Cardiovascular Engineering
Assessing the Impact of Inlet Waveform Boundary Conditions on Hemodynamic Parameters in Thoracic Aorta Aneurysms: A Patient-specific Sensitivity Analysis
Yu Xuan Huang
Undergraduate Student
University of Toronto
MARKHAM, Ontario, Canada
Farshad Tajeddini
PhD Student
University of Toronto, United States
David Romero
Senior Research Associate
University of Toronto, United States
Jennifer Chung
Surgeon-Investigator
Division of Cardiovascular Surgery, University Health Network, University of Toronto, United States
Cristina Amon
Professor
University of Toronto, United States
Thoracic aortic aneurysm (TAA) poses a substantial risk of life-threatening complications for patients, including aneurysm rupture and dissection. Research suggests a potential link between hemodynamic forces, specifically wall shear stress (WSS) and the pathogenesis of TAA, and consequently proposes WSS as a surrogate biomarker. Computational fluid dynamics (CFD) is a well-established method for quantifying hemodynamic parameters, but concerns arise about its reliability due to inlet boundary condition (BC) selection. Although 4D Magnetic Resonance Imaging (4D MRI) allows for the definition of spatio-temporal patient-specific BCs, it is limited by its relatively poor resolution and global accessibility. Thus, in the absence of patient-specific BCs, predicting hemodynamic conditions in the ascending aorta or arch becomes challenging due to the complex effects of valve closure on the flow jet entering the ascending aorta. Prior research indicates that the spatial distribution and precise temporal variation of inflow BC have minimal influence on the descending aorta. Hence, in the absence of 4D MRI, selecting a generic waveform for the inlet BC and scaling it by cardiac output (CO) can still provide a good prediction of hemodynamics in this region. Nonetheless, CO measurements are not consistently provided for patients. In such scenarios, anthropometric data could serve as a valuable tool to establish a correlation between CO and average anthropometric data from a population and to study the effect of this approximation. In this regard, we present a study to analyze the sensitivity of hemodynamic parameters in ascending aorta, arch, and descending aorta, under various inlet waveform scenarios.
The acquisition of 4D MRI data from three age and gender-matched TAA patients was carried out at the University Health Network, following their informed consent. The geometry of the thoracic aorta was segmented and reconstructed for each patient by utilizing ITK-SNAP, leveraging the information from DICOM format MRI files. The resulting geometry underwent surface smoothing techniques utilizing MeshLab and SolidWork, and was imported into COMSOL, where a tetrahedral mesh was generated consisting of approximately 2.5 million elements. The mesh incorporated ten layers of prism-shaped inflation to accurately capture the influence of boundary layer effects. Transient simulations with rigid-wall conditions were then carried out using five different inlet waveform BC scenarios, including: 1) Patient-specific 3D profile (4D MRI), 2) Patient-specific inlet velocity waveform, 3) first generic waveform scaled by CO approximation using anthropometric data, 4) second generic waveform scaled by CO approximation using anthropometric data, 5) a pressure waveform. The three-element Windkessel outlet BCs were calibrated by utilizing flow rate data from each scenario, as well as pressure measurements. The calibration process involved splitting the flow based on the cross-sectional areas of the branching vessels. The sensitivity of near-wall hemodynamic parameters of interest, specifically Time-Averaged Wall Shear Stress (TAWSS), Oscillatory Shear Index (OSI), and Relative Residence Time (RRT), was evaluated following the simulations.
The use of anthropometric data to estimate CO results in an error of up to 19% with respect to CO obtained from 4D MRI. In ascending aorta and arch, the difference in hemodynamic parameters (TAWSS, OSI, and RRT) can surpass 100% for scenarios 2) to 5) compared to scenario 1). In the descending aorta, while the first generic waveform exhibits only a 10% difference in TAWSS and the influence of the inlet BC is relatively less prominent, there is still an observable 35% difference in TAWSS between the second generic waveform and 4D MRI, even when scaled by the CO approximation derived from anthropometric data.
Neglecting the spatio-temporal distribution of velocity at the aortic root when predicting hemodynamics in the ascending aorta and arch can lead to inaccurate results, not only in terms of magnitude but also in the location of high or low WSS. This is mainly due to the proximity of the ascending aorta and arch to the cardiac valves which contributes to a complex velocity profile and flow jet angle. In the descending aorta, although the hemodynamic parameters are less influenced by the spatial distribution of the inlet BC, certain temporal variations, such as the occurrence of reverse flows, still hold significance. Even with an accurate estimation of CO, overestimating or underestimating these temporal variations can introduce noticeable errors in the results. Thus, it is essential to precisely account for these features to ensure the reliability of the conclusions drawn from the study.
If the area of interest lies in the ascending aorta or the aortic arch, it is advisable to utilize 4D MRI to obtain the inlet BC. When considering the descending aorta, if the focus is on the overall distribution of hemodynamic parameters, any generic waveform scaled by CO can be utilized. However, for accurate predictions of both the magnitude and spatial distribution of hemodynamic parameters in the descending aorta, it is crucial to consider not only CO but also certain flow waveform features such as the characteristic of reverse flow.