Robust Two-Point Attenuation Coefficient Estimation Using Kullback-Leibler-type Divergence for Ultrasound Shear Wave Elastography.

Ultrasound shear wave elastography (SWE) is a noninvasive technique for assessing the mechanical properties of soft tissues. In SWE, an acoustic radiation force (ARF) generates shear waves that propagate laterally and are tracked over space and time. Tissue viscoelasticity is commonly evaluated using shear wave phase velocity dispersion and attenuation. This work proposes a robust, model-free method to estimate the shear wave attenuation coefficient, α0, termed the two-point attenuation coefficient estimate using Kullback-Leibler-type divergence (2PACK). The method quantifies spectral differences between source and attenuated signals at their time-domain peak responses using Kullback-Leibler-type divergence, without requiring rheological assumptions or parametric spectral modeling. The method was validated using analytical viscoelastic models, custom-made tissue-mimicking phantoms, and ex vivo bovine liver, and compared with the two-point frequency-shift power (2P-FSP) method. In tissue-mimicking phantoms, both methods showed low coefficient-of-variation (CV) values (5-31%) at higher push/detection voltage levels (≥30 V). At lower voltages (<30 V), the 2P-FSP method became highly variable, with CV values exceeding 100%. In contrast, 2PACK kept the variability below 56%. In ex vivo bovine liver, both methods produced similar ROI variability (9-12%). For full field-of-view estimates, 2PACK maintained low variability (10-11%), whereas 2P-FSP showed markedly higher variability at one location (greater than 100%).