J A Evans

Dependence of the velocity and attenuation of ultrasound in bone on the mineral content

Measurements of the attenuation and velocity of ultrasound from 0.3 to 0.8 MHz have been performed on a number of bovine cancellous bone samples. The influence of bone mineral content has been isolated by measuring the acoustic properties of the samples at various stages of demineralization resulting from controlled nitric acid attack. The correlation coefficient r, between the attenuation at different frequencies and bone density was found to be in the range 0.68-0.97. Broadband ultrasonic attenuation (BUA) was also calculated and produced r values between 0.84 and 0.99. The velocity measurements indicated a correlation greater than 0.97 in all cases. Thus velocity is the parameter most sensitive to changes in bone mineral density alone. Attenuation and BUA are less well correlated presumably because of a sensitivity to minor structural change.

Ultrasonic attenuation and velocity in bone.

The measurement of attenuation and velocity of ultrasound in cancellous bovine femora has been studied. The dependence of both attenuation between 0.2 and 0.8 MHZ and velocity on the bone density has been measured. The results show a correlation coefficient of around 0.5 for attenuation and density, and a value of roughly 0.85 for velocity and density. The clinical consequences for the use of low frequency ultrasound as a diagnostic tool in bone disease are discussed.

The effect of bone structure on ultrasonic attenuation and velocity

The relationship between the structure of bovine cancellous bone, and its ultrasonic propagation parameters is investigated by means of a novel technique involving the application of large static loads, thereby changing the porosity in a controlled manner. The results show that for frequencies in the range 0.4 to 1 MHz, porosity decreases up to 35% are associated with a reduction in attenuation of up to 500%, whereas the velocity increases by roughly 35% for the same changes. The data taken overall suggest that in determining the ultrasonic attenuation coefficient at these frequencies, the amount of material in a given bone section is significantly less important than the distribution of that material.

On the measurement of the velocity of ultrasound in the os calcis using short pulses

Abstract Objectives: The aim of this study was compare several methods for the measurement of velocity of ultrasound in the os calcis and to evaluate the extent to which dispersion is present and influential. Methods: The phase velocity of ultrasound in a series of 10 fixed os calces was measured by performing phase spectrum analyses on short pulses. The group velocity in these samples was also calculated from the short pulses using the slope of the phase velocity as a function of frequency between 600 and 800 kHz. Additional to these two methods, a statistical method, the cross correlation technique, was used to calculate the mean pulse velocity. Results: Negative dispersion was found in all the os calces using the slope of the phase velocity as a function of frequency. Consequently, the group velocity at any frequency was found to be less than the phase velocity at that frequency. When the cross correlation technique was applied, the velocity values obtained were slightly greater than the phase velocity values for most of the bones at 750 kHz. This has also been shown using a computer simulation. The attenuation coefficient at 700 kHz was also measured on these os calces. The values obtained correlated better with the velocity using the cross correlation technique ( r 2 = 95.7%) and the phase velocity ( r 2 = 95.0%), than with the group velocity calculated from the slope of the phase velocity ( r 2 = 29.5%). The coefficients of correlation between the normalised broadband ultrasound attenuation and the velocities were similar to those given above. Conclusions: We have shown that all three alternative methods are more accurate than currently used time-domain methods found in commercial systems.

Implementing a newly proposed Monte Carlo based small field dosimetry formalism for a comprehensive set of diode detectors

