Effects of tissue anisotropy on the spectral characteristics of ultrasonic backscatter measured with a clinical imaging system.

In this paper, we report the effects of inherent tissue anisotropy on the spectral properties of backscattered ultrasound when measured with a commercially-available imaging system. We insonified five specimens of bovine tendon immersed in a water tank and rotated in 10 degrees increments while being imaged with a Hewlett-Packard Sonos 1500 system. The backscattered RF signals corresponding to each angle of insonification were digitized and the spectral characteristics of the backscattered ultrasound were determined. The mean anisotropy, defined as the average difference between values at perpendicular and parallel insonification, for band-limited estimates of backscattered power, centroid frequency, upper-band to lower-band power ratio, and upper-band to total-band power ratio were found to be 24.6 +/- 1.1 dB, 142 +/- 27 kHz, 32 +/- 13%, and 22 +/- 5%, respectively (mean +/- SE). The magnitude of each of these backscatter spectral parameters was larger at perpendicular insonification compared with the corresponding values at parallel insonification, consistent with previous measurements of the inherent anisotropy of ultrasonic attenuation and backscatter in tissue.

Comparison of integrated backscatter values obtained with acoustic densitometry with values derived from spectral analysis of digitized signals from a clinical imaging system.

Time-domain-based integrated backscatter values obtained with the use of acoustic densitometry (AD) were compared with values determined from a spectral-based analysis of the radio-frequency (RF) signals with a modified Hewlett-Packard Sonos 1500 imaging system. Integrated backscatter images of five specimens of bovine tendon were acquired in the AD acquisition mode, and the corresponding signals related to the backscattered RF were digitized for each angle of insonification as the specimens were rotated in 10-degree increments. The integrated backscatter images were analyzed with the AD analysis package, and the corresponding values determined from the RF power spectra were obtained from the digitized ultrasonic signals. Good agreement was found between the two methods over the entire range of measured values. The mean anisotropy in the measured integrated backscatter (mean +/- standard error) was found to be 27 +/- 2 dB for time-domain-based analysis and 25 +/- 2 dB for RF spectral-based analysis.