Littérature scientifique sur le sujet « ULA-OP 256 »

Methods of increasing complexity are currently being proposed for ultrasound (US) echographic signal processing. Graphics Processing Unit (GPU) resources allowing massive exploitation of parallel computing are ideal candidates for these tasks. Many high-performance US instruments, including open scanners like ULA-OP 256, have an architecture based only on Field-Programmable Gate Arrays (FPGAs) and/or Digital Signal Processors (DSPs). This paper proposes the implementation of the embedded NVIDIA Jetson Xavier AGX module on board ULA-OP 256. The system architecture was revised to allow the introduction of a new Peripheral Component Interconnect Express (PCIe) communication channel, while maintaining backward compatibility with all other embedded computing resources already on board. Moreover, the Input/Output (I/O) peripherals of the module make the ultrasound system independent, freeing the user from the need to use an external controlling PC.

The spread of high frame rate and 3-D imaging techniques has raised pressing requirements for ultrasound systems. In particular, the processing power and data transfer rate requirements may be so demanding to hinder the real-time (RT) implementation of such techniques. This paper first analyzes the general requirements involved in RT ultrasound systems. Then, it identifies the main bottlenecks in the receiving section of a specific RT scanner, the ULA-OP 256, which is one of the most powerful available open scanners and may therefore be assumed as a reference. This case study has evidenced that the “star” topology, used to digitally interconnect the system’s boards, may easily saturate the data transfer bandwidth, thus impacting the achievable frame/volume rates in RT. The architecture of the digital scanner was exploited to tackle the bottlenecks by enabling a new “ring“ communication topology. Experimental 2-D and 3-D high-frame-rate imaging tests were conducted to evaluate the frame rates achievable with both interconnection modalities. It is shown that the ring topology enables up to 4400 frames/s and 510 volumes/s, with mean increments of +230% (up to +620%) compared to the star topology.

The assessment of the velocity of blood flowing in the carotid, in modern clinical practice, represents an important exam performed both in emergency situations and as part of scheduled screenings. It is typically performed by an expert sonographer who operates a complex and costly clinical echograph. Unfortunately, in developing countries, in rural areas, and even in crowded modern cities, the access to this exam can be limited by the lack of suitable personnel and ultrasound equipment. The recent availability of low-cost, handheld devices has contributed to solving part of the problem, but a wide access to the exam is still hampered by the lack of expert sonographers. In this work, an automated procedure is presented with the hope that, in the near future, it can be integrated into a low-cost, handheld instrument that is also suitable for self-measurement, for example, as can be done today with the finger oximeter. The operator should only place the probe on the neck, transversally with respect to the common tract of the carotid. The system, in real-time, automatically locates the vessel lumen, places the sample volume, and performs an angle-corrected velocity measurement of the common carotid artery peak velocity. In this study, the method was implemented for testing on the ULA-OP 256 scanner. Experiments on flow phantoms and volunteers show a performance in sample volume placement similar to that achieved by expert operators, and an accuracy and repeatability of 3.2% and 4.5%, respectively.

Abstract Background As many cardiac diseases do not lead to symptoms under resting conditions, subjecting the patient to a stress (e.g. exercise) test is clinical routine to reveal disease. Quantification of cardiac function using speckle tracking echocardiography (STE) during such stress test remains a challenge not only as overall image quality typically drops but also as the heart rate increases which results in an unfavourable heart-rate-to-frame-rate ratio as required for robust tracking. Moreover, conventional STE runs at relatively low frame rate thereby under-resolving peak strain rate (SR) values – which is particularly true at high heart rates – while this biomarker has been shown to better disclose stress-induced disease than strain using Tissue Doppler Imaging (TDI)-based strain methodologies. Purpose The aim of this study was therefore to test whether HFR STE could reveal the strain / SR response described previously in the TDI literature. Methods Stress echocardiography was performed in 25 healthy volunteers at four different stress levels (baseline, 25%, 50% and 66% of maximum workload) using a supine bike. Apical 4-chamber views were recorded using the ULA-OP 256 experimental scanner running a HFR sequence based on the compounding of 6 diverging wave transmits to achieve a frame rate of 833Hz. A myocardial contour was manually drawn on the reconstructed images and tracked during the cardiac cycle by a custom-made 2D HFR STE algorithm (DOI: 10.1109/TUFFC.2020.2985451). Then, strain and strain rate curves were computed from which systolic (S) and diastolic (early (E) and late (A)) peak values, as well as the short-lived isovolumic relaxation peak (IVR) were extracted. Finally, these values were compared amongst the different stress levels using a single factor ANOVA. Results Some datasets had to be discarded as the contour was not properly tracked at a visual assessment of an expert cardiologist. Tracking was feasible in 92%, 98%, 80% and 64% of cases, for baseline, 25%, 50% and 66% of maximum workload, respectively. The decreasing feasibility with exercise level is in line with what is seen clinically, as volunteers' breathing is faster and heavier with increasing exercise level making the imaging more challenging. The extracted strain and SR curves showed a physiological pattern (Fig. 1a). As shown in Fig. 1b, the global systolic strain response was biphasic, showing a significant increase at low stress level but then reaching a plateau. In contrast, all SR indices linearly increased at each stress level. Conclusion Strain rate clinical markers extracted with HFR STE are concordant with what was reported in TDI literature. These findings show that HFR STE allows to assess cardiac function adequately during stress echocardiography. Funding Acknowledgement Type of funding sources: Foundation. Main funding source(s): FWO - Research Foundation Flanders

