Inter-system comparison of site-matched ultrasonic measurements of the calcaneus in-vitro

Effects of phase cancellation and receiver aperture size on broadband ultrasonic attenuation for trabecular bone in vitro

Phase cancellation in ultrasound due to large receiver size has been proposed as a contributing factor to the inaccuracy of estimating broadband ultrasound attenuation (BUA), which is used to characterize bone quality. Transducers with aperture size ranging from 2 to 5 mm have been used in previous attempts to study the effect of phase cancellation. However, these receivers themselves are susceptible to phase cancellation because aperture size is close to one center wavelength (about 3 mm at 500 KHz in water). This study uses an ultra small receiver (aperture size: 0.2 mm) in conjunction with a newly developed two-dimensional (2-D) synthetic array system to investigate the effects of phase cancellation and receiver aperture size on BUA estimations of bone tissue. In vitro ultrasound measurements were conducted on 54 trabecular bone samples (harvested from sheep femurs) in a confocal configuration with a focused transmitter and synthesized focused receivers of different aperture sizes. Phase sensitive (PS) and phase insensitive (PI) detections were performed. The results show that phase cancellation does have a significant effect on BUA. The normalized BUA (nBUA) with PS is 8.1% higher than PI nBUA while PI BUA is well correlated with PS BUA. Receiver aperture size also influences the BUA reading for both PI and PS detection and smaller receiver aperture tends to result in higher BUA readings. The results also indicate that the receiver aperture size used in the confocal configuration with PI detection should at least equal the aperture of the transmitter to capture most of the energy redistributed by the interference and diffraction from the trabecular bone.

Development of a method for ultrasound-guided placement of pedicle screws

Abstract-Many forms of spinal fusion involve the placement of long screws through the pedicles into the vertebral body. During the procedure, there is substantial risk of damage to vital neural and vascular structures due to the limited visibility of anatomic landmarks and high anatomic variability. As an alternative to current guidance systems, we have investigated the feasibility of performing ultrasound imaging through cancellous bone for the purpose of pedicle screw guidance. Quantitative ultrasonic characterization and A-mode imaging of seven defatted vertebral cancellous bone specimens was performed along the craniocaudal axis in water with unfocused, 1-MHz and 3.5- MHz broadband transducers. The center frequency attenuation increased considerably from 10.5 +/- 4.6 dB/cm at 1 MHz to 24.1 +/- 7.2 dB/cm at 3.5 MHz, while the speed of sound exhibited moderate positive dispersion, increasing from 1489 +/- 4.7 m/s at 1 MHz to 1494 +/- 4.2 m/s at 3.5 MHz. Despite the high attenuation and large specimen thickness (1.0-1.9 cm), A-mode imaging through cancellous bone to detect an aluminum reflector was possible in 83.2% and 70.1% of the cases at 1 MHz and 3.5 MHz, respectively. Specimen boundaries were identifiable with clinically sufficient average accuracy of 1.1 mm and 0.9 mm in the 1 MHz and 3.5 MHz A-mode images, respectively.

In vitro acoustic waves propagation in human and bovine cancellous bone

The acoustic behavior of cancellous bone with regard to its complex poroelastic nature has been investigated. The existence of two longitudinal modes of propagation is demonstrated in both bovine and human cancellous bone. Failure to take into account the presence of these two waves may result in inaccurate material characterization.

INTRODUCTION

Acoustic wave propagation is now a commonly used nondestructive method for cancellous bone characterization. However, wave propagation in this material may be affected by fluid-solid interactions inherent to its poroelastic nature, resulting in two different longitudinal waves. This phenomenon has been demonstrated in previous studies and is in agreement with Biot's theory. The purpose of this paper is to extend these findings to human trabecular bone and to thoroughly investigate these two waves.

MATERIALS AND METHODS

Sixty human and 14 bovine cancellous bone cubic specimens were tested in vitro in three different directions using an immersion acoustic transmission method. Original procedures were developed to quantify both velocity and attenuation characteristics of each wave. In term of attenuation, a modified broadband ultrasound attenuation (BUA), describing the rate of change of the frequency-dependent attenuation, was defined for each wave (FDUA).

RESULTS

Both waves were identified in most of the specimens. The fast wave velocities demonstrated a negative linear correlation with porosity (1500-2300 m/s, R2 = 0.44, p < 10(-3)), whereas the slow wave velocities exhibited two different behaviors: (1) a first set of data clearly dependent on porosity showing a positive linear correlation (1150-1450 m/s, R2 = 0.26, p < 10(-3)) and (2) a second group independent on porosity. The fast wave FDUA (20-140 dB/cmMHz) showed a parabolic behavior and reached a maximum for 75% porosity (second degree relationship R2 = 0.41,p < 10(-3)), whereas a positive linear behavior was observed for the slow wave FDUA (15-40 dB/cmMHz; R2 = 0.15, p < 10(-2)).

CONCLUSIONS

Existence of two wave propagation modes were demonstrated in human cancellous bone. Our data suggest that, in some cases, the amplitude of the slow wave is much larger than the amplitude of the fast wave. For this reason, care should be taken when using measurement systems that incorporate simple threshold detection because the fast wave could remain undetected. Moreover, failure to consider the presence of these two waves could result in an inaccurate quantification of cancellous bone physical properties.

On the ultrasonic attenuation and its frequency dependence in the os calcis assessed with a multielement receiver

Measurements of broadband ultrasound attenuation (BUA) on a single sample using different devices have previously shown significant variations. We have used a new 3x3 array transducer to allow the evaluation of three new approaches to the BUA algorithms and the extent to which the variations might be reduced. Ten human os calces were measured using the array. The attenuation and its slope, the BUA, were calculated using three methods: 1. A phase sensitive approach, summing the individual signals from the multielement array in the time domain (TD); 2. a phase insensitive approach, summing the signals in the frequency domain (FD); 3. averaging of the individually measured attenuation and BUA values (AV). The TD and AV approaches resulted in slightly larger values for both attenuation (approximately 4%) and BUA (approximately 7%) than with the FD method. The differences between the TD and FD method may be due to phase cancellation. However, the three sets of results were not significantly different (p<0.05), which suggests that measurements on commercial equipment using large receiver apertures are not strongly affected by phase cancellation, averaging or scattering. Nevertheless, it is not clear if the differences observed are of clinical relevance, which should be investigated in the future. Our study shows that the 3 approaches are strongly related (r2> or =0.94), suggesting that translation may be possible between methods.