This PDF file contains the front matter associated with SPIE Proceedings Volume 10139, including the Title Page, Copyright information, Table of Contents, and Conference Committee listing.

The cardiac elastography aims at identification of non-transmural infarctions. Two displacement estimation methods in such an application using synthetic ultrasonic data are studied. Reference was obtained from Finite Element Modelling. Models had the form of half of an ellipsoid with 15 mm wall thickness. The homogenous model, models with transmural and nontransmural inclusion were designed. Deformation of the models was simulated using Abaqus. Ultrasonic data of LAX and SAX views were generated using Field II. Radial (dR) and lateral (dL) displacements were estimated using a 2D correlation search with 2D stretching (2DCS) and B-spline (BS) method. Strains were estimated using least squares estimator. Mean Absolute Error (MAE) of the dR in the LAX view was approx. 6[μm] for 2DCS and 8[μm] for BS, that of the dL 30 and 24[μm] respectively. MAE of the second component of the principal strain (epsilon)₂ was 0.10 and 0.14[%], respectively. Corresponding values for SAX view were 7, 10, 42, 52[μm] and 0.47 and 1.08[%]. In the LAX view both estimation methods result in the (epsilon)₂ behavior coherent with the presence of the inclusion, with the 2DCS results closer to the reference. In the SAX view the BS approach results in high errors of the estimate. The (epsilon)₂ profiles, LAX view, show minor discrepancies with respect to the reference and show the effect of the inclusion. The (epsilon)₂ profiles, SAX view, obtained from displacements estimated using the BS method strongly deviate from the reference. Block matching performs better in application to the local strain estimation.

Ultrasound elastography entails imaging mechanical properties of tissue and is therefore of significant clinical importance. In elastography, two frames of radio-frequency (RF) ultrasound data that are obtained while the tissue is undergoing deformation, and the time-delay estimate (TDE) between the two frames is used to infer mechanical properties of tissue. TDE is a critical step in elastography, and is challenging due to noise and signal decorrelation. This paper presents a novel and robust technique TDE using all samples of RF data simultaneously. We assume tissue deformation can be approximated by an affine transformation, and hence call our method ATME (Affine Transformation Model Elastography). The affine transformation model is utilized to obtain initial estimates of axial and lateral displacement fields. The affine transformation only has six degrees of freedom (DOF), and as such, can be efficiently estimated. A nonlinear cost function that incorporates similarity of RF data intensity and prior information of displacement continuity is formulated to fine-tune the initial affine deformation field. Optimization of this function involves searching for TDE of all samples of the RF data. The optimization problem is converted to a sparse linear system of equations, which can be solved in real-time. Results on simulation are presented for validation. We further collect RF data from in-vivo patellar tendon and medial collateral ligament (MCL), and show that ATME can be used to accurately track tissue displacement.

Speckle tracking is a common method for non-rigid tissue motion analysis in 3D echocardiography, where unique texture patterns are tracked through the cardiac cycle. However, poor tracking often occurs due to inherent ultrasound issues, such as image artifacts and speckle decorrelation; thus regularization is required. Various methods, such as optical flow, elastic registration, and block matching techniques have been proposed to track speckle motion. Such methods typically apply spatial and temporal regularization in a separate manner. In this paper, we propose a joint spatiotemporal regularization method based on an adaptive dictionary representation of the dense 3D+time Lagrangian motion field. Sparse dictionaries have good signal adaptive and noise-reduction properties; however, they are prone to quantization errors. Our method takes advantage of the desirable noise suppression, while avoiding the undesirable quantization error. The idea is to enforce regularization only on the poorly tracked trajectories. Specifically, our method 1.) builds data-driven 4-dimensional dictionary of Lagrangian displacements using sparse learning, 2.) automatically identifies poorly tracked trajectories (outliers) based on sparse reconstruction errors, and 3.) performs sparse reconstruction of the outliers only. Our approach can be applied on dense Lagrangian motion fields calculated by any method. We demonstrate the effectiveness of our approach on a baseline block matching speckle tracking and evaluate performance of the proposed algorithm using tracking and strain accuracy analysis.

Elastography comprises a set of modalities that image the biomechanical properties of soft tissues for disease detection and diagnosis. Quasi-static ultrasound elastography, in particular, tracks sub-surface displacements resulting from an applied surface force. The local displacement information and measured surface loads may be used to compute a parametric summary of biomechanical properties; however, the inverse problem is under- determined, limiting most techniques to estimating a single linear-elastic parameter. We previously described a new method to develop mechanical models using a combination of computational mechanics and machine learning that circumvents the limitations associated with the inverse problem. The Autoprogressive method weaves together finite element analysis and artificial neural networks (ANNs) to develop empirical models of mechanical behavior using only measured force-displacement data. We are extending that work by incorporating spatial information with the material properties. Previously, the ANNs accepted only a strain vector input and computed the corresponding stress, meaning any spatial information was encoded in the finite element mesh. Now, using a pair of ANNs working in tandem with spatial coordinates included as part of the input, these new Cartesian ANNs are able to learn the spatially varying mechanical behavior of complex media. We show that a single Cartesian ANN is able to describe the same mechanical behavior of an object that previously required at least two ANNs. Furthermore, we show the new ANNs can learn complex material property distributions and reconstruct images of the Young’s modulus distribution, not merely classify, filter, or otherwise process an existing image. For the first time, we present results using Cartesian neural networks within the Autoprogressive Method to form elastic modulus images.

The pulse wave velocity (PWV) is considered one of the most important clinical parameters to evaluate CV risk, vascular adaptation, etc. There has been substantial work attempting to measure the PWV in peripheral vessels using ultrasound (US). This paper presents a fully automatic algorithm for PWV estimation from the human carotid using US sequences acquired with a Logic E9 scanner (modified for RF data capture) and a 9L probe. Our algorithm samples the pressure wave in time by tracking wall displacements over the sequence, and estimates the PWV by calculating the temporal shift between two sampled waves at two distinct locations. Several recent studies have utilized similar ideas along with speckle tracking tools and high frame rate (above 1 KHz) sequences to estimate the PWV. To explore PWV estimation in a more typical clinical setting, we used focused-beam scanning, which yields relatively low frame rates and small fields of view (e.g., 200 Hz for 16.7 mm filed of view). For our application, a 200 Hz frame rate is low. In particular, the sub-frame temporal accuracy required for PWV estimation between locations 16.7 mm apart, ranges from 0.82 of a frame for 4m/s, to 0.33 for 10m/s. When the distance is further reduced (to 0.28 mm between two beams), the sub-frame precision is in parts per thousand (ppt) of the frame (5 ppt for 10m/s). As such, the contributions of our algorithm and this paper are: 1. Ability to work with low frame-rate ( 200Hz) and decreased lateral field of view. 2. Fully automatic segmentation of the wall intima (using raw RF images). 3. Collaborative Speckle Tracking of 2D axial and lateral carotid wall motion. 4. Outlier robust PWV calculation from multiple votes using RANSAC. 5. Algorithm evaluation on volunteers of different ages and health conditions.