In the last decades, ultrasound imaging systems have become more and more popular thanks to their capability to investigate tissues in safe, cost effective, and non-invasive way. Their role in diagnostic imaging has become fundamental in several medical specialties, thanks also to the introduction of advanced echographic systems fostered by the efforts of several research laboratories around the world. Such efforts are more frequently based on the use of special research scanners, characterized by flexible hardware and programmable software and firmware. These features have been demonstrated ideal for the implementation and test of new methods, such as high frame-rate (HFR) imaging, color flow imaging (CFI), vector Doppler imaging, and 3 D imaging. Especially HFR and 3 D imaging have recently attracted great interest, but they are technically demanding since they involve either the formation of thousands of images per second, or the use of 2-D probes having a large number of elements. Therefore, great challenges must be faced for effective real-time implementation of 3 D and HFR imaging methods. My PhD activity aimed to implement and test advanced 2 D and 3 D ultrasound imaging modalities on an open research scanner called ULA OP 256. In the first part of my work, a new ultrasound imaging modality called Virtual real-time (VRT) was introduced through the modification of the firmware and software of the research scanner ULA-OP 256. With this modality, during a real-time (RT) investigation, the scanner initially acquires and stores in its memory up to 20 s of raw echo data. On user demand, the scanner can be switched to VRT mode: the stored data are re processed by the same resources used in RT but at different (typically lower) rates and, possibly, with different processing algorithms and parameters. In this way, contingent difficulties of image interpretation (especially in presence of rapidly moving phenomena), or possible computational limitations imposed by hardware during continuous RT processing can be overcome. The VRT modality has been demonstrated useful in different applications, for example, to implement a high-PRF version of the Multiline vector Doppler (MLVD) method, and a High- rame-rate CFI method, characterized by enhanced temporal and spatial resolution. The second part of my work included the software upgrade of ULA OP 256; it enabled the use of 2 D probes and the implementation of 3 D scanning methods. The ULA OP 256 can now be coupled to 2 D probes with arbitrary geometries, including matrix and sparse arrays. Furthermore, the scanner is now capable of simultaneously imaging multiple planes with programmable rotational angles. Novel approaches based on a sparse spiral array probe have been implemented and tested for different applications. For example, bi-plane imaging was evaluated for robust flow mediated dilation exams. Real-time 3 D spectral Doppler analysis was also performed. Here, two planes with programmable rotational angles were scanned to produce corresponding B-Mode images, over which multiple Doppler lines could be arbitrarily set to obtain the relative Multigate spectral Doppler (MSD) profiles. Finally, the last part of my work was specifically dedicated to the technical problems involved by HFR 3 D imaging. The management of (several) hundreds of transducer elements of a 2 D probe yields a huge amount of echo data: this makes complex and computationally expensive the processing of data volumes including thousands of lines, especially if performed at HFR. As a case report, the requirements of the main processing stages involved in ULA OP 256 receiver have been thoroughly investigated to detect and, possibly, solve the main bottlenecks. The study has evidenced that the star architecture that digitally interconnects the eight front-end boards of ULA OP 256 may frequently encounter data transfer bandwidth saturation that limits the overall performance in terms of frame/volume rate. A new architecture for data transfer has been proposed and shown effective to reduce the bandwidth requirements and thus, increase the performance of the scanner.