Extracting and summarizing white matter hyperintensities using supervised segmentation methods in Alzheimer's disease risk and aging studies

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.

Application of machine learning in measurement of ageing and geriatric diseases: a systematic review

BACKGROUND

As the ageing population continues to grow in many countries, the prevalence of geriatric diseases is on the rise. In response, healthcare providers are exploring novel methods to enhance the quality of life for the elderly. Over the last decade, there has been a remarkable surge in the use of machine learning in geriatric diseases and care. Machine learning has emerged as a promising tool for the diagnosis, treatment, and management of these conditions. Hence, our study aims to find out the present state of research in geriatrics and the application of machine learning methods in this area.

METHODS

This systematic review followed Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines and focused on healthy ageing in individuals aged 45 and above, with a specific emphasis on the diseases that commonly occur during this process. The study mainly focused on three areas, that are machine learning, the geriatric population, and diseases. Peer-reviewed articles were searched in the PubMed and Scopus databases with inclusion criteria of population above 45 years, must have used machine learning methods, and availability of full text. To assess the quality of the studies, Joanna Briggs Institute's (JBI) critical appraisal tool was used.

RESULTS

A total of 70 papers were selected from the 120 identified papers after going through title screening, abstract screening, and reference search. Limited research is available on predicting biological or brain age using deep learning and different supervised machine learning methods. Neurodegenerative disorders were found to be the most researched disease, in which Alzheimer's disease was focused the most. Among non-communicable diseases, diabetes mellitus, hypertension, cancer, kidney diseases, and cardiovascular diseases were included, and other rare diseases like oral health-related diseases and bone diseases were also explored in some papers. In terms of the application of machine learning, risk prediction was the most common approach. Half of the studies have used supervised machine learning algorithms, among which logistic regression, random forest, XG Boost were frequently used methods. These machine learning methods were applied to a variety of datasets including population-based surveys, hospital records, and digitally traced data.

CONCLUSION

The review identified a wide range of studies that employed machine learning algorithms to analyse various diseases and datasets. While the application of machine learning in geriatrics and care has been well-explored, there is still room for future development, particularly in validating models across diverse populations and utilizing personalized digital datasets for customized patient-centric care in older populations. Further, we suggest a scope of Machine Learning in generating comparable ageing indices such as successful ageing index.

Simple topological task-based functional connectivity features predict longitudinal behavioral change of fluid reasoning in the RANN cohort

Recent attention has been given to topological data analysis (TDA), and more specifically persistent homology (PH), to identify the underlying shape of brain network connectivity beyond simple edge pairings by computing connective components across different connectivity thresholds (see Sizemore et al., 2019). In the present study, we applied PH to task-based functional connectivity, computing 0-dimension Betti (B0) curves and calculating the area under these curves (AUC); AUC indicates how quickly a single connected component is formed across correlation filtration thresholds, with lower values interpreted as potentially analogous to lower whole-brain system segregation (e.g., Gracia-Tabuenca et al., 2020). One hundred sixty-three participants from the Reference Ability Neural Network (RANN) longitudinal lifespan cohort (age 20-80 years) were tested in-scanner at baseline and five-year follow-up on a battery of tests comprising four domains of cognition (i.e., Stern et al., 2014). We tested for 1.) age-related change in the AUC of the B0 curve over time, 2.) the predictive utility of AUC in accounting for longitudinal change in behavioral performance and 3.) compared system segregation to the PH approach. Results demonstrated longitudinal age-related decreases in AUC for Fluid Reasoning, with these decreases predicting longitudinal declines in cognition, even after controlling for demographic and brain integrity factors; moreover, change in AUC partially mediated the effect of age on change in cognitive performance. System segregation also significantly decreased with age in three of the four cognitive domains but did not predict change in cognition. These results argue for greater application of TDA to the study of aging.

Proteomic clusters underlie heterogeneity in preclinical Alzheimer's disease progression

Heterogeneity in progression to Alzheimer's disease (AD) poses challenges for both clinical prognosis and clinical trial implementation. Multiple AD-related subtypes have previously been identified, suggesting differences in receptivity to drug interventions. We identified early differences in preclinical AD biomarkers, assessed patterns for developing preclinical AD across the amyloid-tau-(neurodegeneration) [AT(N)] framework, and considered potential sources of difference by analysing the CSF proteome. Participants (n = 10) enrolled in longitudinal studies at the Knight Alzheimer Disease Research Center completed four or more lumbar punctures. These individuals were cognitively normal at baseline. Cerebrospinal fluid measures of amyloid-β (Aβ)42, phosphorylated tau (pTau181), and neurofilament light chain (NfL) as well as proteomics values were evaluated. Imaging biomarkers, including PET amyloid and tau, and structural MRI, were repeatedly obtained when available. Individuals were staged according to the amyloid-tau-(neurodegeneration) framework. Growth mixture modelling, an unsupervised clustering technique, identified three patterns of biomarker progression as measured by CSF pTau181 and Aβ42. Two groups (AD Biomarker Positive and Intermediate AD Biomarker) showed distinct progression from normal biomarker status to having biomarkers consistent with preclinical AD. A third group (AD Biomarker Negative) did not develop abnormal AD biomarkers over time. Participants grouped by CSF trajectories were re-classified using only proteomic profiles (AUCAD Biomarker Positive versus AD Biomarker Negative = 0.857, AUCAD Biomarker Positive versus Intermediate AD Biomarkers = 0.525, AUCIntermediate AD Biomarkers versus AD Biomarker Negative = 0.952). We highlight heterogeneity in the development of AD biomarkers in cognitively normal individuals. We identified some individuals who became amyloid positive before the age of 50 years. A second group, Intermediate AD Biomarkers, developed elevated CSF ptau181 significantly before becoming amyloid positive. A third group were AD Biomarker Negative over repeated testing. Our results could influence the selection of participants for specific treatments (e.g. amyloid-reducing versus other agents) in clinical trials. CSF proteome analysis highlighted additional non-AT(N) biomarkers for potential therapies, including blood-brain barrier-, vascular-, immune-, and neuroinflammatory-related targets.

Semisupervised white matter hyperintensities segmentation on MRI

This study proposed a semisupervised loss function named level-set loss (LSLoss) for cerebral white matter hyperintensities (WMHs) segmentation on fluid-attenuated inversion recovery images. The training procedure did not require manually labeled WMH masks. Our image preprocessing steps included biased field correction, skull stripping, and white matter segmentation. With the proposed LSLoss, we trained a V-Net using the MRI images from both local and public databases. Local databases were the small vessel disease cohort (HKU-SVD, n = 360) and the multiple sclerosis cohort (HKU-MS, n = 20) from our institutional imaging center. Public databases were the Medical Image Computing Computer-assisted Intervention (MICCAI) WMH challenge database (MICCAI-WMH, n = 60) and the normal control cohort of the Alzheimer's Disease Neuroimaging Initiative database (ADNI-CN, n = 15). We achieved an overall dice similarity coefficient (DSC) of 0.81 on the HKU-SVD testing set (n = 20), DSC = 0.77 on the HKU-MS testing set (n = 5), and DSC = 0.78 on MICCAI-WMH testing set (n = 30). The segmentation results obtained by our semisupervised V-Net were comparable with the supervised methods and outperformed the unsupervised methods in the literature.

