Radio Frequency Data Research Articles

A Systemic Study on the Performance of Different Quantitative Ultrasound Imaging Techniques for Microwave Ablation Monitoring of Liver

Many quantitative ultrasound imaging techniques based on tissue properties have been established to ensure the safety and efficacy of thermal ablation treatment. However, which quantitative ultrasound imaging technique has superior monitoring properties has not yet been revealed. In this study, we investigated and compared the performance of envelope data distribution-dependent ultrasonic Nakagami, envelope data distribution-independent horizontally normalized Shannon entropy (hNSE), and differential attenuation coefficient intercept (DACI) based on frequency domain imaging technique during liver thermal ablation treatments. Quantitative ultrasound parameter images were constructed simultaneously from radiofrequency (RF) data collected during the process of microwave ablation (MWA) treatments in <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">ex vivo</i> and <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">in vivo</i> porcine livers. Contrast-to-noise ratios (CNR) were calculated to evaluate the contrast resolution of different imaging techniques. The mean CNR value of hNSE images was 6.06 ± 2.34 dB, being significantly higher than that of Nakagami images ( <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">p</i> <0.05, statistical) and DACI images ( <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">p</i> <0.05, statistical). Furthermore, the receiver operating characteristic (ROC) was used to evaluate the thermal lesion detection ability. The area under the curve (AUC) of B-mode, Nakagami, hNSE and DACI was 0.823, 0.839, 0.86 and 0.825, respectively. Additionally, our data analysis exhibited that hNSE has the excellent property of ablation area detections in the present study, whose correlation coefficient is 0.91. In conclusion, our study suggests that hNSE imaging, which does not require the specific data distribution has the best performance in thermal ablation detections and is a promising imaging method for further clinical research and application.

Secure Industrial IoT Systems via RF Fingerprinting Under Impaired Channels With Interference and Noise

Industrial IoT-enabled critical infrastructures are susceptible to cyber attacks due to their mission-critical deployment. To ensure security by design, radio frequency (RF)-based security is considered an effective way for wirelessly monitored or actuated critical infrastructures. For this purpose, this paper presents a novel augmentation-driven deep learning approach to analyze unique transmitter fingerprints and determine the legitimacy of a user device or transmitter. An RF fingerprinting model is susceptible to various channel and environmental conditions that impact the learning performance of a machine/deep learning model. As data gathering cannot always be considered a feasible alternative, efficient solutions that can tackle the impact of varying propagation channels on learning performance are emergent. This work aims to shed light on the RF fingerprinting problem from a different angle when 4G, 5G, and WiFi data samples are collected from different transmitters by proposing a fine-grained augmentation approach to improve the learning performance of a deep learning model. This work also proposes an enhanced classifier structure following the fine-grained augmentation approach. Results of experiments, conducted on the POWDER dataset, demonstrate promising RF fingerprinting performance when training data are augmented in a waveform-specific fine-grained manner. Thus, the RF identification accuracy can be boosted to 97.84% on unseen RF data instances from our previously published work where we had achieved an accuracy of 87.94% using tapped delay line (TDL)/clustered delay line (CDL)-based augmentation approach. The paper also presents a sensitivity analysis of the fine-grained approach concerning different signal-to-noise-ratio (SNR), signal-to-interference-ratio (SIR) levels (20 dB and 30 dB), and signal-to-interference-plus-noise-ratio (SINR) levels (15 dB, 25 dB). The sensitivity analysis exhibits that it achieves 85.78% accuracy at 20 dB SIR on both Day 1 (train) and Day 2 (test) data. In addition, it achieves 92.37% accuracy even at 20 dB SNR on Day 2 data from POWDER dataset. Furthermore, it achieves 84.95% accuracy at 15 dB SINR on Day 2 data. Hence, these results exhibit the resiliency of the fine-grained augmentation approach against interference and noise.

Quantitative Ultrasound for Evaluation of Tumour Response to Ultrasound-Microbubbles and Hyperthermia.

