TISSUE SCATTERER ANALYSIS WITH SPECTRAL REDUNCANCY

Drexel University, Philadelphia PA

Investigators

Linked publications & trials

Abstract

The continuation of this research project aims at extracting new and relevant tissue information from spectra of broadband ultrasonic echoes. The ultimate goal of this project is to improve the detection capabilities of malignant tissues with ultrasound imaging. A unique aspect of this project is its focus on characterizing Fourier phase and magnitude of rf ultrasonic tissue echoes with spectral redundancy methods, such as those based on the spectral autocorrelation function (also referred to as the generalized spectrum). Previous work has shown that estimates using the generalized spectrum can reliably reveal the nature of tissue regularity in the presence of diffuse scatterers over relatively small (several millimeters) segments of tissue. The good convergence properties of spectral correlation methods have allowed for creating relatively high resolution parametric images that reveal spatial changes in tissue structure parameters, such as the mean scatterer spacing and echo specularity. Future work focuses on understanding and applying new generalized spectrum features such as the specular scatterer size, dispersion, multiple scatterer spacings, and the degree of regularity for both resolvable and unresolvable spacings. While other methods have been used to extract resolvable scatterer spacing information, the new features utilize the unique way in which the generalized spectrum presents information as a function of frequency. In addition, statistical tests and significance levels will be used in analyzing generalized spectra for more automatic and repeatable classifications of tissue regions. B-scans on liver and breast tissues will be examined to determine changes in spectral redundancy parameters for normal, benign and malignant tissue regions. Databases will be established to statistically examine the nature and consistency with which malignant neoplasms disrupt the normal tissue structures. The success of the proposed research will enhance capabilities for diagnostic ultrasound, particularly for early detection of cancer.