Cerebral microbleeds (CMBs) are small haemorrhages nearby blood vessels. They have been recognized as important diagnostic biomarkers for many cerebrovascular diseases and cognitive dysfunctions. In current clinical routine, CMBs are manually labelled by radiologists but this procedure is laborious, time-consuming, and error prone. In this paper, we propose a novel automatic method to detect CMBs from magnetic resonance (MR) images by exploiting the 3D convolutional neural network (CNN). Compared with previous methods that employed either low-level hand-crafted descriptors or 2D CNNs, our method can take full advantage of spatial contextual information in MR volumes to extract more representative high-level features for CMBs, and hence achieve a much better detection accuracy. To further improve the detection performance while reducing the computational cost, we propose a cascaded framework under 3D CNNs for the task of CMB detection. We first exploit a 3D fully convolutional network (FCN) strategy to retrieve the candidates with high probabilities of being CMBs, and then apply a well-trained 3D CNN discrimination model to distinguish CMBs from hard mimics. Compared with traditional sliding window strategy, the proposed 3D FCN strategy can remove massive redundant computations and dramatically speed up the detection process. We constructed a large dataset with 320 volumetric MR scans and performed extensive experiments to validate the proposed method, which achieved a high sensitivity of 93.16% with an average number of 2.74 false positives per subject, outperforming previous methods using low-level descriptors or 2D CNNs by a significant margin. The proposed method, in principle, can be adapted to other biomarker detection tasks from volumetric medical data.

The term vascular cognitive impairment (VCI) was introduced around the start of the new millennium and refers to the contribution of vascular pathology to any severity of cognitive impairment, ranging from subjective cognitive decline and mild cognitive impairment to dementia. Although vascular pathology is common in elderly individuals with cognitive decline, pure vascular dementia (that is, dementia caused solely by vascular pathology) is uncommon. Indeed, most patients with vascular dementia also have other types of pathology, the most common of which is Alzheimer disease (specifically, the diffuse accumulation of amyloid-β plaques and neurofibrillary tangles composed of tau). At present, the main treatment for VCI is prevention by treating vascular diseases and other risk factors for VCI, such as hypertension and diabetes mellitus. Despite the current paucity of disease-modifying pharmacological treatments, we foresee that eventually, we might be able to target specific brain diseases to prevent cognitive decline and dementia.

Background/Aims: To evaluate the psychometric properties of the Hong Kong Montreal Cognitive Assessment (HK-MoCA) in patients with cerebral small vessel disease (SVD). Methods: 40 SVD patients and 40 matched controls were recruited. Concurrent and criterion validity, inter-rater and test-retest reliability, internal consistency of the HK-MoCA were examined and clinical observations were made. Results: Performance on the HK-MoCA was significantly predicted by both executive (β = 0.23, p = 0.013) and non-executive (β = 0.64, p < 0.001) composite scores. It differentiated SVD patients from controls (area under the curve = 0.81, p < 0.001) with an optimal cutoff at 21/22. Reliability, internal consistency and clinical utility were good. Conclusion: The HK-MoCA is a useful cognitive screening instrument for use in SVD patients.

No abstract

Background and Purpose-Cerebral microbleeds (MBs) detected by gradient-echo MRI are considered evidence of advanced microangiopathy with potential for further bleeding. The goal of this study was to determine whether the presence of MBs is a risk factor for subsequent intracerebral hemorrhage among patients with acute ischemic stroke. Methods-We prospectively examined patients hospitalized with acute cerebral infarction with gradient-echo T2*-weighted MRI for the presence of MBs. We recorded demographics, medical history, and stroke severity. Patients were then followed up for the development of stroke, other vascular events, and death. Results-One hundred twenty-one consecutive patients with a mean age of 67.96Ϯ10.97 years were recruited. MBs were present in 43 patients (35.5%). During follow-up of 27.15Ϯ11.68 months, 16 patients had recurrent stroke. There was no difference between patients with or without MB for the development of ischemic stroke (5 and 6 respectively, Pϭ0.841). However, 4 patients (9.3%) with MBs and 1 patient (1.3%) without an MB had intracerebral hemorrhage during follow-up (Pϭ0.053). Of the 5 patients who developed subsequent intracerebral hemorrhages, 3 were treated with aspirin and 2 with anticoagulation. Two of the intracerebral hemorrhages occurred in the site where asymptomatic MBs were found at baseline. Conclusions-MBs appear to be a risk factor for subsequent intracerebral hemorrhage among patients with ischemic stroke in this small cohort of Chinese stroke patients. A large cohort study is required to confirm this observation.

No abstract

Background and Purpose The NINDS-CSN vascular cognitive impairment (VCI) Harmonization working group proposed a brief cognitive protocol for screening of VCI. We investigated the validity, reliability and feasibility of the Montreal Cognitive Assessment 5-minute protocol (MoCA 5-min protocol) administered over the telephone. Methods Four items examining attention, verbal learning and memory, executive functions/language and orientation were extracted from the MoCA to form the MoCA 5-min protocol. One hundred and four patients with stroke or TIA, including 53 with normal cognition (CDR 0) and 51 with cognitive impairment (CDR 0.5 or 1), were administered the MoCA in clinic and a month later, the MoCA 5-min protocol over the telephone. Results Administration of the MoCA 5-min protocol took five minutes over the telephone. Total score of the MoCA 5-min protocol correlated negatively with age (r=-0.36, p<0.001) and positively with years of education (r=0.41, p<0.001) but not with gender (rho=0.03, p=0.773). Total scores of the MoCA and MoCA 5-min protocol were highly correlated (r=0.87, p<0.001). The MoCA 5-min protocol performed equally well as the MoCA in differentiating patients with cognitive impairment from those without (AUC for MoCA 5-min protocol=0.78; MoCA=0.74, p>0.05 for difference; Cohen's d for group difference 0.801.13). It differentiated cognitively impaired patients with executive domain impairment from those without (AUC=0.89, p<0.001; Cohen's d=1.7 for group difference). 30-day test-retest reliability was excellent (Intraclass correlation coefficient=0.89). Conclusions The MoCA 5-min protocol is a free, valid and reliable cognitive screen for stroke and TIA. It is brief and highly feasible for telephone administration.

Chronic brain changes including WMCs, MTLA, and AD pathology are associated with incident dementia after stroke/TIA. Interventions targeting these chronic brain changes may reduce burden of vascular cognitive impairment.