Local Density Random Walk Model Preprocessing for Local Shear Wave Viscoelastographic Estimation

Shear wave (SW) elastography is an ultrasound technique that provides quantitative tissue elasticity and viscosity measurements by imaging the tissue response to an applied excitation. In particular, the phase difference method allows for local viscoelasticity estimation through the dispersion curve using the phases from acquired signals at two laterally-spaced pixels in SW elastography. However, this method is sensitive to measurement noise in the SW particle displacement signals. To address this, we propose adopting the local density random walk model to fit the measured noisy SW particle displacement signals. Local elasticity and viscosity are then estimated from the phase difference of the fit signals at two neighboring pixels. Elasticity and viscosity were estimated in six customized phantoms by the phase difference method with and without the proposed preprocessing by model fitting. A decreased interquartile range was obtained by model fitting in both the elasticity and viscosity estimates across all phantoms. In addition, the Levene test indicated a significant difference (p-value < 0.05) between the original and proposed methods in the variance of the estimated elasticity and viscosity. These findings highlight the reduced variability observed in pixel-based viscoelastic estimations with the proposed method, indicating its enhanced robustness.