PURPOSE The goal of this work was to implement a recently proposed small field dosimetry formalism [Alfonso et al., Med. Phys. 35(12), 5179-5186 (2008)] for a comprehensive set of diode detectors and provide the required Monte Carlo generated factors to correct measurement. METHODS Jaw collimated square small field sizes of side 0.5, 0.6, 0.7, 0.8, 0.9, 1.0, and 3.0 cm normalized to a reference field of 5.0 cm × 5.0 cm were used throughout this study. Initial linac modeling was performed with electron source parameters at 6.0, 6.1, and 6.2 MeV with the Gaussian FWHM decreased in steps of 0.010 cm from 0.150 to 0.100 cm. DOSRZnrc was used to develop models of the IBA stereotactic field diode (SFD) as well as the PTW T60008, T60012, T60016, and T60017 field diodes. Simulations were run and isocentric, detector specific, output ratios (OR(det)) calculated at depths of 1.5, 5.0, and 10.0 cm. This was performed using the following source parameter subset: 6.1 and 6.2 MeV with a FWHM = 0.100, 0.110, and 0.120 cm. The source parameters were finalized by comparing experimental detector specific output ratios with simulation. Simulations were then run with the active volume and surrounding materials set to water and the replacement correction factors calculated according to the newly proposed formalism. RESULTS In all cases, the experimental field size widths (at the 50% level) were found to be smaller than the nominal, and therefore, the simulated field sizes were adjusted accordingly. At a FWHM = 0.150 cm simulation produced penumbral widths that were too broad. The fit improved as the FWHM was decreased, yet for all but the smallest field size worsened again at a FWHM = 0.100 cm. The simulated OR(det) were found to be greater than, equivalent to and less than experiment for spot size FWHM = 0.100, 0.110, and 0.120 cm, respectively. This is due to the change in source occlusion as a function of FWHM and field size. The corrections required for the 0.5 cm field size were 0.95 (± 1.0%) for the SFD, T60012 and T60017 diodes and 0.90 (± 1.0%) for the T60008 and T60016 diodes-indicating measured output ratios to be 5% and 10% high, respectively. Our results also revealed the correction factors to be the same within statistical variation at all depths considered. CONCLUSIONS A number of general conclusions are evident: (1) small field OR(det) are very sensitive to the simulated source parameters, and therefore, rigorous Monte Carlo linac model commissioning, with respect to measurement, must be pursued prior to use, (2) backscattered dose to the monitor chamber should be included in simulated OR(det) calculations, (3) the corrections required for diode detectors are design dependent and therefore detailed detector modeling is required, and (4) the reported detector specific correction factors may be applied to experimental small field OR(det) consistent with those presented here.

Reproducibility of ultrasound measurement of the abdominal aorta

Abdominal aortic aneurysm (AAA) screening and surveillance programmes use ultrasound imaging to measure the anteroposterior (AP) diameter of the infrarenal aorta. The aim of this study was to examine potential observer bias and variability in ultrasound measurements.

The influence of porosity and pore size on the ultrasonic properties of bone investigated using a phantom material

Ultrasonic propagation in bone has been investigated using the Leeds Ultrasonic Bone Phantom Material. Phantoms were produced with different porosities in the range of 45−83% and pore sizes of 1.3 and 0.6 mm. The phase velocity at 600 kHz was found to follow a second-order polynomial as a function of porosity. Phase velocity values between 1545 and 2211 m s−1 were measured and found to be largely independent of pore size for a given porosity. The slope of the phase velocity as a function of frequency (dispersion) decreases with increasing porosity. The values obtained from samples having different pore sizes were also similiar. The attenuation coefficient and normalized broadband ultrasonic attenuation (nBUA) reached a maximum at about 50%. The normalized attenuation ranged from 6 to 25 dB cm−1 over the porosity range available and consistently showed higher values for the larger pore size. Similarly, the nBUA values were found to be between 14 and 53 dB MHz−1 cm−1, with the values for the larger pore size being roughly 10 dB MHz−1 cm−1 greater than those for the smaller pore size. These findings demonstrate that the Leeds phantom can be used to investigate the effect of structural changes in bone and to aid the understanding of quantitative ultrasound. The results support the assumption that the velocity in trabecular bone is not dependent on pore size but is influenced by the mechanical properties of the bones constituents and the overall framework, whereas the attenuation and BUA are also influenced by structure.

