G. Berger

Frequency dependence of ultrasonic backscattering in cancellous bone: Autocorrelation model and experimental results

The goal of this study is to model the frequency dependence of the ultrasonic backscatter coefficient in cancellous bone. A twofold theoretical approach has been adopted: the analytical theoretical model of Faran for spherical and cylindrical elastic scatterers, and the scattering model for weakly scattering medium in which the backscatter coefficient is related to the autocorrelation function of the propagating medium. The ultrasonic backscatter coefficient was measured in 19 bone specimens (human calcaneae) in the frequency range of 0.4-1.2 MHz. The autocorrelation function was computed from the three-dimensional (3D) microarchitecture measured using synchrotron radiation microtomography. Good agreement was found between the frequency dependence of the experimental (f3.38+/-0.31) and autocorrelation modeled (f3.48+/-0.26) backscatter coefficients. The results based on Faran theory (cylindrical Faran model: f2.89+/-0.06 and spherical Faran model: f3.91+/-0.04) show qualitative agreement with experimental data. The good prediction obtained by modeling the backscatter coefficient using the autocorrelation function of the medium opens interesting prospects for the investigation of the influence of bone microarchitecture on ultrasonic scattering.

Assessment of articular cartilage and subchondral bone : Subtle and progressive changes in experimental osteoarthritis using 50 MHz echography in vitro

The main objectives of this work were to demonstrate the potential of 50 MHz echography for assessing initial and progressive morphological and structural changes of articular cartilage and bone developed in an experimental model of osteoarthritis (OA). Degenerative lesions were induced in rat knees by the unilateral intra‐articular injection of a 3 mg dose of mono‐iodo‐acetic acid. To assess the lesion progression, the animals (n = 30) were sacrificed at different time intervals up to 8 weeks after the injection. Three‐dimensional echographic data were acquired in vitro on patellar cartilage and bone at various stages of the remodeling process using a scanning ultrasound microscope. Changes involving the OA cartilage characteristics are discussed relative to those of the contralateral control joint which received a placebo. Images of control cartilage showed a smooth hyperechoic articular surface and an echoic matrix. The cartilage thickness was 266 ± 44 μm (mean ± SD) in the central region of the tissue. The precision of ultrasonic thickness measurements was better than 1.3%. First changes in cartilage internal structure and subchondral bone appeared on ultrasound images 3 days after the injection and were even more evident by day 7. They resulted in a slight thinning of the cartilage, a 30% increase of its internal structure echogenicity, and the appearance of echoic zones in subchondral bone. Histologic findings confirmed chondrocyte depletion and degeneration, decrease of matrix proteoglycans, and fibrovascular connective tissue proliferation at the subchondral plate. Progressive and severe lesions at both bone and cartilage surface and internal structure were assessed and correlated to histologic features. These results show that high resolution echography is sensitive to subtle and progressive osteochondral remodeling. This technique has the potential to be used for intra‐articular quantitative imaging and assessment of early changes in bone and cartilage structure associated with natural human disease.

Parametric (integrated backscatter and attenuation) images constructed using backscattered radio frequency signals (25–56 MHz) from human aortae in vitro

Quantitative ultrasonic tissue characterization using backscattered high-frequency intravascular ultrasound could provide a basis for the objective identification of lesions in vivo. Representation of local measurements of quantitative ultrasonic parameters in a conventional image format should facilitate their interpretation and thus increase their clinical utility. Toward this goal, the apparent integrated backscatter, the slope of attenuation (25-56 MHz) and the value of the attenuation on the linear fit at 37.5 MHz were measured using the backscattered radio frequency signals from in vitro human aortae. Local estimations of these ultrasonic parameters from both normal and atherosclerotic aortic segments were displayed in a B-scan format. The morphological features of these parametric images corresponded well to features of histological images of the same regions. The attenuation from 25-56 MHz of seven segments of the medial layer (both with and without overlying atheroma) were measured using the multinarrow-band backscatter method. The average attenuation in the media at 24 degrees C +/- 3 degrees C was 45 +/- 16 dB/cm at 25 MHz and 102 +/- 13 dB/cm at 50 MHz. This work represents progress toward the development of quantitative imaging methods for intravascular applications.

Assessment of the relationship between broadband ultrasound attenuation and bone mineral density at the calcaneus using BUA imaging and DXA.

