detection approach to identify chronic brain infarcts on MRI

The performance of current machine learning methods to detect heterogeneous pathology is limited by the quantity and quality of pathology in medical images. A possible solution is anomaly detection; an approach that can detect all abnormalities by learning how 'normal' tissue looks like. In this work, we propose an anomaly detection method using a neural network architecture for the detection of chronic brain infarcts on brain MR images. The neural network was trained to learn the visual appearance of normal appearing brains of 697 patients. We evaluated its performance on the detection of chronic brain infarcts in 225 patients, which were previously labeled. Our proposed method detected 374 chronic brain infarcts (68% of the total amount of brain infarcts) which represented 97.5% of the total infarct volume. Additionally, 26 new brain infarcts were identified that were originally missed by the radiologist during radiological reading. Our proposed method also detected white matter hyperintensities, anomalous calcifications, and imaging artefacts. This work shows that anomaly detection is a powerful approach for the detection of multiple brain abnormalities, and can potentially be used to improve the radiological workflow efficiency by guiding radiologists to brain anomalies which otherwise remain unnoticed.

Automatic brain lesion segmentation on standard magnetic resonance images: a scoping review

OBJECTIVES

Medical image analysis practices face challenges that can potentially be addressed with algorithm-based segmentation tools. In this study, we map the field of automatic MR brain lesion segmentation to understand the clinical applicability of prevalent methods and study designs, as well as challenges and limitations in the field.

DESIGN

Scoping review.

SETTING

Three databases (PubMed, IEEE Xplore and Scopus) were searched with tailored queries. Studies were included based on predefined criteria. Emerging themes during consecutive title, abstract, methods and whole-text screening were identified. The full-text analysis focused on materials, preprocessing, performance evaluation and comparison.

RESULTS

Out of 2990 unique articles identified through the search, 441 articles met the eligibility criteria, with an estimated growth rate of 10% per year. We present a general overview and trends in the field with regard to publication sources, segmentation principles used and types of lesions. Algorithms are predominantly evaluated by measuring the agreement of segmentation results with a trusted reference. Few articles describe measures of clinical validity.

CONCLUSIONS

The observed reporting practices leave room for improvement with a view to studying replication, method comparison and clinical applicability. To promote this improvement, we propose a list of recommendations for future studies in the field.

Relation between physical activity and cerebral small vessel disease: A nine-year prospective cohort study

BACKGROUND AND AIMS

Given the unexplored potential of physical activity to reduce the progression of cerebral small vessel disease (cSVD, the purpose of this study was to prospectively (across nine-year follow-up) examine the relation between (baseline) physical activity and the (clinical and imaging) consequences of the whole spectrum of cerebral small vessel disease.

METHODS

Five hundred and three patients with cerebral small vessel disease from the RUNDMC study were followed for nine years. Physical activity was assessed using a questionnaire in 2006, 2011, and 2015. Clinical events (i.e. all-cause mortality, cerebrovascular events (by stroke subtype)) were collected with a structured questionnaire. Patients underwent magnetic resonance imaging scanning for the assessment of magnetic resonance imaging markers of cerebral small vessel disease (i.e. white matter hyperintensities, lacunes, and microbleeds) and microstructural integrity of the white matter at three timepoints.

RESULTS

The mean age at baseline was 66 (SD 9.0) years; 44% were women. A higher baseline physical activity level was independently associated with a lower all-cause mortality (HR: 0.69, 95%CI: 0.49-0.98, p = 0.03) and incidence of cerebrovascular disease (HR: 0.58, 95%CI: 0.36-0.96, p = 0.03). However, we found no relation between physical activity and incident lacunar stroke or progression of magnetic resonance imaging markers of cerebral small vessel disease.

CONCLUSIONS

Whilst regular physical activity was not related to the progression of magnetic resonance imaging markers of cerebral small vessel disease across a nine-year follow-up, results from our study prove that high levels of physical activity in patients with cerebral small vessel disease are associated with a lower all-cause mortality and lower incidence of cerebrovascular events.

Orthostatic hypotension is not associated with small vessel disease progression or cognitive decline

INTRODUCTION

Cerebral hypoperfusion is thought to play an important role in the etiology of cerebral small vessel disease (SVD). Orthostatic hypotension (OH) is assumed to be a cause of cerebral hypoperfusion by causing recurrent hypoperfusion episodes, and might thus be related to progression of SVD. Here, we investigated whether presence of OH is associated with the progression of SVD MRI-markers and cognitive decline over a time period of 9 years in a cohort of sporadic SVD patients.

METHODS

This study included SVD patients from the RUN DMC study, a prospective longitudinal single-center cohort study. In total, 503 patients were included at baseline (2006), from whom 351 participated at first follow-up (2011), and 293 at second follow-up (2015). During all visits, patients underwent MRI and cognitive testing. Association between presence of OH at baseline and progression of SVD-markers on MRI and cognitive decline over time was estimated using linear mixed-effects models.

RESULTS

Of the 503 patients who participated at baseline, 46 patients (9.1%) had OH. Cross-sectional analysis of the baseline data showed that OH was associated with higher white matter hyperintensity (WMH) volume (β = 0.18, p = 0.03), higher mean diffusivity (MD; β = 0.02, p = 0.002), and with presence of microbleeds (OR 2.37 95% CI 1.16-4.68). Longitudinally, OH was however not associated with a progression of total WMH volume (β = -0.17, p = 0.96) or with higher MD (β = -0.001, p = 0.49). There was no association between OH and cognitive performance, both at baseline and over time.

CONCLUSION

In this longitudinal observational study, there was no evidence that presence of OH is associated with progression of SVD-markers or cognitive decline over time. Our findings indicate that OH may not be causally related to SVD progression over time.

