Cardiovascular Engineering
Comparing Different Outflow Boundary Conditions for Predicting Blood Flow in Pulmonary Hypertensive Patients Using Computational Fluid Dynamics
Robert Rider
Undergraduate Student
Duquesne University
Oakmont, Pennsylvania, United States
Fatemeh Bahmani
Postdoc
East Carolina University
Greenville, North Carolina, United States
Stephanie George
Associate Professor
East Carolina University, United States
Pulmonary Hypertension (PH) is defined by mean pulmonary arterial pressure (mPAP) greater than 20 mmHg [1]. There are five groups of PH, but the most common of those is Group 2, or pulmonary venous hypertension (PVH) [2]. PVH is caused primarily by left ventricular dysfunction due to left heart disease or valvular disease [3]. PVH is distinguishable from Group 1 PH, or pulmonary arterial hypertension (PAH), by a pre-capillary wedge pressure (PCWP) greater than 15 mmHg and a pulmonary vascular resistance (PVR) less than or equal to 3 Woods units [2]. However, a subset of PVH patients mimics the pre-capillary vasculature seen in PAH patients [2]. Treatment of PAH could negatively exacerbate or worsen left heart disease or left heart failure in PVH patients [2]. An echocardiogram is often used to screen for PH by assessing vascular structure and estimating systolic pulmonary arterial pressure (sPAP) [1]. If the estimated sPAP value is greater than or equal to 30 mmHg, a patient is referred to right heart catheterization (RHC) for final diagnosis [4]. RHC is a very invasive and expensive procedure but is needed to distinguish between PAH and PVH [5]. Computational fluid dynamics (CFD), in combination with magnetic resonance imaging (MRI), may be a useful alternative to invasive RHC. However, simulation results are dependent on selection of appropriate boundary conditions. This study compared CFD simulation results with different outflow boundary conditions assigned to the pulmonary artery.
MRI was used to obtain images of the pulmonary artery of one patient confirmed to have PH based on RHC. Phase contrast MR images were used to create a patient-specific velocity inlet waveform. Materialise Mimics (Materialise, Inc., Plymouth, MI) was used to segment portions of the MRIs to recreate a 3-D pulmonary artery. The geometry was exported to ANSYS Workbench 2022 R2 (ANSYS, Inc., Canonsburg, PA) to create a mesh composed of 1,361,879 triangular elements. Laminar blood flow was assumed with blood viscosity of 0.0035 kg/m.s and blood density of 1060 kg/m3. A no-slip condition and rigid assumption was applied to the fluid wall. The MR-derived velocity inlet waveform was applied as a blunt profile to the inlet. Simulations were run for three cardiac cycles with 765-time steps and 0.00382 seconds time step size in ANSYS Fluid Flow (FLUENT). Zero-pressure gradient, split flow, and constant resistance conditions were applied to left and right branches of the pulmonary artery separately. Zero-pressure gradient assigned pressure values of zero to all outlets. Split flow divided initial flow rate between outlets based on cross-sectional areas directing more flow to larger outlets and less flow towards smaller outlets. Constant resistance used constant pressure values calculated by split flow time-averaged flow rates multiplied by patient-specific PVR. Wall shear stress (WSS), velocity volume, and pressure volume contours were created at peak systole of the third cardiac cycle. Ansys Ensight was used to calculate oscillatory shear index (OSI) and Time-Averaged wall shear stress (TAWSS) contours.
Results and Discussion: The specific values for each outflow condition are shown in Table 1. RHC values of the subject were PCWP of 24 mmHg, mPAP of 33 mmHg, and PVR of 2.7 Woods. The zero pressure maximum WSS was 48.7% less than that of the split flow condition and 46.4% less than the maximum WSS of the constant resistance condition. Decreased WSS is known to be an indicator of PH. The zero-pressure gradient maximum TAWSS was 26.5% less than that of the split flow conditions and 84.8% less than the maximum TAWSS of the constant resistance condition. The high difference in values occurs due to zero pressure boundary condition using zero pressure while constant resistance uses higher values of pressure at the outlets. The maximum OSI value for the zero-pressure gradient was 2.41% lower than that of the split flow value and 1.65% lower than that of the constant resistance maximum OSI value. Therefore, the OSI contours do not show significant variation for any outflow condition. The split flow condition uses initial flow rate without assuming specific outlet pressures, making this the most physiologically relevant condition. The WSS gets higher towards the outlets due to the higher velocity gradient that occurs within the pulmonary artery as shown in Figures 1, 2, and 3. (Note the scales vary for each outflow condition)
Conclusion: The selection of outflow boundary condition impacts measures such as TAWSS, Max WSS, and Max OSI. The zero-pressure gradient outlet does not account for differences in downstream resistance between the outlets. The constant resistance or constant pressure condition does not allow for time-varying pressures in the outlet which is not physiological. The split flow condition may be the most physiologically relevant as it allows for time varying pressures at the outlets. Future work could investigate other outflow boundary conditions such as Windkessel models or other lumped parameter models. Further studies that include more subjects would enable correlation between MR/CFD parameters and RHC measures potentially leading to a less invasive, less expensive way of diagnosing PH patients.
This material is based upon work supported by the National Science Foundation under Grant No. 1950507. Any opinions, findings and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.
[1] Pearce, Daniel. Fluid-Structure Interaction Modeling of Pulmonary Artery Blood Flow in End-Stage Renal Disease Patients. 2020 East Carolina University, Master Thesis.
[2] Al-Omary, Mohammed S., et al. “Pulmonary Hypertension Due to Left Heart Disease.” Hypertension, vol. 75, no. 6, 2020, pp. 1397-1408, https://doi.org/10.1161/hypertensionaha.119.14330
[3] Dadfaray, Sina et al. “Differentiating pulmonary arterial and pulmonary venous hypertension and the implications for therapy.” Congestive heart failure (Greenwich, Conn.) vol. 16,6 (2010): 287-91. doi:10.1111/j.1751-7133.2010.00192.x
[4] Rabidou, Jake. Pulmonary Artery Hemodynamics Using MRI & CFD. 2017. East Carolina University, Master Thesis.
[5] Zombrano, Byron A., et al. “Patient Specific Computational Analysis of Hemodynamics and Wall Mechanics and Their Interactions in Pulmonary Arterial Hypertension.” Frontiers in Bioengineering and Biotechnology, vol. 8, 2021, https://doi.org/10.3389/fbioe.2020.611149
[6] Tang, Beverky T., et al. “Wall Shear Stress Is Decreased in the Pulmonary Arteries of Patients with Pulmonary Arterial Hypertension: An Image-based, Computational Fluid Dynamics.” Pulmonary Circulation, vol. 2, no. 4, 2012, pp. 470-476, https://doi.org/10.4103/2045-8932.105035