It is proposed that additional analyses of the Cholesterol Lowering Atherosclerosis Study (CLAS) coronary data base be carried out by the California Institute of Technology (Caltech). A series of new coronary artery roughness measures tuned to coronary segment size will be recomputed from vessel diameter data. A new class of roughness measures based on fractal mathematics will also be computed. The power of segment-size-specific roughness and fractal derived measures to discriminate drug treated and control subjects will be compared with conventional computer derived measures such as percent lesion stenosis. All new roughness measures will be correlated with on-trial changes in lipid, lipoprotein cholesterol and apolipoprotein levels. The CLAS coronary angiographic data base includes 25 computer derived descriptors for 10,000 individual coronary artery image segments at baseline and two years from a 50% random sample of subjects (41 treatment, 40 control). CLAS investigators are now publishing evidence that percent stenosis measures show significant (p=.003) colestipol-niacin therapy benefit in confirmation of previous CLAS reports of film evaluation by human panel readers. This data also shows significant (p=.02) improvement in a roughness measure, TD50, derived from the early original postmortem femoral angiographic calibration studies and later approximately scaled to coronary dimensions. The new roughness measures under study will be computed from raw coronary edge point data already in the CLAS data base. It is hypothesized that coronary vessel edge roughness measures specifically tuned to coronary dimensions or based on fractal mathematics will offer improved computerized endpoints with potential for reducing the sample size of future clinical trials.
| Shircore, A M; Mack, W J; Selzer, R H et al. (1995) Compensatory vascular changes of remote coronary segments in response to lesion progression as observed by sequential angiography from a controlled clinical trial. Circulation 92:2411-8 |