Neutron Imaging and Learning Algorithms: New Perspectives in Cultural Heritage Applications

Recently, learning algorithms such as Convolutional Neural Networks have been successfully applied in different stages of data processing from the acquisition to the data analysis in the imaging context. The aim of these algorithms is the dimensionality of data reduction and the computational effort, to find benchmarks and extract features, to improve the resolution, and reproducibility performances of the imaging data. Currently, no Neutron Imaging combined with learning algorithms was applied on cultural heritage domain, but future applications could help to solve challenges of this research field. Here, a review of pioneering works to exploit the use of Machine Learning and Deep Learning models applied to X-ray imaging and Neutron Imaging data processing is reported, spanning from biomedicine, microbiology, and materials science to give new perspectives on future cultural heritage applications.

Physical activity moderates the association between white matter hyperintensity burden and cognitive change

Objective

Greater physical activity (PA) could delay cognitive decline, yet the underlying mechanisms remain unclear. White matter hyperintensity (WMH) burden is one of the key brain pathologies that have been shown to predict faster cognitive decline at a late age. One possible pathway is that PA may help maintain cognition by mitigating the detrimental effects of brain pathologies, like WMH, on cognitive change. This study aims to examine whether PA moderates the association between WMH burden and cognitive change.

Materials and methods

This population-based longitudinal study included 198 dementia-free adults aged 20-80 years. Leisure-time physical activity (LTPA) was assessed by a self-reported questionnaire. Occupational physical activity (OPA) was a factor score measuring the physical demands of each job. Total physical activity (TPA) was operationalized as the average of z-scores of LTPA and OPA. Outcome variables included 5-year changes in global cognition and in four reference abilities (fluid reasoning, processing speed, memory, and vocabulary). Multivariable linear regression models were used to estimate the moderation effect of PA on the association between white matter hyperintensities and cognitive change, adjusting for age, sex, education, and baseline cognition.

Results

Over approximately 5 years, global cognition (p < 0.001), reasoning (p < 0.001), speed (p < 0.001), and memory (p < 0.05) scores declined, and vocabulary (p < 0.001) increased. Higher WMH burden was correlated with more decline in global cognition (Spearman's rho = -0.229, p = 0.001), reasoning (rho = -0.402, p < 0.001), and speed (rho = -0.319, p < 0.001), and less increase in vocabulary (rho = -0.316, p < 0.001). Greater TPA attenuated the association between WMH burden and changes in reasoning (βTPA^*WMH = 0.029, 95% CI = 0.006-0.052, p = 0.013), speed (βTPA^*WMH = 0.035, 95% CI = -0.004-0.065, p = 0.028), and vocabulary (βTPA^*WMH = 0.034, 95% CI = 0.004-0.065, p = 0.029). OPA seemed to be the factor that exerted a stronger moderation on the relationship between WMH burden and cognitive change.

Conclusion

Physical activity may help maintain reasoning, speed, and vocabulary abilities in face of WMH burden. The cognitive reserve potential of PA warrants further examination.

Altered local gyrification and functional connectivity in type 2 diabetes mellitus patients with mild cognitive impairment: A pilot cross-sectional small-scale single center study

Aims

This research aimed to explore alterations in the local gyrification index (GI) and resting-state functional connectivity (RSFC) in type 2 diabetes mellitus (T2DM) patients with mild cognitive impairment (MCI).

Methods

In this study, 126 T2DM patients with MCI (T2DM-MCI), 154 T2DM patients with normal cognition (T2DM-NC), and 167 healthy controls (HC) were recruited. All subjects underwent a battery of neuropsychological tests. A multimodal approach combining surface-based morphometry (SBM) and seed-based RSFC was used to determine the structural and functional alterations in patients with T2DM-MCI. The relationships among the GI, RSFC, cognitive ability, and clinical variables were characterized.

Results

Compared with the T2DM-NC group and HC group, T2DM-MCI patients showed significantly reduced GI in the bilateral insular cortex. Decreased RSFC was found between the left insula and right precuneus, and the right superior frontal gyrus (SFG). The altered GI was correlated with T2DM duration, global cognition, and episodic memory. The mediation effects of RSFC on the association between GI and cognition were not statistically significant.

Conclusion

Our results suggest that GI may serve as a novel neuroimaging biomarker to predict T2DM-related MCI and help us to improve the understanding of the neuropathological effects of T2DM-related MCI.

Mediterranean Diet and White Matter Hyperintensity Change over Time in Cognitively Intact Adults

Current evidence on the impact of Mediterranean diet (MeDi) on white matter hyperintensity (WMH) trajectory is scarce. This study aims to examine whether greater adherence to MeDi is associated with less accumulation of WMH. This population-based longitudinal study included 183 cognitively intact adults aged 20−80 years. The MeDi score was obtained from a self-reported food frequency questionnaire; WMH was assessed by 3T MRI. Multivariable linear regression was used to estimate the effect of MeDi on WMH change. Covariates included socio-demographic factors and brain markers. Moderation effects by age, gender, and race/ethnicity were examined, followed by stratification analyses. Among all participants, WMH increased from baseline to follow-up (mean difference [follow-up-baseline] [standard deviation] = 0.31 [0.48], p < 0.001). MeDi adherence was negatively associated with the increase in WMH (β = −0.014, 95% CI = −0.026−−0.001, p = 0.034), adjusting for all covariates. The association between MeDi and WMH change was moderated by age (young group = reference, p-interaction[middle-aged × MeDi] = 0.075, p-interaction[older × MeDi] = 0.037). The association between MeDi and WMH change was observed among the young group (β = −0.035, 95% CI = −0.058−−0.013, p = 0.003), but not among other age groups. Moderation effects by gender and race/ethnicity did not reach significance. Greater adherence to MeDi was associated with a lesser increase in WMH over time. Following a healthy diet, especially at younger age, may help to maintain a healthy brain.

Grey and white matter atrophy one year after stroke aphasia

Dynamic whole-brain changes occur following stroke, and not just in association with recovery. We tested the hypothesis that the presence of a specific behavioural deficit after stroke would be associated with structural decline (atrophy) in the brain regions supporting the affected function, by examining language deficits post-stroke. We quantified whole-brain structural volume changes longitudinally (3–12 months) in stroke participants with (N = 32) and without aphasia (N = 59) as assessed by the Token Test at 3 months post-stroke, compared with a healthy control group (N = 29). While no significant difference in language decline rates (change in Token Test scores from 3 to 12 months) was observed between groups and some participants in the aphasic group improved their scores, stroke participants with aphasia symptoms at 3 months showed significant atrophy (>2%, P = 0.0001) of the left inferior frontal gyrus not observed in either healthy control or non-aphasic groups over the 3–12 months period. We found significant group differences in the inferior frontal gyrus volume, accounting for age, sex, stroke severity at baseline, education and total intracranial volume (Bonferroni-corrected P = 0.0003). In a subset of participants (aphasic N = 14, non-aphasic N = 36, and healthy control N = 25) with available diffusion-weighted imaging data, we found significant atrophy in the corpus callosum and the left superior longitudinal fasciculus in the aphasic compared with the healthy control group. Language deficits at 3 months post-stroke are associated with accelerated structural decline specific to the left inferior frontal gyrus, highlighting that known functional brain reorganization underlying behavioural improvement may occur in parallel with atrophy of brain regions supporting the language function.

