Yassin Labyed

Ultrasound time-reversal MUSIC imaging with diffraction and attenuation compensation

Time-reversal imaging with multiple signal classification (TR-MUSIC) is an algorithm for imaging point-like scatterers embedded in a homogeneous and non-attenuative medium. We generalize this algorithm to account for the attenuation in the medium and the diffraction effects caused by the finite size of the transducer elements. The generalized algorithm yields higher-resolution images than those obtained with the original TR-MUSIC algorithm. We evaluate the axial and lateral resolutions of the images obtained with the generalized algorithm when noise corrupts the recorded signals and show that the axial resolution is degraded more than the lateral resolution. The TR-MUSIC algorithm is valid only when the number of point-like targets in the imaging plane is fewer than the number of transducer elements used to interrogate the medium. We remedy this shortcoming by dividing the imaging plane into subregions and applying the TR-MUSIC algorithm to the windowed backscattered signals corresponding to each subregion. The images of all subregions are then combined to form the total image. Imaging results of numerical and phantom data show that when the number of scatterers within each subregion is much smaller than the number of transducer elements, the windowing method yields super-resolution images with accurate scatterer localization. We use computer simulations and tissue-mimicking phantom data acquired with a real-time synthetic-aperture ultrasound system to illustrate the algorithms presented in the paper.

Super-resolution ultrasound imaging using a phase-coherent MUSIC method with compensation for the phase response of transducer elements

Time-reversal with multiple signal classification (TR-MUSIC) is an imaging method for locating point-like targets beyond the classic resolution limit. In the presence of noise, however, the super-resolution capability of TR-MUSIC is diminished. Recently a new method, phase-coherent MUSIC (PC-MUSIC), was developed. This algorithm modifies TR-MUSIC to make use of phase information from multiple frequencies to reduce noise effects and preserve the super resolution. PC-MUSIC however, ignores the phase response of the transducer elements. In this paper, we account for the phase response of the transducer elements in the derivation of the PC-MUSIC algorithm. Unfortunately, the phase response of the transducer elements may not be known beforehand. We develop an experimental method to estimate this response using measured signals scattered from a glass microsphere embedded in a tissue-mimicking phantom with a homogeneous background medium of a known sound speed. We use numerical simulations to illustrate that the maximum resolution achieved with PC-MUSIC is limited by the transducer bandwidth and the signal-to-noise ratio. We perform experiments on tissue-mimicking phantoms and compare images obtained with different imaging modalities, including X-ray mammography, synthetic-aperture ultrasound imaging, TR-MUSIC, and PC-MUSIC. We demonstrate the significantly improved resolving power of PC-MUSIC.

Ultrasound Time-Reversal MUSIC Imaging of Extended Targets

Ultrasound time-reversal imaging with multiple signal classification (TR-MUSIC) can produce images with subwavelength spatial resolution when the targets are point scatterers. In this experimental study, we evaluate the performance of the TR-MUSIC algorithm when the interrogated medium contains extended targets that cannot be considered as point scatterers, i.e., the size of the targets is on the order of the ultrasound wavelength or larger. We construct four tissue-mimicking phantoms, each of which contains glass spheres of a given size. We show that the quality of the phantom images obtained using the TR-MUSIC algorithm decreases with increasing sphere size. However, significant improvement is achieved when the image plane is divided into subregions, where each subregion is imaged separately. In this method, the TR-MUSIC calculations are performed on the windowed backscattered signals originating from each subregion. Our study demonstrates that the TR-MUSIC algorithm with time windowing can accurately locate extended targets but cannot provide the shape and reflectivity of the targets. We scan an inhomogeneous commercial tissue-mimicking phantom using an investigational synthetic-aperture ultrasound system, and show that the TR-MUSIC algorithm is capable of detecting small targets with high spatial resolution in inhomogeneous media.

Detecting breast microcalcifications using super-resolution ultrasound imaging: a clinical study

Imaging breast microcalcifications is crucial for early detection and diagnosis of breast cancer. It is challenging for current clinical ultrasound to image breast microcalcifications. However, new imaging techniques using data acquired with a synthetic-aperture ultrasound system have the potential to significantly improve ultrasound imaging. We recently developed a super-resolution ultrasound imaging method termed the phase-coherent multiple-signal classification (PC-MUSIC). This signal subspace method accounts for the phase response of transducer elements to improve image resolution. In this paper, we investigate the clinical feasibility of our super-resolution ultrasound imaging method for detecting breast microcalcifications. We use our custom-built, real-time synthetic-aperture ultrasound system to acquire breast ultrasound data for 40 patients whose mammograms show the presence of breast microcalcifications. We apply our super-resolution ultrasound imaging method to the patient data, and produce clear images of breast calcifications. Our super-resolution ultrasound PC-MUSIC imaging with synthetic-aperture ultrasound data can provide a new imaging modality for detecting breast microcalcifications in clinic without using ionizing radiation.

