Improved diagnostics through quantitative ultrasound imaging

Conventional B-mode imaging in ultrasound consists of displaying the log-compressed envelope of the backscattered signal. While clinical ultrasonic B-mode images have good spatial resolution, i.e., better than a millimeter, the contrast resolution of ultrasonic B-mode images is typically low. However, additional information is contained in the ultrasonic backscattered signal, which can be used to create images related to tissue microstructure. Because diagnosis of disease is typically based on histological examination of tissue microstructure, the ability to quantify and describe tissue microstructure through ultrasound may result in improved diagnostic capabilities of ultrasound. Tissue-mimicking phantoms and animal models of breast cancer were used to assess the ability of novel ultrasonic imaging techniques to quantify microstructure. Four parameters were extracted from the ultrasonic backscattered signal and related to the microstructure. The effective scatterer diameter (ESD) and the effective acoustic concentration (EAC) parameters were based on modeling the frequency dependence of the backscatter. The k parameter (which quantifies the periodicity of scatterer locations) and the μ parameter (which estimates the number of scatterers per resolution cell) were based on modeling the statistics of the backscattered envelope. Images constructed with these parameters resulted in an increase in contrast between diseased tissue and normal tissues but at the expense of spatial resolution. Specifically, in simulation, quantitative ultrasound (QUS) increased the contrast-to-noise ratio (CNR) between targets and background by more than 10 times in some cases. Statistically significant differences were observed between three kinds of tumors using the ESD, EAC, and k parameters. QUS imaging was also improved with the addition of coded excitation. A novel coded excitation technique was used that improved the variance of estimates over conventional pulsing methods, e.g, the variance of ESD estimates were reduced by a factor of up to 10.