Impaired functional network properties contribute to white matter hyperintensity related cognitive decline in patients with cerebral small vessel disease

Background

White matter hyperintensity (WMH) is one of the typical neuroimaging manifestations of cerebral small vessel disease (CSVD), and the WMH correlates closely to cognitive impairment (CI). CSVD patients with WMH own altered topological properties of brain functional network, which is a possible mechanism that leads to CI. This study aims to identify differences in the characteristics of some brain functional network among patients with different grades of WMH and estimates the correlations between these different brain functional network characteristics and cognitive assessment scores.

Methods

110 CSVD patients underwent 3.0 T Magnetic resonance imaging scans and neuropsychological cognitive assessments. WMH of each participant was graded on the basis of Fazekas grade scale and was divided into two groups: (A) WMH score of 1–2 points (n = 64), (B) WMH score of 3–6 points (n = 46). Topological indexes of brain functional network were analyzed using graph-theoretical method. T-test and Mann–Whitney U test was used to compare the differences in topological properties of brain functional network between groups. Partial correlation analysis was applied to explore the relationship between different topological properties of brain functional networks and overall cognitive function.

Results

Patients with high WMH scores exhibited decreased clustering coefficient values, global and local network efficiency along with increased shortest path length on whole brain level as well as decreased nodal efficiency in some brain regions on nodal level (p < 0.05). Nodal efficiency in the left lingual gyrus was significantly positively correlated with patients' total Montreal Cognitive Assessment (MoCA) scores (p < 0.05). No significant difference was found between two groups on the aspect of total MoCA and Mini-mental State Examination (MMSE) scores (p > 0.05).

Conclusion

Therefore, we come to conclusions that patients with high WMH scores showed less optimized small-world networks compared to patients with low WMH scores. Global and local network efficiency on the whole-brain level, as well as nodal efficiency in certain brain regions on the nodal level, can be viewed as markers to reflect the course of WMH.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12880-022-00769-7.

Hippocampal subfields atrophy contribute more to cognitive impairment in middle-aged patients with type 2 diabetes rather than microvascular lesions

AIMS

Neurodegeneration and microvascular lesions are related to cognitive impairment in type 2 diabetes mellitus (T2DM). We aimed to use the volume of hippocampal subfields and the burden of white matter hyperintensities (WMH) as neurodegeneration and microangiopathy markers, respectively, to investigate their potential associations with cognitive impairment in T2DM patients.

METHODS

A total of 76 T2DM patients and 45 neurologically unimpaired normal controls were enrolled between February 2016 to August 2018. All participants underwent structural magnetic resonance imaging (MRI) and Montreal Cognitive Assessment (MoCA). The T2DM patients were divided into the T2DM without mild cognitive impairment (T2noMCI) group (n = 44) and the T2DM with mild cognitive impairment (T2MCI) group (n = 32) according to MoCA scores. We used automatic brain segmentation and quantitative technique to assess the volume of twelve hippocampal subfields and WMH on MRI. We used age, sex, education, and total intracranial volume as our covariates and the Bonferroni method for multiple comparison correction.

RESULTS

Both the T2MCI group and T2noMCI group showed significant hippocampal subfields atrophy compared to the controls, which were mainly in the left hippocampal tail, left CA1, bilateral molecular layer, bilateral dentate gyrus, and bilateral CA4 (all p < 0.0042). No significant differences in the volume of total WMH, deep-WMH, and periventricular-WMH were found among the three groups. The HbA1c levels were significantly negatively correlated with hippocampal atrophy, and the MoCA scores were positively correlated with bilateral hippocampal volume in T2DM patients and all samples. Mediation analyses demonstrated that the association of HbA1c levels with cognitive function was mediated by hippocampal subfields volume.

CONCLUSION

Widespread hippocampal atrophies across the subfields have been found in middle-aged T2DM patients, which was positively correlated with the MoCA scores and negatively correlated with the HbA1c levels. The association of HbA1c levels with cognitive function was mediated by some crucial hippocampal subfields volume. In middle-aged patients with T2DM, the neurodegeneration is more strongly associated with cognitive impairment than microvascular lesions. Trail Registeration This study was registered on Clinical-Trails.gov (NCT02738671).

Investigating the microstructural properties of normal-appearing white matter (NAWM) preceding conversion to white matter hyperintensities (WMHs) in stroke survivors

Using advanced diffusion MRI, we aimed to assess the microstructural properties of normal-appearing white matter (NAWM) preceding conversion to white matter hyperintensities (WMHs) using 3-tissue diffusion signal compositions in ischemic stroke. Data were obtained from the Cognition and Neocortical Volume After Stroke (CANVAS) study. Diffusion-weighted MR and high-resolution structural brain images were acquired 3- (baseline) and 12-months (follow-up) post-stroke. WMHs were automatically segmented and longitudinal assessment at 12-months was used to retrospectively delineate NAWM voxels at baseline converting to WMHs. NAWM voxels converting to WMHs were further dichotomized into either: "growing" WMHs if NAWM adhered to existing WMH voxels, or "isolated de-novo" WMHs if NAWM was unconnected to WMH voxels identified at baseline. Microstructural properties were assessed using 3-tissue diffusion signal compositions consisting of white matter-like (WM-like: T_W), gray matter-like (GM-like: T_G), and cerebrospinal fluid-like (CSF-like: T_C) signal fractions. Our findings showed that NAWM converting to WMHs already exhibited similar changes in tissue compositions at baseline to WMHs with lower T_W and increased T_C (fluid-like, i.e. free-water) and T_G compared to persistent NAWM. We also found that microstructural properties of persistent NAWM were related to overall WMH burden with greater free-water content in patients with high WMH load. These findings suggest that NAWM preceding conversion to WMHs are accompanied by greater fluid-like properties indicating increased tissue water content. Increased GM-like properties may indicate a more isotropic microstructure of tissue reflecting a degree of hindered diffusion in NAWM regions vulnerable to WMH development. These results support the usefulness of microstructural compositions as a sensitive marker of NAWM vulnerability to WMH pathogenesis.

Socioeconomic Status Mediates Racial Differences Seen Using the AT(N) Framework

OBJECTIVES

African Americans are at greater risk for developing Alzheimer's disease (AD) dementia than non-Hispanic whites. In addition to biological considerations (eg, genetic influences and comorbid disorders), social and environmental factors may increase the risk of AD dementia. This paper (1) assesses neuroimaging biomarkers of amyloid (A), tau (T), and neurodegeneration (N) for potential racial differences and (2) considers mediating effects of socioeconomic status (SES) and measures of small vessel and cardiovascular disease on observed race differences.

METHODS

Imaging measures of AT(N) (amyloid and tau positron emission tomography [PET]) structural magnetic resonance imaging (MRI), and resting state functional connectivity (rs-fc) were collected from African American (n = 131) and white (n = 685) cognitively normal participants age 45 years and older. Measures of small vessel and cardiovascular disease (white matter hyperintensities [WMHs] on MRI, blood pressure, and body mass index [BMI]) and area-based SES were included in mediation analyses.