White matter hyperintensities at critical crossroads for executive function and verbal abilities in small vessel disease

The presence of white matter lesions in patients with cerebral small vessel disease (SVD) is among the main causes of cognitive decline. We investigated the relation between white matter hyperintensity (WMH) locations and executive and language abilities in 442 SVD patients without dementia with varying burden of WMH. We used Stroop Word Reading, Stroop Color Naming, Stroop Color‐Word Naming, and Category Fluency as language measures with varying degrees of executive demands. The Symbol Digit Modalities Test (SDMT) was used as a control task, as it measures processing speed without requiring language use or verbal output. A voxel‐based lesion–symptom mapping (VLSM) approach was used, corrected for age, sex, education, and lesion volume. VLSM analyses revealed statistically significant clusters for tests requiring language use, but not for SDMT. Worse scores on all tests were associated with WMH in forceps minor, thalamic radiations and caudate nuclei. In conclusion, an association was found between WMH in a core frontostriatal network and executive‐verbal abilities in SVD, independent of lesion volume and processing speed. This circuitry underlying executive‐language functioning might be of potential clinical importance for elderly with SVD. More detailed language testing is required in future research to elucidate the nature of language production difficulties in SVD.

Serum Neurofilament Light Chain Is Associated with Incident Lacunes in Progressive Cerebral Small Vessel Disease

Background and Purpose

Serum neurofilament light (NfL)-chain is a circulating marker for neuroaxonal injury and is also associated with severity of cerebral small vessel disease (SVD) cross-sectionally. Here we explored the association of serum-NfL with imaging and cognitive measures in SVD longitudinally.

Methods

From 503 subjects with SVD, baseline and follow-up magnetic resonance imaging (MRI) was available for 264 participants (follow-up 8.7±0.2 years). Baseline serum-NfL was measured by an ultrasensitive single-molecule-assay. SVD-MRI-markers including white matter hyperintensity (WMH)-volume, mean diffusivity (MD), lacunes, and microbleeds were assessed at both timepoints. Cognitive testing was performed in 336 participants, including SVD-related domains as well as global cognition and memory. Associations with NfL were assessed using linear regression analyses and analysis of covariance (ANCOVA).

Results

Serum-NfL was associated with baseline WMH-volume, MD-values and presence of lacunes and microbleeds. SVD-related MRI- and cognitive measures showed progression during follow-up. NfL-levels were associated with future MRI-markers of SVD, including WMH, MD and lacunes. For the latter, this association was independent of baseline lacunes. Furthermore, NfL was associated with incident lacunes during follow-up (P=0.040). NfL-levels were associated with future SVD-related cognitive impairment (processing speed: β=–0.159; 95% confidence interval [CI], –0.242 to –0.068; P=0.001; executive function β=–0.095; 95% CI, –0.170 to –0.007; P=0.033), adjusted for age, sex, education, and depression. Dementia-risk increased with higher NfL-levels (hazard ratio, 5.0; 95% CI, 2.6 to 9.4; P<0.001), however not after adjusting for age.

Conclusions

Longitudinally, serum-NfL is associated with markers of SVD, especially with incident lacunes, and future cognitive impairment affecting various domains. NfL may potentially serve as an additional marker for disease monitoring and outcome in SVD, potentially capturing both vascular and neurodegenerative processes in the elderly.

Structural network changes in cerebral small vessel disease

OBJECTIVES

To investigate whether longitudinal structural network efficiency is associated with cognitive decline and whether baseline network efficiency predicts mortality in cerebral small vessel disease (SVD).

METHODS

A prospective, single-centre cohort consisting of 277 non-demented individuals with SVD was conducted. In 2011 and 2015, all participants were scanned with MRI and underwent neuropsychological assessment. We computed network properties using graph theory from probabilistic tractography and calculated changes in psychomotor speed and overall cognitive index. Multiple linear regressions were performed, while adjusting for potential confounders. We divided the group into mild-to-moderate white matter hyperintensities (WMH) and severe WMH group based on median split on WMH volume.

RESULTS

The decline in global efficiency was significantly associated with a decline in psychomotor speed in the group with severe WMH (β=0.18, p=0.03) and a trend with change in cognitive index (β=0.14, p=0.068), which diminished after adjusting for imaging markers for SVD. Baseline global efficiency was associated with all-cause mortality (HR per decrease of 1 SD 0.43, 95% CI 0.23 to 0.80, p=0.008, C-statistic 0.76).

CONCLUSION

Disruption of the network efficiency, a metric assessing the efficiency of network information transfer, plays an important role in explaining cognitive decline in SVD, which was however not independent of imaging markers of SVD. Furthermore, baseline network efficiency predicts risk of mortality in SVD that may reflect the global health status of the brain in SVD. This emphasises the importance of structural network analysis in the context of SVD research and the use of network measures as surrogate markers in research setting.

Performance of five automated white matter hyperintensity segmentation methods in a multicenter dataset

White matter hyperintensities (WMHs) are a common manifestation of cerebral small vessel disease, that is increasingly studied with large, pooled multicenter datasets. This data pooling increases statistical power, but poses challenges for automated WMH segmentation. Although there is extensive literature on the evaluation of automated WMH segmentation methods, such evaluations in a multicenter setting are lacking. We performed WMH segmentations in sixty patients scanned on six different magnetic resonance imaging (MRI) scanners (10 patients per scanner) using five freely available and fully-automated WMH segmentation methods (Cascade, kNN-TTP, Lesion-TOADS, LST-LGA and LST-LPA). Different MRI scanner vendors and field strengths were included. We compared these automated WMH segmentations with manual WMH segmentations as a reference. Performance of each method both within and across scanners was assessed using spatial and volumetric correspondence with the reference segmentations by Dice's similarity coefficient (DSC) and intra-class correlation coefficient (ICC) respectively. We found the best performance, both within and across scanners, for kNN-TTP, followed by LST-LPA and LST-LGA, with worse performance for Lesion-TOADS and Cascade. Our findings can serve as a guide for choosing a method and also highlight the importance to further improve and evaluate consistency of methods in a multicenter setting.