The purposes of this study was to determine the relationship between broadband ultrasound attenuation (BUA) and bone mineral density (BMD) measured at different regions of the calcaneus with identical site-matched regions of interest (ROIs). Dual-energy X-ray absorptiometry (DXA) measurements of the calcaneus and BUA imaging were performed in 30 women (15 premenopausal and 15 postmenopausal). Four square ROIs were located in the great tuberosity and one square ROI in the foramen calcaneus. A ROI adapted to the shape and size of the whole calcaneus was also considered. All ROIs were analyzed three times with both techniques to minimize intra-observer variability. The correlation coefficient between attenuation and frequency was used as an index of BUA measurement error. Before accepting a measurement of BUA in inhomogeneous material, it could be useful to map the spatial variations of the measurement error. In all ROIs we found the BUA and BMD were strongly related (r=0.78−0.91,p<0.001). The correlation between BUA and BMD was slightly higher in the inferior part of the posterior tuberosity than in the superior part and in the foramen calcaneus. The very high correlation between attenuation and frequency found in all ROIs (r=0.99) suggests that measurement errors of propagation were probably not significant. Ultrasound imaging yields the opportunity for studying the spatial acoustic properties in the calcaneus and their relation to bone mass or structural parameters provided by independent imaging techniques. BUA measured with current transmission techniques reflects mainly bone mass, and microarchitecture to a smaller extent.

Ultrasound parametric imaging of the calcaneus:In vivo results with a new device

An ultrasound transmission scanning system was constructed to makein vivo parametric images of the acoustic properties of the heel. Broadband ultrasonic attenuation (BUA) images were obtained in transmit mode by using a pair of broadband focused transducers (center frequency 0.5 MHz, diameter 29 mm, focus 50 mm) immersed in a water bath at room temperature. With these characteristics, the theoretical beam width at the focus was approximately 5 mm. The total duration of the acquisition period was 3 minutes. Comparison of the BUA image and the radiograph of the calcaneus was possible in one case and showed that all the anatomical details could be identified. The images reported here demonstrate the wide range of BUA found in both the whole bone and within a ROI centered in the posterior part of the bone thus reinforcing the idea of tremendous heterogeneity of the acoustic properties of bone. This suggests that the accurate control of the position of the measurement site is of the utmost importance for between-subject comparison and for repetitive measurements. We proposed a new method, the likelihood image, as an efficient way of highlighting the regions of the image suspected to be subject to waveform distortion. It could be used to guide the selection of the optimal measurement site. Our results suggest that ultrasound parametric imaging has the potential for enhancing the current ultrasound technique by (1) allowing reproducible, repetitive measurements, (2) permitting the selection of similar optimal measurement sites in all subjects, and (3) avoiding accuracy erros due to wave-form distortion.

Analysis of the axial transmission technique for the assessment of skeletal status

Ultrasonic wave propagation in human cortical bone has been investigated in vitro using the so-called axial transmission technique. This technique, which relies on velocity measurement of the first arriving signal, has been used in earlier investigations to study bone status during fracture healing or osteoporosis. Two quasi-point-source elements, one transmitter and one receiver (central frequency 0.5 MHz), were used to generate a wide ultrasonic beam, part of which strikes the sample surface at the longitudinal critical angle, and to receive the signals reflected from the sample surface. The analysis of the field reflected from a fluid-solid interface for an incident spherical wave predicts the existence of a lateral wave propagating along the sample surface at a velocity close to the longitudinal velocity, in addition to the ordinary reflected wave and vibration modes. The transducer-sample and the transmitter-receiver distances were chosen such that the lateral wave is the first arriving signal. Validation of the measuring technique was performed on test materials and was followed by experiments on human cortical bones. Experimental results (arrival time and velocity) strongly suggest that the first detected signal corresponds to the lateral wave predicted by theory.

Quantitative ultrasound imaging at the calcaneus using an automatic region of interest.

A new approach to measuring bone properties at the calcaneus using ultrasound parametric imaging has recently emerged. However, an additional source of observer-related error is the substantial regional variations in the pattern of ultrasound parameters. The contribution of intra-observer and inter-observer variability to the coefficient of variation can be eliminated using an algorithm which selects the region of interest (ROI) completely automatically. The objective of the present study was the clinical assessment of an automatic ROI for both broadband ultrasonic attenuation (BUA) and speed of sound (SOS) measurement using ultrasound parametric imaging. The automatic ROI was defined as the circular region of lowest attenuation in the posterior tuberosity of the calcaneus. We have tested this algorithm using clinical images of the calcaneus from 265 women. Mean coefficients of variation were 1.6% (95% confidence interval 1.4%−1.9%) and 0.26% (95% confidence interval 0.23%−0.32%) for BUA and SOS respectively (standardized CV was 2.1% for BUA and 2.6% for SOS). Z-scores in an osteoporotic group were −0.61 and −0.52 for BUA and SOS respectively. In healthy women, the age-related decline was −0.50 dB/ MHz per year (0.7%/year) for BUA and −1.2 m/s per year (0.08%/year) for SOS. In the subgroup of healthy postmenopausal women, using stepwise multiple regression, we found that BUA was predicted best by years since menopause (YSM) and weight, with overall modelr2=0.28; SOS was predicted best by YSM only (r2=0.21). Neither the range of biological variation of ultrasound parameters nor the clinical value were affected by the choice of the region of lowest attenuation for measurement. The automatic procedure was totally independent of operator interaction, therefore excluding loss of precision due to intra- or inter-observer variability. The results showed the high precision and robustness of the procedure. These factors make this approach viable for routine clinical use.

