Quantitative ultrasound imaging of in vivo breast tumors

The main goal of this dissertation is to perform quantitative ultrasound imaging (QUS) of in vivo breast tumors in order to provide new tools to differentiate benign from malignant breast masses. The studies outlined in Chapters 3-4 demonstrate that attenuation and backscatter coefficients (BSC) of phantoms can be estimated from the radiofrequency echo data acquired by a clinical ultrasound imaging system, when a well-calibrated reference phantom is used to account for dependencies of echo signals on system settings and transducer properties. Chapter 5 presents attenuation and backscatter coefficients for an animal breast tumor model and demonstrates the feasibility of producing accurate QUS parameters regardless of imaging system used. The study in Chapter 5 was validated in Chapter 6 using a phantom that mimics scanning conditions encountered in the animal model study. A least squares method (LSM) was introduced to estimate attenuation over inhomogeneous paths between an ultrasound transducer and a mass and was verified using phantoms. The LSM was also applied in QUS studies of human breast masses. Preliminary data on attenuation, BSCs, and effective scatterer diameters (ESD) within human breast masses were derived from a subject population undergoing breast biopsies. The attenuation in the mass was estimated from the product of the attenuation coefficient and antero-posterior dimension and was compared to relative acoustic shadowing criteria reported as "posterior echo features" in the BI-RADS lexicon. The BSC averaged over frequency (ABSC) provides information of the echogenicity of an object. The ABSC was compared to the "echo pattern" descriptor of the BI-RADS and showed a potential to quantify the echogenicity within masses. A plot of the ABSC and attenuation coefficient showed potential to differentiate between fibroadenomas and carcinomas with a linear discriminant. In addition, most fibroadenomas exhibited a wider distribution of ESD estimates over the ROI than carcinomas. Chapter 9 presents an assessment of accuracy of attenuation measurements by the reference phantom method when the tissue sound speed differs from that of the reference.;These results demonstrate that a great progress has been made in the effort to develop QUS technology to improve breast ultrasound specificity.