eICAB: A novel deep learning pipeline for Circle of Willis multiclass segmentation and analysis

BACKGROUND

The accurate segmentation, labeling and quantification of cerebral blood vessels on MR imaging is important for basic and clinical research, yet results are not generalizable, and often require user intervention. New methods are needed to automate this process.

PURPOSE

To automatically segment, label and quantify Circle of Willis (CW) arteries on Magnetic Resonance Angiography images using deep convolutional neural networks.

MATERIALS AND METHODS

MRA images were pooled from three public and private databases. A total of 116 subjects (mean age 56 years ± 21 [standard deviation]; 72 women) were used to make up the training set (N=101) and the testing set (N=15). In each image, fourteen arterial segments making up or surrounding the CW were manually annotated and validated by a clinical expert. Convolutional neural network (CNN) models were trained on a training set to be finally combined in an ensemble to develop eICAB. Model performances were evaluated using (1) quantitative analysis (dice score on test set) and (2) qualitative analysis (external datasets, N=121). The reliability was assessed using multiple MRAs of healthy participants (ICC of vessel diameters and volumes on test-retest).

RESULTS

Qualitative analysis showed that eICAB correctly predicted the large, medium and small arteries in 99±0.4%, 97±1% and 88±7% of all images, respectively. For quantitative assessment, the average dice score coefficients for the large (ICAs, BA), medium (ACAs, MCAs, PCAs-P2), and small (AComm, PComm, PCAs-P1) vessels were 0.76±0.07, 0.76±0.08 and 0.41±0.27, respectively. These results were similar and, in some cases, statistically better (p<0.05) than inter-expert annotation variability and robust to image SNR. Finally, test-retest analysis showed that the model yielded high diameter and volume reliability (ICC=0.99).

CONCLUSION

We have developed a quick and reliable open-source CNN-based method capable of accurately segmenting and labeling the CW in MRA images. This method is largely independent of image quality. In the future, we foresee this approach as a critical step towards fully automated analysis of MRA databases in basic and clinical research.

The Associations Between Intracranial Stenosis, Brain Amyloid-beta, and Cognition in a Memory Clinic Sample

BACKGROUND

Intracranial stenosis (ICS) and brain amyloid-beta (Aβ) have been associated with cognition and dementia. We aimed to investigate the association between ICS and brain Aβ and their independent and joint associations with cognition.

METHODS

We conducted a cross-sectional study of 185 patients recruited from a memory clinic. ICS was measured on 3-dimensional time-of-flight magnetic resonance angiography and defined as stenosis ≥50%. Brain Aβ was measured with [ 11 C] Pittsburgh compound B-positron emission tomography imaging. Cognition was assessed with a locally validated neuropsychological battery.

RESULTS

A total of 17 (9.2%) patients had ICS, and the mean standardized uptake value ratio was 1.4 (±0.4 SD). ICS was not significantly associated with brain Aβ deposition. ICS was significantly associated with worse global cognition (β: -1.26, 95% CI: -2.25; -0.28, P =0.013), executive function (β: -1.04, 95% CI: -1.86; -0.22, P =0.015) and visuospatial function (β: -1.29, 95% CI: -2.30; -0.27, P =0.015). Moreover, in ICS patients without dementia (n=8), the presence of Aβ was associated with worse performance on visuomotor speed.

CONCLUSIONS

ICS was significantly associated with worse cognition and showed interaction with brain Aβ such that patients with both pathologies performed worse on visuomotor speed specifically in those without dementia. Further studies may clarify if ICS and brain Aβ deposition indeed have a synergistic association with cognition.

Resolving inflammatory links between myocardial infarction and vascular dementia

Myocardial infarction is associated with increased risk for vascular dementia. In both myocardial infarction and vascular dementia, there is evidence that elevated inflammatory biomarkers are associated with worsened clinical outcomes. Myocardial infarction leads to a systemic inflammatory response, which may contribute to recruitment or activation of myeloid cells, including monocytes, microglia, and perivascular macrophages, within the central nervous system. However, our understanding of the causative roles for these cells linking cardiac injury to the development and progression of dementia is incomplete. Herein, we provide an overview of inflammatory cellular and molecular links between myocardial infarction and vascular dementia and discuss strategies to resolve inflammation after myocardial infarction to limit neurovascular injury.

Intracranial Atherosclerotic Stenosis is Related to Post-stroke Cognitive Impairment: A Cross-sectional Study of Minor Stroke

BACKGROUND

Intracranial Atherosclerotic Stenosis (ICAS) is an important risk factor for cognitive impairment. However, it is unclear whether patients with ICAS are more likely to develop cognitive impairment after an acute, non-disabling ischemic stroke (minor stroke).

OBJECTIVE

We aimed to investigate the association between ICAS and post-stroke cognitive impairment.

METHODS

In this cross-sectional study, patients with acute, non-disabling ischemic stroke underwent two cognitive tests and imaging evaluation for ICAS, within two weeks after the stroke. To determine the association between ICAS and post-stroke cognitive impairment, we performed a multivariate logistic regression analysis adjusted for several demographic and vascular risk factors.