RESULTS

Compared to white participants, African American participants had greater neurodegeneration, as measured by decreased cortical volumes (Cohen's f2 = 0.05, p < 0.001). SES mediated the relationship between race and cortical volumes. There were no significant race effects for amyloid, tau, or rs-fc signature.

INTERPRETATION

Modifiable factors, such as differences in social contexts and resources, particularly area-level SES, may contribute to observed racial differences in AD. Future studies should emphasize collection of relevant psychosocial factors in addition to the development of intentional diversity and inclusion efforts to improve the racial/ethnic and socioeconomic representativeness of AD studies. ANN NEUROL 2021;89:254-265.

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.

Deep-Learning-Based Segmentation and Localization of White Matter Hyperintensities on Magnetic Resonance Images

White matter magnetic resonance hyperintensities of presumed vascular origin, which could be widely observed in elderly people, and has significant importance in multiple neurological studies. Quantitative measurement usually relies heavily on manual or semi-automatic delineation and intuitive localization, which is time-consuming and observer-dependent. Current automatic quantification methods focus mainly on the segmentation, but the spatial distribution of lesions plays a vital role in clinical diagnosis. In this study, we implemented four segmentation algorithms and compared the performances quantitatively and qualitatively on two open-access datasets. The location-specific analysis was conducted sequentially on 213 clinical patients with cerebral ischemia and lacune. The experimental results suggest that our deep-learning-based model has the potential to be integrated into the clinical workflow.

Microstructural degeneration and cerebrovascular risk burden underlying executive dysfunction after stroke

Executive dysfunction affects 40% of stroke patients, but is poorly predicted by characteristics of the stroke itself. Stroke typically occurs on a background of cerebrovascular burden, which impacts cognition and brain network structural integrity. We used structural equation modelling to investigate whether measures of white matter microstructural integrity (fractional anisotropy and mean diffusivity) and cerebrovascular risk factors better explain executive dysfunction than markers of stroke severity. 126 stroke patients (mean age 68.4 years) were scanned three months post-stroke and compared to 40 age- and sex-matched control participants on neuropsychological measures of executive function. Executive function was below what would be expected for age and education level in stroke patients as measured by the organizational components of the Rey Complex Figure Test, F(3,155) = 17, R2 = 0.25, p < 0.001 (group significant predictor at p < 0.001) and the Trail-Making Test (B), F(3,157) = 3.70, R2 = 0.07, p < 0.01 (group significant predictor at p < 0.001). A multivariate structural equation model illustrated the complex relationship between executive function, white matter integrity, stroke characteristics and cerebrovascular risk (root mean square error of approximation = 0.02). Pearson's correlations confirmed a stronger relationship between executive dysfunction and white matter integrity (r = - 0.74, p < 0.001), than executive dysfunction and stroke severity (r = 0.22, p < 0.01). The relationship between executive function and white matter integrity is mediated by cerebrovascular burden. White matter microstructural degeneration of the superior longitudinal fasciculus in the executive control network better explains executive dysfunction than markers of stroke severity. Executive dysfunction and incident stroke can be both considered manifestations of cerebrovascular risk factors.

Serum Glial Fibrillary Acidic Protein (GFAP) Is a Marker of Disease Severity in Frontotemporal Lobar Degeneration

Background: It is still unknown if serum glial fibrillary acidic protein (GFAP) is a useful marker in frontotemporal lobar degeneration (FTLD). Objective: To assess the diagnostic and prognostic value of serum GFAP in a large cohort of patients with FTLD. Methods: In this retrospective study, performed on 406 participants, we measured serum GFAP concentration with an ultrasensitive Single molecule array (Simoa) method in patients with FTLD, Alzheimer’s disease (AD), and in cognitively unimpaired elderly controls. We assessed the role of GFAP as marker of disease severity by analyzing the correlation with clinical variables, neurophysiological data, and cross-sectional brain imaging. Moreover, we evaluated the role of serum GFAP as a prognostic marker of disease survival. Results: We observed significantly higher levels of serum GFAP in patients with FTLD syndromes, except progressive supranuclear palsy, compared with healthy controls, but not compared with AD patients. In FTLD, serum GFAP levels correlated with measures of cognitive dysfunction and disease severity, and were associated with indirect measures of GABAergic deficit. Serum GFAP concentration was not a significant predictor of survival. Conclusion: Serum GFAP is increased in FTLD, correlates with cognition and GABAergic deficits, and thus shows promise as a biomarker of disease severity in FTLD.

Comparison and validation of seven white matter hyperintensities segmentation software in elderly patients

•

The comparison used 207 images from both research and clinical datasets.

•

When retrained, NicMSlesion, a convolutional network, was the most accurate.

•

Performance of this deep learning method severely dropped on clinical routine data.

•

On clinical routine data, regression and clustering methods were the top-ranked methods.

•

SLS was the most robust to artifacted images, and BIANCA to scanners variability.

Background

Manual segmentation is currently the gold standard to assess white matter hyperintensities (WMH), but it is time consuming and subject to intra and inter-operator variability.

Purpose

To compare automatic methods to segment white matter hyperintensities (WMH) in the elderly in order to assist radiologist and researchers in selecting the most relevant method for application on clinical or research data.

Material and Methods

We studied a research dataset composed of 147 patients, including 97 patients from the Alzheimer's Disease Neuroimaging Initiative (ADNI) 2 database and 50 patients from ADNI 3 and a clinical routine dataset comprising 60 patients referred for cognitive impairment at the Pitié-Salpêtrière hospital (imaged using four different MRI machines). We used manual segmentation as the gold standard reference. Both manual and automatic segmentations were performed using FLAIR MRI. We compared seven freely available methods that produce segmentation mask and are usable by a radiologist without a strong knowledge of computer programming: LGA (Schmidt et al., 2012), LPA (Schmidt, 2017), BIANCA (Griffanti et al., 2016), UBO detector (Jiang et al., 2018), W2MHS (Ithapu et al., 2014), nicMSlesion (with and without retraining) (Valverde et al., 2019, Valverde et al., 2017). The primary outcome for assessing segmentation accuracy was the Dice similarity coefficient (DSC) between the manual and the automatic segmentation software. Secondary outcomes included five other metrics.

Results

A deep learning approach, NicMSlesion, retrained on data from the research dataset ADNI, performed best on this research dataset (DSC: 0.595) and its DSC was significantly higher than that of all others. However, it ranked fifth on the clinical routine dataset and its performance severely dropped on data with artifacts. On the clinical routine dataset, the three top-ranked methods were LPA, SLS and BIANCA. Their performance did not differ significantly but was significantly higher than that of other methods.

Conclusion

This work provides an objective comparison of methods for WMH segmentation. Results can be used by radiologists to select a tool.