In recent years, many research studies have been carried out on ultrasound computed tomography (USCT) for improving the detection and management of breast cancer. This paper investigates a signal pre-processing method based on frequency-shift low-pass filtering (FSLF) and least mean square adaptive filtering (LMSAF) for USCT image quality enhancement (proposed in our previous work). FSLF is designed base on Zoom Fast Fourier Transform algorithm (ZFFT) for processing the ultrasound signals in the frequency domain, while LMSAPF is based on the least mean square (LMS) algorithm in the time domain. Through the combination of the two filters, the ultrasound image is expected to have less noises and artifacts, and higher resolution and contrast. The proposed method was verified with the radio-frequency (RF) data of the nylon threads and the breast phantom captured by the USCT system developed in the Medical Ultrasound Laboratory. Experimental results show that the reconstructed images of nylon threads by the proposed method had narrower main lobe width and lower side lobe level comparing to the delay-and-sum (DAS). The background noises and artifacts could also be efficiently restrained. The reconstructed image of breast phantom by the proposed method had a higher resolution and the contrast ratio (CR) could be enhanced for about 12dB to 18dB at different region of interest (ROI).

In this work, we construct a multi-frequency accelerating strategy for the contrast source inversion (CSI) method using pulse data in the time domain. CSI is a frequency-domain inversion method for ultrasound waveform tomography that does not require the forward solver through the process of reconstruction. Several prior researches show that the CSI method has a good performance of convergence and accuracy in the low-center-frequency situation. In contrast, utilizing the high-center-frequency data leads to a high-resolution reconstruction but slow convergence on large numbers of grid. Our objective is to take full advantage of all low frequency components from pulse data with the high-center-frequency data measured by the diagnostic device. First we process the raw data in the frequency domain. Then multi-frequency accelerating strategy helps restart CSI in the current frequency using the last iteration result obtained from the lower frequency component. The merit of multi- frequency accelerating strategy is that computational burden decreases at the first few iterations. Because the low frequency component of dataset computes on the coarse grid with assuming a fixed number of points per wavelength. In the numerical test, the pulse data were generated by the K-wave simulator and have been processed to meet the computation of the CSI method. We investigate the performance of the multi-frequency and single-frequency reconstructions and conclude that the multi-frequency accelerating strategy significantly enhances the quality of the reconstructed image and simultaneously reduces the average computational time for any iteration step.

Frequency-domain ultrasound waveform tomography is a promising method for the visualization and characterization of breast disease. It has previously been shown to accurately reconstruct the sound speed distributions of breasts of varying densities. The reconstructed images show detailed morphological and quantitative information that can help differentiate different types of breast disease including benign and malignant lesions. The attenuation properties of an ex vivo phantom have also been assessed. However, the reconstruction algorithms assumed a 2D geometry while the actual data acquisition process was not. Although clinically useful sound speed images can be reconstructed assuming this mismatched geometry, artifacts from the reconstruction process exist within the reconstructed images. This is especially true for registration across different modalities and when the 2D assumption is violated. For example, this happens when a patient’s breast is rapidly sloping. It is also true for attenuation imaging where energy lost or gained out of the plane gets transformed into artifacts within the image space. In this paper, we will briefly review ultrasound waveform tomography techniques, give motivation for pursuing the 3D method, discuss the 3D reconstruction algorithm, present the results of 3D forward modeling, show the mismatch that is induced by the violation of 3D modeling via numerical simulations, and present a 3D inversion of a numerical phantom.

Ex vivo studies using our ultrasound waveform attenuation algorithm have shown promising results for detection and characterization of lesions of different types. Our preliminary in vivo study shows that the waveform attenuation image has much higher resolution and can better delineate breast lesions boundaries than the corresponding ray-based attenuation image. In this study, we preprocessed our time domain waveforms acquired with a ring array and explored the directional transducer beam pattern to better match calculated wave fields with respect to the acquired wave fields. We have applied waveform attenuation to in vivo data and compared the resulting waveform attenuation images with the ray-based counterparts to assess the resolution and accuracy of the waveform attenuation reconstruction.

Accurate reconstruction of the initial pressure distribution in photoacoustic computed tomography (PACT), in general, requires knowledge of the sound speed distribution of the medium. However, the sound speed distri- bution is often unknown, and estimating both the sound speed and initial pressure from PACT measurements alone is unstable. An alternative is to estimate the sound speed from ultrasound computed tomography (USCT) measurements. This approach fails to exploit the acoustic information in the PACT measurements and may require many USCT measurements to accurately reconstruct the sound speed. Here, we propose a joint recon- struction method where the sound speed and initial pressure distributions are simultaneously estimated from combined PACT/USCT measurements. This approach effectively overcomes the instability of the PACT joint reconstruction problem while requiring few USCT measurements.

Photoacoustic computed tomography (PACT) is an emerging computed imaging modality that exploits optical contrast and ultrasonic detection principles to form images of the photoacoustically induced initial pressure distribution within tissue. The PACT reconstruction problem corresponds to a time-domain inverse source problem, where the initial pressure distribution is recovered from the measurements recorded on an aperture outside the support of the source. A major challenge in transcranial PACT brain imaging is to compensate for aberrations in the measured data due to the propagation of the photoacoustic wavefields through the skull. To properly account for these effects, a wave equation-based inversion method should be employed that can model the heterogeneous elastic properties of the medium. In this study, an iterative image reconstruction method for 3D transcranial PACT is developed based on the elastic wave equation. To accomplish this, a forward model based on a finite-difference time-domain discretization of the elastic wave equation is established. Subsequently, gradient-based methods are employed for computing penalized least squares estimates of the initial source distribution that produced the measured photoacoustic data. The developed reconstruction algorithm is validated and investigated through computer-simulation studies.

Temporal enhanced ultrasound (TeUS) is an imaging approach where a sequence of temporal ultrasound data is acquired and analyzed for tissue typing. Previously, in a series of in vivo and ex vivo studies we have demonstrated that, this approach is effective for detecting prostate and breast cancers. Evidences derived from our experiments suggest that both ultrasound-signal related factors such as induced heat and tissue-related factors such as the distribution and micro-vibration of scatterers lead to tissue typing information in TeUS. In this work, we simulate mechanical micro-vibrations of scatterers in tissue-mimicking phantoms that have various scatterer densities reflecting benign and cancerous tissue structures. Finite element modeling (FEM) is used for this purpose where the vertexes are scatterers representing cell nuclei. The initial positions of scatterers are determined by the distribution of nuclei segmented from actual digital histology scans of prostate cancer patients. Subsequently, we generate ultrasound images of the simulated tissue structure using the Field II package resulting in a temporal enhanced ultrasound. We demonstrate that the micro-vibrations of scatterers are captured by temporal ultrasound data and this information can be exploited for tissue typing.

Contrast Enhanced Ultrasound (CEUS) is a sensitive imaging technique to assess tissue vascularity, that can be useful in the quantification of different perfusion patterns. This can be particularly important in the early detection and staging of arthritis. In a recent study we have shown that a Gamma-variate can accurately quantify synovial perfusion and it is flexible enough to describe many heterogeneous patterns. Moreover, we have shown that through a pixel-by-pixel analysis the quantitative information gathered characterizes more effectively the perfusion. However, the SNR ratio of the data and the nonlinearity of the model makes the parameter estimation difficult. Using classical non-linear-leastsquares (NLLS) approach the number of unreliable estimates (those with an asymptotic coefficient of variation greater than a user-defined threshold) is significant, thus affecting the overall description of the perfusion kinetics and of its heterogeneity. In this work we propose to solve the parameter estimation at the pixel level within a Bayesian framework using Variational Bayes (VB), and an automatic and data-driven prior initialization. When evaluating the pixels for which both VB and NLLS provided reliable estimates, we demonstrated that the parameter values provided by the two methods are well correlated (Pearson’s correlation between 0.85 and 0.99). Moreover, the mean number of unreliable pixels drastically reduces from 54% (NLLS) to 26% (VB), without increasing the computational time (0.05 s/pixel for NLLS and 0.07 s/pixel for VB). When considering the efficiency of the algorithms as computational time per reliable estimate, VB outperforms NLLS (0.11 versus 0.25 seconds per reliable estimate respectively).