Standardized Assessment of Automatic Segmentation of White Matter Hyperintensities and Results of the WMH Segmentation Challenge

Quantification of cerebral white matter hyperintensities (WMH) of presumed vascular origin is of key importance in many neurological research studies. Currently, measurements are often still obtained from manual segmentations on brain MR images, which is a laborious procedure. The automatic WMH segmentation methods exist, but a standardized comparison of the performance of such methods is lacking. We organized a scientific challenge, in which developers could evaluate their methods on a standardized multi-center/-scanner image dataset, giving an objective comparison: the WMH Segmentation Challenge. Sixty T1 + FLAIR images from three MR scanners were released with the manual WMH segmentations for training. A test set of 110 images from five MR scanners was used for evaluation. The segmentation methods had to be containerized and submitted to the challenge organizers. Five evaluation metrics were used to rank the methods: 1) Dice similarity coefficient; 2) modified Hausdorff distance (95th percentile); 3) absolute log-transformed volume difference; 4) sensitivity for detecting individual lesions; and 5) F1-score for individual lesions. In addition, the methods were ranked on their inter-scanner robustness; 20 participants submitted their methods for evaluation. This paper provides a detailed analysis of the results. In brief, there is a cluster of four methods that rank significantly better than the other methods, with one clear winner. The inter-scanner robustness ranking shows that not all the methods generalize to unseen scanners. The challenge remains open for future submissions and provides a public platform for method evaluation.

Longitudinal changes in rich club organization and cognition in cerebral small vessel disease

Cerebral small vessel disease (SVD) is considered the most important vascular contributor to the development of cognitive impairment and dementia. There is increasing awareness that SVD exerts its clinical effects by disrupting white matter connections, predominantly disrupting connections between rich club nodes, a set of highly connected and interconnected regions. Here we examined the progression of disturbances in rich club organization in older adults with SVD and their associations with conventional SVD markers and cognitive decline. We additionally investigated associations of baseline network measures with dementia. In 270 participants of the RUN DMC study, we performed diffusion tensor imaging (DTI) and cognitive assessments longitudinally. Rich club organization was examined in structural networks derived from DTI followed by deterministic tractography. Global efficiency (p<0.05) and strength of rich club connections (p<0.001) declined during follow-up. Decline in strength of peripheral connections was associated with a decline in overall cognition (β=0.164; p<0.01), psychomotor speed (β=0.151; p<0.05) and executive function (β=0.117; p<0.05). Baseline network measures were reduced in participants with dementia, and the association between WMH and dementia was causally mediated by global efficiency (p = =0.037) and peripheral connection strength (p = =0.040). SVD-related disturbances in rich club organization progressed over time, predominantly in participants with severe SVD. In this study, we found no specific role of rich club connectivity disruption in causing cognitive decline or dementia. The effect of WMH on dementia was mediated by global network efficiency and the strength of peripheral connections, suggesting an important role for network disruption in causing cognitive decline and dementia in older adults with SVD.

Use of Flutemetamol F 18-Labeled Positron Emission Tomography and Other Biomarkers to Assess Risk of Clinical Progression in Patients With Amnestic Mild Cognitive Impairment

Importance

Patients with amnestic mild cognitive impairment (aMCI) may progress to clinical Alzheimer disease (AD), remain stable, or revert to normal. Earlier progression to AD among patients who were β-amyloid positive vs those who were β-amyloid negative has been previously observed. Current research now accepts that a combination of biomarkers could provide greater refinement in the assessment of risk for clinical progression.

Objective

To evaluate the ability of flutemetamol F 18 and other biomarkers to assess the risk of progression from aMCI to probable AD.

Design, Setting, and Participants

In this multicenter cohort study, from November 11, 2009, to January 16, 2014, patients with aMCI underwent positron emission tomography (PET) at baseline followed by local clinical assessments every 6 months for up to 3 years. Patients with aMCI (365 screened; 232 were eligible) were recruited from 28 clinical centers in Europe and the United States. Physicians remained strictly blinded to the results of PET, and the standard of truth was an independent clinical adjudication committee that confirmed or refuted local assessments. Flutemetamol F 18-labeled PET scans were read centrally as either negative or positive by 5 blinded readers with no knowledge of clinical status. Statistical analysis was conducted from February 19, 2014, to January 26, 2018.

Interventions

Flutemetamol F 18-labeled PET at baseline followed by up to 6 clinical visits every 6 months, as well as magnetic resonance imaging and multiple cognitive measures.

Main Outcomes and Measures

Time from PET to probable AD or last follow-up was plotted as a Kaplan-Meier survival curve; PET scan results, age, hippocampal volume, and aMCI stage were entered into Cox proportional hazards logistic regression analyses to identify variables associated with progression to probable AD.

Results

Of 232 patients with aMCI (118 women and 114 men; mean [SD] age, 71.1 [8.6] years), 98 (42.2%) had positive results detected on PET scan. By 36 months, the rates of progression to probable AD were 36.2% overall (81 of 224 patients), 53.6% (52 of 97) for patients with positive results detected on PET scan, and 22.8% (29 of 127) for patients with negative results detected on PET scan. Hazard ratios for association with progression were 2.51 (95% CI, 1.57-3.99; P < .001) for a positive β-amyloid scan alone (primary outcome measure), 5.60 (95% CI, 3.14-9.98; P < .001) with additional low hippocampal volume, and 8.45 (95% CI, 4.40-16.24; P < .001) when poorer cognitive status was added to the model.

Conclusions and Relevance

A combination of positive results of flutemetamol F 18-labeled PET, low hippocampal volume, and cognitive status corresponded with a high probability of risk of progression from aMCI to probable AD within 36 months.