High-resolution imaging with a real-time synthetic aperture ultrasound system: a phantom study

It is difficult for ultrasound to image small targets such as breast microcalcifications. Synthetic aperture ultrasound imaging has recently developed as a promising tool to improve the capabilities of medical ultrasound. We use two different tissueequivalent phantoms to study the imaging capabilities of a real-time synthetic aperture ultrasound system for imaging small targets. The InnerVision ultrasound system DAS009 is an investigational system for real-time synthetic aperture ultrasound imaging. We use the system to image the two phantoms, and compare the images with those obtained from clinical scanners Acuson Sequoia 512 and Siemens S2000. Our results show that synthetic aperture ultrasound imaging produces images with higher resolution and less image artifacts than Acuson Sequoia 512 and Siemens S2000. In addition, we study the effects of sound speed on synthetic aperture ultrasound imaging and demonstrate that an accurate sound speed is very important for imaging small targets.

TR-MUSIC inversion of the density and compressibility contrasts of point scatterers

Time-reversal imaging with multiple signal classification (TR-MUSIC) is a super-resolution ultrasound imaging method for detecting point scatterers. This algorithm assumes that there is no contrast between the density of the point targets and that of the background medium, and that ultrasound scattering is caused only by the compressibility contrast. We modify the TR-MUSIC algorithm to account for ultrasound scattering from point targets with both density and compressibility contrasts. In addition, we develop an inversion method for estimating the density and compressibility contrasts of point scatterers with known locations. This approach is an extension of the inversion method previously developed by Devaney et al. for estimating the scattering strengths of point targets that have no density contrasts relative to the background medium. We use numerical phantom data to demonstrate that our new TR-MUSIC inversion algorithm can reliably estimate the density and compressibility contrasts of point scatterers. The estimates of these properties could be used for distinguishing breast calcifications from other tissue scatterers.

Detecting small targets using windowed time-reversal MUSIC imaging: A phantom study

Time reversal with multiple signal classification (TR-MUSIC) is a super-resolution imaging method for detecting targets smaller than the ultrasound wavelength. This method is valid only when the number of targets is fewer than the number of transducer elements used to interrogate the medium. We develop a windowed TR-MUSIC imaging method for obtaining high-resolution images even when the number of targets exceeds the number of transducer elements. Our method is based on dividing the imaging plane into sub-regions and applying the TR-MUSIC algorithm to the windowed backscattered signals corresponding to each sub-region. The images of all sub-regions are then combined to form the total image. We use tissue-mimicking phantom data acquired with a synthetic-aperture ultrasound system to demonstrate the significantly improved quality and resolution of the images obtained with the windowed TR-MUSIC method, particularly when the number of targets are larger than the number of transducer elements.

Toward real-time bent-ray breast ultrasound tomography using GPUs

Breast ultrasound tomography is a promising imaging modality that has the potential to improve the diagnosis and screening of breast cancer. We develop a bent-ray ultrasound tomography algorithm to reconstruct sound-speed images of the breast. We investigate the acceleration of the algorithm using graphical processing units (GPUs). We adapt the algorithmic steps of ultrasound bent-ray tomography to a GPU cluster, and use multi-GPU scaling to speed up the computation. Our results show that it is very promising to use a GPU cluster with multiple GPU cards to achieve nearly real-time tomographic reconstruction.

Attenuation Compensation and Estimation

Estimating the losses of ultrasound signal energy with propagation depth as a function of frequency is essential for quantifying tissue properties. Specifically, ultrasound attenuation is used to correct for spectral distortion prior to estimating quantitative ultrasound parameters to assess the tissue. Ultrasound attenuation can also be used independently to characterize the tissue. In this chapter, we review the primary algorithms for estimating both the local attenuation within a region of interest as well as the total attenuation between a region of interest and an ultrasound source. The strengths and weaknesses of each algorithm are also discussed.

Detection of breast microcalcifications using synthetic-aperture ultrasound

Ultrasound could be an attractive imaging modality for detecting breast microcalcifications, but it requires significant improvement in image resolution and quality. Recently, we have used tissue-equivalent phantoms to demonstrate that synthetic-aperture ultrasound has the potential to detect small targets. In this paper, we study the in vivo imaging capability of a real-time synthetic-aperture ultrasound system for detecting breast microcalcifications. This LANLs (Los Alamos National Laboratorys) custom built synthetic-aperture ultrasound system has a maximum frame rate of 25 Hz, and is one of the very first medical devices capable of acquiring synthetic-aperture ultrasound data and forming ultrasound images in real time, making the synthetic-aperture ultrasound feasible for clinical applications. We recruit patients whose screening mammograms show breast microcalcifications, and use LANLs synthetic-aperture ultrasound system to scan the regions with microcalcifications. Our preliminary in vivo patient imaging results demonstrate that synthetic-aperture ultrasound is a promising imaging modality for detecting breast microcalcifications.