Hardware-Independent Deep Signal Processing: A Feasibility Study in Echocardiography.

Deep learning (DL) models have emerged as alternative methods to conventional ultrasound (US) signal processing, offering the potential to mimic signal processing chains, reduce inference time, and enable the portability of processing chains across hardware. This paper proposes a DL model that replicates the fine-tuned BMode signal processing chain of a high-end US system and explores the potential of using it with a different probe and a lower-end system. A deep neural network was trained in a supervised manner to map raw beamformed in-phase and quadrature component data into processed images. The dataset consisted of 30,000 cardiac image frames acquired using the GE HealthCare Vivid E95 system with the 4Vc-D matrix array probe. The signal processing chain includes depth-dependent bandpass filtering, elevation compounding, frequency compounding, and image compression and filtering. The results indicate that a lightweight DL model can accurately replicate the signal processing chain of a commercial scanner for a given application. Evaluation on a 15 patient test dataset of about three thousand image frames gave a structural similarity index measure of 98.56 ± 0.49. Applying the DL model to data from another probe showed equivalent or improved image quality. This indicates that a single DL model may be used for a set of probes on a given system that targets the same application, which could be a cost-effective tuning and implementation strategy for vendors. Further, the DL model enhanced image quality on a Verasonics dataset, suggesting the potential to port features from high-end US systems to lower-end counterparts.

Automated Measurement of Ovary Development in Atlantic Salmon Using Deep Learning.

OBJECTIVE: Salmon breeding companies control the egg stripping period through environmental change, which triggers the need to identify the state of maturation. Ultrasound imaging of the salmon ovary is a proven non-invasive tool for this purpose; however, the process is laborious, and the interpretation of the ultrasound scans is subjective. Real-time ultrasound image segmentation of Atlantic salmon ovary provides an opportunity to overcome these limitations. However, several application challenges need to be addressed to achieve this goal. These challenges include the potential for false-positive and false-negative predictions, accurate prediction of attenuated lower ovary parts and resolution of inconsistencies in predicted ovary shape. METHODS: We describe an approach designed to tackle these obstacles by employing targeted pre-training of a modified U-Net, capable of performing both segmentation and classification. In addition, a variational autoencoder (VAE) and generative adversarial network (GAN) were incorporated to rectify shape inconsistencies in the segmentation output. To train the proposed model, a data set of Atlantic salmon ovaries throughout two maturation periods was recorded. RESULTS: We then tested our model and compared its performance with that of conventional and novel U-Nets. The method was also tested in a salmon on-site ultrasound examination setting. The results of our application indicate that our method is able to efficiently segment salmon ovary with an average Dice score of 0.885 per individual in real-time. CONCLUSION: These results represent a competitive performance for this specific application, which enables us to design an automated system for smart monitoring of maturation state in Atlantic salmon.

The higher order leaky Lamb wave sensitivity of a notch in a fluid-immersed plate.

We present an ultrasonic method of detecting cracks in pipelines based on using normally incident transducers in a pitch-catch setup, which can only excite Lamb modes of higher order than the fundamental modes A0 and S0 commonly used in crack detection applications. By excitation and measurements of the Lamb modes S1, S2, and A3, in a steel plate immersed in fluid with and without a notch (to emulate a crack), the performance of the modes towards crack detection is quantified by assessing whether it returns a high leaky component and whether the notch has a large impact on the leaky component. In order to narrow the scope of measurements necessary to investigate notch sensitivity for different system parameters, and to potentially optimize the system setup, we present a computationally efficient theoretical model based on angular spectrum method (ASM) and the theoretical sensitivity kernel formulation from the field of seismology that accounts for a scatterer in the wave path between the transmitter and receiver. The model is compared against measurements, which show that the frequency components of the S2 mode has both the largest leaky frequency component in the given setup and the largest sensitivity at a frequency close to the maximum leaky frequency such that a difference caused by the notch is easily measured. By using the measurements and the validation calculation as baseline reference, we calculate the expected S2 mode sensitivity and leaky components for larger plate thicknesses and larger standoffs, which exemplifies how the model can be applied in measurement system design and optimization.

A Very Large Cardiac Channel Data Database (VLCD) Used to Evaluate Global Image Coherence (GIC) as an In Vivo Image Quality Metric.

