Although more than one million coronary arteriograms are performed in the U.S. annually, the potential of the angiogram to predict risk or outcome has probably not been full realized. In the past several years, there has been evolving evidence that atherosclerotic plaque surface morphology, rather than the degree of luminal obstruction, predicts unstable ischemic syndromes. Pathologic and angioscopic studies have shown that plaque disruption (fissuring, ulceration, or hemorrhage, with or without thrombosis) is present in an overwhelming percentage of patients presenting with sudden death, myocardial infarction or unstable angina. Under ideal angiographic imaging conditions, different lesion morphologies have different and sometimes characteristic appearances that can be recognized subjectively. Morphologic classification is highly dependent upon the quality of the angiographic image where noise, contrast, resolution, and projection--as well as the size of the vessel and lesion may limit the degree of observer confidence. These are the rational for this proposal where the major objective is to optimize angiographic detection and classification of coronary artery lesion morphology. Digital image processing will be used to address four long term goals: (1) to develop a basic understanding of visual perception of dynamically displayed cineangiograms; (2) to define the accuracy of morphological classification of coronary stenoses and how accuracy depends upon image enhancement and display; (3) to develop imaging and display techniques for maximizing this accuracy; and (4) to perform clinical validation of morphological classification. Detailed preliminary data are presented and a program of research is described which uses computer simulations, post-mortem histologic validation, and human angioscopic studies to quantify observer performance for angiographic detection of morphologic features. If successful, this grant will extend image signal detection theory to include cine, or dynamic, image display and develop specific and reliable methods to optimize the coronary angiograms for the detection and extractior of anatomic parameters of atherosclerotic plaque surface morphology. Thus, the potential health related impact of this research may be a more accurate means of identifying high risk coronary lesions and thus improve clinical decision making in patients with coronary artery disease.
| Eigler, N L; Eckstein, M P; Mahrer, K N et al. (1994) Improving detection of coronary morphological features from digital angiograms. Effect of stenosis-stabilized display. Circulation 89:2700-9 |