Thermal monitoring for ablation therapy has high demands for preserving healthy tissues while removing malignant ones completely. Various methods have been investigated. However, exposure to radiation, cost-effectiveness, and inconvenience hinder the use of X-ray or MRI methods. Due to the non-invasiveness and real-time capabilities of ultrasound, it is widely used in intraoperative procedures. Ultrasound thermal monitoring methods have been developed for affordable monitoring in real-time. We propose a new method for thermal monitoring using an ultrasound element. By inserting a Lead-zirconate-titanate (PZT) element to generate the ultrasound signal in the liver tissues, the single travel time of flight is recorded from the PZT element to the ultrasound transducer. We detect the speed of sound change caused by the increase in temperature during ablation therapy. We performed an ex vivo experiment with liver tissues to verify the feasibility of our speed of sound estimation technique. The time of flight information is used in an optimization method to recover the speed of sound maps during the ablation, which are then converted into temperature maps. The result shows that the trend of temperature changes matches with the temperature measured at a single point. The estimation error can be decreased by using a proper curve linking the speed of sound to the temperature. The average error over time was less than 3 degrees Celsius for a bovine liver. The speed of sound estimation using a single PZT element can be used for thermal monitoring.

Ultrasound Computer Tomography (USCT) is a promising new imaging system for breast cancer diagnosis. An essential step before further processing is to remove the water background from the reconstructed images. In this paper we present a fully-automated image segmentation method based on three-dimensional active contours. The active contour method is extended by applying gradient vector flow and encoding the USCT aperture characteristics as additional weighting terms. A surface detection algorithm based on a ray model is developed to initialize the active contour, which is iteratively deformed to capture the breast outline in USCT reflection images. The evaluation with synthetic data showed that the method is able to cope with noisy images, and is not influenced by the position of the breast and the presence of scattering objects within the breast. The proposed method was applied to 14 in-vivo images resulting in an average surface deviation from a manual segmentation of 2.7 mm. We conclude that automated segmentation of USCT reflection images is feasible and produces results comparable to a manual segmentation. By applying the proposed method, reproducible segmentation results can be obtained without manual interaction by an expert.

An artefact has recently been reported [1,2] in the estimation of the lateral blood velocity using speckle tracking. This artefact shows as a net velocity bias in presence of strong spatial velocity gradients such as those that occur at the edges of the filling jets in the heart. Even though this artifact has been found both in vitro and in simulated data, its causes are still undescribed. Here we demonstrate that a potential source of this artefact can be traced to smaller errors in the beamforming setup. By inserting a small offset in the beamforming delay, one can artificially create a net lateral movement in the speckle in areas of high velocity gradient. That offset does not have a strong impact in the image quality and can easily go undetected.

Ultrafast 3D transesophageal echocardiographic (TEE) imaging, combined with 3D echo particle image velocimetry (ePIV), would be ideal for tracking the complex blood flow patterns in the heart. We are developing a miniature pediatric matrix TEE transducer that employs micro-beamforming (μBF) and allows high framerate in 3D. In this paper, we assess the feasibility of 3D ePIV with a high frame rate, small aperture transducer and the influence of the micro-beamforming technique. We compare the results of 3D ePIV on simulated images using the μBF transducer and an idealized, fully sampled (FS) matrix transducer. For the two transducers, we have simulated high-framerate imaging of an 8.4mm diameter artery having a known 4D velocity field. The simulations were performed in FieldII. 1000 3D volumes, at a rate of 1000 volumes/sec, were created using a single diverging transmission per volume. The error in the 3D velocity estimation was measured by comparing the ePIV results of both transducers to the ground truth. The results on the simulated volumes show that ePIV can estimate the 4D velocity field of the arterial phantom using these small-aperture transducers suitable for pediatric 3D TEE. The μBF transducer (RMSE 44.0%) achieved comparable ePIV accuracy to that of the FS transducer (RMSE 42.6%).

Echocardiography is the most commonly used image modality in cardiology, assessing several aspects of cardiac viability. The importance of cardiac hemodynamics and 4D blood flow motion has recently been highlighted, however such assessment is still difficult using routine echo-imaging. Instead, combining imaging with computational fluid dynamics (CFD)-simulations has proven valuable, but only a few models have been applied clinically. In the following, patient-specific CFD-simulations from transthoracic dobutamin stress echocardiography have been used to analyze the left ventricular 4D blood flow in three subjects: two with normal and one with reduced left ventricular function. At each stress level, 4D-images were acquired using a GE Vivid E9 (4VD, 1.7MHz/3.3MHz) and velocity fields simulated using a presented pathway involving endocardial segmentation, valve position identification, and solution of the incompressible Navier-Stokes equation. Flow components defined as direct flow, delayed ejection flow, retained inflow, and residual volume were calculated by particle tracing using 4th-order Runge-Kutta integration. Additionally, systolic and diastolic average velocity fields were generated. Results indicated no major changes in average velocity fields for any of the subjects. For the two subjects with normal left ventricular function, increased direct flow, decreased delayed ejection flow, constant retained inflow, and a considerable drop in residual volume was seen at increasing stress. Contrary, for the subject with reduced left ventricular function, the delayed ejection flow increased whilst the retained inflow decreased at increasing stress levels. This feasibility study represents one of the first clinical applications of an echo-based patient-specific CFD-model at elevated stress levels, and highlights the potential of using echo-based models to capture highly transient flow events, as well as the ability of using simulation tools to study clinically complex phenomena. With larger patient studies planned for the future, and with the possibility of adding more anatomical features into the model framework, the current work demonstrates the potential of patient-specific CFD-models as a tool for quantifying 4D blood flow in the heart.

We present methods to optimize the setup of a 3D ultrasound tomography scanner for breast cancer detection. This approach provides a systematic and quantitative tool to evaluate different designs and to optimize the con- figuration with respect to predefined design parameters. We consider both, time-of-flight inversion using straight rays and time-domain waveform inversion governed by the acoustic wave equation for imaging the sound speed. In order to compare different designs, we measure their quality by extracting properties from the Hessian operator of the time-of-flight or waveform differences defined in the inverse problem, i.e., the second derivatives with respect to the sound speed. Spatial uncertainties and resolution can be related to the eigenvalues of the Hessian, which provide a good indication of the information contained in the data that is acquired with a given design. However, the complete spectrum is often prohibitively expensive to compute, thus suitable approximations have to be developed and analyzed. We use the trace of the Hessian operator as design criterion, which is equivalent to the sum of all eigenvalues and requires less computational effort. In addition, we suggest to take advantage of the spatial symmetry to extrapolate the 3D experimental design from a set of 2D configurations. In order to maximize the quality criterion, we use a genetic algorithm to explore the space of possible design configurations. Numerical results show that the proposed strategies are capable of improving an initial configuration with uniformly distributed transducers, clustering them around regions with poor illumination and improving the ray coverage of the domain of interest.

