Search

The arrangement of cardiac fibers determines mechanical and electrical properties of the heart and can be altered due to pathology. Hence, the non-invasive assessment of fiber characteristics is of interest from both a pathophysiologic and a diagnostic point of view. Recently, we proposed and validated, by computer simulations, a theoretical framework to have more insight on 3D ultrasound Backscatter Tensor Imaging. The theory showed that spatial coherence (SC) maps of echo signals, across the probe aperture, not only carry information on fiber direction, but also on fiber size and pitch. The aim of this study was to experimentally validate these findings and conduct preliminary in vivo tests. Experiments were conducted on a purposely designed and built phantom consisting of several sets of parallel nylon wires, having different diameter (S) and pitch (P). Finally, preliminary tests were conducted on the biceps of healthy volunteers. Experiments confirmed the simulation results: averaged over all acquisitions, the main-to-secondary lobe distance linearly correlated with P (R2=94%), while the value of the SC at lag 1 (an estimate of the main lobe width) linearly correlated with S (R2=48.6%). In vivo results demonstrated the feasibility of extracting microstructural information of the tissue by the analysis of SC maps.

International audience; 3D fast/ultrasound imaging has emerged in the last years in research but still suffers from its poor image quality. Indeed, using plane/diverge waves does not permit to insonify the medium with sufficient energy at each point to get a good signal-to-noise ratio, contrast-to-noise-ratio or even resolution. On the other hand, coded excitation is currently used to increase signal-to-noise ratio and penetration depth. In this work, the objective is to combine 3D fast/ultrafast imaging with coded excitation to achieve better image quality at a high acquisition rate. Promising experimental results are obtained from both wire and cyst phantoms using a chirp excitation signal. The contrast-to-noise ratio and signal-to-noise ratio were improved by 4 dB and 2 dB respectively by the proposed method in comparison to the conventional way to do 3D imaging using a standard transmit. The improvement of the axial resolution of about 17% is the third important result obtained by the developed method still in comparison with the classical method. Experimental results show that an effective implementation on a research scanner of 3D coded excitation using plane wave imaging is possible.

In the context of acute myocardial infarction, we propose that Low Energy Ultrasound (LEUS) exposures might attenuate the detrimental effects of ischemia and reperfusion injury. Specifically, our goal is to quantify and monitor the effects of ultrasound using an in vitro cardiac cell model of ischemia-reperfusion. The study was conducted using a mono-layer cell model (H9C2 cardiomyoblasts) exposed to a prolonged hypoxia-reoxygenation (HR) challenge. Two groups were formed: i). HR: cells submitted to hypoxia followed by reoxygenation period, and ii). HR+PostCond-LEUS: LEUS applied repeatedly for 20 min starting at the onset of reoxygenation. Cell death was evaluated by flow cytometry for each group at the end of US exposure. Cell viability was clearly improved in the HR+PostCond-LEUS group, at all time points during reoxygenation. This study suggests a potential protective effect of LEUS on cardiomyoblasts exposed to a prolonged hypoxia-reoxygenation insult. More complex in vitro models exploring potential protective mechanisms (SAFE and RISK signaling pathways and in vivo models will be required for a better comprehension of the underlying mechanisms.

The necessity of frame rate enhancement is particularly important in 3D echocardiography since full-volume acquisitions at a quality comparable to 2D imaging require increasing the number of transmissions. Despite the potential of Multi Line Transmission (MLT) in providing high frame rates while preserving the image quality, this method has a non-negligible limitation: the presence of cross-talks artefacts. The objective of this study was to investigate experimentally if cross-talk reduction can be achieved by replacing the typical alignment of the transmissions parallel to the element lines or columns by an alignment along the transverse diagonal of the transducer. The results showed that the diagonal transmission provides a better axial resolution than the linear one. Additionally, as the principal cause of cross-talk artefacts is the interference between the main lobe and the side lobes of parallel transmissions, we measured the values of these lobes in the two cases. The main lobe was up to 5% higher than the side lobe for the linear transmission and up to 25 % higher for the diagonal one. However, the cross talk reduction was achieved with the cost of Contrast to Noise Ratio (CNR), which was three times lower for the diagonal transmission compared to the linear one.