Objectives: Prior study has demonstrated the implementation of quantitative ultrasound (QUS) for determining the therapy response in breast tumour patients. Several QUS parameters quantified from the tumour region showed a significant correlation with the patient's clinical and pathological response. In this study, we aim to identify if there exists such a link between QUS parameters and changes in tumour morphology due to combined ultrasound-stimulated microbubbles (USMB) and hyperthermia (HT) using the breast xenograft model (MDA-MB-231). Method: Tumours grown in the hind leg of severe combined immuno-deficient mice were treated with permutations of USMB and HT. Ultrasound radiofrequency data were collected using a 25 MHz array transducer, from breast tumour-bearing mice prior and post-24-hour treatment. Result: Our result demonstrated an increase in the QUS parameters the mid-band fit and spectral 0-MHz intercept with an increase in HT duration combined with USMB which was found to be reflective of tissue structural changes and cell death detected using haematoxylin and eosin and terminal deoxynucleotidyl transferase dUTP nick end labelling stain. A significant decrease in QUS spectral parameters was observed at an HT duration of 60 minutes, which is possibly due to loss of nuclei by the majority of cells as confirmed using histology analysis. Morphological alterations within the tumour might have contributed to the decrease in backscatter parameters. Conclusion: The work here uses the QUS technique to assess the efficacy of cancer therapy and demonstrates that the changes in ultrasound backscatters mirrored changes in tissue morphology.

UAV-Based Volumetric Measurements toward Radio Environment Map Construction and Analysis.

Unmanned aerial vehicle (UAV)-empowered communications have gained significant attention in recent years due to the promise of agile coverage provision for a large number of various mobile nodes on the ground and in three-dimensional (3D) space. Consequently, there is a need for efficient spectrum utilization in these dense aerial networks, which is characterized through radio environment maps (REMs), the construction of which is an important research area. Nevertheless, due to the difficult collection of radio frequency (RF) data, there are limited works that are based on real-world measurement campaigns. This paper presents a novel experimental setup that includes a constellation of three UAVs, the communication signals of which are measured by a software-defined radio (SDR) mounted on a separate UAV. It follows a trajectory that defines the REM's two-dimensional (2D) area on a plane, executed at four altitudes, to extend the REM to 3D. The measurements are then processed and their features (received mean power level, average difference of the mean power, percentage of meaningful correlations) are analyzed in the temporal, spatial, and frequency domains to determine the utilization of a 20 MHz band in the 2.4 GHz spectrum, as well as their variation with altitude. This analysis provides a base for research in reducing the amount of measurements (by identifying the regions of low and of high interest) and spectrum occupancy prediction for UAV-based communication coexistence.

Ultrasound Imaging With a Flexible Probe Based on Element Array Geometry Estimation Using Deep Neural Network.

Conventionally, ultrasound (US) diagnosis is performed using hand-held rigid probes. Such devices are difficult to be used for long-term monitoring because they need to be continuously pressed against the body to remove the air between the probe and body. Flexible probes, which can deform and effectively adhere to the body, are a promising technology for long-term monitoring applications. However, owing to the flexible element array geometry, the reconstructed image becomes blurred and distorted. In this study, we propose a flexible probe U.S. imaging method based on element array geometry estimation from radio frequency (RF) data using a deep neural network (DNN). The input and output of the DNN are the RF data and parameters that determine the element array geometry, respectively. The DNN was first trained from scratch with simulation data and then fine-tuned with in vivo data. The DNN performance was evaluated according to the element position mean absolute error (MAE) and the reconstructed image quality. The reconstructed image quality was evaluated with peak-signal-to-noise ratio (PSNR) and mean structural similarity (MSSIM). In the test conducted with simulation data, the average element position MAE was 0.86 mm, and the average reconstructed image PSNR and MSSIM were 20.6 and 0.791, respectively. In the test conducted with in vivo data, the average element position MAE was 1.11 mm, and the average reconstructed image PSNR and MSSIM were 19.4 and 0.798, respectively. The average estimation time was 0.045 s. These results demonstrate the feasibility of the proposed method for long-term real-time monitoring using flexible probes.

Gait variability analysis using continuous RF data streams of human activity

The detection and rapid response to critical health events, such as falls, as well as monitoring of balance and gait for fall risk assessment are important to ensuring positive outcomes, longevity, and quality-of-life, especially in older adults. Mobility limitations affect more than one-third of adults and older adults with a history of falls have exhibited increased gait variability relative to younger adults. However, quantitative gait analysis in laboratories equipped with gold standard sensors, are expensive and may not be easily accessible for aging and rural communities. Thus, radio frequency (RF) sensing is an emerging technology for in-home health monitoring because they are non-contact devices that are effective in the dark and do not acquire private visual recordings. This work proposes a technique for the estimation of step-time variability from continuous streams of RF data acquired while a person does daily activities. The proposed approach is first validated through comparison with a Vicon motion capture system in a lab environment, after which the proposed approaches are used to analyze a continuous stream of daily activity data acquired from 5 different participants. Our results show that RF sensing is a promising, effective tool for gait variability assessment in natural, in-home environments. • Step-time estimation from continuous RF data streams of sequential activity. • Segmentation, human activity classification, and gait variability assessment. • Enables gait analysis and fall risk assessment in home environments. • Radar enables non-intrusive, remote assessment of natural gait.

