On-Axis Acoustic Radiation Force-Based Elasticity Measurement in Homogeneous and Layered, Skin-Mimicking Phantoms

Shear wave elasticity imaging has been difficult in inhomogeneous tissue, such as skin, where thin layers cause bias in shear wave speed estimates. Instead, we propose an FEM-based look-up table to estimate shear modulus at the on-axis location rather than at lateral locations. We estimate the time-to-peak on-axis displacement at each depth and use a look-up table to estimate shear modulus. We generated a stiffness look-up table using a 3D FEM model coupled to Field-II simulations. We simulated shear moduli of 1-13kPa using a CL15-7 transducer, 1.1cm axial push focus, 8.9MHz push frequency, push F/3, 10.4MHz tracking frequency, and plane wave tracking. The displacements were estimated using an advanced Bayesian displacement estimator. Here we test the on-axis stiffness estimates in homogeneous and layered polyvinyl-alcohol (PVA) phantoms. Before the phantoms were sliced into thin layers, we computed five independent shear wave speeds at the same depth that would be sliced and took the median. After slicing and layering the phantoms, we tested the on-axis method and compared it to shear wave speed-derived shear modulus estimates. The presliced homogeneous PVA phantoms had a mean shear modulus of 5.56+/-0.76kPa and 5.46+1.06kPa for the on-axis method and shear wave speed-derived moduli, respectively. The mean shear moduli for the on-axis method and shear wave speed-derived estimates in the thin layer were 5.92+/-1.15kPa and 6.50+/-0.3IkPa, respectively. The shear wave speed-derived estimates had a bias of -1.04kPa and the on-axis method had a bias of -0.46kPa. The on-axis method had a larger variance in shear modulus estimates, but a smaller bias than shear wave speed-derived estimates.