International audience; Background, Motivation and ObjectiveThe emergence of ultrafast imaging allowed new insights into cardiac deformation/motion analysis that enabled new advancements in clinical diagnosis. Several studies showed that a good compromise between the temporal and the spatial resolution can be obtained by transmitting multiple focused beams (MLT) simultaneously. However, the currentimplementation of MLT in 3D cannot be used in dynamic conditions as it was obtained synthetically, by summing the Single Line Transmit (SLT) events before beamforming [Ortega et al., TUFFC, 2016]. The objective of this study is to evaluate for the first time the performance of MLT in 3D ultrasound under dynamic conditions. Statement of Contribution/MethodsThe 3D images were acquired by using four Verasonics research scanners, each of them controlling 8-by-32 elements of a 32-by-32, 3 MHz, Vermon ultrasound transducer. The scan volume was created by stitching 30 triangular sectors (each of angular width = 40°) along the azimuth angular direction. For each triangular sector, 9 transmission events were used for insonifying a full sector by sending simultaneously three focused beams. Thus, the total number of transmissions was 270 corresponding to 27 (elevational) x 30 (azimuthal) focal beams. The Pulse Repetition Frequency (PRF) was set to 2250. Our in vitro model was a tissue mimicking spinning disc having a diameter of 11 cm whose speed could be controlled by a step motor. The disk was placed at approximately 3 cm away from transducer and the image depth was 6.8 cm. Doppler velocities were estimated from the IQ signals using a 2D auto-correlator and a packet length of 7.Results/DiscussionThis study demonstrates the feasibility of Doppler velocity estimation for 3D MLT images. The estimated values presented in the image were in agreement with the expected velocities since the relative mean error was 3.5%. We obtained a three-fold increase in volume rate compared with focus imaging. Analyzing further different MLT/MLA configurations would allow increasing the frequency of volumes and obtaining the optimal compromise between high volume rate and quality color Doppler.

Echocardiography is the most used modality for the evaluation of cardiac function. To obtain a time-resolved volumetric quantification of cardiac motion, ultrafast 3-D imaging is required. Ultrafast ultrasound imaging with diverging waves has demonstrated its potential for clinical 2-D echocardiography. It has been shown that MoCo (motion compensation) strategies based on a triangle transmit arrangement could produce high-contrast cardiac B-mode images at 500 fps [1]. With the purpose of developing high-volume-rate 3-D echocardiography, we introduced a transmit sequence enabling volumetric MoCo for spherical diverging wave imaging.

This work shows the feasibility of performing ultrafast 3D ultrasound imaging by producing diverging waves (DW) with 2D sparse arrays. The 3D volumes were experimentally acquired by individually driving the 1024 elements of a full 32×32 matrix array. The volumes obtained with different sparse configurations were compared to those obtained with two references arrays: the full 32×32 array and a dense array of 716 elements where obtained by de-activating the corner elements of the full array. The comparison criteria were the lateral resolution using the full width at half maximum (FWHM) and the contrast to noise ratio (CNR), measured on the images obtained by scanning a grayscale phantom. The results show that ultrafast 3D ultrasound imaging can be performed with a reduced number of channels using 2D sparse arrays. However, there is still space to find the optimal 2D sparse array configurations to best transmit 3D DWs and potential improvements on image quality may be achieved with dedicated optimization e.g. cost function to homogenize DW wave front.

Ultrafast imaging based on diverging wave (DW) is an active area of research in ultrasound sectorial acquisition because of its capacity of reaching high frame rate. Recently, we have introduced a 3D Fourier-based formalism for the reconstruction of 3D sectorial images with DW insonifications and validated with numerical simulations [Zhang et al. IUS 2016]. The proposed method can provide comparable image quality as DAS but with a much lower computational complexity. This study aims to experimentally verify these theoretical conclusions in vitro.