Ultrasound image quality is of utmost importance for a clinician to reach a correct diagnosis. Conventionally, image quality is evaluated using metrics to determine the contrast and resolution. These metrics require localization of specific regions and targets in the image such as a region of interest (ROI), a background region, and/or a point scatterer. Such objects can all be difficult to identify in in-vivo images, especially for automatic evaluation of image quality in large amounts of data. Using a matrix array probe, we have recorded a Very Large cardiac Channel data Database (VLCD) to evaluate coherence as an in vivo image quality metric. The VLCD consists of 33280 individual image frames from 538 recordings of 106 patients. We also introduce a global image coherence (GIC), an in vivo image quality metric that does not require any identified ROI since it is defined as an average coherence value calculated from all the data pixels used to form the image, below a preselected range. The GIC is shown to be a quantitative metric for in vivo image quality when applied to the VLCD. We demonstrate, on a subset of the dataset, that the GIC correlates well with the conventional metrics contrast ratio (CR) and the generalized contrast-to-noise ratio (gCNR) with R = 0.74 ( ) and R = 0.62 ( ), respectively. There exist multiple methods to estimate the coherence of the received signal across the ultrasound array. We further show that all coherence measures investigated in this study are highly correlated ( 0.9 and ) when applied to the VLCD. Thus, even though there are differences in the implementation of coherence measures, all quantify the similarity of the signal across the array and can be averaged into a GIC to evaluate image quality automatically and quantitatively.

Aberration correction in 2D echocardiography.

Background: An aberration correction algorithm has been implemented and demonstrated in an echocardiographic clinical trial using two-dimensional (2D) imaging. The method estimates and compensates arrival time errors between different sub-aperture processor (SAP) signals in a matrix array probe. Methods: Five standard views of channel data cineloops were recorded from 22 patients (11 male and 11 female) resulting in a total of 116 cineloops. The channel data were processed with and without the aberration correction algorithm, allowing for side-by-side comparison of images processed from the same channel data cineloops. Results: The aberration correction algorithm improved image quality, as quantified by a coherence metric, in all 7,380 processed frames. In a blinded and left-right-randomized side-by-side evaluation, four cardiologists (two experienced and two in training) preferred the aberration corrected cineloops in 97% of the cases. The clinicians reported that the corrected cineloops appeared sharper with better contrast and less noise. Many structures like valve leaflets, chordae, endocardium, and endocardial borders appeared narrower and more clearly defined in the aberration corrected images. An important finding is that aberration correction improves contrast between the endocardium and ventricle cavities for every processed image. The gain difference was confirmed by the cardiologists in their feedback and quantified with a median global gain difference estimate between the aberration-corrected and non-corrected images of 1.2 dB. Conclusions: The study shows the potential value of aberration correction in clinical echocardiography. Systematic improvement of images acquired with state-of-art equipment was observed both with quantitative metrics of image quality and clinician preference.

Effect of bubble size on ultrasound backscatter from bubble clouds in the context of gas kick detection in boreholes.

Early detection of gas influx in boreholes while drilling is of significant interest to drilling operators. Several studies suggest a good correlation between ultrasound backscatter/attenuation and gas volume fraction (GVF) in drilling muds, and thereby propose methods for quantification of GVF in boreholes. However, the aforementioned studies neglect the influence of bubble size, which can vary significantly over time. This paper proposes a model to combine existing theories for ultrasound backscatter from bubbles depending on their size, viz. Rayleigh scattering for smaller bubbles, and specular reflection for larger bubbles. The proposed model is demonstrated using simulations and experiments, where the ultrasound backscatter is evaluated from bubble clouds of varying bubbles sizes. It is shown that the size and number of bubbles strongly influence ultrasound backscatter intensity, and it is correlated to GVF only when the bubble size distribution is known. The information on bubble size is difficult to obtain in field conditions causing this correlation to break down. Consequently, it is difficult to reliably apply methods based on ultrasound backscatter, and by extension its attenuation, for the quantification of GVF during influx events in a borehole. These methods can however be applied as highly sensitive detectors of gas bubbles for GVF [Formula: see text]1 vol[Formula: see text].

ASM and finite beam description of the excited leaky Lamb wave fields in a fluid-immersed plate.

An angular spectrum method (ASM) full-wave description of stress and energy density in a fluid-immersed plate for the optimization of leaky Lamb wave applications is presented. It models the case when leaky Lamb waves are generated by an external finite transmitter in the immersion fluid, and can calculate the associated stress, energy density, and other field variables within the plate. The normal component of the stress tensor and the energy density are compared against calculations in COMSOL with good agreement, but with some differences due to the two methods. The spatial field of the stress is analyzed using the angular spectrum (plane wave) representation of the stress, which is also used as a reference to exemplify the discrepancies between a pure plane wave approach in leaky Lamb wave applications and the spatial fields that accounts for diffraction and dispersion effects. Comments on the insight that the spatial fields within the plate may provide towards NDT/SHM applications are also given, along with a discussion on why the derivation and implementation of the ASM model is valuable when compared against a benchmarked, ready-to-use software such as COMSOL.