In B-mode images from dual-sided ultrasound, it has been shown that by delineating structures suspected of being relatively homogeneous, one can enhance limited angle tomography to produce speed of sound images in the same view as X-ray Digital Breast Tomography (DBT). This could allow better breast cancer detection and discrimination, as well as improved registration of the ultrasound and X-ray images, because of the similarity of SOS and X-ray contrast in the breast. However, this speed of sound reconstruction method relies strongly on B-mode or other reflection mode segmentation. If that information is limited or incorrect, artifacts will appear in the reconstructed images. Therefore, the iterative speed of sound reconstruction algorithm has been modified in a manner of simultaneously utilizing the image segmentations and removing most artifacts. The first step of incorporating a priori information is solved by any nonlinearnonconvex optimization method while artifact removal is accomplished by employing the fast split Bregman method to perform total-variation (TV) regularization for image denoising. The proposed method was demonstrated in simplified simulations of our dual-sided ultrasound scanner. To speed these computations two opposed 40-element ultrasound linear arrays with 0.5 MHz center frequency were simulated for imaging objects in a uniform background. The proposed speed of sound reconstruction method worked well with both bent-ray and full-wave inversion methods. This is also the first demonstration of successful full-wave medical ultrasound tomography in the limited angle geometry. Presented results lend credibility to a possible translation of this method to clinical breast imaging.

The reconstruction of acoustic attenuation maps for transmission Ultrasound Computed Tomography (USCT) based on the standard least-squares full wave inversion method requires the accurate knowledge of the sound speed map in the region under study. Any deviation in the reconstructed speed maps creates a very significant bias in the attenuation map, as the standard least-squares misfit function is more sensitive to time misalignments than to amplitude differences of the signals. In this work, we propose a generalized misfit function which includes an additional term that accounts for the amplitude differences between the measured and the estimated signals. The functional gradients used to minimize the proposed misfit function were obtained using an adjoint field formulation and the fractional Laplacian wave equation. The forward and backward wave propagation was obtained with the parallelized GPU version of the software k-Wave and the optimization was performed with a line search method. A numerical phantom simulating breast tissue and synthetic noisy data were used to test the performance of the proposed misfit function. The attenuation was reconstructed based on a converged speed map. An edge-preserving regularization method based on total variation was also implemented. To quantify the quality of the results, the mean values and their standard deviations in several regions of interest were analyzed and compared to the reference values. The proposed generalized misfit function decreases considerably the bias in the attenuation map caused by the deviations in the speed map in all the regions of interest analyzed.

Waveform inversion methods can produce high-resolution reconstructed sound speed images for ultrasound computed tomography; however, they are very computational expensive. Source encoding methods can reduce this computational cost by formulating the image reconstruction problem as a stochastic optimization problem. Here, we solve this optimization problem by the regularized dual averaging method instead of the more commonly used stochastic gradient descent. This new optimization method allows use of non-smooth regularization functions and treats the stochastic data fidelity term in the objective function separately from the deterministic regularization function. This allows noise to be mitigated more effectively. The method further exhibits lower variance in the estimated sound speed distributions across iterations when line search methods are employed.

Synthetic Aperture Focusing Technique (SAFT) allows fast data acquisition and optimally focused images. The computational burden for 3D imaging is large as for each voxel the delay for each acquired A-scan has to be calculated, e.g. O(N⁵) for N³ voxels and N² A-scans. For 3D reconstruction of objects which are large in terms of the wavelength, e.g. ≥ (100 λ)³, the computation of one volume takes several days on a current multicore PC. If the 3D distribution of the speed of sound is applied to correct the delays, the computation time increases further. In this work a time of flight interpolation based GPU implementation (TOFI-SAFT) is presented which accelerates our previous GPU implementation of speed of sound corrected SAFT by a factor of 7 to 16 min. with only minor reduction of image quality.

A new ultrasound-guided breast biopsy technique is proposed. The technique utilizes conventional ultrasound guidance coupled with a high frequency embedded ultrasound array located within the biopsy needle to improve the accuracy in breast cancer diagnosis.1 The array within the needle is intended to be used to detect micro- calcifications indicative of early breast cancers such as ductal carcinoma in situ (DCIS). Backscattering analysis has the potential to characterize tissues to improve localization of lesions. This paper describes initial results of the application of backscattering analysis of breast biopsy tissue specimens and shows the usefulness of high frequency ultrasound for the new biopsy related technique. Ultrasound echoes of ex-vivo breast biopsy tissue specimens were acquired by using a single-element transducer with a bandwidth from 41 MHz to 88 MHz utilizing a UBM methodology, and the backscattering coefficients were calculated. These values as well as B-mode image data were mapped in 2D and matched with each pathology image for the identification of tissue type for the comparison to the pathology images corresponding to each plane. Microcalcifications were significantly distinguished from normal tissue. Adenocarcinoma was also successfully differentiated from adipose tissue. These results indicate that backscattering analysis is able to quantitatively distinguish tissues into normal and abnormal, which should help radiologists locate abnormal areas during the proposed ultrasound-guided breast biopsy with high frequency ultrasound.

Epidural anesthesia is one of the most commonly used and yet challenging techniques employed for pain management and anesthesia delivery. The major complications of this procedure are due to accidental dural puncture, with an incidence of 1-3%, which could lead to both temporary and irreversible permanent neurological complications. Needle placement under ultrasound (US) guidance has received increasing interest for improving needle placement accuracy. However, poor needle visibility in US, difficulties in displaying relevant anatomical structure such as dura mater due to attenuation and bone shadowing, and image interpretation variability among users pose significant hurdles for any US guidance system. As a result, US guidance for epidural injections has not been widely adopted for everyday use for the performance of neuraxial blocks. The difficulties in localizing the ligamentum flavum and dura with respect to the needle tip can be addressed by integrating A-mode US, provided by a single-element transducer at the needle tip, into the B-mode US guidance system. We have taken the first steps towards providing such a guidance system. Our goal is to improve the safety of this procedure with minimal changes to the clinical workflow. This work presents the design and development of a 20 MHz single-element US transducer housed at the tip of a 19 G needle hypodermic tube, which can fit inside an epidural introducer needle. In addition, the results from initial transducer characterization tests and performance evaluation of the transducer in a euthanized porcine model are provided.

Treatment for gynaecological cancers often includes brachytherapy; in particular, in high-dose-rate (HDR) interstitial brachytherapy, hollow needles are inserted into the tumour and surrounding area through a template in order to deliver the radiation dose. Currently, there is no standard modality for visualizing needles intra-operatively, despite the need for precise needle placement in order to deliver the optimal dose and avoid nearby organs, including the bladder and rectum. While three-dimensional (3D) transrectal ultrasound (TRUS) imaging has been proposed for 3D intra-operative needle guidance, anterior needles tend to be obscured by shadowing created by the template’s vaginal cylinder. We have developed a 360-degree 3D transvaginal ultrasound (TVUS) system that uses a conventional two-dimensional side-fire TRUS probe rotated inside a hollow vaginal cylinder made from a sonolucent plastic (TPX). The system was validated using grid and sphere phantoms in order to test the geometric accuracy of the distance and volumetric measurements in the reconstructed image. To test the potential for visualizing needles, an agar phantom mimicking the geometry of the female pelvis was used. Needles were inserted into the phantom and then imaged using the 3D TVUS system. The needle trajectories and tip positions in the 3D TVUS scan were compared to their expected values and the needle tracks visualized in magnetic resonance images. Based on this initial study, 360-degree 3D TVUS imaging through a sonolucent vaginal cylinder is a feasible technique for intra-operatively visualizing needles during HDR interstitial gynaecological brachytherapy.