RESULTS

Of the 164 patients with minor stroke in this study, 98 (59.76%) were diagnosed with poststroke cognitive impairment (Montreal Cognitive Assessment score<26). After adjusting for potential confounders, we found that patients with ICAS were more likely to develop cognitive impairment after an acute, non-disabling ischemic stroke, compared to patients without ICAS (Odds Ratio: 2.13; 95% Confidence Interval: 1.07-4.26), and underperformed in the tests of visuospatial and executive function.

CONCLUSION

In this cross-sectional study of a population that has experienced a minor stroke, our findings demonstrated a positive association between ICAS and post-stroke cognitive impairment.

The effect of intracranial stenosis on cognitive decline in a memory clinic cohort

BACKGROUND AND PURPOSE

Intracranial stenosis (ICS) is a risk factor for cognitive impairment and dementia in cross-sectional studies. However, data examining the effect of ICS on cognitive decline are limited. We investigated the effect of ICS on cognition over a period of 3 years in a memory clinic cohort.

METHODS

Patients were recruited from the National University Hospital in Singapore. Data were collected using a standardised questionnaire, physical examination, and 3-T magnetic resonance imaging (MRI) at baseline. ICS was defined as arterial narrowing that exceeded 50% of the luminal diameter in any intracranial vessel. Cognition was measured at baseline and annually for 3 years using the Mini-Mental State Examination, the Montreal Cognitive Assessment, and a detailed neuropsychological test battery. The association between ICS and cognitive decline was analysed using generalised estimating equations.

RESULTS

A total of 364 patients were included in the analysis. The mean (±SD) age was 71.9 (±8.0) years, and 164 (45.1%) patients were male. A total of 66 (18.1%) patients had ICS. ICS was associated with worse executive function (β = -0.37, 95% confidence interval = -0.68 to -0.05, p = 0.022) and modified the effect of follow-up time on memory (p = 0.005) and visuomotor speed (p = 0.047). These results remained significant after controlling for demographics, overall diagnosis, cardiovascular risk factors, and MRI markers of cerebrovascular disease.

CONCLUSIONS

Intracranial stenosis was independently associated with worse executive function across all time points, and cognitive decline in memory and visuomotor speed over 3 years of follow-up. This suggests that ICS may be a useful indicator of vascular brain damage leading to cognitive decline and may warrant consideration of antiatherosclerotic treatment in clinical trials.

Impact of cerebral large-artery disease and blood flow in the posterior cerebral artery territory on cognitive function

OBJECTIVES

The purpose of this study was to elucidate the association of cerebral large artery disease (CLAD) with cerebral blood flow (CBF) in the posterior cerebral artery (PCA) territory and cognitive performance.

METHOD

We prospectively registered patients with CLAD who had internal carotid or middle cerebral artery (MCA) with the degree of stenosis ≥50%. Automated brain segmentation was used to quantify CBF in the thalamus, hippocampus, and PCA and MCA territories. We measured cognitive function of patients using the Wechsler Memory Scale Revised (WMS-R), the Mini-Mental State Examination (MMSE), and the Montreal Cognitive Assessment. Patients were divided into 3 groups according to CBF of the cortical and subcortical PCA territory.

RESULTS

There were 60 patients included in this study. The degree of stenosis was significantly correlated with CBF in the PCA territory (Γ = 0.35, P = .006) and hippocampus (Γ = 0.34, P = .008). Verbal memory, general memory, and reproduction on WMS-R and MMSE were significantly reduced areas with low CBF in the PCA territory compared with areas with middle and high CBF.

CONCLUSIONS

CBF of the PCA territory was significantly inversely correlated with the degree of stenosis in CLAD patients. Low CBF of the PCA territory was significantly associated with reduced cognitive and memory functions.

Intracranial stenosis in cognitive impairment and dementia

Intracranial stenosis is a common vascular lesion observed in Asian and other non-Caucasian stroke populations. However, its role in cognitive impairment and dementia has been under-studied. We, therefore, examined the association of intracranial stenosis with cognitive impairment, dementia and their subtypes in a memory clinic case-control study, where all subjects underwent detailed neuropsychological assessment and 3 T neuroimaging including three-dimensional time-of-flight magnetic resonance angiography. Intracranial stenosis was defined as ≥50% narrowing in any of the intracranial arteries. A total of 424 subjects were recruited of whom 97 were classified as no cognitive impairment, 107 as cognitive impairment no dementia, 70 vascular cognitive impairment no dementia, 121 Alzheimer's Disease, and 30 vascular dementia. Intracranial stenosis was associated with dementia (age/gender/education - adjusted odds ratios (OR): 4.73, 95% confidence interval (CI): 1.93-11.60) and vascular cognitive impairment no dementia (OR: 3.98, 95% CI: 1.59-9.93). These associations were independent of cardiovascular risk factors and MRI markers. However, the association with Alzheimer's Disease and vascular dementia became attenuated in the presence of white matter hyperintensities. Intracranial stenosis is associated with vascular cognitive impairment no dementia independent of MRI markers. In Alzheimer's Disease and vascular dementia, this association is mediated by cerebrovascular disease. Future studies focusing on perfusion and functional markers are needed to determine the pathophysiological mechanism(s) linking intracranial stenosis and cognition so as to identify treatment strategies.

