Identification of Cerebral Small Vessel Disease Using Multiple Instance Learning
Cerebral small vessel disease (SVD) is a common cause of ageing-associated physical and cognitive impairment. Identifying SVD is important for both clinical and research purposes but is usually dependent on radiologists’ evaluation of brain scans. Compute
- PDF / 561,652 Bytes
- 8 Pages / 439.363 x 666.131 pts Page_size
- 40 Downloads / 242 Views
Imperial College London, London, United Kingdom Northwick Park Hospital, London, United Kingdom Imperial College Healthcare NHS Trust, London, United Kingdom 2
3
Abstract. Cerebral small vessel disease (SVD) is a common cause of ageing-associated physical and cognitive impairment. Identifying SVD is important for both clinical and research purposes but is usually dependent on radiologists’ evaluation of brain scans. Computer tomography (CT) is the most widely used brain imaging technique but for SVD it shows a low signal-to-noise ratio, and consequently poor inter-rater reliability. We therefore propose a novel framework based on multiple instance learning (MIL) to distinguish between absent/mild SVD and moderate/severe SVD. Intensity patches are extracted from regions with high probability of containing lesions. These are then used as instances in MIL for the identification of SVD. A large baseline CT dataset, consisting of 590 CT scans, was used for evaluation. We achieved approximately 75% accuracy in classifying two different types of SVD, which is high for this challenging problem. Our results outperform those obtained by either standard machine learning methods or current clinical practice.
1
Introduction
The World Health Organization (WHO) states that stroke is the second major cause of death in the world during 2000 and 2012. Stroke, which is a cerebrovascular accident, is the loss of brain function caused by the lack of blood supply [16]. It may lead to long-term disability. Ischemic stroke and hemorrhagic stroke are two different categories of strokes that require different treatments [6]. Ischemic stroke accounts for approximately 80% of all strokes [7]. Intravenous thrombolysis with recombinant tissue plasminogen activator (rt-PA) is the recommended therapy for acute ischemic stroke that reduces severe disability but causes deterioration due to symptomatic intracranial hemorrhage (SICH) in approximately 6% [18]. [4] demonstrates that cerebral SVD is associated with increased risk of ischemic stroke. Hypertensive SVD is the most common mechanism of hemorrhagic stroke [6]. In order to reduce the rate of SICH, which is associated with the worst outcome of stroke, management of SVD is pivotal. Cerebral SVD refers to a group of pathological aetiologies that affect the brain [12]. However, in this c Springer International Publishing Switzerland 2015 N. Navab et al. (Eds.): MICCAI 2015, Part I, LNCS 9349, pp. 523–530, 2015. DOI: 10.1007/978-3-319-24553-9_64
524
L. Chen et al.
Fig. 1. Examples of CT images of the brain: (a) normal brain appearance, (b) brain with mild cerebral SVD, (c) brain with moderate SVD, and (d) cerebrum with severe SVD. The red arrows point out where the lesions are.
paper we will use the term to describe ischemic consequences of white matter (WM) lesions. Figure 1 presents examples of cerebrums with different kinds of SVD. Advanced neuroimaging techniques have been widely used in the diagnosis of stroke. It is normally recommended that patients should undergo either magnetic reson
Data Loading...