Ocular surface microcirculation is better preserved with pulsatile versus continuous flow during cardiopulmonary bypass-An experimental pilot.

BACKGROUND: Non-pulsatile cardiopulmonary bypass (CPB) may induce microvascular dysregulation. In piglets, we compared ocular surface microcirculation during pulsatile versus continuous flow (CF) bypass. METHODS: Ocular surface microcirculation in small tissue volumes (~0.1 mm3 ) at limbus (high metabolic rate) and bulbar conjunctiva (low metabolic rate) was examined in a porcine model using computer assisted video microscopy and diffuse reflectance spectroscopy, before and after 3 and 6 h of pulsatile (n = 5 piglets) or CF (n = 3 piglets) CPB. Functional capillary density, capillary flow velocity and microvascular oxygen saturation were quantified. RESULTS: At limbus, velocities improved with pulsatility (p < 0.01) and deteriorated with CF (p < 0.01). In bulbar conjunctiva, velocities were severely reduced with CF (p < 0.01), accompanied by an increase in capillary density (p < 0.01). Microvascular oxygen saturation decreased in both groups. CONCLUSION: Ocular surface capillary densities and flow patterns are better preserved with pulsatile versus CF during 6 h of CPB in sleeping piglets.

Quantification of ocular surface microcirculation by computer assisted video microscopy and diffuse reflectance spectroscopy.

In piglets we tested the applicability of digital video microscopy and diffuse reflectance spectroscopy for non-invasive assessments of limbal and bulbar conjunctival microcirculation. A priori we postulated that the metabolic rate is higher in limbal as compared to bulbar conjunctiva, and that this difference is reflected in microvascular structure or function between the two locations. Two study sites, Oslo University Hospital (OUH), Norway and Cleveland Clinic (CC), USA, used the same video microscopy and spectroscopy techniques to record limbal and bulbar microcirculation in sleeping piglets. Recordings were analyzed with custom-made software to quantify functional capillary density, capillary flow velocity and microvascular oxygen saturation in measuring volumes of approximately 0.1 mm3. The functional capillary density was higher in limbus than in bulbar conjunctiva at both study sites (OUH: 18.1 ± 2.9 versus 12.2 ± 2.9 crossings per mm line, p < 0.01; CC: 11.3 ± 3.0 versus 7.1 ± 2.8 crossings per mm line, p < 0.01). Median categorial capillary blood flow velocity was higher in bulbar as compared with limbal recordings (CC: 3 (1-3) versus 1 (0-3), p < 0.01). Conjunctival microvascular oxygen saturation was 88 ± 5.9% in OUH versus 94 ± 7.5% in CC piglets. Non-invasive digital video microscopy and diffuse reflectance spectroscopy can be used to obtain data from conjunctival microcirculation in piglets. Limbal conjunctival microcirculation has a larger capacity for oxygen delivery as compared with bulbar conjunctiva.

Row-Column-Based Coherence Imaging Using a 2-D Array Transducer: A Row-Based Implementation.

Reverberations from tissues around the heart often result in cluttered echocardiograms with reduced diagnostic value. As a consequence, some patients must undergo more expensive and, in some cases, invasive imaging modalities. Coherence-based beamforming can suppress the effect of incoherent reverberations compared with the coherent signal. In some cases, these incoherent reverberations are received by only a part of the aperture. However, the coherence-based techniques, when used on a 1-D array transducer, do not take this into account. We propose an extension of coherence imaging method when using a 2-D array transducer and test a row-based implementation of this extension on two in vitro scenarios and four in vivo cases. The results show that the proposed method improves the lateral resolution compared with the (already improved) resolution with conventional coherence imaging. Furthermore, it gives up to 28% increase in generalized contrast-to-noise ratio (gCNR) (improved detection probability) when incoherent reverberations are partly received by the transducer in the elevation direction.

The Generalized Contrast-to-Noise Ratio: A Formal Definition for Lesion Detectability.

