Performance evaluation of compounding and directional beamforming techniques for carotid strain imaging using plane wave transmissions

Carotid strain imaging in 3D is not possible with conventional focused imaging, because the frame rate is too low. Plane wave ultrasound provides sufficiently high frame rates, albeit at the cost of image quality, especially in the off-axis direction due to the lack of focusing. Multiple techniques have been developed to cope with the low off-axis image quality when performing 2D (and in future 3D) motion estimation: cross correlation with directional beamforming (with or without RF (coherent) compounding) and displacement compounding. This study compares the precision of these techniques using linear array ultrasound data of a pulsating concentric homogeneous artery simulated using Field II. The transducer (f_c = 9 MHz, pitch = 197.9 μm, 192 elements, f_s = 180 MHz) transmitted plane waves at 3 sequentially alternating angles (0°, +θ, -θ) at a PRF of 2 kHz. Simulations were repeated for θ ranging from 1° to 20° with increments of 1°. Displacements were estimated for frame intervals of 1/15th s, tracked, and cumulated from diastole to systole using either displacement compounding, or directional beamforming optionally enhanced by RF compounding. 1D directional beamforming with RF compounding and 2D displacement compounding with θ = ~20° performed equally and best with a relative root-mean-squared error of ~2% with respect to the analytical solution. The mean and standard deviation of the estimated motion direction for 2D displacement compounding with θ = 20° was 0.03° +/- 1.43°. Since displacement compounding requires no assumptions regarding the motion direction, this technique seems the best option for plane wave carotid strain imaging.