Microbubble detection on ultrasound imaging by utilizing phase patterned waves

Objective.Super-resolution ultrasonography offers the advantage of visualization of intricate microvasculature, which is crucial for disease diagnosis. Mapping of microvessels is possible by localizing microbubbles (MBs) that act as contrast agents and tracking their location. However, there are limitations such as the low detectability of MBs and the utilization of a diluted concentration of MBs, leading to the extension of the acquisition time. We aim to enhance the detectability of MBs to reduce the acquisition time of acoustic data necessary for mapping the microvessels.Approach.We propose utilizing phase patterned waves (PPWs) characterized by spatially patterned phase distributions in the incident beam to achieve this. In contrast to conventional ultrasound irradiation methods, this irradiation method alters bubble interactions, enhancing the oscillation response of MBs and generating more significant scattered waves from specific MBs. This enhances the detectability of MBs, thereby enabling the detection of MBs that were undetectable by the conventional method. The objective is to maximize the overall detection of bubbles by utilizing ultrasound imaging with additional PPWs, including the conventional method. In this paper, we apply PPWs to ultrasound imaging simulations considering bubble-bubble interactions to elucidate the characteristics of PPWs and demonstrate their efficacy by employing PPWs on MBs fixed in a phantom by the experiment.Main results.By utilizing two types of PPWs in addition to the conventional ultrasound irradiation method, we confirmed the detection of up to 93.3% more MBs compared to those detected using the conventional method alone.Significance.Ultrasound imaging using additional PPWs made it possible to increase the number of detected MBs, which is expected to improve the efficiency of bubble detection.

Hepatocellular Carcinoma Automatic Diagnosis within CEUS and B-Mode Ultrasound Images Using Advanced Machine Learning Methods

Hepatocellular Carcinoma (HCC) is the most common malignant liver tumor, being present in 70% of liver cancer cases. It usually evolves on the top of the cirrhotic parenchyma. The most reliable method for HCC diagnosis is the needle biopsy, which is an invasive, dangerous method. In our research, specific techniques for non-invasive, computerized HCC diagnosis are developed, by exploiting the information from ultrasound images. In this work, the possibility of performing the automatic diagnosis of HCC within B-mode ultrasound and Contrast-Enhanced Ultrasound (CEUS) images, using advanced machine learning methods based on Convolutional Neural Networks (CNN), was assessed. The recognition performance was evaluated separately on B-mode ultrasound images and on CEUS images, respectively, as well as on combined B-mode ultrasound and CEUS images. For this purpose, we considered the possibility of combining the input images directly, performing feature level fusion, then providing the resulted data at the entrances of representative CNN classifiers. In addition, several multimodal combined classifiers were experimented, resulted by the fusion, at classifier, respectively, at the decision levels of two different branches based on the same CNN architecture, as well as on different CNN architectures. Various combination methods, and also the dimensionality reduction method of Kernel Principal Component Analysis (KPCA), were involved in this process. These results were compared with those obtained on the same dataset, when employing advanced texture analysis techniques in conjunction with conventional classification methods and also with equivalent state-of-the-art approaches. An accuracy above 97% was achieved when our new methodology was applied.

Blood flow patterns estimation in the left ventricle with low-rate 2D and 3D dynamic contrast-enhanced ultrasound

BACKGROUND AND OBJECTIVE

Left ventricle (LV) dysfunction always occurs at early heart-failure stages, producing variations in the LV flow patterns. Cardiac diagnostics may therefore benefit from flow-pattern analysis. Several visualization tools have been proposed that require ultrafast ultrasound acquisitions. However, ultrafast ultrasound is not standard in clinical scanners. Meanwhile techniques that can handle low frame rates are still lacking. As a result, the clinical translation of these techniques remains limited, especially for 3D acquisitions where the volume rates are intrinsically low.

METHODS

To overcome these limitations, we propose a novel technique for the estimation of LV blood velocity and relative-pressure fields from dynamic contrast-enhanced ultrasound (DCE-US) at low frame rates. Different from other methods, our method is based on the time-delays between time-intensity curves measured at neighbor pixels in the DCE-US loops. Using Navier-Stokes equation, we regularize the obtained velocity fields and derive relative-pressure estimates. Blood flow patterns were characterized with regard to their vorticity, relative-pressure changes (dp/dt) in the LV outflow tract, and viscous energy loss, as these reflect the ejection efficiency.