In the last 30 years, the contrast-to-noise ratio (CNR) has been used to estimate the contrast and lesion detectability in ultrasound images. Recent studies have shown that the CNR cannot be used with modern beamformers, as dynamic range alterations can produce arbitrarily high CNR values with no real effect on the probability of lesion detection. We generalize the definition of CNR based on the overlap area between two probability density functions. This generalized CNR (gCNR) is robust against dynamic range alterations; it can be applied to all kind of images, units, or scales; it provides a quantitative measure for contrast; and it has a simple statistical interpretation, i.e., the success rate that can be expected from an ideal observer at the task of separating pixels. We test gCNR on several state-of-the-art imaging algorithms and, in addition, on a trivial compression of the dynamic range. We observe that CNR varies greatly between the state-of-the-art methods, with improvements larger than 100%. We observe that trivial compression leads to a CNR improvement of over 200%. The proposed index, however, yields the same value for compressed and uncompressed images. The tested methods showed mismatched performance in terms of lesion detectability, with variations in gCNR ranging from -0.08 to +0.29. This new metric fixes a methodological flaw in the way we study contrast and allows us to assess the relevance of new imaging algorithms.

Methods for reverberation suppression utilizing dual frequency band imaging.

Reverberations impair the contrast resolution of diagnostic ultrasound images. Tissue harmonic imaging is a common method to reduce these artifacts, but does not remove all reverberations. Dual frequency band imaging (DBI), utilizing a low frequency pulse which manipulates propagation of the high frequency imaging pulse, has been proposed earlier for reverberation suppression. This article adds two different methods for reverberation suppression with DBI: the delay corrected subtraction (DCS) and the first order content weighting (FOCW) method. Both methods utilize the propagation delay of the imaging pulse of two transmissions with alternating manipulation pressure to extract information about its depth of first scattering. FOCW further utilizes this information to estimate the content of first order scattering in the received signal. Initial evaluation is presented where both methods are applied to simulated and in vivo data. Both methods yield visual and measurable substantial improvement in image contrast. Comparing DCS with FOCW, DCS produces sharper images and retains more details while FOCW achieves best suppression levels and, thus, highest image contrast. The measured improvement in contrast ranges from 8 to 27 dB for DCS and from 4 dB up to the dynamic range for FOCW.

Mechanisms of the ultrasound-mediated intracellular delivery of liposomes and dextrans.

The mechanism involved in the ultrasoundenhanced intracellular delivery of fluorescein-isothiocyanate (FITC)-dextran (molecular weight 4 to 2000 kDa) and liposomes containing doxorubicin (Dox) was studied using HeLa cells and an ultrasound transducer at 300 kHz, varying the acoustic power. The cellular uptake and cell viability were measured using flow cytometry and confocal microscopy. The role of endocytosis was investigated by inhibiting clathrin- and caveolae-mediated endocytosis, as well as macropinocytosis. Microbubbles were found to be required during ultrasound treatment to obtain enhanced cellular uptake. The percentage of cells internalizing Dox and dextran increased with increasing mechanical index. Confocal images and flow cytometric analysis indicated that the liposomes were disrupted extracellularly and that released Dox was taken up by the cells. The percentage of cells internalizing dextran was independent of the molecular weight of dextrans, but the amount of the small 4-kDa dextran molecules internalized per cell was higher than for the other dextrans. The inhibition of endocytosis during ultrasound exposure resulted in a significant decrease in cellular uptake of dextrans. Therefore, the improved uptake of Dox and dextrans may be a result of both sonoporation and endocytosis.

Effect of ultrasound parameters on the release of liposomal calcein.

The ultrasound exposure parameters that maximize drug release from dierucoyl-phosphatidylcholine (DEPC)-based liposomes were studied using two transducers operating at 300 kHz and 1 MHz. Fluorescent calcein was used as a model drug, and the release from liposomes in solution was measured using a spectrophotometer. The release of calcein was more efficient at 300 kHz than at 1 MHz, with thresholds of peak negative pressures of 0.9 MPa and 1.9 MPa, respectively. Above this threshold, the release increased with increasing peak negative pressure, mechanical index (MI), and duty cycle. The amount of drug released followed first-order kinetics and increased with exposure time to a maximal release. To increase the release further, the MI had to be increased. The results demonstrate that the MI and the overall exposure time are the major parameters that determine the drug's release. The drug's release is probably due to mechanical (cavitation) rather than thermal effects, and that was also confirmed by the detection of hydroxide radicals.

Transmit beamforming for optimal second-harmonic generation.

A simulation study of transmit ultrasound beams from several transducer configurations is conducted to compare second-harmonic imaging at 3.5 MHz and 11 MHz. Second- harmonic generation and the ability to suppress near field echoes are compared. Each transducer configuration is defined by a chosen f-number and focal depth, and the transmit pressure is estimated to not exceed a mechanical index of 1.2. The medium resembles homogeneous muscle tissue with nonlinear elasticity and power-law attenuation. To improve computational efficiency, the KZK equation is utilized, and all transducers are circular-symmetric. Previous literature shows that second-harmonic generation is proportional to the square of the transmit pressure, and that transducer configurations with different transmit frequencies, but equal aperture and focal depth in terms of wavelengths, generate identical second-harmonic fields in terms of shape. Results verify this for a medium with attenuation f1. For attenuation f1.1, deviations are found, and the high frequency subsequently performs worse than the low frequency. The results suggest that high frequencies are less able to suppress near-field echoes in the presence of a heterogeneous body wall than low frequencies.