Wave Intensity Analysis (WIA) can provide parameters representative of the interaction between the vascular network and the heart. It has been already demonstrated that WIA-derived biomarkes have a quantitative physiological meaning. Aim of this study was to develop an image process algorithm for performing non-invasive WIA in mice and correlate commonly used cardiac function parameters with WIA-derived indexes. Sixteen wild-type male mice (8 weeks-old) were imaged with high-resolution ultrasound (Vevo 2100). Abdominal aorta and common carotid pulse wave velocities (PWVabd, PWVcar) were obtained processing B-Mode and PW-Doppler images and employed to assess WIA. Amplitudes of the first (W1abd, W1car) and the second (W2abd, W2car) local maxima and minimum (Wb_abd,Wb_car) were evaluated; areas under the negative part of the curve were also calculated (NA_abd, NA_car). Cardiac output (CO), ejection fraction (EF) fractional shortening (FS) and stroke volume (SV) were estimated; strain analysis provided strain and strain rate values for longitudinal, radial and circumferential directions (LS, LSR, RS, RSR, CS, CSR). Isovolumetric relaxation time (IVRT) was calculated from mitral inflow PW-Doppler images; IVRT values were normalized for cardiac cycle length. W1abd was correlated with LS (R=0.65) and LSR (R=0.59), while W1_car was correlated with CO (R=0.58), EF (R=0.72), LS (R=0.65), LSR (R=0.89), CS (R=0.71), CSR (R=0.70). Both W2_abd and W2_car were not correlated with IVRT. Carotid artery WIA-derived parameters are more representative of cardiac function than those obtained from the abdominal aorta. The described US-based method can provide information about cardiac function and cardio-vascular interaction simply studying a single vascular site.

Worldwide, 99% of all maternal deaths occur in low-resource countries. Ultrasound imaging can be used to detect maternal risk factors, but requires a well-trained sonographer to obtain the biometric parameters of the fetus. One of the most important biometric parameters is the fetal Head Circumference (HC). The HC can be used to estimate the Gestational Age (GA) and assess the growth of the fetus. In this paper we propose a method to estimate the fetal HC with the use of the Obstetric Sweep Protocol (OSP). With the OSP the abdomen of pregnant women is imaged with the use of sweeps. These sweeps can be taught to somebody without any prior knowledge of ultrasound within a day. Both the OSP and the standard two-dimensional ultrasound image for HC assessment were acquired by an experienced gynecologist from fifty pregnant women in St. Luke’s Hospital in Wolisso, Ethiopia. The reference HC from the standard two-dimensional ultrasound image was compared to both the manually measured HC and the automatically measured HC from the OSP data. The median difference between the estimated GA from the manual measured HC using the OSP and the reference standard was -1.1 days (Median Absolute Deviation (MAD) 7.7 days). The median difference between the estimated GA from the automatically measured HC using the OSP and the reference standard was -6.2 days (MAD 8.6 days). Therefore, it can be concluded that it is possible to estimate the fetal GA with simple obstetric sweeps with a deviation of only one week.

Ultrasound tomography (UST) is an emerging modality that can offer quantitative measurements of breast density. Recent breakthroughs in UST image reconstruction involve the use of a waveform reconstruction as opposed to a raybased reconstruction. The sound speed (SS) images that are created using the waveform reconstruction have a much higher image quality. These waveform images offer improved resolution and contrasts between regions of dense and fatty tissues. As part of a study that was designed to assess breast density changes using UST sound speed imaging among women undergoing tamoxifen therapy, UST waveform sound speed images were then reconstructed for a subset of participants. These initial results show that changes to the parenchymal tissue can more clearly be visualized when using the waveform sound speed images. Additional quantitative testing of the waveform images was also started to test the hypothesis that waveform sound speed images are a more robust measure of breast density than ray-based reconstructions. Further analysis is still needed to better understand how tamoxifen affects breast tissue.

Freehand three-dimensional (3D) ultrasound imaging enables low-cost and flexible 3D scanning of arbitrary-shaped organs, where the operator can freely move a two-dimensional (2D) ultrasound probe to acquire a sequence of tracked cross-sectional images of the anatomy. Often, the acquired 2D ultrasound images are irregularly and sparsely distributed in the 3D space. Several 3D reconstruction algorithms have been proposed to synthesize 3D ultrasound volumes based on the acquired 2D images. A challenging task during the reconstruction process is to preserve the texture patterns in the synthesized volume and ensure that all gaps in the volume are correctly filled. This paper presents an adaptive kernel regression algorithm that can effectively reconstruct high-quality freehand 3D ultrasound volumes. The algorithm employs a kernel regression model that enables nonparametric interpolation of the voxel gray-level values. The kernel size of the regression model is adaptively adjusted based on the characteristics of the voxel that is being interpolated. In particular, when the algorithm is employed to interpolate a voxel located in a region with dense ultrasound data samples, the size of the kernel is reduced to preserve the texture patterns. On the other hand, the size of the kernel is increased in areas that include large gaps to enable effective gap filling. The performance of the proposed algorithm was compared with seven previous interpolation approaches by synthesizing freehand 3D ultrasound volumes of a benign breast tumor. The experimental results show that the proposed algorithm outperforms the other interpolation approaches.

There has been a great deal of research into ultrasound tomography for breast imaging over the past 35 years. Few successful attempts have been made to reconstruct high-resolution images using transmission ultrasound. To this end, advances have been made in 2D and 3D algorithms that utilize either time of arrival or full wave data to reconstruct images with high spatial and contrast resolution suitable for clinical interpretation. The highest resolution and quantitative accuracy result from inverse scattering applied to full wave data in 3D. However, this has been prohibitively computationally expensive, meaning that full inverse scattering ultrasound tomography has not been considered clinically viable. Here we show the results of applying a nonlinear inverse scattering algorithm to 3D data in a clinically useful time frame. This method yields Quantitative Transmission (QT) ultrasound images with high spatial and contrast resolution. We reconstruct sound speeds for various 2D and 3D phantoms and verify these values with independent measurements. The data are fully 3D as is the reconstruction algorithm, with no 2D approximations. We show that 2D reconstruction algorithms can introduce artifacts into the QT breast image which are avoided by using a full 3D algorithm and data. We show high resolution gross and microscopic anatomic correlations comparing cadaveric breast QT images with MRI to establish imaging capability and accuracy. Finally, we show reconstructions of data from volunteers, as well as an objective visual grading analysis to confirm clinical imaging capability and accuracy.

Synthetic transmit aperture beamforming is an increasingly used method to improve resolution in biomedical ultrasound imaging. Synthetic aperture sequential beamforming (SASB) is an implementation of this concept which features a relatively low computation complexity. Moreover, it can be implemented in a dual-stage architecture, where the first stage only applies simple single receive-focused delay-and-sum (srDAS) operations, while the second, more complex stage is performed either locally or remotely using more powerful processing. However, like traditional DAS-based beamforming methods, SASB is susceptible to inaccurate speed-of-sound (SOS) information. In this paper, we show how SOS estimation can be implemented using the srDAS beamformed image, and integrated into the dual-stage implementation of SASB, in an effort to obtain high resolution images with relatively low-cost hardware. Our approach builds on an existing per-channel radio frequency data-based direct estimation method, and applies an iterative refinement of the estimate. We use this estimate for SOS compensation, without the need to repeat the first stage beamforming. The proposed and previous methods are tested on both simulation and experimental studies. The accuracy of our SOS estimation method is on average 0.38% in simulation studies and 0.55% in phantom experiments, when the underlying SOS in the media is within the range 1450-1620 m/s. Using the estimated SOS, the beamforming lateral resolution of SASB is improved on average 52.6% in simulation studies and 50.0% in phantom experiments.

