Automated morphometric analysis of coronary artery lesions: an extension of quantitative coronary arteriography

The authors describe an extension to standard quantitative coronary artery analysis which provides automated objective quantitative evaluation of morphologic features of coronary-artery lesions. Standard edge detection has been enhanced with a sequential edge-linking algorithm to provide increased spatial sampling of arterial borders. These edges are analyzed using previously developed shape-analysis techniques. Morphometric parameters describing edge features such as the presence of thrombi, lumen roughness, and the abruptness of the stenosis are extracted and quantified. The performance of the method was evaluated by substituting mathematical functions for artery edges. An initial clinical study showed significant differences in morphometric parameters measured in patients with stable angina as compared to patients with unstable angina. This work demonstrates that shaped analysis of adequately sampled arterial borders provides clinically meaningful morphometric information