RESULTS

We evaluated the proposed method on 18 patients (9 responders and 9 non-responders) who underwent cardiac resynchronization therapy (CRT). After CRT, the responder group evidenced a significant (p<0.05) increase in vorticity and peak dp/dt, and a non-significant decrease in viscous energy loss. No significant difference was found in the non-responder group. Relative feature variation before and after CRT evidenced a significant difference (p<0.05) between responders and non-responders for vorticity and peak dp/dt. Finally, the method feasibility is also shown with 3D DCE-US.

CONCLUSIONS

Using the proposed method, adequate visualization and quantification of blood flow patterns are successfully enabled based on low-rate DCE-US of the LV, facilitating the clinical adoption of the method using standard ultrasound scanners. The clinical value of the method in the context of CRT is also shown.

QR Based Despeckling Approach for Medical Ultrasound Images

BACKGROUND

An approach based on QR decomposition, to remove speckle noise from medical ultrasound images, is presented in this paper.

METHODS

The speckle noisy image is segmented into small overlapping blocks. A global covariance matrix is calculated by averaging the corresponding covariances of the blocks. QR decomposition is applied to the global covariance matrix. To filter out speckle noise, the first subset of orthogonal vectors of the Q matrix is projected onto the signal subspace. The proposed approach is compared with five benchmark techniques; Homomorphic Wavelet Despeckling (HWDS), Speckle Reducing Anisotropic Diffusion (SRAD), Frost, Kuan and Probabilistic Non-Local Mean (PNLM).

RESULTS AND CONCLUSION

When applied to different simulated and real ultrasound images, the QR based approach has secured maximum despeckling performance while maintaining optimal resolution and edge detection, and that is regardless of image size or nature of speckle; fine or rough.

Groundwater contamination sources identification based on kernel extreme learning machine and its effect due to wavelet denoising technique

Measurements of contaminant concentrations inevitably contain noise because of accidental and systematic errors. However, groundwater contamination sources identification (GCSI) is highly dependent on the data measurements, which directly affect the accuracy of the identification results. Thus, in the present study, the wavelet hierarchical threshold denoising method was employed to denoise concentration measurements and the denoised measurements were then used for GCSI. A 0-1 mixed-integer nonlinear programming optimization model (0-1 MINLP) based on a kernel extreme learning machine (KELM) was applied to identify the location and release history of a contamination source. The results showed the following. (1) The wavelet hierarchical threshold denoising method was not very effective when applied to concentration measurements observed every 2 months (the number of measurements is small and relatively discrete) compared with those obtained every 2 days (the number of measurements is large and relatively continuous). (2) When the concentration measurements containing noise were employed for GCSI, the identifications results were further from the true values when the measurements contained more noise. The approximation of the identification results to the true values improved when the denoised concentration measurements were employed for GCSI. (3) The 0-1 MINLP based on the surrogate KELM model could simultaneously identify the location and release history of contamination sources, as well reducing the computational load and decreasing the calculation time by 96.5% when solving the 0-1 MINLP.

Ultrasound Sub-pixel Motion-tracking Method with Out-of-plane Motion Detection for Precise Vascular Imaging

Ultrasound vascularity imaging provides important information for differential diagnosis of tumors. Peak-hold (PH) is a useful technique for precisely imaging small vessels by selecting a maximum brightness in each pixel through the frames obtained sequentially. To use PH successfully one needs motion compensation to reduce image blur, but out-of-plane motion cannot be avoided. To address this problem, we developed a sub-pixel motion-tracking method with out-of-plane motion detection (OPMD). It is a combination of the sum of the absolute differences (SAD) method and the Kanade-Lucas-Tomasi method and can be accurately applied to various motions. The value from OPMD (γ) is defined as a statistical value obtained from the distribution of residual values in the SAD procedure with the obtained frames. The value is ideally 0, and the frames having large γ are removed from the PH procedure. The accuracy of the proposed tracking method was found by a simulation study to be approximately 20 μm. We also found, through a phantom experiment, that the value of γ sensitively increased enough to detect out-of-plane motion. Most important, γ begins to increase before tracking errors occur. This suggests that OPMD can be used to predict tracking errors and effectively remove frames from the PH procedure. An in vivo experiment with a rabbit showed that the PH image obtained with motion tracking clearly revealed peripheral vessels that were blurred in the PH image obtained without motion tracking. We also found that the image quality becomes better when OPMD was used to remove frames including out-of-plane motion.

