Medical imaging of the heart :quantitative analysis of three-dimensional echocardiographic images

Abstract

Accurate, reproducible determination of cardiac chamber volume, especially left ventricular (LV) volume, is important for clinical assessment, risk stratification, selection of therapy, and serial monitoring of patients with cardiovascular disease. Echocardiography is the most widely used imaging modality in the clinical diagnosis of left ventricular function abnormalities. In the last 15 years, developments in real time three-dimensional echocardiography (RT3DE) have achieved superior accuracy and reproducibility compared with conventional two-dimensional echocardiography (2DE) for measurement of left ventricular volume and function. However, RT3DE suffers from the limitations inherent to the ultrasonic imaging modality and the cost of increased effort of data handling and image analysis. There were two aims of this research project. Firstly, it aimed to develop different new semi-automated algorithms for LV endocardial surface delineation, LV volume and EF quantification from clinical RT3DE images. Secondly, through assessing and comparing the performance of these algorithms in the aspects of accuracy and reproducibility, this project aimed to investigate what factors in real time 3D echo images influenced the performance of each algorithm, so that advantages and drawbacks of 3D echo images can be better understood. The basic structure of the content of this thesis is as follows: Chapter 1 introduces the background and the aims of this project. Chapter 2 describes the development of the new semi-automated algorithms. Chapter 3 to Chapter 6 presents the four studies designed to assess and compare the accuracy and reproducibility of each algorithm. These studies were the balloon phantom study, the tissue-mimicking phantom study, the clinical cardiac magnetic resonance images study and the clinical contrast enhanced 3D stress echo images study. Chapter 7 summarises all these studies, draws conclusions, and describes future work. In conclusion, it has been shown that the semi-automated algorithms can measure LV volume and EF quantitatively in clinical 3D echo images. To achieve better accuracy and reproducibility, 3D echo images should be analysed from all three dimensions.