P3E-5 Attenuation Estimations from B-Mode Image Data Derived using Bandpass Filtering

Abstract

Frequency-dependent ultrasound attenuation in soft tissue is an important parameter for discriminating normal from diseased tissues. A previously described video signal analysis method computes the ratio of mean echo intensities from a sample to that from a reference phantom scanned using identical transducer settings. The relative echogenicity versus depth yields information on the attenuation of the sample. A problem with VSA, however, is that the relative echogenicity versus depth curves derived in this manner are distorted when broadband pulses common in pulse-echo ultrasound systems are utilized. In this paper, we present results that show the accuracy of attenuation estimates computed from image data derived after narrowband bandpass filtering of the backscattered radiofrequency echo signals at different center frequencies around the center frequency of the broadband echo signal. Envelope or B-mode signals derived from the narrow bandpass filtered radiofrequency data are processed to estimate attenuation using the video signal analysis method. Since higher frequency components are more severely attenuated than lower frequencies, even with a 1 MHz bandwidth filter, narrow band video signal analysis data for center frequencies above the center frequency of the pulse yielded significant underestimation of the attenuation, particularly at large depths. On the other band, filtered signals with center frequencies lower than or at the center frequency of the broadband pulse produced unbiased attenuation estimates. Local attenuation coefficient estimates using the video signal analysis method with narrow bandpass filtering demonstrate comparable results to other existing estimation methods, while maintaining computational efficiency for real-time implementation