Correlation of ultrasonic attenuation (30 to 50 MHz) and constituents of atherosclerotic plaque

The ultrasonic integrated attenuation and the slope of attenuation (30-50 MHz) were measured in vitro at 20 degrees +/- 2 degrees C using radio frequency signals backscattered from human aortae. Ultrasonic measurements and histologic classifications were made in a total of 124 local regions from 58 independent segments of aortae. Values of the integrated attenuation were significantly higher in collagen-lipidic (142 +/- 51 dB cm-1, n = 18), and lipidic regions (139 +/- 53 dB cm-1, n = 11) compared to regions of normal media (97 +/- 20 dB cm-1, n = 44) and dense collagen (107 +/- 33 dB cm-1, n = 43). The most elevated integrated attenuation values were observed in calcified regions (245 +/- 93 dB cm-1, n = 8). The slope of attenuation was significantly higher in lipidic than in normal media (p = 0.002), dense collagen (p = 0.0007) or collagenlipidic (p = 0.04) regions. The correlation between attenuation and local tissue composition was used to establish ranges of values of integrated attenuation that are most likely to indicate specific tissue types. Images of the local tissue type were constructed. Comparison of these quantitative images with the corresponding histologic sections demonstrates that attenuation measurements offer promise for the in vivo characterization of plaque structure and composition.

Ultrasound Measurement on the Calcaneus: Influence of Immersion Time and Rotation of the Foot

Abstract: The aim of this study was to evaluate the influence of immersion time and rotation of the heel around the leg axis on the reproducibility of measurements using an ultrasound bone imaging scanner (UBIS) with a temperature-controlled water bath. Measurements were obtained in 10 men, 11 premenopausal women and 10 postmenopausal women. The right foot of all subjects was scanned 12 times with an interval of 3 min between each scan. The first 10 measurements, performed to study the effect of immersion time, were taken without intermediate repositioning. Measurements 11 and 12 were also taken without removing the foot, but the calcaneus position was varied by an angle of ±2.5° with respect to the reference position used during the first 10 measurements. Prolonged immersion of the heel led to a decrease in the variations of quantitative ultrasound (QUS) parameters between successive measurements. Following rapid variations with immersion time, the QUS parameters reached a plateau. The effect of immersion time on measurement error remained significant until the first two or three scans for broadband ultrasound attenuation (BUA) and until the sixth or seventh scans for speed of sound (SOS). The variation in BUA was more pronounced (p<0.05) for the group of postmenopausal women (20.7% change in BUA between the first and tenth scans; p<0.005) than for the group of premenopausal women (6.8% change in BUA between the first and tenth scans; p<0.005). The variations in SOS were similar in the two groups (0.8% variation; p<0.005). The impact of immersion time was smaller for men than for women [2.5%, (p<0.01) and 0.4% (p<0.005) of the change between the first and tenth scans for BUA and SOS respectively]. On the whole, the measurement errors due to rotation of the heel were lower than those caused by immersion time. The variations were significant only in men and premenopausal women. Both immersion time and rotation of the heel may play a role in the precision of QUS and should be carefully standardized, particularly for longitudinal studies. In addition, following these results we have adopted a standardized protocol to derive the technique reproducibility in groups of premenopausal and postmenopausal women. The coefficients of variation were 1.1% for BUA and 0.1% for SOS in premenopausal women, and 1.4% for BUA and 0.13% for SOS in postmenopausal women.

Automatic three-dimensional reconstruction and characterization of articular cartilage from high-resolution ultrasound acquisitions.

Three-dimensional (3D) high-resolution ultrasonography has proved to be useful for in vitro assessment of cartilage remodeling due to osteoarthritis. The diagnosis is performed by computation of the mean thickness of the cartilage, which reveals hypertrophy or thinning, and by 3D reconstruction of the data, which provides essential information about the size, extent, and localization of the lesion. In both cases, preliminary segmention of the cartilage is necessary. This article proposes an algorithm for automatic segmentation of the cartilage from 3D ultrasonic acquisitions of the rat patella, which includes the detection of the cartilage surface and the cartilage/bone interface. The method was designed on the assumption of regularity and smoothness of the interfaces. The use of a global threshold was sufficient to separate the patella area from the background. The cartilage/bone interface was detected by selection of regions of interest (ROIs) encompassing the interface, followed by the detection of the interface within these ROIs using the graph theory. The method was applied to 162 samples. The detection accuracy was judged to be very good or good in 99% of the cases for the cartilage surface and in 86% of the cases for the cartilage/bone interface. The mean cartilage thickness value in the central part of the patella obtained from the automatic detection method was compared to that obtained manually. The coefficient of correlation between the two measurements was 0.92. These results show that our method is reliable. Thus, fast processing of a large number of acquisitions and a more complete analysis of the cartilage surface become possible.