Estimating Left Main Bifurcation Geometry: Are Geometric Models Reliable in Diseased Arteries?

Accurate characterization of coronary bifurcation geometry is essential for planning and optimizing percutaneous coronary interventions. Although several geometric models have been proposed to predict coronary bifurcation dimensions, their performance in diseased vessels remains insufficiently explored.

To assess the accuracy and precision of three geometric prediction models (Murray, Finet, and Huo-Kassab) in estimating left main coronary artery (LMCA) bifurcation diameters, using quantitative coronary angiography (QCA) and intravascular ultrasound (IVUS) as reference standards.

This retrospective, single-center study included 13 patients undergoing IVUS evaluation of the LMCA, left anterior descending artery, and left circumflex artery. Bifurcation diameters predicted by each geometric model were compared with measurements obtained by QCA and IVUS. Accuracy, precision, and model ratios were calculated for each comparison. Agreement between predicted and measured values was further assessed using Bland-Altman analysis.

When compared with QCA, the Murray and Huo-Kassab models demonstrated higher accuracy (0.06 mm and -0.03 mm, respectively) and similar precision (0.95 mm for both) than the Finet model (accuracy -0.35 mm, precision 0.96 mm). In comparisons with IVUS, the Murray model showed the best performance, with an accuracy of -0.49 mm and a precision of 0.84 mm. Bland-Altman analysis indicated superior agreement for the Murray model, particularly when IVUS was used as the reference standard.

In this population of patients with moderately diseased LMCA bifurcations, the Murray and Huo-Kassab models provided more accurate geometric predictions than the Finet model. Overall, the Murray model demonstrated the best performance, especially in IVUS-based assessments.