Discrimination of human arterial pathology by computer processing of angiograms for serial assessment of atherosclerosis change
✍ Scribed by Samuel H. Brooks; Donald W. Crawford; Robert H. Selzer; David H. Blankenhorn; Robert Barndt Jr.
- Publisher
- Elsevier Science
- Year
- 1978
- Tongue
- English
- Weight
- 742 KB
- Volume
- 11
- Category
- Article
- ISSN
- 0010-4809
No coin nor oath required. For personal study only.
✦ Synopsis
Femoral angiograms were made in 21 cadavers under conditions which simulated clinical films. Then these arteries were excised, laid open, and color photographed along side of corresponding casts of the arterial lumen. Twenty-seven segments, 5 cm long, were classified as exemplary of normal vessel, uncomplicated stenosis, or hemorrhagic ulceration. The corresponding angiogram images of these arterial segments were digitized and processed to yield outlines of the vessel edges, and from these, computer measures were developed which described how these outlines deviated from ideal ones. Six of these measures were selected by means of stepwise discriminant analysis to distinguish the type of disease along the 27 exemplary segments, and 96% were correctly class&& Computer processing of angiograms shows promise for discriminating the nature of atherosclerotic lesions.
We have developed a technique which involves computer processing of angiograms to assess the degree of atherosclerosis (I, 2). This method was designed primarily for clinical use; but to provide adequate standardization it has been calibrated against autopsy findings using the International Atherosclerosis Project (IAP) grading procedure as a standard (3). A clinical feasibility trial, now completed, indicated that changes in relatively early atherosclerosis could be detected in living man. (4). In its current stage of development the computer method is most readily applicable to intermediate stages of atherosclerosis. This paper projects means to expand the method's capability for evaluation of later stages of atherosclerosis through lesion discrimination.
During calibration of our method we became aware of conceptual difficulties which arise when a grade such as the IAP is to be used as a single metric to calibrate procedures planned for serial evaluation. The IAP grading plan was devised as a definitive measure of atherosclerosis determined at autopsy which could be