Deep Complex-Valued Convolutional Neural Network for Drone Recognition Based on RF Fingerprinting

Drone-aided ubiquitous applications play important roles in our daily lives. Accurate recognition of drones is required in aviation management due to their potential risks and disasters. Radiofrequency (RF) fingerprinting-based recognition technology based on deep learning (DL) is considered an effective approach to extracting hidden abstract features from the RF data of drones. Existing deep learning-based methods are either high computational burdens or have low accuracy. In this paper, we propose a deep complex-valued convolutional neural network (DC-CNN) method based on RF fingerprinting for recognizing different drones. Compared with existing recognition methods, the DC-CNN method has a high recognition accuracy, fast running time, and small network complexity. Nine algorithm models and two datasets are used to represent the superior performance of our system. Experimental results show that our proposed DC-CNN can achieve recognition accuracies of 99.5% and 74.1%, respectively, on four and eight classes of RF drone datasets.

Optically Powered and Controlled Drones Using Optical Fibers for Airborne Base Stations

Mobile communication services are crucial during emergency disasters and temporary events, and future mobile communication systems should be able to provide such services. Airborne base stations using drones are highly effective as stand-in base stations in areas where the ground base stations are inoperable or at temporary event sites. However, it is difficult for conventional drones to provide mobile communication services without interruption due to flight time limitations caused by their limited battery capacity. Thus, a drone drive with a non-interrupted power supply is urgently needed. In this study, we developed an airborne base station that enables drones to be driven and maneuvered by optical fibers. We simultaneously transmitted radio frequency (RF) data signals for the airborne base station and control signals for the drone and evaluated the transmission performances of the RF signals and the controllability of the drone. Furthermore, we conducted a flight experiment on a medium-sized drone powered by optical fibers.

Deep learning for hetero–homo conversion in channel-domain for phase aberration correction in ultrasound imaging

Echo imaging in ultrasound computed tomography (USCT) using the synthetic aperture technique is performed with the assumption that the speed of sound is constant in the system. However, tissue heterogeneity causes a mismatch between the predicted arrival time and the actual arrival time of the echo signal, which will result in phase aberration, leading to the quality degradation of the reconstructed B-mode image. The conventional correction methods that use the correlation of each different channel require the presence of strong point scatterers and involve the problem of local solutions due to excessive correction. In this study, we propose a novel approach to correcting the signal distortion due to sound speed heterogeneity using a deep neural network (DNN). The DNN was trained to convert the distorted radio frequency (RF) inputs for the heterogeneous medium to the distortion-free RF outputs for the homogeneous medium. The network with U-net architecture using ResNet-34 as a backbone was trained using the hetero–homo corresponding channel-domain RF data generated via numerical simulations. The trained network performed phase aberration correction in the channel-domain RF, with the B-mode images reconstructed with the corrected RF demonstrating a higher contrast and an improved resolution compared with uncorrected cases. It was also demonstrated that the DNN model is robust to both varied reflection intensities and varied sound speed heterogeneities. The successful results demonstrated that the proposed DNN-based method is effective for phase aberration correction in US imaging.

3-D H-scan ultrasound imaging of relative scatterer size using a matrix array transducer and sparse random aperture compounding