Nonlinear propagation delay and pulse distortion resulting from dual frequency band transmit pulse complexes.

A method of acoustic imaging is discussed that potentially can improve the diagnostic capabilities of medical ultrasound. The method, given the name second order ultrasound field imaging, is achieved by the processing of the received signals from transmitted dual frequency band pulse complexes with at least partly overlapping high frequency (HF) and low frequency (LF) pulses. The transmitted HF pulses are used for image reconstruction whereas the transmitted LF pulses are used to manipulate the elastic properties of the medium observed by the HF imaging pulses. In the present paper, nonlinear propagation effects observed by a HF imaging pulse due to the presence of a LF manipulation pulse is discussed. When using dual frequency band transmit pulse complexes with a large separation in center frequency (e.g., 1:10), these nonlinear propagation effects are manifested as a nonlinear HF propagation delay and a HF pulse distortion different from conventional harmonic distortion. In addition, with different transmit foci for the HF and LF pulses, nonlinear aberration will occur.

Nonlinear propagation acoustics of dual-frequency wide-band excitation pulses in a focused ultrasound system.

In this article, acoustic propagation effects of dual-frequency wide-band excitation pulses in a focused ultrasound system are demonstrated in vitro. A designed and manufactured dual-frequency band annular array capable of transmitting 0.9/7.5 MHz center frequency wide-band pulses was used for this purpose. The dual-frequency band annular array, has been designed using a bi-layer piezo-electric stack. Water tank measurements demonstrate the function of the array by activating the low- and high-frequency layers individually and simultaneously. The results show that the array works as intended. Activating the low- and high-frequency layers individually, results in less than -50 dB signal level from the high- and low-frequency layers respectively. Activating both layers simultaneously, produce a well defined dual-frequency pulse. The presence of the low-frequency pulse leads to compression, expansion, and a time delay of the high-frequency pulse. There is a phase shift between the low- and high-frequency pulse as it propagates from the array to the focus. This makes the latter described effects also dependent on the array configuration. By varying the low-frequency pressure, a shift of up to 0.5 MHz in center frequency of a 8.0 MHz transmitted high-frequency pulse is observed at the array focus. The results demonstrate the high propagation complexity of dual-frequency pulses.

Utilizing dual frequency band transmit pulse complexes in medical ultrasound imaging.

A method of acoustic imaging that potentially can improve the diagnostic capabilities of medical ultrasound is presented. The method, given the name SURF (Second order UltRasound Field) imaging, is achieved by processing the received signals from transmitted dual frequency band pulse complexes with at least partly overlapping high frequency (HF) and low frequency (LF) pulses. The transmitted HF pulses are used for image reconstruction, whereas the transmitted LF pulses are used to manipulate the elastic properties of the medium observed by the HF imaging pulses. The present paper discusses fundamental concepts in relation to the use of dual frequency band pulse complexes for medical ultrasound imaging.

Transmit beams adapted to reverberation noise suppression using dual-frequency SURF imaging.

A method that uses dual-frequency pulse complexes of widely separated frequency bands to suppress noise caused by multiple scattering or multiple reflections in medical ultrasound imaging is presented. The excitation pulse complexes are transmitted to generate a second order ultrasound field (SURF) imaging synthetic transmit beam. This beam has reduced amplitude near the transducer, which illustrates the multiple scattering suppression ability of the imaging method. Field simulations solving a nonlinear wave equation are used to calculate SURF imaging beams, which are compared with beams for pulse inversion (PI) and fundamental imaging. In addition, a combined SURF and PI beam generation is described and compared with the beams mentioned above. A quality ratio, relating the energy within the near-field to that within the imaging region, is defined and used to score the multiple scattering and multiple reflection suppression abilities when imaging with the different beams. The realized combined SURF-PI beam scores highest, followed by SURF, PI (that score equally well), and the fundamental. The amplitude in the imaging region and therefore also the SNR is highest for the fundamental followed by SURF, PI, and SURF-PI. The work hence indicates that when substituting PI for SURF, one may trade increased SNR into use of increased imaging frequencies without loss of multiple scattering and multiple reflection noise suppression.