Full-waveform shear wave elastography for imaging tumors

Abstract

Shear wave elastography (SWE) is a method of reconstructing the stiffness of soft biological tissues by matching the observed and the simulated wavefields using an inverse optimization scheme. SWE reconstruction algorithms can be classified into two main categories, local and global methods. Global methods consider more complete physics of the waves, i.e., refraction and scattering, and, thus, have the potential to better characterize the heterogeneity of the domain. These approaches require full-waveform inversion (FWI), which is computationally expensive. More importantly, due to highly nonlinear nature, FWI has the limitation of converging to local minima, leading to erroneous reconstruction. In this work, we address this issue and propose a cost functional that not only reduces the nonlinearity of the FWI but also results in a reconstruction algorithm that is independent of push amplitudes, less sensitive to the initial guess, and has a better convergence behavior compared to the classical least-squares cost functional. In addition, we propose to utilize only a fraction of the measurements with the eventual goal of 3D reconstruction of tumors using limited ultrasound measurements. In this talk, we will present the details of the underlying formulation and examples showing the effectiveness of the proposed method.