Ultrasound shear wave elastography is a medical imaging technique where mechanical properties (e.g., shear modulus) of soft tissue are imaged with the goal of diagnosing disease. However, different sources of noise or error often lead to poor images that exacerbate uncertainty of medical diagnosis based on these techniques. Sources of uncertainty include scatterer positions and strengths (e.g., emanating from the cellular structure of tissue), other surrounding tissue not included in the image, and electronic noise, among others. In this work, we are developing Optimal Experimental Design (OED) techniques based on the simulation of the ultrasound imaging process and particle tracking to identify controls that maximize the quality of elastography images. Specifically, we propose a general formulation for OED where the observation error depends on both uncertain quantities and the variables to be inferred (e.g., shear moduli). Furthermore, we pose a nonlinear OED problem where the main design variables are the angles used in plane wave compounding for ultrasound motion tracking of propagating shear waves. The goal of the OED problem is to find the number and values of compounding angles that minimize bias and variance in the estimated shear modulus image. We will demonstrate our approach through various numerical examples.
Michael et al. (2025) studied this question.