A simulation setup to optimize particle flow velocimetry

Particle Flow Velocimetry (PFV) has been introduced as a new ultrasound methodology to measure two-dimensional intraventricular flow patterns. It can potentially provide important new information on cardiac hemodynamics and function but how to optimize the whole process (such as frame rate, line density, contrast concentration) is still not clear. The aim of this study was therefore to build a simulation environment allowing to optimize this methodology by combining computational fluid dynamics (CFD) and ultrasound simulations. A 2D model of the left ventricular (LV) geometry was generated and meshed in order to be used as input to commercially available CFD software. An analytic description of a typical ventricular inflow velocity profile (showing an early and atrial filling phase) was used as a boundary condition at the inlet of the LV model and the dynamic flow field was simulated. Point scatterers were subsequently put at random positions within the model and their positions were updated over time based on the simulated flow field. From this dynamic scatterer field, ultrasound data could subsequently be obtained using a convolution-based model previously introduced by our lab. In order to test the simulation setup, RF signals from both a PW Doppler acquisition in the inlet portion of the model and a 2D color Doppler image sequence were simulated. For the PW Doppler spectrogram, the normalized RMSE of the estimated velocities relative to the CFD reference was 3.09%. Moreover, good qualitative agreement was found between the CFD and the color Doppler measurement. In conclusion, a simulation setup was constructed and shown to work correctly. It will be useful for optimizing PFV and for developing flow tracking methods in echocardiography further.