Ultrasonic Characterization of Cell Pellet Biophantoms and Tumors Using Quantitative Ultrasound Models By

ii ABSTRACT This dissertation investigates the ultrasonic scattering from tumors. Tumors are complex acoustic scattering media and to date there has not been an adequate acoustic scattering model that fits them well. The tumor scattering mechanisms are studied herein by a step-wise approach: starting from simple (low-concentration cell pellet biophantoms) to moderately complex (high-concentration cell pellets) and to significantly complex media (actual tumors). The comparison between the backscattering coefficients (BSC) of lower-and higher-concentration cell pellet biophantoms suggest that high concentration of cells could lead to correlation in cell positions, causing additional scattering that may be modeled by the structure function. The structure function for the high-concentration cell pellet biophantom was isolated by comparing the BSC of the high-concentration biophantom to that of a very low-concentration biophantom where the cell spatial positions are assumed to be random. Polydisperse structure function models that take into account the polydispersity of scatterer radii are introduced and shown to be consistent with the structure function curves measured from BSC. The two-dimensional (2D) structure function is estimated from the histological sections, and shown to have a similar shape to the theoretical structure function and to that calculated from BSC. The comparison between cell pellets and tumors of the same cell type suggests that the anatomic details in tumors must be taken into account for modeling iii purposes, in addition to the scattering from cells. Also, histology studies suggest that the structure functions in tumors are slightly different from those in cell pellets: the tumor cell spatial arrangement is slightly more random compared to cell pellets. The effect of the structure function on parameter estimation is discussed. Further work is shown to be required for modeling the tumor structure function. Additionally, the comparison between different tumor types shows that ultrasound backscattering is sensitive to unique tumor structures. The EHS tumor has a distribution of clustered cells and shows a different BSC and structure function pattern than the tumors that have a homogenous distribution of cells. A scattering model is developed to detect the clustering feature.