Attention convolutional neural network for accurate segmentation and quantification of lesions in ischemic stroke disease

Ischemic stroke lesion and white matter hyperintensity (WMH) lesion appear as regions of abnormally signal intensity on magnetic resonance image (MRI) sequences. Ischemic stroke is a frequent cause of death and disability, while WMH is a risk factor for stroke. Accurate segmentation and quantification of ischemic stroke and WMH lesions are important for diagnosis and prognosis. However, radiologists have a difficult time distinguishing these two types of similar lesions. A novel deep residual attention convolutional neural network (DRANet) is proposed to accurately and simultaneously segment and quantify ischemic stroke and WMH lesions in the MRI images. DRANet inherits the advantages of the U-net design and applies a novel attention module that extracts high-quality features from the input images. Moreover, the Dice loss function is used to train DRANet to address data imbalance in the training data set. DRANet is trained and evaluated on 742 2D MRI images which are produced from the sub-acute ischemic stroke lesion segmentation (SISS) challenge. Empirical tests demonstrate that DRANet outperforms several other state-of-the-art segmentation methods. It accurately segments and quantifies both ischemic stroke lesion and WMH. Ablation experiments reveal that attention modules improve the predictive performance of DRANet.

Contrast-enhanced T1-weighted image radiomics of brain metastases may predict EGFR mutation status in primary lung cancer

Identification of EGFR mutations is critical to the treatment of primary lung cancer and brain metastases (BMs). Here, we explored whether radiomic features of contrast-enhanced T1-weighted images (T1WIs) of BMs predict EGFR mutation status in primary lung cancer cases. In total, 1209 features were extracted from the contrast-enhanced T1WIs of 61 patients with 210 measurable BMs. Feature selection and classification were optimized using several machine learning algorithms. Ten-fold cross-validation was applied to the T1WI BM dataset (189 BMs for training and 21 BMs for the test set). Area under receiver operating characteristic curves (AUC), accuracy, sensitivity, and specificity were calculated. Subgroup analyses were also performed according to metastasis size. For all measurable BMs, random forest (RF) classification with RF selection demonstrated the highest diagnostic performance for identifying EGFR mutation (AUC: 86.81). Support vector machine and AdaBoost were comparable to RF classification. Subgroup analyses revealed that small BMs had the highest AUC (89.09). The diagnostic performance for large BMs was lower than that for small BMs (the highest AUC: 78.22). Contrast-enhanced T1-weighted image radiomics of brain metastases predicted the EGFR mutation status of lung cancer BMs with good diagnostic performance. However, further study is necessary to apply this algorithm more widely and to larger BMs.

Three-tissue compositional analysis reveals in-vivo microstructural heterogeneity of white matter hyperintensities following stroke

White matter hyperintensities (WMHs) are frequently observed on brain scans of older individuals and are associated with cognitive impairment and vascular brain burden. Recent studies have shown that WMHs may only represent an extreme end of a diffuse pathological spectrum of white matter (WM) degeneration. The present study investigated the microstructural characteristics of WMHs using an advanced diffusion MRI modelling approach known as Single-Shell 3-Tissue Constrained Spherical Deconvolution (SS3T-CSD), which provides information on different tissue compartments within each voxel. The SS3T-CSD method may provide complementary information in the interpretation of pathological tissue through the tissue-specific microstructural compositions of WMHs. Data were obtained from stroke patients enrolled in the Cognition and Neocortical Volume After Stroke (CANVAS) study, a study examining brain volume and cognition after stroke. WMHs were segmented using an automated method, based on fluid attenuated inversion recovery (FLAIR) images. Automated tissue segmentation was used to identify normal-appearing white matter (NAWM). WMHs were classified into juxtaventricular, periventricular and deep lesions, based on their distance from the ventricles (3-10 ​mm). We aimed to compare in stroke participants the microstructural composition of the different lesion classes of WMHs and compositions of NAWM to assess the in-vivo heterogeneity of these lesions. Results showed that the 3-tissue composition significantly differed between WMHs classes and NAWM. Specifically, the 3-tissue compositions for juxtaventricular and periventricular WMHs both exhibited a relatively greater fluid-like (free water) content, which is compatible with a presence of interstitial fluid accumulation, when compared to deep WMHs. These findings provide evidence of microstructural heterogeneity of WMHs in-vivo and may support new insights for understanding the role of WMH development in vascular neurodegeneration.

Pervasive White Matter Fiber Degeneration in Ischemic Stroke

Background and Purpose- We examined if ischemic stroke is associated with white matter degeneration predominantly confined to the ipsi-lesional tracts or with widespread bilateral axonal loss independent of lesion laterality. Methods- We applied a novel fixel-based analysis, sensitive to fiber tract-specific differences within a voxel, to assess axonal loss in stroke (N=104, 32 women) compared to control participants (N=40, 15 women) across the whole brain. We studied microstructural differences in fiber density and macrostructural (morphological) changes in fiber cross-section. Results- In participants with stroke, we observed significantly lower fiber density and cross-section in areas adjacent, or connected, to the lesions (eg, ipsi-lesional corticospinal tract). In addition, the changes extended beyond directly connected tracts, independent of the lesion laterality (eg, corpus callosum, bilateral inferior fronto-occipital fasciculus, right superior longitudinal fasciculus). Conclusions- We conclude that ischemic stroke is associated with extensive neurodegeneration that significantly affects white matter integrity across the whole brain. These findings expand our understanding of the mechanisms of brain volume loss and delayed cognitive decline in stroke.

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.

Disrupted functional and structural connectivity within default mode network contribute to WMH-related cognitive impairment

•

Disconnective DMN contribute to impaired cognition

Aims

The prevalence of white matter hyperintensities (WMH) rises dramatically with aging. Both the progression of WMH and changing patterns of default mode network (DMN) have been proven to be closely associated with cognitive function. The present study hypothesized that changes in functional connectivity and structural connectivity of DMN contributed to WMH related cognitive impairment.

Methods

A total of 116 subjects were enrolled from the Cerebral Small Vessel Disease Register in Drum Tower Hospital of Nanjing University, and were distributed across three categories according to Fazekas rating scale: WMH I (n = 57), WMH II (n = 34), and WMH III(n = 25). All participants underwent neuropsychological tests and multimodal MRI scans, including diffusion tensor imaging and resting-state fMRI imaging. The alterations of functional connectivity and structural connectivity within the DMN were further explored.

Results

Age and hypertension were risk factors for WMH progression. Subjects with a higher WMH burden displayed higher DMN functional connectivity in the medial frontal gyrus, while lower DMN functional connectivity in the thalamus. After adjusting for aging, gender, and education, the increased DMN functional connectivity in the medial frontal gyrus, and the increased mean diffusivity of the white matter tracts between the hippocampus and posterior cingulate cortex were independent indicators of worse performance in memory. Moreover, the decreased DMN functional connectivity in the thalamus and increased mean diffusivity of the white matter tracts between the thalamus and posterior cingulate cortex were independent risk factors for a slower processing speed.

Conclusion

The changes in functional connectivity and structural connectivity within the DMN attributed to WMH progression were responsible for the development of cognitive impairment.

The Modulatory Effect of Cerebrovascular Burden in Response to Cognitive Stimulation in Healthy Ageing and Mild Cognitive Impairment

Background

Cerebrovascular burden is a common pathology in mild cognitive impairment (MCI) and Alzheimer's disease (AD), with an additive impact on cognitive functioning. Despite being proposed as a potential moderator of cholinesterase inhibiting drug therapy, there is a paucity of evidence investigating the impact of cerebrovascular pathology on responsiveness to cognitive interventions.

