Effects of wall motion reconstruction in patient-specific CFD simulations of left atrial blood flow
Please login to view abstract download link
The left atrial appendage (LAA) is a common site of intracardiac thrombosis because it tends to accumulate stagnant blood, increasing the risk of cardioembolic stroke. Patient-specific computational fluid dynamics (CFD) simulations offer promise to evaluate the risk of LAA thrombosis [2]. However, they require time-dependent models of the cardiac walls and their motion, which are commonly obtained from medical images with limited temporal resolution (O(10) frames per cycle). Since CFD solvers use significantly finer temporal resolutions, the wall motion obtained from medical images is heavily interpolated to match the CFD resolution. The effects of this approach are not well understood, given the paucity of high-resolution wall motion data to perform validation studies. We address this need by performing CFD simulations using the wall motion obtained from high-resolution, patient-specific electromechanical (EM) cardiac tissue models [1]. We first perform a reference CFD simulation with boundary conditions defined by the wall motion obtained from the full-resolution EM results. Subsequently, we run the CFD solver with progressively coarser temporal interpolations of the EM model wall motion. By varying the number of retained EM frames and the interpolation details, we study the sensitivity of the results to the wall motion reconstruction. In the presentation, we will discuss how motion reconstruction affects the CFD results, in particular, the blood residence time in the LAA, which has been shown to correlate with thrombosis risk.