The measurement of the velocity of ultrasound in fixed trabecular bone using broadband pulses and single-frequency tone bursts

The velocity of ultrasound in a series of 10 fixed os calces was measured using both short pulses and 750 kHz tonebursts. The values obtained from the pulse measurements differed from the toneburst values by up to 16% depending on the selection of the zero-crossing point used as a reference in the pulse measurements. It is demonstrated that the discrepancy between the values is itself a function of the frequency-dependent attenuation in the propagating medium and this is confirmed by theoretical modelling. The toneburst results are also compared with measurements using a cross-correlation technique, and these show a close agreement.

A PHANTOM FOR QUANTITATIVE ULTRASOUND OF TRABECULAR BONE

The propagation mechanisms of ultrasound in trabecular bone are poorly understood and have been the subject of extended debate; also, the reproducibility of ultrasonic measurements on bone in vivo using commercial ultrasound heel-scanning devices is such that the interpretation of the obtained data is difficult. In this paper we describe recent developments in the production of a bone-mimicking material which is well suited to the task of routine monitoring of commercial ultrasound bone scanners. The material, based on a standard epoxy resin is fabricated to a pre-determined porosity value by the inclusion of a marrow-mimicking material thereby introducing a known and controlled mean pore size. Measurements of the velocity and attenuation of the material have been performed over a range of porosity values from 10% to 80% in the frequency range 500-900 kHz; also, broadband ultrasonic attenuation (BUA) values have been obtained from commercial equipment. The material displays velocities in the range 1844-3118 m s(-1) and attenuation ranging from 7.0 to 17.7 dB cm(-1) at 500 kHz.

Monte Carlo modelling of diode detectors for small field MV photon dosimetry: detector model simplification and the sensitivity of correction factors to source parameterization

The goal of this work was to examine the use of simplified diode detector models within a recently proposed Monte Carlo (MC) based small field dosimetry formalism and to investigate the influence of electron source parameterization has on MC calculated correction factors. BEAMnrc was used to model Varian 6 MV jaw-collimated square field sizes down to 0.5 cm. The IBA stereotactic field diode (SFD), PTW T60016 (shielded) and PTW T60017 (un-shielded) diodes were modelled in DOSRZnrc and isocentric output ratios (OR(fclin)(detMC)) calculated at depths of d = 1.5, 5.0 and 10.0 cm. Simplified detector models were then tested by evaluating the percent difference in (OR(fclin)(detMC)) between the simplified and complete detector models. The influence of active volume dimension on simulated output ratio and response factor was also investigated. The sensitivity of each MC calculated replacement correction factor (k(fclin,fmsr)(Qclin,Qmsr)), as a function of electron FWHM between 0.100 and 0.150 cm and energy between 5.5 and 6.5 MeV, was investigated for the same set of small field sizes using the simplified detector models. The SFD diode can be approximated simply as a silicon chip in water, the T60016 shielded diode can be modelled as a chip in water plus the entire shielding geometry and the T60017 unshielded diode as a chip in water plus the filter plate located upstream. The detector-specific (k(fclin,fmsr)(Qclin,Qmsr)), required to correct measured output ratios using the SFD, T60016 and T60017 diode detectors are insensitive to incident electron energy between 5.5 and 6.5 MeV and spot size variation between FWHM = 0.100 and 0.150 cm. Three general conclusions come out of this work: (1) detector models can be simplified to produce OR(fclin)(detMC) to within 1.0% of those calculated using the complete geometry, where typically not only the silicon chip, but also any high density components close to the chip, such as scattering plates or shielding material is necessary to be included in the model, (2) diode detectors of smaller active radius require less of a correction and (3) (k(fclin,fmsr)(Qclin,Qmsr)) is insensitive to the incident the electron energy and spot size variations investigated. Therefore, simplified detector models can be used with acceptable accuracy within the recently proposed small field dosimetry formalism.