Image-Based Quantification Workflow for Coronary Morphology: A Tool for Use in Next-Generation Bifurcation Stent Design
Date of Award
Master of Science (MS)
LaDisa Jr., John F.
Coronary artery disease (CAD) occurs in ~200,000 bifurcation lesions annually. Treatment of CAD near bends and bifurcations is challenging and a preferred strategy for bifurcation lesions has yet to be established. However, a favorable treatment option may be elucidated by a more thorough understanding of vessel morphology as well as local hemodynamic alterations caused by current stenting approaches. Computational modeling of human arteries offers an attractive way to investigate the relationships between geometry, hemodynamics and vascular disease. Recent developments also make it possible to perform analysis on realistic geometries acquired noninvasively.
The objective of this work was twofold. The first aim was to build on previous work in this area by quantifying hemodynamic alterations introduced by treatment of an idealized coronary bifurcation using several approaches that involve multiple stents. Each model was created using combined computer aided design techniques and computational fluid dynamics (CFD) analysis tools. Resting and hyperemic blood flow conditions were also studied to determine the severity of local hemodynamic alterations and for comparison to previous results. Indices of time-averaged wall shear stress (TAWSS) and oscillatory shear index (OSI) were quantified for four idealized computational models. The luminal surface exposed to low TAWSS was similar in the main vessel (MV) for all models. Greatest differences were noted between un-stented versus stented side branch vessels (ex. rest: 1% vs. 35%). Sites of elevated OSI (>0.1) were minimal, except under hyperemia conditions in the MV (10% surface area). Flow disturbances were quantified for each provisional technique used, illustrating how stents protruding in main vessels impact flow profiles. Stents without kissing balloon dilation had abnormal flow disturbances, but showed decreased percentage of area exposed to areas of low WSS.
A second aim of this work was to design a robust and unbiased method to quantify vessel morphology and representative trends for three bifurcation sites prone to CAD. Computational models of these sites were generated using computed topography images from 22 patients. Models were used to query geometric characteristics from each bifurcation site including area, length, eccentricity, taper, curvature and bifurcation angles. Post-processing was accomplished by a combination of statistical methods and clustering analysis. Vessel length and area were significantly different within and between bifurcation sites. The left main coronary artery (LCA) bifurcation was significantly different from its two daughter bifurcations (left anterior descending and left circumflex arteries). Specifically vessel area and length were significantly different both between and within bifurcation sites. The daughter bifurcation sites were similar for all characteristics. Vessel area and length proved to be the most useful properties for identifying trends within a particular bifurcation site. The outcome of this work provides a workflow for characterizing coronary bifurcations and a strong foundation for elucidating common parameters from normal, healthy coronary arteries.
Collectively these results from idealized and patient-specific coronary bifurcations offer additional insight into the impact of current treatment approaches and characteristics associated with current stenting techniques. Flow disturbances and local hemodynamic changes have been quantified for provisional techniques currently used. These methods and results may ultimately be useful in the design of next-generation bifurcation stents.