Search

This paper describes work aimed at combining 3D ultrasound with full-field digital mammography via a semi-automatic prototype ultrasound scanning mechanism attached to the digital mammography system gantry. Initial efforts to obtain high x-ray and ultrasound image quality through a compression paddle are proving successful. Registration between the x-ray mammogram and ultrasound image volumes is quite promising when the breast is stably compressed. This prototype system takes advantage of many synergies between the co-registered digital mammography and pulse-echo ultrasound image data used for breast cancer detection and diagnosis. In addition, innovative combinations of advanced US and X-ray applications are being implemented and tested along with the basic modes. The basic and advanced applications are those that should provide relatively independent information about the breast tissues. Advanced applications include x-ray tomosynthesis, for 3D delineation of mammographic structures, and non-linear elasticity and 3D color flow imaging by ultrasound, for mechanical and physiological information unavailable from conventional, non-contrast x-ray and ultrasound imaging.

An investigation was performed to determine whether the sound-attenuation-insuspensions theory of Allegra and Hawley can be used to explain the compressional (longitudinal wave) attenuation of ultrasonically tissue-mimicking materials commonly used in phantoms for testing the performance of medical ultrasound systems. These materials are composed of microscopic graphite particles suspended in a gel. The theory was first tested using materials containing spherical glass beads instead of graphite particles because these materials more closely fit the geometric conditions assumed in the theory. For the glass bead type materials as well as the graphite particle type materials, the attenuation coefficients predicted using the Allegra and Hawley model agreed rather well with experimental measurements over the diagnostic frequency range.

This paper reports the development of an improved three-dimensional computer simulation model for evaluation of ultrasonic imaging systems. This model was used to successfully evaluate a signal processing method for improving lesion detection in ultrasound imaging. Linear processing of the rf signal amplitudes from a limited region of tissue was compared with the logarithmic compression employed by most commercial scanners. Two lesions were simulated by spherical distributions of scatterers having backscatter coefficients greater than the scatterers in the surrounding medium. Linear processing improved the differential contrast by a factor of about two. The simulation is based on the three-dimensional distribution of acoustic frequency spectra in a transducer beam and integration of scattered pulses from a corresponding three-dimensional array of scatterers. The simulation reported in previous papers depended upon physical measurement of the impulse response of a transducer. An original contribution described briefly herein, and in more detail in a companion article, is the addition of a model of the transducer's pulse waveform generation. Another new addition is the definition of a specific lesion detection task for objective assessment of a change in image quality following perturbation of some system parameter.

Experimental results for the angular distribution of scattered ultrasound in water from a 0.635 mm diameter steel sphere embedded in agar were obtained for interrogating frequencies of 1.0, 2.25, 3.5 and 5.0 MHz. The experimental results compared favorably with theoretical predictions based on the work of Faran. The results for both theory and experiment are presented in the form of the Differential Scattering Cross Section (DSC) for scattering angles from 170 degrees to 44 degrees. The implications of this study on calculating the expected scattering behavior of gelatin based ultrasound phantoms are briefly discussed.

A three-dimensional model for production of gray scale texture in ultrasound B-mode images is described. The model computes time-dependent echo signals resulting from scattering of acoustic pulses by particles randomly distributed in an attenuating medium and transforms these signals into a gray scale image. Specific transducer and pulser-receiver characteristics are accounted for, as well as the three-dimensional nature of the problem, without loss of computational efficiency. The model generates texture that closely corresponds to that found experimentally in ultrasound images of tissue-mimicking phantoms. The dependence of the texture upon the depth of the region that was scanned and on the characteristics of the transducer-receiver system were clearly demonstrated. Good agreement between theory and experiment was found for the texture in phantoms containing simulated spherical low-scatter tumors.

When modeling or simulating an ultrasonic pulse-echo system it is necessary to know the transducer waveforms that are the input to the system. In general, accurate transducer modeling is complex and waveform calculation is computationally intensive. Because of this, investigators often assume these waveforms to be simple sinusoids, modulated by Gaussian or exponential envelopes. However, these latter type pulses do not properly represent the complex response of modern piezoelectric transducers. In this paper, a simple set of equations is presented which can approximate the behavior of a number of common transducer configurations. In the first step of the procedure, a simple but accurate model is employed to calculate the step responses of quarter-wave-matched and backed piezoelectric transducers, assuming open circuit conditions. Effects of electrical terminations and transmit/receive bandwidth limitations are accounted for, approximately, by cascaded filtering. Emitted pressure waveforms and echo voltage waveforms synthesized by this simple procedure are very similar to measured waveforms from real transducers. Under most conditions, the method is only an approximation because the model erroneously assumes isolation between the electrical and acoustic networks. However, under many conditions the model is sufficiently accurate to predict actual transducer performance.

A three-dimensional model for gray scale texture in ultrasound B-mode images has been extended to enable absolute quantitative predictions of echo signal levels and texture characteristics. The quantitative aspects of the model are tested with a phantom in which the ultrasonic scattering properties as well as the ultrasonic attenuation coefficient and the speed of sound are known as functions of frequency. Good agreement between theoretical predictions and experimental results was found in the overall brightness of B-mode images, in pixel value histograms, and in an autocorrelation function analysis of images. Passing quantitative tests is additional evidence that the model is a reliable tool for computer simulations involving many important facets of ultrasound B-mode imaging and tissue characterization.