Clear visualization of hyper-echoic targets such as bone surfaces in ultrasound imaging is desirable for accurate registration of ultrasound to other imaging modalities or statistical models. Challenges such as strong reverberation, off-axis reflections and speed-of-sound variation reduce the contrast and resolution of the targets. In this paper, we propose a phase-factor based beamforming method which applies the Hilbert transform on delay compensated channel data across the receive aperture. The accumulated phase is then calculated and utilized as the weight in the beamforming output. Using this method, the reverberation artifact of a point target is reduced about 9 dB compared with other beamforming methods such as delay and sum, Wiener, minimum variance and coherent-factor based beamforming in a point target phantom study. The resolution and contrast of hyper-echoic targets are also improved in a vertebrae phantom study.

In our previous studies, we demonstrated that our aperture domain model-based clutter suppression algorithm improved image quality of in vivo B-mode data obtained from focused transmit beam sequences. Our approach suppresses off-axis clutter and reverberation and tackles limitations of related algorithms because it preserves RF channel signals and speckle statistics. We call the algorithm aperture domain model image reconstruction (ADMIRE). We previously focused on reverberation suppression, but ADMIRE is also effective at suppressing off-axis clutter. We are interested in how ADMIRE performs on plane wave sequences and the impact of AD- MIRE applied before and after synthetic beamforming of steered plane wave sequences. We employed simulated phantoms using Field II and tissue-mimicking phantoms to evaluate ADMIRE applied to plane wave sequencing. We generated images acquired from plane waves with and without synthetic aperture synthesis and measured contrast and contrast-to-noise ratio (CNR). For simulated cyst images formed from single plane waves, the contrast for delay-and-sum (DAS) and ADMIRE are 15.64 dB and 28.34 dB, respectively, while the CNR are 1.76 dB and 3.90 dB, respectively. Based on these findings, ADMIRE improves plane wave image quality. We also applied ADMIRE to resolution phantoms having a point target at 3 cm depth on-axis, simulating the point spread functions from data obtained from 1 and 75 steered plane waves, along with linear scan at focus of 3 and 4 cm depth. We then examined the outcome of applying ADMIRE before and after synthetic aperture processing. Finally, we applied this to an in vivo carotid artery.

In this paper we present and analyse a technique for applying minimum variance distortionless response (MVDR) beamforming to a coherent plane-wave compounding (CPWC) acquisition system. In the past, this has been done using a spatial smoothing approach that reduces the effective size of the receive aperture and degrades the image resolution. In this paper, we apply the MVDR algorithms in a novel way to the acquired data from the individual transducer elements, before any summation or other compounding. This enables us to propose a new approach for estimation of the covariance matrix that decorrelates the coherence among the components at all the different acquisition angles. This results in a new approach to receive beamforming for CPWC acquisition. The new beamformer is demonstrated on imaging data acquired with a research scanner. We find the new beamformer offers substantial improvements over the DAS method. It also significantly outperforms the previously published MVDR/CPWC beamformer on phantom studies where the signal from the main target is dominated by noise and interference. These improvements motivate further study in this new approach for enhancing image quality.

Based on coherent plane-wave compounding (CPWC) and synthetic aperture sequential beamforming (SASB) we propose a two stage CPWC beamformer able to reduce the number of channels between probe and the scanner by a factor of 4 without a significant loss in image quality. Further reduction, of up to a factor of 18, is possible at the cost of a drop in contrast ratio of 2.3 dB, but lateral resolution remains almost unchanged. The propose method may be significantly simpler to implement in the probe than SASB, but it may also increase thermal noise significantly.

Three dimensional (3D) ultrasound imaging is becoming a standard mode for medical ultrasound diagnoses. Conventional 3D ultrasound imaging is mostly scanned either by using a two dimensional matrix array or by motorizing a one dimensional array in the elevation direction. However, the former system is not widely assessable due to its cost, and the latter one has limited resolution and field-of-view in the elevation axis. Here, we propose a 3D ultrasound imaging system based on the synthetic tracked aperture approach, in which a robotic arm is used to provide accurate tracking and motion. While the ultrasound probe is moved by a robotic arm, each probe position is tracked and can be used to reconstruct a wider field-of-view as there are no physical barriers that restrict the elevational scanning. At the same time, synthetic aperture beamforming provides a better resolution in the elevation axis. To synthesize the elevational information, the single focal point is regarded as the virtual element, and forward and backward delay-andsum are applied to the radio-frequency (RF) data collected through the volume. The concept is experimentally validated using a general ultrasound phantom, and the elevational resolution improvement of 2.54 and 2.13 times was measured at the target depths of 20 mm and 110 mm, respectively.

Ultrasound elastography has become an important procedure that provides information about the tissue dynamics and may help on the detection of tissue abnormalities. Therefore, motion estimation in a sequence of ultrasound acquisition is crucial to the quality of this information. We propose a novel algorithm to perform speckle tracking, which consists in an implementation of 2D Block Matching with two enhancements: sub-pixel linear interpolation and displacement propagation, which are able to increase resolution, reduce computation time and prevent kernel mismatching errors. This method does not require any additional hardware and provide real-time information. The proposed technique was evaluated using four different numerical phantoms and its results were compared with the results from standard 2D block matching and optical flow. The proposed method outperformed the other two methods, providing an average error of 0.98 pixels, while standard 2D block matching and optical flow presented an average error of 2.50 and 10.03 pixels, respectively. The proposed algorithm was also assessed with four different physical phantoms and a qualitative comparison showed that the proposed technique presented results that were compatible to the results from the built-in elastography mode of the ultrasound equipment (Ultrasonix Touch).

Accurate estimation of myocardial motion based on ultrasound imaging is of great value for evaluation of cardiac function. Typically, myocardium undergoes complex motion and deformation including shear deformation and rotation. Thus a compression model is insufficient for investigating the performance of different algorithms. In this study, simulated shearing and rotating models are used to study the performance of optical flow (OF) and block matching (BM) methods based on ultrasound radio-frequency (RF) data. A deforming model was simulated with applied axial shear strains of 2- 6%, respectively. In addition, a rotating model was simulated with rotation angles of 0.5°-4°, respectively. Axial strains of 0%, 1% and 2% were also applied to these two models to study the influence of applied strain on the estimation of axial shear strain and rotation. To quantify the estimation performance, the root mean square error (RMSE) was used as the evaluation criterion. The results show that OF has lower RMSEs of the estimated displacement, strain and rotation angle than BM, especially at large axial shear strains and rotation angles. For the shearing model, the RMSEs of axial strains, lateral strains, and axial shear strains are reduced by up to 95.5%, 70.3% and 90.0%, respectively. For the rotating model, the RMSEs of axial strains, lateral strains, and rotation angles are reduced by up to 96.9%, 93.4% and 89.7%, respectively. OF is proved to outperform BM and thus is recommended to be used for shear strain and rotation estimation. The validations of phantom and in-vivo experiments are still required.

The cardiac elastography evolves to enable local strain estimation and identification of non-transmural infarctions. Below we compare the strain values obtained using EchoPAC in physical left ventricular phantoms made of PVA with results of the Finite Element Modelling (FEM) studies on their counterparts. Models had the form of half of an ellipsoid with 15 mm wall thickness. The homogenous model, transmural inclusion model and nontransmural inclusion (5mm thickness) model were designed. The inclusions were located in the mid segment. The material of the ventricle in the FEM studies was modeled as a hyperelastic, isotropic one. The material parameters came from measurements of the PVA samples for the homogenous case and were extrapolated to obtain stiffer inclusions. The model was deformed by applying 36 kPa pressure load to its inner surface. Peak systolic strain values were close to those observed in healthy subjects. A dedicated setup, the Vivid 6 scanner, probe M4S-RS and EchoPAC BT13 software were used in experiments. The values of strains from FEM models were averaged over nodes corresponding to the layers used in the EchoPAC software. The circumferential strain (CS) values from the FEM simulation and the physical experiment are qualitatively very close and correlate well with the clinical data. The experimental CS results also agree with expectations in terms of slope across the wall and effect of the inclusion. Segmental radial strains obtained from EchoPAC and FEM are close. The proposed approach (phantoms, setup) may be used for development of methods for identification of nontransmural infarctions.