Contrast-Enhanced Ultrasound Quantification: From Kinetic Modeling to Machine Learning

Ultrasound contrast agents (UCAs) have opened up immense diagnostic possibilities by combined use of indicator dilution principles and dynamic contrast-enhanced ultrasound (DCE-US) imaging. UCAs are microbubbles encapsulated in a biocompatible shell. With a rheology comparable to that of red blood cells, UCAs provide an intravascular indicator for functional imaging of the (micro)vasculature by quantitative DCE-US. Several models of the UCA intravascular kinetics have been proposed to provide functional quantitative maps, aiding diagnosis of different pathological conditions. This article is a comprehensive review of the available methods for quantitative DCE-US imaging based on temporal, spatial and spatiotemporal analysis of the UCA kinetics. The recent introduction of novel UCAs that are targeted to specific vascular receptors has advanced DCE-US to a molecular imaging modality. In parallel, new kinetic models of increased complexity have been developed. The extraction of multiple quantitative maps, reflecting complementary variables of the underlying physiological processes, requires an integrative approach to their interpretation. A probabilistic framework based on emerging machine-learning methods represents nowadays the ultimate approach, improving the diagnostic accuracy of DCE-US imaging by optimal combination of the extracted complementary information. The current value and future perspective of all these advances are critically discussed.

Investigation of Microbubble Detection Methods for Super-Resolution Imaging of Microvasculature

Ultrasound super-resolution techniques use the response of microbubble (MB) contrast agents to visualize the microvasculature. Techniques that localize isolated bubble signals first require detection algorithms to separate the MB and tissue responses. This work explores the three main MB detection techniques for super-resolution of microvasculature. Pulse inversion (PI), differential imaging (DI), and singular value decomposition (SVD) filtering were compared in terms of the localization accuracy, precision, and contrast-to-tissue ratio. MB responses were simulated based on the properties of Sonovue and using the Marmottant model. Nonlinear propagation through tissue was modeled using the k-Wave software package. For the parameters studied, the results show that PI is most appropriate for low frequency applications, but also most dependent on transducer bandwidth. SVD is preferable for high frequency acquisition where localization precision on the order of a few microns is possible. PI is largely independent of flow direction and speed compared to SVD and DI, so is appropriate for visualizing the slowest flows and tortuous vasculature. SVD is unsuitable for stationary MBs and can introduce a localization error on the order of hundreds of microns over the speed range 0-2 mm/s and flow directions from lateral (parallel to probe) to axial (perpendicular to probe). DI is only suitable for flow rates >0.5 mm/s or as flow becomes more axial. Overall, this study develops an MB and tissue nonlinear simulation platform to improve understanding of how different MB detection techniques can impact the super-resolution process and explores some of the factors influencing the suitability of each.

SUSHI: Sparsity-Based Ultrasound Super-Resolution Hemodynamic Imaging

Identifying and visualizing vasculature within organs and tumors has major implications in managing cardiovascular diseases and cancer. Contrast-enhanced ultrasound scans detect slow-flowing blood, facilitating noninvasive perfusion measurements. However, their limited spatial resolution prevents the depiction of microvascular structures. Recently, super-localization ultrasonography techniques have surpassed this limit. However, they require long acquisition times of several minutes, preventing the detection of hemodynamic changes. We present a fast super-resolution method that exploits sparsity in the underlying vasculature and statistical independence within the measured signals. Similar to super-localization techniques, this approach improves the spatial resolution by up to an order of magnitude compared to standard scans. Unlike super-localization methods, it requires acquisition times of only tens of milliseconds. We demonstrate a temporal resolution of ~25 Hz, which may enable functional super-resolution imaging deep within the tissue, surpassing the temporal resolution limitations of current super-resolution methods, e.g., in neural imaging. The subsecond acquisitions make our approach robust to motion artifacts, simplifying in vivo use of super-resolution ultrasound.

A Temporal and Spatial Analysis Approach to Automated Segmentation of Microbubble Signals in Contrast-Enhanced Ultrasound Images: Application to Quantification of Active Vascular Density in Human Lower Limbs

Contrast-enhanced ultrasound (CEUS) using microbubble contrast agents has shown great promise in visualising and quantifying active vascular density. Most existing approaches for vascular density quantification using CEUS are calculated based on image intensity and are susceptible to confounding factors and imaging artefact. Poor reproducibility is a key challenge to clinical translation. In this study, a new automated temporal and spatial signal analysis approach is developed for reproducible microbubble segmentation and quantification of contrast enhancement in human lower limbs. The approach is evaluated in vitro on phantoms and in vivo in lower limbs of healthy volunteers before and after physical exercise. In this approach, vascular density is quantified based on the relative areas microbubbles occupy instead of their image intensity. Temporal features of the CEUS image sequences are used to identify pixels that contain microbubble signals. A microbubble track density (MTD) measure, the ratio of the segmented microbubble area to the whole tissue area, is calculated as a surrogate for active capillary density. In vitro results reveal a good correlation (r² = 0.89) between the calculated MTD measure and the known bubble concentration. For in vivo results, a significant increase (129% in average) in the MTD measure is found in lower limbs of healthy volunteers after exercise, with excellent repeatability over a series of days (intra-class correlation coefficient = 0.96). This compares to the existing state-of-the-art approach of destruction and replenishment analysis on the same patients (intra-class correlation coefficient ≤0.78). The proposed new approach shows great potential as an accurate and highly reproducible clinical tool for quantification of active vascular density.