Risk of Nursing Home Admission in Cerebral Small Vessel Disease

Background and Purpose- Since cerebral small vessel disease (SVD) is associated with cognitive and motor impairment and both might ultimately lead to nursing home admission, our objective was to investigate the association of SVD markers with nursing home admission. Methods- The RUN DMC study (Radboud University Nijmegen Diffusion Tensor and Magnetic Resonance Cohort) is a prospective cohort of 503 independent living individuals with SVD. Date of nursing home admission was retrieved from the Dutch municipal personal records database. Risk of nursing home admission was calculated using a competing risk analysis, with mortality as a competing risk. Results- During follow-up (median 8.7 years, interquartile range 8.5-8.9), 31 participants moved to a nursing home. Before nursing home admission, 19 participants were diagnosed with dementia, 6 with parkinsonism, and 10 with stroke. Participants with the lowest white matter volume had an 8-year risk of nursing home admission of 13.3% (95% CI, 8.6-18.9), which was significantly different from participants with middle or highest white matter volume (respectively, 4.8% [95% CI, 2.3-8.8] and 0%; P<0.001). After adjusting for baseline age and living condition, the association of white matter volume and total brain volume with nursing home admission was significant, with, respectively, hazard ratios of 0.88 [95% CI, 0.84-0.95] ( P value 0.025) and 0.92 [95% CI, 0.85-0.98] ( P<0.001) per 10 mL. The association of white matter hyperintensities and lacunes with nursing home admission was not significant. Conclusions- This study demonstrates that in SVD patients, independent from age and living condition, a lower white matter volume and a lower total brain volume is associated with an increased risk of nursing home admission. Nursing home admission is a relevant outcome in SVD research since it might be able to combine both cognitive and functional consequences of SVD in 1 outcome.

The role of small diffusion-weighted imaging lesions in cerebral small vessel disease

OBJECTIVE

To investigate the prevalence of asymptomatic diffusion-weighted imaging-positive (DWI+) lesions in individuals with cerebral small vessel disease (SVD) and identify their role in the origin of SVD markers on MRI.

METHODS

We included 503 individuals with SVD from the Radboud University Nijmegen Diffusion Tensor and Magnetic Resonance Imaging Cohort (RUN DMC) study (mean age 65.6 years [SD 8.8], 56.5% male) with 1.5T MRI in 2006 and, if available, follow-up MRI in 2011 and 2015. We screened DWI scans (n = 1,152) for DWI+ lesions, assessed lesion evolution on follow-up fluid-attenuated inversion recovery, T1 and T2* images, and examined the association between DWI+ lesions and annual SVD progression (white matter hyperintensities [WMH], lacunes, microbleeds).

RESULTS

We found 50 DWI+ lesions in 39 individuals on 1,152 DWI (3.4%). Individuals with DWI+ lesions were older (p = 0.025), more frequently had a history of hypertension (p = 0.021), and had a larger burden of preexisting SVD MRI markers (WMH, lacunes, microbleeds: all p < 0.001) compared to individuals without DWI+ lesions. Of the 23 DWI+ lesions with available follow-up MRI, 14 (61%) evolved into a WMH, 8 (35%) resulted in a cavity, and 1 (4%) was no longer visible. Presence of DWI+ lesions was significantly associated with annual WMH volume increase and yearly incidence of lacunes and microbleeds (all p < 0.001).

CONCLUSION

Over 3% of individuals with SVD have DWI+ lesions. Although DWI+ lesions play a role in the progression of SVD, they may not fully explain progression of SVD markers on MRI, suggesting that other factors than acute ischemia are at play.

Adiposity is related to cerebrovascular and brain volumetry outcomes in the RUN DMC study

OBJECTIVE

Adiposity predictors, body mass index (BMI), waist circumference (WC), and blood leptin and total adiponectin levels were associated with components of cerebral small vessel disease (CSVD) and brain volumetry in 503 adults with CSVD who were ≥50 years of age and enrolled in the Radboud University Nijmegen Diffusion Tensor and Magnetic Resonance Imaging Cohort (RUN DMC).

METHODS

RUN DMC participants were followed up for 9 years (2006-2015). BMI, WC, brain imaging, and dementia diagnoses were evaluated at baseline and follow-up. Adipokines were measured at baseline. Brain imaging outcomes included CSVD components, white matter hyperintensities, lacunes, microbleeds, gray and white matter, hippocampal, total brain, and intracranial volumes.

RESULTS

Cross-sectionally among men at baseline, higher BMI, WC, and leptin were associated with lower gray matter and total brain volumes, and higher BMI and WC were associated with lower hippocampal volume. At follow-up 9 years later, higher BMI was cross-sectionally associated with lower gray matter volume, and an obese WC (>102 cm) was protective for ≥1 lacune or ≥1 microbleed in men. In women, increasing BMI and overweight or obesity (BMI ≥25 kg/m² or WC >88 cm) were associated with ≥1 lacune. Longitudinally, over 9 years, a baseline obese WC was associated with decreasing hippocampal volume, particularly in men, and increasing white matter hyperintensity volume in women and men.

CONCLUSIONS

Anthropometric and metabolic adiposity predictors were differentially associated with CSVD components and brain volumetry outcomes by sex. Higher adiposity is associated with a vascular-neurodegenerative spectrum among adults at risk for vascular forms of cognitive impairment and dementias.

Evaluation of Enhanced Learning Techniques for Segmenting Ischaemic Stroke Lesions in Brain Magnetic Resonance Perfusion Images Using a Convolutional Neural Network Scheme

Magnetic resonance (MR) perfusion imaging non-invasively measures cerebral perfusion, which describes the blood's passage through the brain's vascular network. Therefore, it is widely used to assess cerebral ischaemia. Convolutional Neural Networks (CNN) constitute the state-of-the-art method in automatic pattern recognition and hence, in segmentation tasks. But none of the CNN architectures developed to date have achieved high accuracy when segmenting ischaemic stroke lesions, being the main reasons their heterogeneity in location, shape, size, image intensity and texture, especially in this imaging modality. We use a freely available CNN framework, developed for MR imaging lesion segmentation, as core algorithm to evaluate the impact of enhanced machine learning techniques, namely data augmentation, transfer learning and post-processing, in the segmentation of stroke lesions using the ISLES 2017 dataset, which contains expert annotated diffusion-weighted perfusion and diffusion brain MRI of 43 stroke patients. Of all the techniques evaluated, data augmentation with binary closing achieved the best results, improving the mean Dice score in 17% over the baseline model. Consistent with previous works, better performance was obtained in the presence of large lesions.

