S23 - Benchtop Models of Patient-Specific Cardiovascular Flow Dynamics

Introduction: : Left Ventricular Assist Devices (LVADs) are mechanical pumps surgically implanted in the heart that redirect blood flow from the left ventricle to the aorta. LVADs provide an alternative to heart transplant for heart failure patients, but exhibit significant complications including a high stroke risk. Blood clots form in regions of blood stasis, which are highly dependent on the geometry and fluid mechanics of each patient. The aim of this study is to recreate the patient-specific (PS) flow dynamics in a benchtop model and assess the patient-specific risk of thrombus (blood clot) formation and stroke.

Materials and

Methods: : In a collaboration with Houston Methodist Hospital, segmented CT scans of heart geometries and conditions were provided from HM3 LVAD recipients and used to reconstruct the PS aorta and left ventricle, which are assembled into a mock circulatory loop. The PS aorta is 3D-printed with a transparent resin and attached to a soft dip-molded silicone left ventricle and a HeartMate3 LVAD (Abbott Laboratories). A HeartMate 3 LVAD was inserted through an apical sleeve and secured with zip ties. The assembly was immersed in a water-filled tank attached to a Windkessel model of the systemic circulation and filled with a viscosity-matched blood analog of 40% glycerol and water. Pressure and flow were measured in several locations during the studies. Flow visualization is performed with a LaVision 3-D Particle Imaging Velocimetry (3-D PIV) system, using 50-µm neutrally buoyant fluorescent particles seeded through the left atrial chamber during imaging. For each flow condition, 3-10 sec of 10-bit digital images (1376x1040pixels) were acquired.Two different heartbeat programs were combined with the LVAD speeds reported in the patient medical records (5, 5.2 and 5.4krpm) to reproduce the target hemodynamics: heart rate of 70-80 beats per minute, mean aortic pressure (MAP) of 89 mmHg and severely depressed cardiac function. At 5krpm, cardiac output was 2.77 L/min and increased to 2.88 L/min at 5.2krpm. No opening of the aortic valve was observed and the contractility was low. A tomographic analysis of the PIV data produced the time-resolved instantaneous 3D velocity vector field v(x,y,z,t) in the complete measurement volume.

Results, Conclusions, and Discussions:: The target values for hemodynamics were attained with PS70 as shown in Table 1. Only slight improvement in pressure and flow was observed with the addition of the patient-specific heart function over the LVAD only conditions. The enlarged LV showed only slight volume change in the bench top model during the cardiac cycle. 3-D flow visualization showed that inflow through the mitral valve was directed towards the anterior wall, and followed along the LV boundary towards the LVAD inflow cannula at the apex. A region of high stasis was observed in the LV outflow tract (Figure 2) and slow flow in the remainder of the LV. With the low pulsatility of contraction, this result would be considered a high potential risk for thrombus. However, more objective flow biomarkers such as residence time and shear-activated potential will be quantified to evidence this potential.

Acknowledgements and/or References (Optional):: The authors gratefully acknowledge the contribution of Houston Methodist to the project, especially Dr. Mahwash Kassi.