H-scan ultrasound (US) is a high-resolution imaging technique for soft tissue characterization. By acquiring data in volume space, H-scan US can provide insight into subtle tissue changes or heterogenous patterns that might be missed using traditional cross-sectional US imaging approaches. In this study, we introduce a 3-dimensional (3-D) H-scan US imaging technology for voxel-level tissue characterization in simulation and experimentation. Using a matrix array transducer, H-scan US imaging was developed to evaluate the relative size of US scattering aggregates in volume space. Experimental data was acquired using a programmable US system (Vantage 256, Verasonics Inc, Kirkland, WA) equipped with a 1024-element (32 × 32) matrix array transducer (Vermon Inc, Tours, France). Imaging was performed using the full array in transmission. Radiofrequency (RF) data sequences were collected using a sparse random aperture compounding technique with 6 different data compounding approaches. Plane wave imaging at five angles was performed at a center frequency of 8 MHz. Scan conversion and attenuation correction were applied. To generate the 3-D H-scan US images, a convolution filter bank (N = 256) was then used to process the RF data sequences and measure the spectral content of the backscattered US signals before volume reconstruction. Preliminary experimental studies were conducted using homogeneous phantom materials embedded with spherical US scatterers of varying diameter, i.e., 27 to 45, 63 to 75, or 106–126 μm. Both simulated and experimental results revealed that 3-D H-scan US images have a low spatial variance when tested with homogeneous phantom materials. Furthermore, H-scan US is considerably more sensitive than traditional B-mode US imaging for differentiating US scatterers of varying size (p = 0.001 and p = 0.93, respectively). Overall, this study demonstrates the feasibility of 3-D H-scan US imaging using a matrix array transducer for tissue characterization in volume space.

Spatiotemporal singular value decomposition for denoising in photoacoustic imaging with a low-energy excitation light source.

Photoacoustic (PA) imaging is an emerging hybrid imaging modality that combines rich optical spectroscopic contrast and high ultrasonic resolution, and thus holds tremendous promise for a wide range of pre-clinical and clinical applications. Compact and affordable light sources such as light-emitting diodes (LEDs) and laser diodes (LDs) are promising alternatives to bulky and expensive solid-state laser systems that are commonly used as PA light sources. These could accelerate the clinical translation of PA technology. However, PA signals generated with these light sources are readily degraded by noise due to the low optical fluence, leading to decreased signal-to-noise ratio (SNR) in PA images. In this work, a spatiotemporal singular value decomposition (SVD) based PA denoising method was investigated for these light sources that usually have low fluence and high repetition rates. The proposed method leverages both spatial and temporal correlations between radiofrequency (RF) data frames. Validation was performed on simulations and in vivo PA data acquired from human fingers (2D) and forearm (3D) using a LED-based system. Spatiotemporal SVD greatly enhanced the PA signals of blood vessels corrupted by noise while preserving a high temporal resolution to slow motions, improving the SNR of in vivo PA images by 90.3%, 56.0%, and 187.4% compared to single frame-based wavelet denoising, averaging across 200 frames, and single frame without denoising, respectively. With a fast processing time of SVD (∼50 µs per frame), the proposed method is well suited to PA imaging systems with low-energy excitation light sources for real-time in vivo applications.

4 days in dry immersion increases arterial wall response to ultrasound wave as measured using radio-frequency signal, comparison with spaceflight data.

Recent studies have reported a significant increase in common carotid artery (CCA) intima media thickness, wall stiffness and reflectivity to ultrasound, in astronauts, after six months of spaceflight. The hypothesis was that 4 days in dry immersion (subjects under bags of water) will be sufficient to change the CCA wall reflectivity to ultrasound similar to what observed after spaceflight. Such response would be quantified using the amplitude of the ultrasound signal returned to the probe by the target concerned. [coefficient of signal return (Rs)]. The Rs for anterior and posterior CCA wall, sternocleidomastoid muscle, intima layer and CCA lumen were calculated from the ultrasound radio frequency (RF) data displayed along each echographic line. After four days of DI, Rs increased in the CCA posterior wall (+15% +/- 10 from pre DI, p < 0.05), while no significant change was observed in the other targets. The observed increase in Rs with DI was approximately half compared to what was observed after six months of space flight (+34% +/- 14). This difference may be explained by dose response (dry immersion only four days in duration). As a marker of tissue-level physical changes, Rs provide complimentary information alongside previously observed CCA wall thickness and stiffness.

