Precision and accuracy of dimensional assessment of luminal contours by commercially available quantitative angiography software as a prerequisite to angiography based FFR and other derived parametrics

Tobe, Akihiro, Miyashita, Kotaro, Revaiah, Pruthvi C., Tsai, Tsung-Ying, Oshima, Asahi, Hu, Shiuan Hao, Sevestre, Emelyne, Garg, Scot, Bourantas, Christos, Girasis, Chrysafios, Wentzel, Jolanda J., Onuma, Yoshinobu, .Serruys, Patrick W. Precision and accuracy of dimensional assessment of luminal contours by commercially available quantitative angiography software as a prerequisite to angiography based ffr and other derived parametrics. Catheterization and Cardiovascular Interventions, https://doi.org/10.1002/ccd.31670

Abstract

Background Accurate dimensional measurements are critical for quantitative coronary angiography (QCA) and serve as the first step in angiography-based fractional flow reserve (FFR) calculations. Aims To compare minimum lumen diameter (MLD) measurements across multiple QCA or anigo-based FFR software programs using phantom models. Methods Fourteen QCA and angio-based FFR programs were evaluated using six plexiglass phantoms, each containing three sequential bifurcations with known true values for the MLD of the proximal main, distal main, and side branch vessels. The accuracy and precision of MLD measurements were assessed by comparing software-measured values with true values across 54 MLD measurement points. No manual correction of the vessel contour was performed. The results of the 14 programs were reported anonymously. Results The mean differences between the measured and true values were small (< 0.1 mm), however, in two angio-based FFR programs, the discrepancies were large (> 0.3 mm). The standard deviations of the differences were approximately 0.1 mm, except in one angio-based FFR program, where it exceeded 0.3 mm. Differences from true values were more pronounced in small (≦ 0.7 mm) compared to large (> 0.7 mm) true MLDs. The reproducibility of measurements was high (Pearson's correlation coefficient > 0.98) across all programs. Conclusion Variations in MLD measurements were observed among different QCA and angio-based FFR programs. These variations may influence diagnostic performance and can seriously impact decisions made solely using angio-based FFR.

Publisher

Wiley

Rights

Attribution-NonCommercial-NoDerivatives 4.0 International

cbn

Collections

Medicine (Scholarly Articles)