Reverberation artifact rejection and masking in arterial ARFI imaging

In arterial ultrasound imaging, interpretation of plaque geometry may be distorted by reverberation artifacts in the lumen. Reverberation and arterial wall signals may be highly correlated, which precludes frequency domain or conventional regression filtering methods; however, decorrelation and rates vary. The variance of time derivatives in 1D axial cross-correlation (CC) and displacement (Disp) measures were exploited to reject the arterial lumen and reverberation artifact. Adaptively determined thresholds were applied to the variance of either individual or spatially clustered ARFI displacement profiles. The resulting binary masks were applied in vivo to six porcine arteries with varying degrees of reverberation artifact. Mask performance was measured as the percentage of the lumen rejected given at least 98% of each arterial wall remaining intact. The highest performing masks used the variance of the 2nd derivative of displacement or correlation with spatial clustering to preserve both arterial wall and soft tissue signals.