Method

The current study uses neuropsychological, neurostructural, and functional connectivity indices to characterise response to a cognitive stimulation paradigm in 25 healthy ageing and 22 MCI participants, to examine the hypothesised detrimental effects of concurrent vascular pathology.

Results

In both healthy ageing and MCI, increased levels of vascular pathology limited the potential for a neuroplastic response to cognitive stimulation. In healthy ageing, participants with lower levels of vascular burden had greater functional connectivity response in the target posterior default mode network. Those with low levels of vascular pathology in the MCI cohort had increased functional connectivity of the right insula and claustrum within the salience network. Burden did not, however, predict cognitive or neuroanatomical changes.

Conclusions

The current research evidences the modulatory effect of cerebrovascular pathology in interventions aimed at re-establishing network connectivity to prevent cognitive deterioration and delay the transition to the dementia stage of AD. Examination of co-occurring vascular pathology may improve precision in targeting treatment to MCI candidates who may respond optimally to such cognitive interventions.

Brain Amyloid PET Tracer Delivery is Related to White Matter Integrity in Patients with Mild Cognitive Impairment

BACKGROUND AND PURPOSE

Amyloid deposition, tau neurofibrillary tangles, and cerebrovascular dysfunction are important pathophysiologic features in Alzheimer's disease. Pittsburgh compound B ([¹¹ C]-PIB) is a positron emission tomography (PET) radiotracer used to quantify amyloid deposition in vivo. In addition, certain models of [¹¹ C]-PIB delivery reflect cerebral blood flow rather than amyloid plaques. As cerebral blood flow and perfusion deficits are associated with white matter pathology, we hypothesized that [¹¹ C]-PIB delivery in white matter regions may reflect white matter integrity.

METHODS

We obtained [¹¹ C]-PIB-PET scans and quantified white matter hyperintensities and global fractional anisotropy on magnetic resonance images as biomarkers of white matter pathology in 34 older participants with mild cognitive impairment with or without a history of major depressive disorder. We analyzed the [¹¹ C]-PIB time-activity curve data with models associated with cerebral blood flow: the early maximum standard uptake value and the relative delivery parameter R1. We used a global white matter region of interest.

RESULTS

Both of the partial-volume corrected PET parameters were correlated with white matter hyperintensities and fractional anisotropy.

CONCLUSION

Future studies are warranted to explore whether [¹¹ C]-PIB PET is a "triple biomarker" that may provide information about amyloid deposition, cerebral blood flow, and white matter pathology.

Diagnosis and prognosis of Alzheimer's disease using brain morphometry and white matter connectomes

Accurate, reliable prediction of risk for Alzheimer's disease (AD) is essential for early, disease-modifying therapeutics. Multimodal MRI, such as structural and diffusion MRI, is likely to contain complementary information of neurodegenerative processes in AD. Here we tested the utility of the multimodal MRI (T1-weighted structure and diffusion MRI), combined with high-throughput brain phenotyping-morphometry and structural connectomics-and machine learning, as a diagnostic tool for AD. We used, firstly, a clinical cohort at a dementia clinic (National Health Insurance Service-Ilsan Hospital [NHIS-IH]; N = 211; 110 AD, 64 mild cognitive impairment [MCI], and 37 cognitively normal with subjective memory complaints [SMC]) to test the diagnostic models; and, secondly, Alzheimer's Disease Neuroimaging Initiative (ADNI)-2 to test the generalizability. Our machine learning models trained on the morphometric and connectome estimates (number of features = 34,646) showed optimal classification accuracy (AD/SMC: 97% accuracy, MCI/SMC: 83% accuracy; AD/MCI: 97% accuracy) in NHIS-IH cohort, outperforming a benchmark model (FLAIR-based white matter hyperintensity volumes). In ADNI-2 data, the combined connectome and morphometry model showed similar or superior accuracies (AD/HC: 96%; MCI/HC: 70%; AD/MCI: 75% accuracy) compared with the CSF biomarker model (t-tau, p-tau, and Amyloid β, and ratios). In predicting MCI to AD progression in a smaller cohort of ADNI-2 (n = 60), the morphometry model showed similar performance with 69% accuracy compared with CSF biomarker model with 70% accuracy. Our comparisons of the classifiers trained on structural MRI, diffusion MRI, FLAIR, and CSF biomarkers showed the promising utility of the white matter structural connectomes in classifying AD and MCI in addition to the widely used structural MRI-based morphometry, when combined with machine learning.

Multi-atlas based detection and localization (MADL) for location-dependent quantification of white matter hyperintensities

The extent and spatial location of white matter hyperintensities (WMH) on brain MRI may be relevant to the development of cognitive decline in older persons. Here, we introduce a new method, known as the Multi-atlas based Detection and Localization (MADL), to evaluate WMH on fluid-attenuated inversion recovery (FLAIR) data. This method simultaneously parcellates the whole brain into 143 structures and labels hyperintense areas within each WM structure. First, a multi-atlas library was established with FLAIR data of normal elderly brains; and then a multi-atlas fusion algorithm was developed by which voxels with locally abnormal intensities were detected as WMH. At the same time, brain segmentation maps were generated from the multi-atlas fusion process to determine the anatomical location of WMH. Areas identified using the MADL method agreed well with manual delineation, with an interclass correlation of 0.97 and similarity index (SI) between 0.55 and 0.72, depending on the total WMH load. Performance was compared to other state-of-the-art WMH detection methods, such as BIANCA and LST. MADL-based analyses of WMH in an older population revealed a significant association between age and WMH load in deep WM but not subcortical WM. The findings also suggested increased WMH load in selective brain regions in subjects with mild cognitive impairment compared to controls, including the inferior deep WM and occipital subcortical WM. The proposed MADL approach may facilitate location-dependent characterization of WMH in older individuals with memory impairment.

•

We proposed a multi-atlas based method for simultaneous detection and location of WMH on FLAIR images.

•

The method generates whole-brain segmentation for location-dependent WMH analysis.

•

The method showed reasonably high detection accuracy in comparison with other methods.

•

Results revealed a selective association between deep brain WMH and subject age.

•

Results suggested increased WMH in the inferior white matter in MCI patients.

Spatial patterns of white matter hyperintensities associated with Alzheimer's disease risk factors in a cognitively healthy middle-aged cohort

Background

White matter hyperintensities (WMH) of presumed vascular origin have been associated with an increased risk of Alzheimer’s disease (AD). This study aims to describe the patterns of WMH associated with dementia risk estimates and individual risk factors in a cohort of middle-aged/late middle-aged individuals (mean 58 (interquartile range 51–64) years old).

Methods

Magnetic resonance imaging and AD risk factors were collected from 575 cognitively unimpaired participants. WMH load was automatically calculated in each brain lobe and in four equidistant layers from the ventricular surface to the cortical interface. Global volumes and regional patterns of WMH load were analyzed as a function of the Cardiovascular Risk Factors, Aging and Incidence of Dementia (CAIDE) dementia risk score, as well as family history of AD and Apolipoprotein E (APOE) genotype. Additional analyses were performed after correcting for the effect of age and hypertension.

