Advanced coronary CT angiography image processing techniques.

摘要

Computer-aided cardiac image analysis obtained by various modalities plays an important role in the early diagnosis and treatment of cardiovascular disease. Numerous computerized methods have been developed to tackle this problem. Recent studies employ sophisticated techniques using available cues from cardiac anatomy such as geometry, visual appearance, and prior knowledge. Especially, visual analysis of three-dimensional (3D) coronary computed tomography angiography (CCTA) remains challenging due to large number of image slices and tortuous character of the vessels. In this thesis, we focus on cardiac applications associated with coronary artery disease and cardiac arrhythmias, and study the related computer-aided diagnosis problems from computed tomography angiography (CCTA). First, in Chapter 2, we provide an overview of cardiac segmentation techniques in all kinds of cardiac image modalites, with the goal of providing useful advice and references. In addition, we describe important clinical applications, imaging modalities, and validation methods used for cardiac segmentation.;In Chapter 3, we propose a robust, automated algorithm for unsupervised computer detection of coronary artery lesions from CCTA. Our knowledge-based algorithm consists of centerline extraction, vessel classification, vessel linearization, lumen segmentation with scan-specific lumen attenuation ranges, and lesion location detection. Presence and location of lesions are identified using a multi-pass algorithm which considers expected or "normal" vessel tapering and luminal stenosis from the segmented vessel. Expected luminal diameter is derived from the scan by automated piecewise least squares line fitting over proximal and mid segments (67%) of the coronary artery considering the locations of the small branches attached to the main coronary arteries. We applied this algorithm to 42 CCTA patient datasets, acquired with dual-source CT, where 21 datasets had 45 lesions with stenosis 25%. The reference standard was provided by visual and quantitative identification of lesions with any stenosis ≥25% by 3 expert observers using consensus reading. Our algorithm identified 43 lesions (93%) confirmed by the expert observers. There were 46 additional lesions detected; 23 out of 46 (50%) of these were less- stenosed lesions. When the artery was divided into 15 coronary segments according to standard cardiology reporting guidelines, per-segment basis, sensitivity was 93% and per-segment specificity was 81%. Our algorithm shows promising results in the detection of obstructive and nonobstructive CCTA lesions.;In Chapter 4, we propose a novel low-radiation dose CCTA denoising algorithm. Our aim in this study was to optimize and validate an adaptive de-noising algorithm based on Block-Matching 3D, for reducing image noise and improving left ventricular assessment, in low-radiation dose CCTA. In this study, we describe the denoising algorithm and its validation, with low-radiation dose coronary CTA datasets from consecutive 7 patients. We validated the algorithm using a novel method, with the myocardial mass from the low-noise cardiac phase as a reference standard, and objective measurement of image noise. After denoising, the myocardial mass was not statistically different by comparison of individual data points by the students' t-test (130.9+/-31.3g in low-noise 70% phase vs 142.1+/-48.8g in the denoised 40% phase, p= 0.23). Image noise improved significantly between the 40% phase and the denoised 40% phase by the students' t-test, both in the blood pool (p-value <0.0001) and myocardium (p-value <0.0001). We optimized and validated an adaptive BM3D denoising algorithm for coronary CTA. This new method reduces image noise and has the potential for improving myocardial function assessment from low-dose coronary CTA.;In Chapter 5, we propose a novel machine learning technique to detect coronary arterial lesions with stenosis ≥25% from CCTA. We proposed an improved automated algorithm for detection of coronary arterial lesions from coronary CT angiography, by adapting a machine learning algorithm on the same data used in Chapter 3, which was described based on [139]. Our structured learning-based algorithm consists of two stages: (1) Dividing each coronary artery into small volume patches, and integrating several quantitative geometric and shape features for coronary arterial lesions in each small volume patch by Support Vector Machine (SVM) algorithm, (2) Applying SVM-based decision fusion algorithm to combine a formula-based analytic method and a learning-based method in the stage (1). We applied this algorithm to 42 CCTA patient datasets, acquired with dual-source CT, where 21 datasets had 45 lesions with stenosis ≥25%. The reference standard was provided by visual and quantitative identification of lesions with any stenosis ≥25% by three expert readers using consensus reading. When the artery was divided into 15 coronary segments according to standard cardiology reporting guidelines, per-segment basis, the sensitivity was 93%, and the specificity was 95% using 10-fold cross-validation. In conclusions, we developed a novel machine learning based algorithm for detection of coronary arterial lesions from CCTA. The proposed structured learning algorithm performed with high sensitivity and high specificity as compared to 3 experienced expert readers.