Embolic Transport in LVAD Outflow: An Experimental Study Using Patient-Specific and Idealized Aortic Models

A left ventricular assist device (LVAD) is a mechanical pump that provides circulatory support as a bridge-to-cardiac transplantation or as a destination therapy in patients with advanced heart failure. A potential adverse event of LVAD support is thrombus ingestion or formation, which may then travel through the device into the cerebral arteries, causing ischemic strokes. Previous numerical simulations of embolus transport within LVAD systems have exhibited inconsistencies in the results in assessing the fate of emboli in LVAD settings. These disparities prompted the development of an experimental framework tailored for a systematic measurement of particle transport in the context of LVADs. In this in vitro study, we utilized a nearly refractive-index-matched time-resolved particle tracking velocimetry (PTV) system to resolve and visualize particle trajectories within each aortic model, complemented by particle image velocimetry (PIV) measurements. We also conducted a meticulous measurement of particle weight in each individual branch by collecting the particles from each outlet. Four LVAD patients, as well as two idealized models of the human aorta, each featuring a cannula grafted at an anastomosis angle of 45 degrees, were considered. Thin-wall high-resolution phantoms of these models were 3D-printed with precision and placed in a flow loop that provided physiological flow conditions. Three different sizes of precision fluorescent beads (neutrally buoyant) with particle-to-cannula diameter ratios of $d_p/D$ = 0.031, 0.053, 0.075 were used to replicate emboli at two clinically relevant flow rates, spanning over 50 experimental cases combined. This systematic investigation reveals that particle distributions largely follow the branchwise flow split, nearly independent of the range of Stokes numbers and inlet Reynolds numbers examined. This finding partially challenges commonly held assumptions in LVAD studies.