Intracranial atherosclerosis and dementia: The Atherosclerosis Risk in Communities (ARIC) Study

OBJECTIVE

To explore the association of intracranial atherosclerotic disease (ICAD) with mild cognitive impairment (MCI) and dementia.

METHODS

From 2011 to 2013, 1,744 participants completed high-resolution vessel wall MRI from the population-based Atherosclerosis Risk in Communities Study by a sampling strategy that allowed weighting back to the cohort. We defined ICAD by plaque features (presence, territory, stenosis, number). Trained clinicians used an algorithm incorporating information from interviews and neuropsychological and neurologic examinations to adjudicate for MCI and dementia. We determined the relative prevalence ratio (RPR) of MCI or dementia after adjusting for risk factors at midlife using multinomial logistic regression.

RESULTS

A total of 601 (34.5%) participants had MCI (mean age ± SD, 76.6 ± 5.2 years), 83 (4.8%) had dementia (79.1 ± 5.3 years), and 857 (49.1%) were current or former smokers. Anterior cerebral artery (ACA) plaque (adjusted RPR 3.81, 95% confidence interval [CI] 1.57-9.23), >2 territories with plaque (adjusted RPR 2.12, 95% CI 1.00-4.49), and presence of stenosis >50% (adjusted RPR 1.92, 95% CI 1.01-3.65) were associated with increased prevalence of dementia in separate models. Posterior cerebral artery plaque was associated with MCI but did not reach statistical significance for dementia (adjusted RPR MCI 1.43, 95% CI 1.04-1.98; adjusted RPR dementia 1.58, 95% CI 0.79-2.85). There were no associations with middle cerebral artery atherosclerotic lesions or cognitive impairment. Many participants had plaque in >1 territory (n = 291, 46%) and participants with ACA plaques (n = 69) had the greatest number of plaques in other territories (mean 6.0, SD 4.4).

CONCLUSIONS

This study demonstrates associations between ICAD and clinical MCI and dementia.

Correlations between homocysteine and grey matter volume in patients with Alzheimer's disease

BACKGROUND

Previous studies have reported that elevated total homocysteine levels are associated with cognitive dysfunction. However, few studies have examined the radiological markers of associated neuropathology in Alzheimer's disease (AD). We hypothesized that elevated levels of homocysteine are associated with cerebral grey matter volume loss. We compared the grey matter in a high homocysteine group and a normal homocysteine group using an optimized voxel-based morphometry.

METHODS

The study included 79 patients with AD who were divided into two groups: a high homocysteine group and a normal homocysteine group. The participants underwent brain magnetic resonance imaging using a standardized protocol and neurocognitive evaluation. Homocysteine tests and other routine laboratory examinations for dementia assessment were carried out in all patients.

RESULTS

There was no significant difference in grey matter volume between the patients with high homocysteine levels and those with normal homocysteine levels. A multiple regression analysis also revealed that the levels of homocysteine were not associated with the grey matter volume in patients with AD. Homocysteine levels were not correlated significantly with Mini-Mental State Examination, Global Deterioration Scale, or Clinical Dementia Rating.

CONCLUSION

Our results showed that elevated homocysteine levels are not associated with reduced cerebral grey matter volume in AD. Larger samples will be needed to assess potential correlations between homocysteine and neuroanatomical pathology in the future.

Comprehensive gene expression profiling reveals synergistic functional networks in cerebral vessels after hypertension or hypercholesterolemia

Atherosclerotic stenosis of cerebral arteries or intracranial large artery disease (ICLAD) is a major cause of stroke especially in Asians, Hispanics and Africans, but relatively little is known about gene expression changes in vessels at risk. This study compares comprehensive gene expression profiles in the middle cerebral artery (MCA) of New Zealand White rabbits exposed to two stroke risk factors i.e. hypertension and/or hypercholesterolemia, by the 2-Kidney-1-Clip method, or dietary supplementation with cholesterol. Microarray and Ingenuity Pathway Analyses of the MCA of the hypertensive rabbits showed up-regulated genes in networks containing the node molecules: UBC (ubiquitin), P38 MAPK, ERK, NFkB, SERPINB2, MMP1 and APP (amyloid precursor protein); and down-regulated genes related to MAPK, ERK 1/2, Akt, 26 s proteasome, histone H3 and UBC. The MCA of hypercholesterolemic rabbits showed differentially expressed genes that are surprisingly, linked to almost the same node molecules as the hypertensive rabbits, despite a relatively low percentage of ‘common genes’ (21 and 7%) between the two conditions. Up-regulated common genes were related to: UBC, SERPINB2, TNF, HNF4A (hepatocyte nuclear factor 4A) and APP, and down-regulated genes, related to UBC. Increased HNF4A message and protein were verified in the aorta. Together, these findings reveal similar nodal molecules and gene pathways in cerebral vessels affected by hypertension or hypercholesterolemia, which could be a basis for synergistic action of risk factors in the pathogenesis of ICLAD.