The velocity of mechanical waves (MW) in the heart reflects myocardial tissue properties. Different wave velocity estimation methods have been proposed, using the slope of the wave projection in M-mode, or based on the gradient of the time-of-flight (TOF) map (gradient method). In this work, we compare these methods using a simulation and 3D wave propagation in vivo. Waves were detected using both clutter filter wave imaging (CFWI) and tissue Doppler imaging (TDI). The effect of pipeline parameters on velocity estimation was studied. Finally, an in vivo investigation was made for healthy controls and patients with aortic stenosis. When the wave propagation was mainly in-plane, all methods yielded similar results, verified using both simulations and in vivo data. However, velocity overestimation occurred due to misalignment between the M-line and the wave propagation direction, and for wave-view misalignment when using the 2D gradient method. The gradient method was sensitive to processing parameters, where smoothing of the TOF map also led to an overestimation of the wave velocities. For our data, CFWI provided the most robust results, however, the choice of filter cutoff influenced the output, which became similar to TDI for high cutoff velocities.

Our study shows that the gradient method can provide similar results as the M-mode slope when the wave propagation is aligned in-plane, and further provide localized wave velocity estimates in 2D and 3D, limited by smoothing requirements. This can be advantageous for mapping heterogeneous tissue properties, and the method can provide valuable clinical insight in the future.