Cardiac strain quantification using ultrasound is an active area of research. Physical left ventricle (LV) model play a significant role in the evaluation and development of myocardium strain imaging techniques. Several LV models have been reported. In spite of increasing interest in LV twist, only one of them implements torsional deformation, but it does not allow long axis views. This work presents a novel prototype of physical LV model and dedicated measurement setup which do not have this limitation. The model was made of Poly(vinyl alcohol) cryogel (PVA-c). The solution for the chamber part was doped with scattering particles to imitate the echogenicity of myocardium and to facilitate automatic segmentation of the chamber wall. The model was mounted in a measurement setup allowing computer controlled linear motion of the basis, rotation of the apex and inflation of the ventricle and the use both ultrasound imaging planes: short (SAX) and long axis (LAX). During preliminary tests RF signals as well as B-mode and M-mode images were acquired. Experiment results confirmed the possibility of forcing controlled deformation of the presented LV model wall, including elongation/shortening in the long axis direction and twist around this axis. In consequence, the model can mimic deformations of the LV wall to a large extent. The chamber wall can be segmented in B-mode images in both projections. The model with the measurement setup may be used in development and validation of a wide range of echocardiographic diagnostic procedures including segmentation, strain estimation and 3D data processing.

Sarcopenia is the degenerative loss of skeletal muscle ability associated with aging. One reason is the increasing of adipose ratio of muscle, which can be estimated by the speed of sound (SOS), since SOSs of muscle and adipose are different (about 7%). For SOS imaging, the conventional bent-ray method iteratively finds ray paths and corrects SOS along them by travel-time. However, the iteration is difficult to converge for soft tissue with bone inside, because of large speed variation. In this study, the bent-ray method is modified to produce SOS images for limb muscle with bone inside. The modified method includes three steps. First, travel-time is picked up by a proposed Akaike Information Criterion (AIC) with energy term (AICE) method. The energy term is employed for detecting and abandoning the transmissive wave through bone (low energy wave). It results in failed reconstruction for bone, but makes iteration convergence and gives correct SOS for skeletal muscle. Second, ray paths are traced using Fermat’s principle. Finally, simultaneous algebraic reconstruction technique (SART) is employed to correct SOS along ray paths, but excluding paths with low energy wave which may pass through bone. The simulation evaluation was implemented by k-wave toolbox using a model of upper arm. As the result, SOS of muscle was 1572.0±7.3 m/s, closing to 1567.0 m/s in the model. For vivo evaluation, a ring transducer prototype was employed to scan the cross sections of lower arm and leg of a healthy volunteer. And the skeletal muscle SOSs were 1564.0±14.8 m/s and 1564.1±18.0 m/s, respectively.

In linear approximation, the formation of a radio-frequency (RF) ultrasound image can be described based on a standard convolution model in which the image is obtained as a result of convolution of the point spread function (PSF) of the ultrasound scanner in use with a tissue reflectivity function (TRF). Due to the band-limited nature of the PSF, the RF images can only be acquired at a finite spatial resolution, which is often insufficient for proper representation of the diagnostic information contained in the TRF. One particular way to alleviate this problem is by means of image deconvolution, which is usually performed in a “blind” mode, when both PSF and TRF are estimated at the same time. Despite its proven effectiveness, blind deconvolution (BD) still suffers from a number of drawbacks, chief among which stems from its dependence on a stationary convolution model, which is incapable of accounting for the spatial variability of the PSF. As a result, virtually all existing BD algorithms are applied to localized segments of RF images. In this work, we introduce a novel method for non-stationary BD, which is capable of recovering the TRF concurrently with the spatially variable PSF. Particularly, our approach is based on semigroup theory which allows one to describe the effect of such a PSF in terms of the action of a properly defined linear semigroup. The approach leads to a tractable optimization problem, which can be solved using standard numerical methods. The effectiveness of the proposed solution is supported by experiments with in vivo ultrasound data.

State-of-the-art echocardiography allows to correctly diagnose most of cardiovascular diseases. An unknown source of clutter, however, hinders the visualization of the heart in some cases. We believe this clutter is caused by the ultrasound beam being partially reflected by the ribs into the elevation direction, so that structures outside the imaging plane are displayed on top of the heart image as clutter noise. We conducted in vitro experiments in a water tank using a synthetic ventricle and pig ribs. By partially blocking the probe with the ribs in the elevation direction, objects outside the imaging plane were rendered in the B-mode image, which confirms that the ribs can behave as specular reflectors. In addition, we succeeded in reproducing clutter noise using a piece of polystyrene to simulate the reflections from the lungs. This indicates that the origin of the clutter noise in echocardiograms can be reverberation coming from the lungs via specular reflection at the ribs.

This paper describes a region growing segmentation algorithm for medical ultrasound images. The algorithm starts with anisotropic diffusion filtering to reduce speckle noise without blurring the edges. Then, region growing is performed starting from a seed point, using a merging criterion that compares intensity gradients to the noise level inside the region. Finally, the boundaries are smoothed using morphological closing. The algorithm was evaluated with two simulated images and eleven phantom images and converged in 10 of them with accurate region delimitation. Preliminary results show that the proposed method can be used for ultrasound image segmentation and does not require previous knowledge of the anatomy of the structures.

Objective: Respiratory motion correction is necessary to quantitative analysis of liver contrast-enhance ultrasound (CEUS) image sequences. However, traditionally manual selecting reference image would affect the accuracy of the respiratory motion correction. Methods First, the original high-dimensional ultrasound gray-level image data was mapped into a two-dimensional space by using Laplacian Eigenmaps (LE). Then, the cluster analysis was adopted using K-means, and the optimal ultrasound reference image could be gotten for respiratory motion correction. Finally, this proposed method was validated on 18 CEUS cases of VX2 tumor in rabbit liver, and the effectiveness of this method was demonstrated. Results After correction, the time-intensity curves extracted from the region of interest of CEUS image sequences became smoother. Before correction, the average of total mean structural similarity (TMSSIM) and the average of mean correlation coefficient (MCC) from image sequences were 0.45±0.11 and 0.67±0.16, respectively. After correction, the two parameters were increased obviously（P<0.001), and were 0.59±0.11 and 0.81±0.11, respectively. The average of deviation valve (DV) from image sequences before correction was 92.16±18.12. After correction, the average was reduced to one-third of the original value. Conclusions: The proposed respiratory motion method could improve the accuracy of the quantitative analysis of CEUS by using the reference image based on the traditionally manual selection. This method is operated simply and has a potential in clinical application.

Heart fiber mechanics can be important predictors in current and future cardiac function. Accurate knowledge of these mechanics could enable cardiologists to provide a diagnosis before conditions progress. Magnetic resonance diffusion tensor imaging (MR-DTI) has been used to determine cardiac fiber orientations. Ultrasound is capable of providing anatomical information in real time, enabling a physician to quickly adjust parameters to optimize image scans. If known fiber orientations from a template heart measured using DTI can be accurately deformed onto a cardiac ultrasound volume, fiber orientations could be estimated for the patient without the need for a costly MR scan while still providing cardiologists valuable information about the heart mechanics. In this study, we apply the method to pig hearts, which are a close representation of human heart anatomy. Experiments from pig hearts show that the registration method achieved an average Dice similarity coefficient (DSC) of 0.819 ± 0.050 between the ultrasound and deformed MR volumes and that the proposed ultrasound-based method is able to estimate the cardiac fiber orientation in pig hearts.

