Novel bifurcation phantoms for validation of quantitative coronary angiography algorithms.

Background: Validation is lacking for two- and three-dimensional (2D and 3D) bifurcation quantitative coronary angiography (QCA) algorithms.

Methods: Six plexiglas phantoms were designed, each of them mimicking a coronary vessel with three successive bifurcations lesions, wherein at least one vessel segment had a percent diameter stenosis (DS) of ≥60%. The five most frequently occurring Medina classes (1,1,1), (1,1,0), (0,1,0), (0,1,1), and (1,0,0) were used in the design. Diameters of the daughter vessels in every bifurcation were dictated by the scaling law of Finet. Lesions were cosinus-shaped in longitudinal view and circular-shaped in cross-sectional view. At the level of the carina, lesions were becoming eccentric, favoring "plaque" at the outer bifurcation walls. Adjacent bifurcation lesions were kept distant by nontapering, stenosis-free segments of ≥10 mm length. The direction of the side branch relative to the main vessel was based on relevant literature. The phantoms were precision manufactured using computer-aided design and machining techniques. Because of the high drilling accuracy (within 10 μm), the 3D luminal surface description of the phantom could be used to determine the true lumen dimensions and bifurcation angle (BA) values of the final geometry.

Results: Our series of bifurcation phantoms comprised 33 narrowed and 21 stenosis-free vessel segments with a mean true minimal lumen diameter (MLD) value of 0.98 ± 0.40 mm (range, 0.53-1.96 mm) and 2.29 ± 0.74 mm (range, 1.40-4.00 mm), respectively. Overall, the mean true values for MLD, reference diameter, and DS were 1.49 ± 0.85 mm, 2.70 ± 0.71 mm, and 40.9% ± 34.2%. The mean true values for the proximal and the distal BA were 123.6° ± 19.0° and 69.6° ± 19.9°, respectively.

Conclusions: Six plexiglas phantoms containing a total of 18 bifurcations lesions with variable anatomy and Medina class were designed and precision manufactured to facilitate the validation of bifurcation QCA algorithms.