Attenuation estimation using a synthetic aperture focusing technique

Abstract

Diffraction effects resulting from radiation of ultrasound sources pose problems for ultrasonic exams. In qualitative ultrasound imaging (i.e., B-mode), problems arising from diffraction are related to loss of image quality, (i.e., spatial resolution and contrast). In quantitative imaging, improper diffraction compensation results in loss of accuracy when estimating parameters such as attenuation coefficients. Although techniques have been developed to correct for diffraction effects for both qualitative and quantitative imaging, these two modalities have largely been treated separately in the specialized literature. The goal of this study is to explore synthetic aperture focusing techniques (SAFT), which have been successfully employed to reduce beamwidth spreading in B-mode imaging, to minimize diffraction effects when estimating attenuation coefficients. The ability of SAFT to produce more accurate attenuation estimates was explored experimentally. A flat circular piston with a 0.25” diameter and 3.5 MHz center frequency was used to collect radiofrequency (rf) data. The imaging targets were two agar phantoms with attenuation coefficients equal to 0.4 dB/cm/MHz (phantom A) and 0.7 dB/cm/MHz (phantom B). Both phantoms were raster scanned over a 4 cm by 4 cm area at 0.75 mm steps (i.e., a dataset of 54 rf lines per 54 scan sections). 3D-SAFT volumes were created using conventional delay-and-sum. For both rf and 3D-SAFT datasets, 54 amplitude profiles were obtained by averaging the envelopes of all data lines within each section. The amplitude profiles were gated using 6 mm overlapping rectangular gates centered within the phantom at five depths between 4.9 and 6.2 cm. Attenuation coefficients at all five different depths and 54 sections were estimated in time domain by dividing the slope of the log-compressed gated amplitude profiles vs. depth by the center frequency of the transducer. The estimated attenuation coefficients for phantom A using rf and 3D-SAFT data were 0.64 +/- 0.22 dB/cm/MHz and 0.42 +/- 0.21 dB/cm/MHz, respectively. For phantom B, the estimates using unfocused radio-frequency and 3D-SAFT data were 1.00 +/- 0.22 dB/cm/MHz and 0.76 +/- 0.24 dB/cm/MHz, respectively. Therefore, the use of unfocused radio-frequency data resulted in consistent over-estimation of attenuation coefficients whereas the estimates using SAFT exhibited high accuracy (i.e., maximum estimation bias of 9% for SAFT data vs. 43% for rf data). The precision was comparable for both approaches. The preliminary experimental results obtained in this work indicate that SAFT has potential for simplifying the ability to correct for diffraction effects when estimating attenuation coefficients. Further, these results suggest beamforming techniques can be valuable tools for improving the accuracy of quantitative ultrasonic imaging. This work was supported by NIH Grants R21-CA139095 and R01-EB008992.