Advances in two-dimensional quantitative coronary angiographic assessment of bifurcation lesions: improved small lumen diameter detection and automatic reference vessel diameter derivation

Abstract

To validate a new two dimensional (2-D) bifurcation quantitative coronary angiography (QCA) software. In the latest edition of the Cardiovascular Angiography Analysis System (CAAS 5.9; Pie Medical Imaging, Maastricht, The Netherlands) video-densitometric information is dynamically integrated into the edge-detection algorithm of 11- and 6-segment models to reduce overestimation of small diameters. Furthermore, automatic reference obstruction analysis was optimised. Values of the minimal lumen diameter (MLD), reference vessel diameter (RVD), percent diameter stenosis (DS) and bifurcation angle (BA) for the different bifurcation segment models were validated against precision manufactured plexiglass phantoms. In anteroposterior views, accuracy and precision (mean difference±SD) of 11- and 6-segment models for MLD were 0.013±0.082 mm vs. 0.003±0.100 mm, for RVD -0.030±0.047 mm vs. -0.029±0.045 mm and for DS -0.48±3.66% vs. -0.11±3.97%. In smaller vessel segments (true MLD <0.7 mm), MLD overestimation was reduced. Inter-observer variability for MLD, RVD and DS for either model was ≤0.052 mm, ≤0.043 mm and ≤2.24%, respectively. Agreement between models for MLD, RVD and DS was ±0.076 mm, ±0.021 mm and ±2.53%, respectively. Accuracy and precision for BA were -2.6±3.5°, and variability was ≤1.2°. Accuracy and precision for diameter-derived parameters were slightly decreased in projections with 30° rotation; BA precision dropped to 6.2°. MLD quantification is improved for true MLD <0.7 mm, resulting in highly accurate and precise diameter measurements over the entire range of phantom diameters. Automatic reference obstruction analysis provides highly accurate, precise and reproducible RVD and DS measurements.