Echocardiography is one of the most widely spread modality for non-invasive 2-D imaging of the heart. However, to fully observe and quantify left ventricular function and morphology, 3-D imaging is required. The observation of the whole left ventricle remains limited by the temporal resolution of conventional 3-D echocardiography. Indeed, the only way to get wide-angle volumetric images is to merge small scan volumes during inspiratory breath hold. This limitation makes difficult 3D speckle tracking echocardiography. Increasing the temporal resolution with ultrafast 3-D echocardiography could enable imaging of the whole myocardium in a single heartbeat. It has recently been shown that 3-D ultrafast ultrasound imaging is possible with spherical diverging waves emitted by a matrix-array probe. As ultrafast ultrasound requires backscattered signals to be combined coherently, this approach is sensitive to high-velocity tissue motion, especially when the heart is investigated. Ultrafast ultrasound imaging with diverging waves has demonstrated its potential for clinical 2-D-echocardiography only when motion compensation (MoCo) is integrating during the compounding process. It has been demonstrated that MoCo strategies based on a triangle transmit arrangement could produce high-contrast 2-D cardiac B-mode images at 500 fps. In this study, we proposed a feasibility study of high-frame-rate 3-D echocardiography based on an innovative MoCo approach using steered spherical diverging waves. The method was validated in vitro with a rotating disk. The sequences were generated with four Verasonics scanners combined to get 1024 channels and a 3-MHz matrix phased array of 32×32 elements.

The emergence of ultrafast imaging allowed new insights into cardiac deformation/motion analysis that enabled new advancements in clinical diagnosis. Several studies showed that a good compromise between the temporal and the spatial resolution can be obtained by transmitting multiple focused beams (MLT) simultaneously. However, the current implementation of MLT in 3D cannot be used in dynamic conditions as it was obtained synthetically, by summing the Single Line Transmit (SLT) events before beamforming [Ortega et al., TUFFC, 2016]. The objective of this study is to evaluate for the first time the performance of MLT in 3D ultrasound under dynamic conditions.

International audience; Background, Motivation and ObjectiveAchieving a high frame rate in echocardiography is highly important for quantifying the short phases of the cardiac cycle that contain valuable information for medical diagnosis. Additionally, the 3D quantitative assessment of the heart would significantly improve the current measurements used in daily clinical routine. Nevertheless, obtaining ultrafast imagesremains a challenge due to the trade-off between the image quality and a high frame rate, especially when volumetric data is acquired. Among the current ultrafast imaging methods, multi-line-transmit imaging (MLT) provides an increased frame rate but in the same time mostly preserves the image quality. However, the current implementation of this methodin 3D proposed by Ortega et al. [IEEE TUFC 2016] is based on generating the MLT data synthetically by summing up the raw data before beamforming. In this paper we present the first real-time implementation of the MLT in 3D ultrasound.Statement of Contribution/MethodsMulti-line-transmission was performed by dividing the angular aperture into three regions. Then, due to practical limitations, three equally spaced focused beams were transmitted simultaneously in the first YZ plane of the first region. Once the transmission for a full YZ plane was completed, the process was repeated for the first YZ plane of the second region. The same steps were followed till the full volume was insonified. Data acquisition was performed using four Verasonics systems synchronized to drive a 32x32 matrix probe. A transmit frequency of 2.97 MHz was used and the images were acquired using a sampling frequency of 11.9 MHz. The focal point was set to 6.7 cm depth along 27 different angles in elevational direction (YZ) and 30 angles in azimuthal direction (XZ) between -20º and 20º.Results/DiscussionThe contrast and the resolution assessment, performed on a CIRS ultrasound phantom (Figure 1) showed a contrast of 6.36 dB and a mean axial (lateral) resolution of 0.9 mm (1.77 mm) measured for different depths. These values are comparable with those obtained for a 3D focused sequence obtained by using the same emission parameters. The results indicate the potential of MLT 3D for achieving high contrast and resolution while increasing the frame rate. This study thus demonstrates the feasibility of 3D MLT in real-time and extends its possible applications to dynamic cardiac imaging.