Apathy is associated with large-scale white matter network disruption in small vessel disease

OBJECTIVE

To investigate whether white matter network disruption underlies the pathogenesis of apathy, but not depression, in cerebral small vessel disease (SVD).

METHODS

Three hundred thirty-one patients with SVD from the Radboud University Nijmegen Diffusion Tensor and Magnetic Resonance Cohort (RUN DMC) study completed measures of apathy and depression and underwent structural MRI. Streamlines reflecting underlying white matter fibers were reconstructed with diffusion tensor tractography. First, path analysis was used to determine whether network measures mediated associations between apathy and radiologic markers of SVD. Next, we examined differences in whole-brain network measures between participants with only apathy, only depression, and comorbid apathy and depression and a control group free of neuropsychiatric symptoms. Finally, we examined regional network differences associated with apathy.

RESULTS

Path analysis demonstrated that network disruption mediated the relationship between apathy and SVD markers. Patients with apathy, compared to all other groups, were impaired on whole-brain measures of network density and efficiency. Regional network analyses in both the apathy subgroup and the entire sample revealed that apathy was associated with impaired connectivity in premotor and cingulate regions.

CONCLUSIONS

Our results suggest that apathy, but not depression, is associated with white matter tract disconnection in SVD. The subnetworks delineated suggest that apathy may be driven by damage to white matter networks underlying action initiation and effort-based decision making.

Detection of white matter lesion regions in MRI using SLIC0 and convolutional neural network

BACKGROUND AND OBJECTIVE

White matter lesions are non-static brain lesions that have a prevalence rate up to 98% in the elderly population. Because they may be associated with several brain diseases, it is important that they are detected as soon as possible. Magnetic Resonance Imaging (MRI) provides three-dimensional data with the possibility to detect and emphasize contrast differences in soft tissues, providing rich information about the human soft tissue anatomy. However, the amount of data provided for these images is far too much for manual analysis/interpretation, representing a difficult and time-consuming task for specialists. This work presents a computational methodology capable of detecting regions of white matter lesions of the brain in MRI of FLAIR modality. The techniques highlighted in this methodology are SLIC0 clustering for candidate segmentation and convolutional neural networks for candidate classification.

METHODS

The methodology proposed here consists of four steps: (1) images acquisition, (2) images preprocessing, (3) candidates segmentation and (4) candidates classification.

RESULTS

The methodology was applied on 91 magnetic resonance images provided by DASA, and achieved an accuracy of 98.73%, specificity of 98.77% and sensitivity of 78.79% with 0.005 of false positives, without any false positives reduction technique, in detection of white matter lesion regions.

CONCLUSIONS

It is demonstrated the feasibility of the analysis of brain MRI using SLIC0 and convolutional neural network techniques to achieve success in detection of white matter lesions regions.

Memory decline in elderly with cerebral small vessel disease explained by temporal interactions between white matter hyperintensities and hippocampal atrophy

White matter hyperintensities (WMH) constitute the visible spectrum of cerebral small vessel disease (SVD) markers and are associated with cognitive decline, although they do not fully account for memory decline observed in individuals with SVD. We hypothesize that WMH might exert their effect on memory decline indirectly by affecting remote brain structures such as the hippocampus. We investigated the temporal interactions between WMH, hippocampal atrophy and memory decline in older adults with SVD. Five hundred and three participants of the RUNDMC study underwent neuroimaging and cognitive assessments up to 3 times over 8.7 years. We assessed WMH volumes semi-automatically and calculated hippocampal volumes (HV) using FreeSurfer. We used linear mixed effects models and causal mediation analyses to assess both interaction and mediation effects of hippocampal atrophy in the associations between WMH and memory decline, separately for working memory (WM) and episodic memory (EM). Linear mixed effect models revealed that the interaction between WMH and hippocampal volumes explained memory decline (WM: β = .067; 95%CI[.024-0.111]; p < .01; EM: β = .061; 95%CI[.025-.098]; p < .01), with better model fit when the WMH*HV interaction term was added to the model, for both WM (likelihood ratio test, χ² [1] = 9.3, p < .01) and for EM (likelihood ratio test, χ² [1] = 10.7, p < .01). Mediation models showed that both baseline WMH volume (β = -.170; p = .001) and hippocampal atrophy (β = 0.126; p = .009) were independently related to EM decline, but the effect of baseline WMH on EM decline was not mediated by hippocampal atrophy (p value indirect effect: 0.572). Memory decline in elderly with SVD was best explained by the interaction of WMH and hippocampal volumes. The relationship between WMH and memory was not causally mediated by hippocampal atrophy, suggesting that memory decline during aging is a heterogeneous condition in which different pathologies contribute to the memory decline observed in elderly with SVD.

Brain atrophy and strategic lesion location increases risk of parkinsonism in cerebral small vessel disease

INTRODUCTION

Incident parkinsonism in patients with comparable cerebral small vessel disease (SVD) burden is not fully explained by presence of SVD alone. We therefore investigated if severity of SVD, SVD location, incidence of SVD and/or brain atrophy plays a role in this distinct development of parkinsonism.

METHODS

Participants were from the RUN DMC study, a prospective cohort of 503 individuals with SVD. Parkinsonism was diagnosed according to the UKPDS brain bank criteria. Fine and Gray method was used to assess the association between SVD and incident parkinsonism. Differences in white matter hyperintensities (WMH) progression and brain atrophy were calculated with a linear mixed effect analysis.

