3-D RECONSTRUCTION AND AUTOMATIC TRACKING OF 2-D ECHOES

点击翻译按钮获取中文摘要

Long term objective: validate and provide new quantitative echocardiographic methods for the study of global and regional ventricular disease. Aim 1: To develop valid and precise measures of left ventricular volume mass, and endocardial surface area in infarcted and normal formalin-fixed and isolated ejecting hearts, intact animals, and man with a minimum of global geometric assumption. This will be accomplished by means of a spark-gap transducer locating system in concert with a two-dimensional echocardiographic contouring system, both developed in this laboratory. The result will be three-dimensional reconstruction of multiple two-dimensional echocardiographic images. For ventricular mass in intact dogs and patients, comparison will be made against the actual mass of the excised ventricle. Patients will consist of those undergoing cardiac transplantation; the mass of the explanted ventricle will be readily determined after surgery. The ability to obtain volume, mass and surface area in this way should provide numerous applications to the study of disease states resulting in regional deformation or dysfunction - such as coronary artery disease - where the utility of methods which rely on major global geometric assumptions would be limited. Aim 2: To extend and validate an existing computer system for automatic tracking of two-dimensional echocardiograms in animals and man. This will be done by a combination of cross-correlation mapping and edge-tracking techniques. Validation in animals will include images obtained in vitro and in vivo followed by clinical validation. Evaluation and validation of the accuracy and speed of automatic contouring will be made in four steps. 1) alignment accuracy and calibration; 2) formalin-fixed dog and pig hearts to provide echocardiograms from which left ventricular cavity and muscle volume can be measured and quantified; 3) studies using isolated supported ejecting dog hearts to compare both manual and automatically contoured echocardiographic data with absolute frame-by-frame volume data determined accurately with a volumetric chamber apparatus, in both infarcted and non-infarcted ventricles; 4) clinical echocardiograms of varying quality will be used to compare manual and automatic contouring techniques. Finally, we will implement the automatic contouring system for use in three-dimensional echocardiographic reconstruction (Aim 1). This system when validated will enable rapid highly consistent analysis of endocardial and epicardial contours, and allow automated acquisition of data necessary for analysis of regional wall motion, thickening, volume, and endocardial surface area.