Coronary heart disease (CHD) is the most prevalent cause of death worldwide. Atherosclerosis which is the condition of plaque buildup on the inside of the coronary artery wall is the main cause of CHD. Rupture of unstable atherosclerotic coronary plaque is known to be the cause of acute coronary syndrome. Vulnerability of atherosclerotic plaque has been related to a large lipid core covered by a fibrous cap. Non-invasive assessment of plaque characterization is necessary due to prognostic importance of early stage identification. The purpose of this study is to use the additional attenuation data provided by dual energy computed tomography (DECT) for plaque characterization. We propose to train supervised learners on pixel values recorded from DECT monochromatic X-ray and material basis pairs images, for more precise classification of fibrous and lipid plaques. The interaction of the pixel values from different image types is taken into consideration, as single pixel value might not be informative enough to separate fibrous from lipid. Organic phantom plaques scanned in a fabricated beating heart phantom were used as ground truth to train the learners. Our results show that support vector machines, artificial neural networks and random forests provide accurate results both on phantom and patient data.
- Supervised learning
- dual energy computed tomography (DECT)
ASJC Scopus subject areas
- Computer Science Applications
- Electrical and Electronic Engineering
- Health Information Management