Abstract 12894: Murine Arterial Wall Strain Estimation With Ultrasound-Based Lagrangian Carotid Strain Imaging and Validation Against Histological Findings

Introduction: Early identification of arterial injury by non-invasive imaging can facilitate research into mechanisms of arterial injury and has the potential to improve patient management. We tested the hypothesis that Lagrangian carotid strain imaging (LCSI) combined with adaptive Bayesian regularization could detect early arterial injury and monitor atherogenesis in a murine model by associating findings with histological examinations. Methods: Radio-frequency (RF) data were collected from the right and left common carotid arteries (CCA) of 10 (5 male and 5 female) ApoE tm1Unc/J mice at 6, 16 and 24 weeks using a high frequency ultrasound system. Cumulative axial, lateral and shear strain images were estimated using LCSI and 3 strain indices: maximum accumulated strain index (MASI), peak mean strain of full ROI index (PMSRI) and strain at peak axial displacement index (SPADI). Mice were euthanized (n=2 at 6 and 16 weeks, n=6 at 24 weeks) for histological examination. Results: We observed sex-specific differences in the temporal variation of strain indices. For male mice, axial PMSRI and SPADI changed significantly from 6 to 24 weeks (mean axial SPADI at 6 = 14.10±6.30 % and 24 weeks = 0.24±6.22%; p=0.0004). For female mice, lateral MASI was significantly higher at 24 than 6 weeks (mean lateral MASI at 6 = 10.26±3.13 % and 24 weeks = 16.42±7.15% p=0.048). Both sexes had strong associations with histological findings. For males, axial PMSRI and SPADI showed significant correlation with the number of elastin fibers (axial PMSRI: r s =0.83, p=0.01; axial SPADI: r s =0.76, p=0.02). For females, lateral and shear MASI had significant correlations with plaque scores (lateral MASI: r s =0.66, p=0.04; shear MASI: r s 0.77, p =0.009). Conclusions: LCSI can quantify carotid arterial wall strain in a murine model non-invasively. Changes in strain were associated with changes in arterial wall microstructure after plaque formation.

Improving breast cancer diagnosis by incorporating raw ultrasound parameters into machine learning

The improved diagnostic accuracy of ultrasound breast examinations remains an important goal. In this study, we propose a biophysical feature-based machine learning method for breast cancer detection to improve the performance beyond a benchmark deep learning algorithm and to furthermore provide a color overlay visual map of the probability of malignancy within a lesion. This overall framework is termed disease-specific imaging. Previously, 150 breast lesions were segmented and classified utilizing a modified fully convolutional network and a modified GoogLeNet, respectively. In this study multiparametric analysis was performed within the contoured lesions. Features were extracted from ultrasound radiofrequency, envelope, and log-compressed data based on biophysical and morphological models. The support vector machine with a Gaussian kernel constructed a nonlinear hyperplane, and we calculated the distance between the hyperplane and each feature’s data point in multiparametric space. The distance can quantitatively assess a lesion and suggest the probability of malignancy that is color-coded and overlaid onto B-mode images. Training and evaluation were performed on in vivo patient data. The overall accuracy for the most common types and sizes of breast lesions in our study exceeded 98.0% for classification and 0.98 for an area under the receiver operating characteristic curve, which is more precise than the performance of radiologists and a deep learning system. Further, the correlation between the probability and Breast Imaging Reporting and Data System enables a quantitative guideline to predict breast cancer. Therefore, we anticipate that the proposed framework can help radiologists achieve more accurate and convenient breast cancer classification and detection.

In Vivo Comparison of Pulse Wave Velocity Estimation Based on Ultrafast Plane Wave Imaging and High-Frame-Rate Focused Transmissions.

Ultrasound-based local pulse wave velocity (PWV) estimation, as a measure of arterial stiffness, can be based on fast focused imaging (FFI) or plane wave imaging (PWI). This study was aimed at comparing the accuracy of in vivo PWV estimation using FFI and PWI. Ultrasound radiofrequency data of carotid arteries were acquired in 14 healthy volunteers (25-57 y) by executing the FFI (12 lines, 7200 Hz) and PWI (128 lines, 2000 Hz) methods consecutively. PWV was derived at two time-reference points, dicrotic notch (DN) and systolic foot (SF), for multiple pressure cycles by fitting a linear function through the positions of the peaks of low-pass filtered wall acceleration curves as a function of time. The accuracy of PWV estimation was determined for various cutoff frequencies (10-200 Hz). No statistically significant difference was observed between PWVs estimated by both approaches. The PWV and R2 at DN were higher, on average, than those at SF (PWV/R2: FFI SF 5.5/0.92, FFI DN 6.1/0.92; PWI SF 5.4/0.89, PWI DN 6.3/0.95). The use of cutoff frequencies between 40 and 80Hz provided the most accurate PWVs. Both methods seemed equally suitable for use in clinical practice, although we have a preference for the PWV at DN given the higher R2 values.