RESULTS

After a median follow-up of 8.6 years, 32 of 501 participants developed parkinsonism (6.4%). The highest WMH load was found in the frontal lobe for both groups. Presence of more than one lacune at baseline was higher in the group who developed parkinsonism, especially in the frontal lobe (22% versus 3%, p < 0.001) and basal ganglia (12.5% versus 1%, p-value <0.001). The annual rate of total brain atrophy was significantly higher for those who developed parkinsonism compared to those who did not (8.7 ml [95%CI 7.1-10.3] and 4.9 ml [95%CI 4.5-5.3], respectively). While WMH progression was not different, incidence of lacunes and microbleeds was higher in the group with parkinsonism.

CONCLUSION

The risk of parkinsonism in patients with SVD is especially increased when WMH and lacunes are present in the frontal lobe. A higher brain atrophy rate might further increase this risk.

Progression of White Matter Hyperintensities Preceded by Heterogeneous Decline of Microstructural Integrity

BACKGROUND AND PURPOSE

White matter hyperintensities (WMH) are frequently seen on neuroimaging of elderly and are associated with cognitive decline and the development of dementia. Yet, the temporal dynamics of conversion of normal-appearing white matter (NAWM) into WMH remains unknown. We examined whether and when progression of WMH was preceded by changes in fluid-attenuated inversion recovery and diffusion tensor imaging values, thereby taking into account differences between participants with mild versus severe baseline WMH.

METHODS

From 266 participants of the RUN DMC study (Radboud University Nijmegen Diffusion Tensor and Magnetic Resonance Imaging Cohort), we semiautomatically segmented WMH at 3 time points for 9 years. Images were registered to standard space through a subject template. We analyzed differences in baseline fluid-attenuated inversion recovery, fractional anisotropy, and mean diffusivity (MD) values and changes in MD values over time between 4 regions: (1) remaining NAWM, (2) NAWM converting into WMH in the second follow-up period, (3) NAWM converting into WMH in the first follow-up period, and (4) WMH.

RESULTS

NAWM converting into WMH in the first or second time interval showed higher fluid-attenuated inversion recovery and MD values than remaining NAWM. MD values in NAWM converting into WMH in the first time interval were similar to MD values in WMH. When stratified by baseline WMH severity, participants with severe WMH had higher fluid-attenuated inversion recovery and MD and lower fractional anisotropy values than participants with mild WMH, in all areas including the NAWM. MD values in WMH and in NAWM that converted into WMH continuously increased over time.

CONCLUSIONS

Impaired microstructural integrity preceded conversion into WMH and continuously declined over time, suggesting a continuous disease process of white matter integrity loss that can be detected using diffusion tensor imaging even years before WMH become visible on conventional neuroimaging. Differences in microstructural integrity between participants with mild versus severe WMH suggest heterogeneity of both NAWM and WMH, which might explain the clinical variability observed in patients with similar small vessel disease severity.

Plasma Aβ (Amyloid-β) Levels and Severity and Progression of Small Vessel Disease

BACKGROUND AND PURPOSE

Cerebral small vessel disease (SVD) is a frequent pathology in aging and contributor to the development of dementia. Plasma Aβ (amyloid β) levels may be useful as early biomarker, but the role of plasma Aβ in SVD remains to be elucidated. We investigated the association of plasma Aβ levels with severity and progression of SVD markers.

METHODS

We studied 487 participants from the RUN DMC study (Radboud University Nijmegen Diffusion Tensor and Magnetic Resonance Imaging Cohort) of whom 258 participants underwent 3 MRI assessments during 9 years. We determined baseline plasma Aβ38, Aβ40, and Aβ42 levels using ELISAs. We longitudinally assessed volume of white matter hyperintensities semiautomatically and manually rated lacunes and microbleeds. We analyzed associations between plasma Aβ and SVD markers by ANCOVA adjusted for age, sex, and hypertension.

RESULTS

Cross-sectionally, plasma Aβ40 levels were elevated in participants with microbleeds (mean, 205.4 versus 186.4 pg/mL; P<0.01) and lacunes (mean, 194.8 versus 181.2 pg/mL; P<0.05). Both Aβ38 and Aβ40 were elevated in participants with severe white matter hyperintensities (Aβ38, 25.3 versus 22.7 pg/mL; P<0.01; Aβ40, 201.8 versus 183.3 pg/mL; P<0.05). Longitudinally, plasma Aβ40 levels were elevated in participants with white matter hyperintensity progression (mean, 194.6 versus 182.9 pg/mL; P<0.05). Both Aβ38 and Aβ40 were elevated in participants with incident lacunes (Aβ38, 24.5 versus 22.5 pg/mL; P<0.05; Aβ40, 194.9 versus 181.2 pg/mL; P<0.01) and Aβ42 in participants with incident microbleeds (62.8 versus 60.4 pg/mL; P<0.05).

CONCLUSIONS

Plasma Aβ levels are associated with both presence and progression of SVD markers, suggesting that Aβ pathology might contribute to the development and progression of SVD. Plasma Aβ levels might thereby serve as inexpensive and noninvasive measure for identifying individuals with increased risk for progression of SVD.

DEWS (DEep White matter hyperintensity Segmentation framework): A fully automated pipeline for detecting small deep white matter hyperintensities in migraineurs

Migraineurs show an increased load of white matter hyperintensities (WMHs) and more rapid deep WMH progression. Previous methods for WMH segmentation have limited efficacy to detect small deep WMHs. We developed a new fully automated detection pipeline, DEWS (DEep White matter hyperintensity Segmentation framework), for small and superficially-located deep WMHs. A total of 148 non-elderly subjects with migraine were included in this study. The pipeline consists of three components: 1) white matter (WM) extraction, 2) WMH detection, and 3) false positive reduction. In WM extraction, we adjusted the WM mask to re-assign misclassified WMHs back to WM using many sequential low-level image processing steps. In WMH detection, the potential WMH clusters were detected using an intensity based threshold and region growing approach. For false positive reduction, the detected WMH clusters were classified into final WMHs and non-WMHs using the random forest (RF) classifier. Size, texture, and multi-scale deep features were used to train the RF classifier. DEWS successfully detected small deep WMHs with a high positive predictive value (PPV) of 0.98 and true positive rate (TPR) of 0.70 in the training and test sets. Similar performance of PPV (0.96) and TPR (0.68) was attained in the validation set. DEWS showed a superior performance in comparison with other methods. Our proposed pipeline is freely available online to help the research community in quantifying deep WMHs in non-elderly adults.