Compound speckle model detects anti-angiogenic tumor response in preclinical nonlinear contrast-enhanced ultrasonography

PURPOSE

This paper proposes a method for analyzing the first-order speckle statistics of nonlinear contrast-enhanced ultrasound images from tumors.

METHODS

Contrast signal intensity is modeled as a compound distribution of exponential probability density functions with a gamma weighting function. The gamma probability weighting function serves as an approximation for log-normally distributed flow velocities in a vascular network. The model was applied to sub-harmonic bolus-injection images acquired from a mouse breast cancer xenograft model treated with murine version bevacizumab.

RESULTS

The area under curve produced using the compound statistical model could more accurately discriminate anti-VEGF-treated tumors from untreated tumors than conventional contrast-enhanced ultrasound image processing. This result was validated with gold standard histological measures of microvascular density. Fractal vessel geometry was estimated using the gamma weighting function and tested against micro-CT perfusion casting. Treated tumors had a significantly lower vascular fractal dimension than control tumors. Vascular complexity estimated using the ultrasound compound statistical model performed similarly to micro-CT fractal dimension for discriminating treated from control tumors.

CONCLUSION

The proposed technique can quantify tumor perfusion and provide an index of vascular complexity, making it a potentially useful addition for clinical detection of vascular normalization in anti-angiogenic trials.

Fractal Dimension of Tumor Microvasculature by DCE-US: Preliminary Study in Mice

Neoangiogenesis, which results in the formation of an irregular network of microvessels, plays a fundamental role in the growth of several types of cancer. Characterization of microvascular architecture has therefore gained increasing attention for cancer diagnosis, treatment monitoring and evaluation of new drugs. However, this characterization requires immunohistologic analysis of the resected tumors. Currently, dynamic contrast-enhanced ultrasound imaging (DCE-US) provides new options for minimally invasive investigation of the microvasculature by analysis of ultrasound contrast agent (UCA) transport kinetics. In this article, we propose a different method of analyzing UCA concentration that is based on the spatial distribution of blood flow. The well-known concept of Mandelbrot allows vascular networks to be interpreted as fractal objects related to the regional blood flow distribution and characterized by their fractal dimension (FD). To test this hypothesis, the fractal dimension of parametric maps reflecting blood flow, such as UCA wash-in rate and peak enhancement, was derived for areas representing different microvascular architectures. To this end, subcutaneous xenograft models of DU-145 and PC-3 prostate-cancer lines in mice, which show marked differences in microvessel density spatial distribution inside the tumor, were employed to test the ability of DCE-US FD analysis to differentiate between the two models. For validation purposes, the method was compared with immunohistologic results and UCA dispersion maps, which reflect the geometric properties of microvascular architecture. The results showed good agreement with the immunohistologic analysis, and the FD analysis of UCA wash-in rate and peak enhancement maps was able to differentiate between the two xenograft models (p < 0.05).

Visualizing the Tumor Microvasculature With a Nonlinear Plane-Wave Doppler Imaging Scheme Based on Amplitude Modulation

Imaging with ultrasonic plane waves enables the combination of Doppler and microbubble contrast-enhanced imaging without compromising the Doppler ensemble length, as is the case for conventional line-by-line imaging, thus maintaining flow sensitivity. This permits the separation of conduit flow in large vessels from the perfusion background and the presentation of velocity estimates in real-time. However, the ability to differentiate perfusion from the tissue signal is limited by the contrast-to-tissue (CTR) ratio achieved with the contrast-enhanced pulsing sequence, independently of the acquisition length. One way to improve the CTR is to use a Doppler sequence based on amplitude modulation instead of one based on pulse inversion. In this work, we discuss how amplitude modulation can be adapted to Doppler processing. We show that amplitude modulation Doppler, like pulse inversion Doppler, can separate the signal of moving tissue from that of moving microbubbles, while achieving a better contrast-to-tissue ratio than pulse inversion Doppler, both in vitro and in vivo. Both amplitude modulation Doppler and pulse inversion Doppler yield similar velocity estimates when the bandwidth of the RF echo is properly compensated. Finally, we demonstrate how amplitude modulation Doppler can be used to reveal both the conduit flow and the capillary perfusion at high frame rates in an in vivo tumor.