Results

The studied cohort showed very low WMH burden (median 1.94 cm³) and 20-year dementia risk estimates (median 1.47 %). Even so, higher CAIDE scores were significantly associated with increased global WMH load. The main drivers of this association were age and hypertension, with hypercholesterolemia and body mass index also displaying a minor, albeit significant, influence. Regionally, CAIDE scores were positively associated with WMH in anterior areas, mostly in the frontal lobe. Age and hypertension showed significant association with WMH in almost all regions analyzed. The APOE-ε2 allele showed a protective effect over global WMH with a pattern that comprised juxtacortical temporo-occipital and fronto-parietal deep white matter regions. Participants with maternal family history of AD had higher WMH load than those without, especially in temporal and occipital lobes.

Conclusions

WMH load is associated with AD risk factors even in cognitively unimpaired subjects with very low WMH burden and dementia risk estimates. Our results suggest that tight control of modifiable risk factors in middle-age/late middle-age could have a significant impact on late-life dementia.

Electronic supplementary material

The online version of this article (10.1186/s13195-018-0460-1) contains supplementary material, which is available to authorized users.

Characteristic changes in the default mode network in hypertensive patients with cognitive impairment

Hypertension has a close affinity to brain degeneration and cognitive decline during the aging process. The default mode network (DMN) is usually affected in various diseases related to cognitive impairment (CI). The present research aimed to explore the alterations in the DMN and its subcomponents in hypertensive patients with and without CI and to investigate the associations between cognitive performance and network abnormalities. Resting-state functional magnetic resonance imaging and neuropsychological tests were performed in 74 subjects, namely, 30 hypertensive patients with normal cognition (HTN-NC), 25 hypertensive patients with CI (HTN-CI), and 19 healthy controls. Seed-based functional connectivity (FC) analysis was performed to identify the DMN patterns. The group differences in the DMN were mainly shown in brain regions related to the core subsystem and the dorsal medial subsystem of the DMN. Post hoc analysis revealed a trend of dissociation among the DMN subsystems in the HTN-NC group. In contrast, the HTN-CI group displayed extensively increased FC in both subsystems. Importantly, increased FC of the dorsal medial subsystem in the HTN-CI patients was associated with poor cognitive performance, such as scores on Mini-Mental State Examination (ρ = -0.438, P = 0.029) and Montreal Cognitive Assessment (ρ = -0.449, P = 0.025). The findings suggest that extensively increased connectivities in the core subsystem and the dorsal media subsystem of the DMN may distinguish hypertension with CI from hypertension with normal cognition. The characteristic change in the dorsal medial subsystem may become an early imaging biomarker for the diagnosis and treatment of cognitive impairment associated with hypertension.

Voxel-Wise Logistic Regression and Leave-One-Source-Out Cross Validation for white matter hyperintensity segmentation

Many algorithms have been proposed for automated segmentation of white matter hyperintensities (WMH) in brain MRI. Yet, broad uptake of any particular algorithm has not been observed. In this work, we argue that this may be due to variable and suboptimal validation data and frameworks, precluding direct comparison of methods on heterogeneous data. As a solution, we present Leave-One-Source-Out Cross Validation (LOSO-CV), which leverages all available data for performance estimation, and show that this gives more realistic (lower) estimates of segmentation algorithm performance on data from different scanners. We also develop a FLAIR-only WMH segmentation algorithm: Voxel-Wise Logistic Regression (VLR), inspired by the open-source Lesion Prediction Algorithm (LPA). Our variant facilitates more accurate parameter estimation, and permits intuitive interpretation of model parameters. We illustrate the performance of the VLR algorithm using the LOSO-CV framework with a dataset comprising freely available data from several recent competitions (96 images from 7 scanners). The performance of the VLR algorithm (median Similarity Index of 0.69) is compared to its LPA predecessor (0.58), and the results of the VLR algorithm in the 2017 WMH Segmentation Competition are also presented.

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 hyperintensity and stroke lesion segmentation and differentiation using convolutional neural networks

White matter hyperintensities (WMH) are a feature of sporadic small vessel disease also frequently observed in magnetic resonance images (MRI) of healthy elderly subjects. The accurate assessment of WMH burden is of crucial importance for epidemiological studies to determine association between WMHs, cognitive and clinical data; their causes, and the effects of new treatments in randomized trials. The manual delineation of WMHs is a very tedious, costly and time consuming process, that needs to be carried out by an expert annotator (e.g. a trained image analyst or radiologist). The problem of WMH delineation is further complicated by the fact that other pathological features (i.e. stroke lesions) often also appear as hyperintense regions. Recently, several automated methods aiming to tackle the challenges of WMH segmentation have been proposed. Most of these methods have been specifically developed to segment WMH in MRI but cannot differentiate between WMHs and strokes. Other methods, capable of distinguishing between different pathologies in brain MRI, are not designed with simultaneous WMH and stroke segmentation in mind. Therefore, a task specific, reliable, fully automated method that can segment and differentiate between these two pathological manifestations on MRI has not yet been fully identified. In this work we propose to use a convolutional neural network (CNN) that is able to segment hyperintensities and differentiate between WMHs and stroke lesions. Specifically, we aim to distinguish between WMH pathologies from those caused by stroke lesions due to either cortical, large or small subcortical infarcts. The proposed fully convolutional CNN architecture, called uResNet, that comprised an analysis path, that gradually learns low and high level features, followed by a synthesis path, that gradually combines and up-samples the low and high level features into a class likelihood semantic segmentation. Quantitatively, the proposed CNN architecture is shown to outperform other well established and state-of-the-art algorithms in terms of overlap with manual expert annotations. Clinically, the extracted WMH volumes were found to correlate better with the Fazekas visual rating score than competing methods or the expert-annotated volumes. Additionally, a comparison of the associations found between clinical risk-factors and the WMH volumes generated by the proposed method, was found to be in line with the associations found with the expert-annotated volumes.

•

Robust, fully automatic white matter hyperintensity and stroke lesion segmentation and differentiation

•

A novel patch sampling strategy used during CNN training that avoids the introduction of erroneous locality assumptions

•

Improved segmentation accuracy in terms of Dice scores when compared to well established state-of-the-art methods

Altered cerebral hemodyamics and cortical thinning in asymptomatic carotid artery stenosis

Cortical thinning is a potentially important biomarker, but the pathophysiology in cerebrovascular disease is unknown. We investigated the association between regional cortical blood flow and regional cortical thickness in patients with asymptomatic unilateral high-grade internal carotid artery disease without stroke. Twenty-nine patients underwent high resolution anatomical and single-delay, pseudocontinuous arterial spin labeling magnetic resonance imaging with partial volume correction to assess gray matter baseline flow. Cortical thickness was estimated using Freesurfer software, followed by co-registration onto each patient's cerebral blood flow image space. Paired t-tests assessed regional cerebral blood flow in motor cortex (supplied by the carotid artery) and visual cortex (indirectly supplied by the carotid) on the occluded and unoccluded side. Pearson correlations were calculated between cortical thickness and regional cerebral blood flow, along with age, hypertension, diabetes and white matter hyperintensity volume. Multiple regression and generalized estimating equation were used to predict cortical thickness bilaterally and in each hemisphere separately. Cortical blood flow correlated with thickness in motor cortex bilaterally (p = 0.0002), and in the occluded and unoccluded sides individually; age (p = 0.002) was also a predictor of cortical thickness in the motor cortex. None of the variables predicted cortical thickness in visual cortex. Blood flow was significantly lower on the occluded versus unoccluded side in the motor cortex (p<0.0001) and in the visual cortex (p = 0.018). On average, cortex was thinner on the side of occlusion in motor but not in visual cortex. The association between cortical blood flow and cortical thickness in carotid arterial territory with greater thinning on the side of the carotid occlusion suggests that altered cerebral hemodynamics is a factor in cortical thinning.