White matter changes and gait decline in cerebral small vessel disease

The relation between progression of cerebral small vessel disease (SVD) and gait decline is uncertain, and diffusion tensor imaging (DTI) studies on gait decline are lacking. We therefore investigated the longitudinal associations between (micro) structural brain changes and gait decline in SVD using DTI. 275 participants were included from the Radboud University Nijmegen Diffusion tensor and Magnetic resonance imaging Cohort (RUN DMC), a prospective cohort of participants with cerebral small vessel disease aged 50-85 years. Gait (using GAITRite) and magnetic resonance imaging measures were assessed during baseline (2006-2007) and follow-up (2011 - 2012). Linear regression analysis was used to investigate the association between changes in conventional magnetic resonance and diffusion tensor imaging measures and gait decline. Tract-based spatial statistics analysis was used to investigate region-specific associations between changes in white matter integrity and gait decline. 56.2% were male, mean age was 62.9 years (SD8.2), mean follow-up duration was 5.4 years (SD0.2) and mean gait speed decline was 0.2 m/s (SD0.2). Stride length decline was associated with white matter atrophy (β = 0.16, p = 0.007), and increase in mean white matter radial diffusivity and mean diffusivity, and decrease in mean fractional anisotropy (respectively, β = - 0.14, p = 0.009; β = - 0.12, p = 0.018; β = 0.10, p = 0.049), independent of age, sex, height, follow-up duration and baseline stride length. Tract-based spatial statistics analysis showed significant associations between stride length decline and fractional anisotropy decrease and mean diffusivity increase (primarily explained by radial diffusivity increase) in multiple white matter tracts, with the strongest associations found in the corpus callosum and corona radiata, independent of traditional small vessel disease markers. White matter atrophy and loss of white matter integrity are associated with gait decline in older adults with small vessel disease after 5 years of follow-up. These findings suggest that progression of SVD might play an important role in gait decline.

Evaluation of a deep learning approach for the segmentation of brain tissues and white matter hyperintensities of presumed vascular origin in MRI

Automatic segmentation of brain tissues and white matter hyperintensities of presumed vascular origin (WMH) in MRI of older patients is widely described in the literature. Although brain abnormalities and motion artefacts are common in this age group, most segmentation methods are not evaluated in a setting that includes these items. In the present study, our tissue segmentation method for brain MRI was extended and evaluated for additional WMH segmentation. Furthermore, our method was evaluated in two large cohorts with a realistic variation in brain abnormalities and motion artefacts. The method uses a multi-scale convolutional neural network with a T1-weighted image, a T2-weighted fluid attenuated inversion recovery (FLAIR) image and a T1-weighted inversion recovery (IR) image as input. The method automatically segments white matter (WM), cortical grey matter (cGM), basal ganglia and thalami (BGT), cerebellum (CB), brain stem (BS), lateral ventricular cerebrospinal fluid (lvCSF), peripheral cerebrospinal fluid (pCSF), and WMH. Our method was evaluated quantitatively with images publicly available from the MRBrainS13 challenge (n = 20), quantitatively and qualitatively in relatively healthy older subjects (n = 96), and qualitatively in patients from a memory clinic (n = 110). The method can accurately segment WMH (Overall Dice coefficient in the MRBrainS13 data of 0.67) without compromising performance for tissue segmentations (Overall Dice coefficients in the MRBrainS13 data of 0.87 for WM, 0.85 for cGM, 0.82 for BGT, 0.93 for CB, 0.92 for BS, 0.93 for lvCSF, 0.76 for pCSF). Furthermore, the automatic WMH volumes showed a high correlation with manual WMH volumes (Spearman's ρ = 0.83 for relatively healthy older subjects). In both cohorts, our method produced reliable segmentations (as determined by a human observer) in most images (relatively healthy/memory clinic: tissues 88%/77% reliable, WMH 85%/84% reliable) despite various degrees of brain abnormalities and motion artefacts. In conclusion, this study shows that a convolutional neural network-based segmentation method can accurately segment brain tissues and WMH in MR images of older patients with varying degrees of brain abnormalities and motion artefacts.

Nonlinear temporal dynamics of cerebral small vessel disease: The RUN DMC study

OBJECTIVE

To investigate the temporal dynamics of cerebral small vessel disease (SVD) by 3 consecutive assessments over a period of 9 years, distinguishing progression from regression.

METHODS

Changes in SVD markers of 276 participants of the Radboud University Nijmegen Diffusion Tensor and Magnetic Resonance Imaging Cohort (RUN DMC) cohort were assessed at 3 time points over 9 years. We assessed white matter hyperintensities (WMH) volume by semiautomatic segmentation and rated lacunes and microbleeds manually. We categorized baseline WMH severity as mild, moderate, or severe according to the modified Fazekas scale. We performed mixed-effects regression analysis including a quadratic term for increasing age.

RESULTS

Mean WMH progression over 9 years was 4.7 mL (0.54 mL/y; interquartile range 0.95-5.5 mL), 20.3% of patients had incident lacunes (2.3%/y), and 18.9% had incident microbleeds (2.2%/y). WMH volume declined in 9.4% of the participants during the first follow-up interval, but only for 1 participant (0.4%) throughout the whole follow-up. Lacunes disappeared in 3.6% and microbleeds in 5.7% of the participants. WMH progression accelerated over time: including a quadratic term for increasing age during follow-up significantly improved the model (p < 0.001). SVD progression was predominantly seen in participants with moderate to severe WMH at baseline compared to those with mild WMH (odds ratio [OR] 35.5, 95% confidence interval [CI] 15.8-80.0, p < 0.001 for WMH progression; OR 5.7, 95% CI 2.8-11.2, p < 0.001 for incident lacunes; and OR 2.9, 95% CI 1.4-5.9, p = 0.003 for incident microbleeds).

