SIGNAL PROCESSING FOR ENHANCED TISSUE IMAGING

Drexel University, Philadelphia PA

Investigators

Linked publications & trials

Abstract

The primary goal of the project is to characterize the tissue based on the ultrasonic echo from the insonated tissue regions. The regions of the tissue will be classified as normal and abnormal based on a set of criteria that will be developed on the basis of the proposed investigations. The proposed research is based on the premise that the ultrasound echo from liver as well as breast tissue may be modeled using the K-distribution. The parameter of the K-distribution appears to hold information on the nature of the tissue, normal or abnormal, or benign or malignant. Further more, the non-Rayleigh nature of the echo could be extended to include the effect of the phase statistics and its relationship to the diseased and normal tissue. These non-Rayleigh aspects of the back-scattered echo, in terms of the envelope and phase information of the B scan image of the tissue, will be extensively used to achieve the major goal of tissue characterization. The efficacy of our techniques will be studied using the standard ROC analyses. Based on the large number of patients as well as multiple scans of the patients, there will be sufficient data sets available to make a definitive evaluation of the methodologies developed in this research. The research will be undertaken using the patient data supplied by the clinical core group from Thomas Jefferson University. The techniques developed through this research will therefore be used to propose strategies for the screening of patients for cancer (breast and liver) by examining the ultrasound B-scans. It is hoped that outcomes of this research will be such that a low cost simpler method could be developed for the screening of patients for cancer using ultrasound, a non-ionizing radiation.