Precision in the estimation of the severity of coronary artery disease is critical for evaluating the long-term effects of therapeutic interventions such as diet, lipid-lowering drugs, coronary bypass surgery, and balloon angioplasty (PTCA), as well as assessing prognosis and choosing the appropriate form of therapy for each individual patient. Visual interpretation of the coronary arteriogram is today the gold standard for defining the pathoanatomy in coronary heart disease. There are inherent limitations, however, in this method of interpretation. Typically, the observer will derive an estimate of the percent of vessel diameter narrowing by looking at an averaging 2 views taken at right angles to each other. A significant error of estimating lumen-size may result when an atherosclerotic plaque is not perfectly round or concentrically located within the lumen; in fact, the majority of lesions are eccentric and have irregular shapes. This proposal is to develop an improved, objective method for 3-dimensional reconstruction of coronary angiographic data. This will be achieved by applying refinements of a novel algorithm to already developed computerized densitometric techniques. The algorithm is rapid, and may readily be applied to a large number of image frames. The-algorithm is shown to provide accurate measurements of cross- sectional shape at very low quantum counts. A new anti=warping transformation is presented and designed to invert precisely the effects of patient and heart motion. The goal of the effort is to develop a real-time algorithm for angiographic data, which will allow for reliable decisions for choosing therapy, and immediate decisions while doing balloon angioplasty. The algorithm is expected to have widespread application in industry.