CONCLUSIONS

SVD progression is nonlinear, accelerating over time, and a highly dynamic process, with progression interrupted by reduction in some, in a population that on average shows progression.

Validation of a Regression Technique for Segmentation of White Matter Hyperintensities in Alzheimer's Disease

Segmentation and volumetric quantification of white matter hyperintensities (WMHs) is essential in assessment and monitoring of the vascular burden in aging and Alzheimer's disease (AD), especially when considering their effect on cognition. Manually segmenting WMHs in large cohorts is technically unfeasible due to time and accuracy concerns. Automated tools that can detect WMHs robustly and with high accuracy are needed. Here, we present and validate a fully automatic technique for segmentation and volumetric quantification of WMHs in aging and AD. The proposed technique combines intensity and location features frommultiplemagnetic resonance imaging contrasts and manually labeled training data with a linear classifier to perform fast and robust segmentations. It provides both a continuous subject specific WMH map reflecting different levels of tissue damage and binary segmentations. Themethodwas used to detectWMHs in 80 elderly/AD brains (ADC data set) as well as 40 healthy subjects at risk of AD (PREVENT-AD data set). Robustness across different scanners was validated using ten subjects from ADNI2/GO study. Voxel-wise and volumetric agreements were evaluated using Dice similarity index (SI) and intra-class correlation (ICC), yielding ICC=0.96 , SI = 0.62±0.16 for ADC data set and ICC=0.78 , SI=0.51±0.15 for PREVENT-AD data set. The proposed method was robust in the independent sample yielding SI=0.64±0.17 with ICC=0.93 for ADNI2/GO subjects. The proposed method provides fast, accurate, and robust segmentations on previously unseen data from different models of scanners, making it ideal to study WMHs in large scale multi-site studies.

Location Sensitive Deep Convolutional Neural Networks for Segmentation of White Matter Hyperintensities

The anatomical location of imaging features is of crucial importance for accurate diagnosis in many medical tasks. Convolutional neural networks (CNN) have had huge successes in computer vision, but they lack the natural ability to incorporate the anatomical location in their decision making process, hindering success in some medical image analysis tasks. In this paper, to integrate the anatomical location information into the network, we propose several deep CNN architectures that consider multi-scale patches or take explicit location features while training. We apply and compare the proposed architectures for segmentation of white matter hyperintensities in brain MR images on a large dataset. As a result, we observe that the CNNs that incorporate location information substantially outperform a conventional segmentation method with handcrafted features as well as CNNs that do not integrate location information. On a test set of 50 scans, the best configuration of our networks obtained a Dice score of 0.792, compared to 0.805 for an independent human observer. Performance levels of the machine and the independent human observer were not statistically significantly different (p-value = 0.06).

Performance comparison of 10 different classification techniques in segmenting white matter hyperintensities in aging

INTRODUCTION

White matter hyperintensities (WMHs) are areas of abnormal signal on magnetic resonance images (MRIs) that characterize various types of histopathological lesions. The load and location of WMHs are important clinical measures that may indicate the presence of small vessel disease in aging and Alzheimer's disease (AD) patients. Manually segmenting WMHs is time consuming and prone to inter-rater and intra-rater variabilities. Automated tools that can accurately and robustly detect these lesions can be used to measure the vascular burden in individuals with AD or the elderly population in general. Many WMH segmentation techniques use a classifier in combination with a set of intensity and location features to segment WMHs, however, the optimal choice of classifier is unknown.

METHODS

We compare 10 different linear and nonlinear classification techniques to identify WMHs from MRI data. Each classifier is trained and optimized based on a set of features obtained from co-registered MR images containing spatial location and intensity information. We further assess the performance of the classifiers using different combinations of MRI contrast information. The performances of the different classifiers were compared on three heterogeneous multi-site datasets, including images acquired with different scanners and different scan-parameters. These included data from the ADC study from University of California Davis, the NACC database and the ADNI study. The classifiers (naïve Bayes, logistic regression, decision trees, random forests, support vector machines, k-nearest neighbors, bagging, and boosting) were evaluated using a variety of voxel-wise and volumetric similarity measures such as Dice Kappa similarity index (SI), Intra-Class Correlation (ICC), and sensitivity as well as computational burden and processing times. These investigations enable meaningful comparisons between the performances of different classifiers to determine the most suitable classifiers for segmentation of WMHs. In the spirit of open-source science, we also make available a fully automated tool for segmentation of WMHs with pre-trained classifiers for all these techniques.

RESULTS

Random Forests yielded the best performance among all classifiers with mean Dice Kappa (SI) of 0.66±0.17 and ICC=0.99 for the ADC dataset (using T1w, T2w, PD, and FLAIR scans), SI=0.72±0.10, ICC=0.93 for the NACC dataset (using T1w and FLAIR scans), SI=0.66±0.23, ICC=0.94 for ADNI1 dataset (using T1w, T2w, and PD scans) and SI=0.72±0.19, ICC=0.96 for ADNI2/GO dataset (using T1w and FLAIR scans). Not using the T2w/PD information did not change the performance of the Random Forest classifier (SI=0.66±0.17, ICC=0.99). However, not using FLAIR information in the ADC dataset significantly decreased the Dice Kappa, but the volumetric correlation did not drastically change (SI=0.47±0.21, ICC=0.95).

CONCLUSION

Our investigations showed that with appropriate features, most off-the-shelf classifiers are able to accurately detect WMHs in presence of FLAIR scan information, while Random Forests had the best performance across all datasets. However, we observed that the performances of most linear classifiers and some nonlinear classifiers drastically decline in absence of FLAIR information, with Random Forest still retaining the best performance.