3D controlled-source electromagnetic inversion in the radio-frequency band

The classical radio-magnetotelluric (RMT) method is nowadays routinely applied to various environmental, engineering, and exploration problems. The technique uses radio transmitters broadcasting in the frequency range of 10 kHz to 1 MHz, and the measurements are carried out in the far field. The well-known disadvantages of RMT are a lack of robust radio transmitters in remote areas; the absence of transmitters broadcasting below 10 kHz, which limits the penetration depth; and a possible low signal-to-noise ratio. To overcome these difficulties, controlled sources can be used (controlled-source RMT [CSRMT]). We extend the CSRMT method to perform measurements not only in the far field but also in the transition zone. In CSRMT practice, it often is challenging to maintain far-field conditions for logistical reasons. Therefore, part of the measured data contains signatures of the source field, which cannot be interpreted with magnetotelluric software. In addition, the source placed directly in the survey area allows us to increase the signal-to-noise ratio and resolution. Such CSRMT in the transition zone is, in fact, a controlled-source electromagnetic method but with full impedance tensor and tipper vector transfer functions. We develop new procedures for the 3D modeling and inversion of the tensor radio-frequency data measured in the transition zone of two perpendicular horizontal electric dipole sources. In this case, the geometry of the source must be considered in the forward modeling. The developed modeling and inversion software is tested on a synthetic 3D model. The 3D resistivity models derived from the real data confirm the geologic settings and are consistent with the available borehole information. Therefore, we conclude that the CSRMT approach extended to include the source field is feasible and that the developed procedures are reliable.

Redefining food safety traceability system through blockchain: findings, challenges and open issues.

In the last few decades, there has been an increase in food safety and traceability issues. To prevent accidents and misconduct, it became essential to establish Food Safety Traceability System (FSTS) to trace the food from producer to consumer. The traceability systems can help track food in supply chains from farms to retail. Numerous technologies such as Radio Frequency Identification (RFID), sensor networks, and data mining have been integrated into traditional food supply chain systems to remove unsafe food products from the chain. But, these are not adequate for the current supply chain market. The emerging technology of blockchain can overcome safety and tracking issues. This can be possible with the help of blockchain features like transparent, decentralized, distributed, and immutable. Most of the previous works missed the discussion of the systematic process and technology involved in implementing the FSTS using blockchain. In this paper, we have discussed an organized state of research of the existing FSTS using blockchain. This survey paper aims to outline a detailed analysis of blockchain technology, FSTS using blockchain, consensus algorithms, security attacks, and solutions. Several survey papers and solutions based on blockchain are included in this research paper. Also, this work discusses some of the open research issues related to FSTS.

Spectral analysis of ultrasound backscatter for non-invasive measurement of plaque composition

Carotid atherosclerotic plaque composition may be an important indication of patient risk for future cerebrovascular events. Ultrasound spectral analysis has the potential to provide a robust measure of plaque composition in vivo if the backscatter transfer function can be sufficiently isolated from the effects of attenuation from overlying tissue, receive and transmit transfer functions from the ultrasound system and transducer, and diffraction. This study examines the usefulness of the nonlinearly generated second harmonic portion of the backscatter signal and the effects of a variety of attenuation compensation techniques for noninvasively characterizing human carotid plaque using spectral analysis and machine learning. Post-beamformed ultrasound backscatter radiofrequency (RF) data were acquired from 6 normal subjects and 119 carotid endarterectomy patients prior to surgery. Plaque obtained following surgery was histologically processed, and regions of interest (ROI) corresponding to homogenous tissue types (fibrous/fibro-lipidic, hemorrhagic and/or necrotic core and calcified) were selected from RF data. Both the harmonic and fundamental power spectra for each ROI was obtained and normalized by data from a uniform phantom (0.5 dB/cm-MHz slope of attenuation). Additional attenuation compensation approaches were compared to simply using the reference phantom: (1) optimum power spectral shift estimation, (2) one-step adventitial, or (3) two-step adventitial. Spectral parameters extracted from both the fundamental and harmonic estimates of the backscatter transfer function of 363 ROI’s from 152 plaque specimens were used to train and test random forest and support vector machine classification models. The best results came from using spectral parameters derived from both the fundamental and second harmonic bands with a predictive accuracy of 65–68%, kappa statistic of 0.49–0.54, and accuracies of 84% for fibrous, 68–74% for hemorrhagic and/or necrotic core, and 78–81% for calcified ROI’s. The result indicated that the nonlinearly generated second harmonic portion of backscatter is useful for carotid plaque tissue characterization and that a reference phantom approach with a 0.5 dB/cm-MHz slope of attenuation works as well as more complicated approaches.