Ultrasound tomography (UST) is an emerging breast imaging modality that can be used to quantitatively measure breast density. However, the sound speed images that are used in this analysis must first be segmented in order to accurately parse any quantitative information. Previously, this segmentation has been done manually, but this is time consuming, especially when dealing with a large number of images that must be masked. An automated masking algorithm has been developed that applies thresholding and morphological operators to UST attenuation images to automatically create masks that separate the breast tissue from the water bath. An initial set of images was tested using this algorithm to fine tune settings and very good agreement was achieved. However, when the optimized settings were applied to a larger dataset of 286 images, the robustness of the algorithm was tested. The manual masks measured a larger volume (921 cm³) than the automated masks (713 cm³), but fortunately, the difference in mean sound speed was much smaller (1449 m/s versus 1448 m/s). A majority of the automated masks (72.7%) had a measured Dice similarity coefficient (DSC) of greater than 0.8 which indicates that there was good to great overlap in the volumes of tissue created by the automated method. This algorithm shows promise to be used as a tool to quickly and effectively measure breast density.

We derived analytical forward and inverse solution of thermoacoustic wave equation for inhomogeneous multi- layered planar and cylindrical mediums with the source distribution existing in all layers. These solutions are applicable for imaging of organs such as breast and brain, whose structures are suitable for multi-layer modelling. For qualitative testing and comparison of the point-spread-functions associated with the homogeneous and layered solutions, we performed numerical simulations. Our simulation results show that the conventional inverse solution based on homogeneous medium assumption, as expected, produces incorrect locations of point sources and significantly increased side lobes, whereas our inverse solution involving the multi-layered medium produces point sources at the correct locations with lower side lobes.

It has been shown recently, that varying the excitation sequence could deliver additional benefits for photoacoustic imaging, for instance, bringing additional information on the sample under study, or reducing the total acquisition time. However, for the typically used solid state laser systems, such modification requires significant increase of the systems’ complexity. We are taking an advantage of high pulse repetition rates that semiconductor laser diodes could offer. That allows the usage of dense pulse bursts with varied number of pulses and inter-pulse delays in the range of the transducer waveform duration to study the effects of the overlay of the single pulse photoacoustic responses. In this study, we conduct a pump-probe experiment, using multi-pulse excitation sequences with varied inter-pulse delays while registering the acoustic response. We show that pulse burst excitation can be beneficial for increasing the registered amplitude and suitable inter-pulse delay values can be obtained from the transducer transfer function, either known or measured. Additionally, we examine the frequency content of the multi-pulse photoacoustic response and show that it is dominated by the pulse repetition rate used. We focus on low central frequency transducers as being widely used for clinical applications.

Optoacoustic imaging is a rapidly developing area of biomedical imaging due its combination of rich optical contrast and ultrasound depth penetration. Just like conventional pulse-echo ultrasound imaging, optoacoustic tomography relies on the use of ultrasound detector arrays with a large number of elements. The precise knowledge of the transducer’s sensitivity is crucial for the prediction of its performance for a given imaging task. Sensitivity characteristics such as the central frequency and bandwidth are routinely characterized. However, this characterization is typically performed solely under normal incidence since the measurement of the angle and frequency depended sensitivity (directivity) is difficult and time consuming with existing ultrasound characterization methods. We present a simple and fast characterization method for broadband directivity measurements of the angular transducer sensitivity based on the optoacoustic effect. The method utilizes a thin absorbing suture in order to generate omnidirectional and broadband optoacoustic signals, which are calibrated using a needle hydrophone. We applied this method to characterize and compare the directivity of a conventional piezoelectric (PZT) transducer to the directivity of a capacitive micromachined ultrasonic (cMUT) transducer. Both technologies showed a similar broadband response at normal incidence and the PZT transducer displayed a more than two times larger signal to noise ratio at normal incidence. However, the cMUT transducer’s sensitivity was significantly less angle-depended and outperformed the PZT’s sensitivity for angles larger than 20°.

Photoacoustic (PA) imaging is a hybrid imaging modality that integrates the strength of optical and ultrasound imaging. Nanosecond (ns) pulsed lasers used in current PA imaging systems are expensive, bulky and they often waste energy. We propose and evaluate, through simulations, the use of a continuous wave (CW) laser whose amplitude is linear frequency modulated (chirp) for PA imaging. The chirp signal provides signal-to-side-lobe ratio (SSR) improvement potential and full control over PA signal frequencies excited in the sample. The PA signal spectrum is a function of absorber size and the time frequencies present in the chirp. A mismatch between the input chirp spectrum and the output PA signal spectrum can affect the compressed pulse that is recovered from cross-correlating the two. We have quantitatively characterized this effect. The k-wave Matlab tool box was used to simulate PA signals in three dimensions for absorbers ranging in size from 0.1 mm to 0.6 mm, in response to laser excitation amplitude that is linearly swept from 0.5 MHz to 4 MHz. This sweep frequency range was chosen based on the spectrum analysis of a PA signal generated from ex-vivo human prostate tissue samples. In comparison, the energy wastage by a ns laser pulse was also estimated. For the chirp methodology, the compressed pulse peak amplitude, pulse width and side lobe structure parameters were extracted for different size absorbers. While the SSR increased 6 fold with absorber size, the pulse width decreased by 25%.

Ultrasound contrast agents (UCAs) such as SonoVue or Optison have been used widely in clinic for contrast-enhanced vascular imaging. However, microbubbles UCAs display limitations in tumor-targeted imaging due to the large sizes, nanoscaled UCAs has consequently attracted increasing attentions. In this work, we synthesized nanobubbles (NBs) by ultrasonic cavitation method, then a fluorescent marker of Alexa Fluor 680 was conjugated to the shell in order to observe the localization of NBs in tumor tissue. Measurement of fundamental characteristics showed that the NBs had homogeneous distribution of mean diameter of 267.9 ± 19.2 nm and polydispersity index of 0.410 ± 0.056. To assess in vivo tumor-selectivity of NBs, we established the rabbits VX2 hepatocellular carcinoma model though surgical implantation method. After the rabbits were intravenous administered of NBs, contrast-enhanced sonograms was observed in the surrounding of VX2 tumor, which showed there are rich capillaries in the tumor periphery. We additionally investigated the toxic of the NBs by hematoxylin-eosin staining. The results indicated that the NBs is a biocompatible non-toxic lipid system. Furthermore, the VX2 tumors and major organs were analyzed using ex vivo fluorescence imaging to confirm the targeted selectivity of NBs, and the results verified that the NBs were capable of targeting VX2 tumor. Confocal laser scanning microscopy examination showed that the NBs can traverse the VX2 tumor capillaries and target to the hepatocellular carcinoma tumor cells. All these results suggested that the newly prepared NBs have a potential application in molecular imaging and tumor-targeting therapy.

In the past years we have perceived within the USCT research community a demand for freely available USCT data sets. Inspired by the idea of Open Science, this collection of data sets could stimulate the collaboration and the exchange of ideas and experiences between USCT researchers. In addition, it may lead to comprehensive comparison of different reconstruction algorithms and their results. Finally, by collecting feedback from the users about data and system architecture, valuable information is gathered for further development of measurement setups. For the above reasons, we have initiated a digital portal with several reference data sets and access scripts under free licenses. To kick off this initiative, we organized a USCT data challenge event at SPIE Medical Imaging 2017.