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.

Reliability of an automatic classifier for brain enlarged perivascular spaces burden and comparison with human performance

In the brain, enlarged perivascular spaces (PVS) relate to cerebral small vessel disease (SVD), poor cognition, inflammation and hypertension. We propose a fully automatic scheme that uses a support vector machine (SVM) to classify the burden of PVS in the basal ganglia (BG) region as low or high. We assess the performance of three different types of descriptors extracted from the BG region in T2-weighted MRI images: (i) statistics obtained from Wavelet transform's coefficients, (ii) local binary patterns and (iii) bag of visual words (BoW) based descriptors characterizing local keypoints obtained from a dense grid with the scale-invariant feature transform (SIFT) characteristics. When the latter were used, the SVM classifier achieved the best accuracy (81.16%). The output from the classifier using the BoW descriptors was compared with visual ratings done by an experienced neuroradiologist (Observer 1) and by a trained image analyst (Observer 2). The agreement and cross-correlation between the classifier and Observer 2 (κ = 0.67 (0.58-0.76)) were slightly higher than between the classifier and Observer 1 (κ = 0.62 (0.53-0.72)) and comparable between both the observers (κ = 0.68 (0.61-0.75)). Finally, three logistic regression models using clinical variables as independent variable and each of the PVS ratings as dependent variable were built to assess how clinically meaningful were the predictions of the classifier. The goodness-of-fit of the model for the classifier was good (area under the curve (AUC) values: 0.93 (model 1), 0.90 (model 2) and 0.92 (model 3)) and slightly better (i.e. AUC values: 0.02 units higher) than that of the model for Observer 2. These results suggest that, although it can be improved, an automatic classifier to assess PVS burden from brain MRI can provide clinically meaningful results close to those from a trained observer.

An Examination of Brain Abnormalities and Mobility in Individuals with Mild Cognitive Impairment and Alzheimer's Disease

Background: Mobility changes are concerning for elderly patients with cognitive decline. Given frail older individuals' vulnerability to injury, it is critical to identify contributors to limited mobility. Objective: To examine whether structural brain abnormalities, including reduced gray matter volume and white matter hyperintensities, would be associated with limited mobility among individuals with cognitive impairment, and to determine whether cognitive impairment would mediate this relationship. Methods: Thirty-four elderly individuals with mild cognitive impairment (MCI) and Alzheimer's disease underwent neuropsychological evaluation, mobility assessment, and structural brain neuroimaging. Linear regression was conducted with predictors including gray matter volume in six regions of interest (ROI) and white matter hyperintensity (WMH) burden, with mobility measures as outcomes. Results: Lower gray matter volume in caudate nucleus was associated with slower speed on a functional mobility task. Higher cerebellar volume was also associated with slower functional mobility. White matter hyperintensity burden was not significantly associated with mobility. Conclusion: Our findings provide evidence for associations between subcortical gray matter volume and speed on a functional mobility task among cognitively impaired individuals.

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

PURPOSE

White matter hyperintensities (WMH) are seen on FLAIR-MRI in several neurological disorders, including multiple sclerosis, dementia, Parkinsonism, stroke and cerebral small vessel disease (SVD). WMHs are often used as biomarkers for prognosis or disease progression in these diseases, and additionally longitudinal quantification of WMHs is used to evaluate therapeutic strategies. Human readers show considerable disagreement and inconsistency on detection of small lesions. A multitude of automated detection algorithms for WMHs exists, but since most of the current automated approaches are tuned to optimize segmentation performance according to Jaccard or Dice scores, smaller WMHs often go undetected in these approaches. In this paper, the authors propose a method to accurately detect all WMHs, large as well as small.

METHODS

A two-stage learning approach was used to discriminate WMHs from normal brain tissue. Since small and larger WMHs have quite a different appearance, the authors have trained two probabilistic classifiers: one for the small WMHs (⩽3 mm effective diameter) and one for the larger WMHs (>3 mm in-plane effective diameter). For each size-specific classifier, an Adaboost is trained for five iterations, with random forests as the basic classifier. The feature sets consist of 22 features including intensities, location information, blob detectors, and second order derivatives. The outcomes of the two first-stage classifiers were combined into a single WMH likelihood by a second-stage classifier. Their method was trained and evaluated on a dataset with MRI scans of 362 SVD patients (312 subjects for training and validation annotated by one and 50 for testing annotated by two trained raters). To analyze performance on the separate test set, the authors performed a free-response receiving operating characteristic (FROC) analysis, instead of using segmentation based methods that tend to ignore the contribution of small WMHs.

RESULTS

Experimental results based on FROC analysis demonstrated a close performance of the proposed computer aided detection (CAD) system to human readers. While an independent reader had 0.78 sensitivity with 28 false positives per volume on average, their proposed CAD system reaches a sensitivity of 0.73 with the same number of false positives.

CONCLUSIONS

The authors have developed a CAD system with all its ingredients being optimized for a better detection of WMHs of all size, which shows performance close to an independent reader.

White Matter Hyperintensities in Older Adults and Motoric Cognitive Risk Syndrome

INTRODUCTION

Motoric cognitive risk (MCR) syndrome is a recently described pre-dementia syndrome characterized by slow gait and cognitive complaints that has been implicated as a predictor of cognitive decline and dementia in older adults. Previous work suggests that cerebrovascular disease is associated with MCR. White matter hyperintensities (WMH) are postulated to be a product of cerebrovascular disease, and have been associated with impaired mobility and impaired cognition. This study aimed to determine if MCR is associated with regional WMH.

METHODS

Two cross-cultural cohorts of non-demented older adults were examined: 174 from a French memory clinic (62.1% male, mean age 70.7 ± 4.3 years) and 184 from an Indian community-dwelling cohort (55.4% male, mean age 66.2 ± 5.2 years). Participants were evaluated for slow gait, cognitive complaints, and regional WMH via MRI (fluid attenuated inversion recovery) FLAIR sequence.

RESULTS

Overall, 20.7% of participants met criteria for MCR, and 72.9% of participants had WMH on FLAIR. WMH in the frontal, parieto-occipital, temporal, basal ganglia, cerebellum, or brainstem were not associated with MCR in either of the two cohorts.

CONCLUSION

WMH was not significantly associated with MCR in this studied sample of participants, suggesting that other cerebrovascular pathophysiological mechanisms, or combination of mechanisms, might underlie MCR.