Evaluation of Takayasu's arteritis activity by ultrasound localization microscopy

Abstract Objectives Use of sulphur hexafluoride microbubbles (MB) with ultrasound can provide arterial wall enhancement. The use of ultrafast imaging with MB super-localization offers the possibility of ultrasound localization microscopy (ULM) and thus a visualization of in vivo carotid vasa vasorum. We aim to perform in vivo Vasa vasorum imaging in the carotid wall in Takayasu's arteritis and to provide a correlation with the disease's activity. Methods Patients with Takayasu arteritis were consecutively included in the national referral centre. Assessment of activity was performed by NIH activity index, with measurement of biological inflammation, morphological evaluation by computed tomography (CT) angiography, and 18-fluorodeoxyglucose CT scan. Ultrafast ultrasound imaging was performed with a 7 MHz central frequency linear probe. A dedicated contrast imaging sequence (plane waves with 8 angles, frame rate 500 Hz, voltage 8 V, duration of 8 s), coupled with the intravenous injection of sulphur hexafluoride MB allow us to store the raw radio frequency data of the MB passage. After a pre-processing step to accumulate frames properly including a frames selection and a movement correction, the MB were localized then tracked, allowing reconstruction of the vasa vasorum. The inflammation was quantified through MB tracked and normalized for inter acquisition comparison. Results 16 patients were included. 5 patients (median age 35.8 [24.5–46.0] years, 3 women) had active disease and 11 patients (37.2 [31.7–47.3] years, 9 women) had quiescent disease. The passage of MB allowed the visualisation of microvessels within the arterial wall for active cases. The number of MB detected per second in the wall was 118 [80–169] for active cases vs. 13 [10–15] for quiescent cases (p=0.0005). Conclusion ULM allows visualization of microvessels within the carotid wall, with significantly greater MB passage in active Takayasu arteritis. ULM provides for the first time a precise visualization in vivo of the vasa vasorum and gives access to quantification of the vascularity of the arterial wall. Funding Acknowledgement Type of funding sources: Foundation. Main funding source(s): Société française de cardiologie

Comparative Study of Raw Ultrasound Data Representations in Deep Learning to Classify Hepatic Steatosis

Adiposity accumulation in the liver is an early-stage indicator of non-alcoholic fatty liver disease. Analysis of ultrasound (US) backscatter echoes from liver parenchyma with deep learning (DL) may offer an affordable alternative for hepatic steatosis staging. The aim of this work was to compare DL classification scores for liver steatosis using different data representations constructed from raw US data. Steatosis in N=31 patients with confirmed or suspected non-alcoholic fatty liver disease was stratified based on fat-fraction cutoff values using magnetic resonance imaging as a reference standard. US radiofrequency (RF) frames (raw data) and clinical B-mode images were acquired. Intermediate image formation stages were modeled from RF data. Power spectrum representations and phase representations were also calculated. Co-registered patches were used to independently train 1-, 2- and 3-D convolutional neural networks (CNNs), and classifications scores were compared with cross-validation. There were 67,800 patches available for 2-D/3-D classification and 1,830,600 patches for 1-D classification. The results were also compared with radiologist B-mode annotations and quantitative ultrasound (QUS) metrics. Patch classification scores (area under the receiver operating characteristic curve [AUROC]) revealed significant reductions along successive stages of the image formation process (p &lt; 0.001). Patient AUROCs were 0.994 for RF data and 0.938 for clinical B-mode images. For all image formation stages, 2-D CNNs revealed higher patch and patient AUROCs than 1-D CNNs. CNNs trained with power spectrum representations converged faster than those trained with RF data. Phase information, which is usually discarded in the image formation process, provided a patient AUROC of 0.988. DL models trained with RF and power spectrum data (AUROC=0.998) provided higher scores than conventional QUS metrics and multiparametric combinations thereof (AUROC=0.986). Radiologist annotations indicated lower hepatic steatosis classification accuracies (Acc=0.914) with respect to magnetic resonance imaging proton density fat fraction that DL models (Acc=0.989). Access to raw ultrasound data combined with artificial intelligence techniques may offer superior opportunities for quantitative tissue diagnostics than conventional sonographic images.