In vitro validation and comparison of different software packages or algorithms for coronary bifurcation analysis using calibrated phantoms: Implications for clinical practice and research of bifurcation stenting

Background: The accuracy and precision of quantitative coronary angiography (QCA) software dedicated for bifurcation lesions compared with conventional single-vessel analysis remains unknown. Furthermore, comparison of different bifurcation analysis algorithms has not been performed. Methods: Six plexiglas phantoms with 18 bifurcations were manufactured with a tolerance < 10 μm. The bifurcation angiograms were analyzed using Cardiovascular Angiography Analysis System (CAAS; Version 5.10, Pie Medical Imaging, Maastricht, The Netherlands) and QAngio XA (Version 7.3, Medis Medical Imaging System BV, Leiden, The Netherlands) software packages. Results: Conventional single-vessel analysis underestimated the reference vessel diameter and percent diameter stenosis in the proximal main vessel while it overestimated these parameters in the distal main vessel and side branch. CAAS software showed better overall accuracy and precision than QAngio XA (with automatic Y- or T-shape bifurcation algorithm selection) for various phantom diameters including minimum lumen diameter (0.012 ± 0.103 mm vs. 0.041 ± 0.322 mm, P = 0.003), reference vessel diameter (-0.050 ± 0.043 mm vs. 0.116 ± 0.610 mm, P = 0.026), and % diameter stenosis (-0.94 ± 4.07 % vs. 1.74 ± 7.49 %, P = 0.041). QAngio XA demonstrated higher minimal lumen diameter, reference vessel diameter, and % diameter stenosis when compared to the actual phantom diameters; however, the accuracy of these parameters improved to a similar level as CAAS when the sole T-shape algorithm in the QAnxio XA was used. Conclusion: The use of the single-vessel QCA method is inaccurate in bifurcation lesions. Both CAAS and QAngio XA (when the T shape is systematically used) bifurcation software packages are suitable for quantitative assessment of bifurcations.