Abstract

Ultrasound is a medical imaging modality with many advantages, such as being real-time, portable, and cost-effective. Nevertheless, it provides a qualitative representation of the human body, which cannot reveal the physical characteristics of the tissue. Quantitative ultrasound (QUS) has evolved as a non-invasive ultrasound-based imaging modality to investigate the acoustic properties of tissue microstructure. It aims at recovering quantitative properties of tissue microstructure by investigating the power spectra of the radio frequency data or statistics of the envelope of the backscattered signal. The accuracy and precision of microstructural properties are necessary for a correct description of tissue microstructure. In particular, spectral-based techniques estimate the frequency-dependent backscatter coefficient (BSC) from the echo signal power spectra after removing attenuation effects from tissues between the transducer and the region of interest. The BSC can be parametrized in terms of a power law function or form factor models. While the power law model is associated with physics, the form factor can inform sub-resolution scatterer features, such as the effective scatterer diameter (ESD) and the acoustic concentration (AC). Common approaches to estimating the ESD and AC are based on minimization strategies of the squared difference between a model spectrum and a measured spectrum (or form factors). A key aspect of ESD and AC estimation is that it accurately and precisely quantifies the scattering properties of tissue. This thesis aims to introduce our novel regularized-based strategies to improve the estimation of the average attenuation, BSC, ESD, and AC. Chapter 2 presents two versions of our proposed method, ALGEBRA, to accurately and precisely estimate average attenuation and BSC in various tissue-mimicking phantoms. The power spectra at each frequency and depth have equal weights in ALGEBRA. However, due to the attenuation, the high-frequency contents of the power spectra at deep regions have a low signal-to-noise ratio. Additionally, the average attenuation varies gradually while the BSC alters markedly in different parts of the tissue. In Chapter 3, we consider these two shortcomings of the ALGEBRA and propose a novel method optimized using alternating direction method of multipliers (ADMM) to estimate the same parameters. Chapters 4 to 6 are dedicated to estimating ESD and AC using dynamic programming (DP) and analytical-based methods. In Chapter 7, we propose a novel approach to estimate the distribution of scatterer sizes instead of reporting a single size to more accurately characterize tissue. In the final chapter, we provide conclusions and future work.