Construction of four-dimensional dynamic simulation tool for assessment of left ventricular myocardial function

碩士 === 中原大學 === 生物醫學工程研究所 === 98 === Patients with cardiovascular disease such as cardiomyopathy and myocardial infarction were usually having abnormal myocardial motion. The cardiac contraction and relaxation information could be obtained speedily and safely by multiple slices CT imaging system wit...

Full description

Bibliographic Details
Main Authors: Ming-Huei Ma, 馬銘徽
Other Authors: Wei-Chih Hu
Format: Others
Language:zh-TW
Published: 2010
Online Access:http://ndltd.ncl.edu.tw/handle/57963150883514355955
Description
Summary:碩士 === 中原大學 === 生物醫學工程研究所 === 98 === Patients with cardiovascular disease such as cardiomyopathy and myocardial infarction were usually having abnormal myocardial motion. The cardiac contraction and relaxation information could be obtained speedily and safely by multiple slices CT imaging system with multiple phases. However, Four Dimension non-invasive cardiac medical images could not provide enough information of regional myocardial motion for physicians. Thus, to assess left ventricular function (LV) and regional myocardial motion, a dynamical simulating tool was constructed in this research. The 3D LV model was constructed from 4D MSCT images. For data pre-processing in this study, firstly, the reposition of the left ventricle in short axis view, the central axis of the 3D LV model was adjusted manually to be aligned with axis that was constructed using central point of short axis LV image from mitral valve to the apex of left ventricle. And, then, the short axis LV images were obtained by re-sampling. Each unit of myocardial motion was calibrated by assuming a rotating function based on volume time curve of LV ejection function to the unit. Thus, the LV motion could be simulated by shortening of long axis and the twist of myocardium using the rate of radius of change by Archimedan Spiral function. To illustrate the regional myocardial motion, the data was portraying with the common accepted 18 myocardial regions on a Bull eyes graphical figure with the lively video display. A high resolution of 900 segments illustration was also provided. The result showed that correlation between simulated and controlled volume-time curve (VTC) was highly correlated. The VTC was 0.99 (r=0.99) and rate of volume change (RVC) was 0.97 (r=0.97). This demonstrated that the high correlation of global LV motion between the controlled data and 3D LV motion model. For the regional LV motion, the correlation between simulated model and controlled motion data of 18 myocardial regions on a completed cardiac cycle was 0.87(r=0.87). The Bland-Altman analyses on the regional myocardial motion and regional ejection fraction were both within ±2STD. This demonstrated that the simulated model and the controlled data at every LV myocardial motion of whole cardiac cycle were similar. These results showed that a designated set of contraction functions was able to simulate the LV contraction that will be producing desired cardio output.