Objective: Described here is the development of an ultrasound matrix transducer prototype for high-frame-rate 3-D intra-cardiac echocardiography. Methods: The matrix array consists of 16 × 18 lead zirconate titanate elements with a pitch of 160 µm × 160 µm built on top of an application-specific integrated circuit that generates transmission signals and digitizes the received signals. To reduce the number of cables in the catheter to a feasible number, we implement subarray beamforming and digitization in receive and use a combination of time-division multiplexing and pulse amplitude modulation data transmission, achieving an 18-fold reduction. The proposed imaging scheme employs seven fan-shaped diverging transmit beams operating at a pulse repetition frequency of 7.7 kHz to obtain a high frame rate. The performance of the prototype is characterized, and its functionality is fully verified. Results: The transducer exhibits a transmit efficiency of 28 Pa/V at 5 cm per element and a bandwidth of 60% in transmission. In receive, a dynamic range of 80 dB is measured with a minimum detectable pressure of 10 Pa per element. The element yield of the prototype is 98%, indicating the efficacy of the manufacturing process. The transducer is capable of imaging at a frame rate of up to 1000 volumes/s and is intended to cover a volume of 70° × 70° × 10 cm. Conclusion: These advanced imaging capabilities have the potential to support complex interventional procedures and enable full-volumetric flow, tissue, and electromechanical wave tracking in the heart.

@en

High-frame-rate volumetric ultrasound imaging is highly desired to enable novel clinical ultrasound applications. However, realizing high-quality volumetric ultrasound imaging at a high frame rates (>500 Hz) is challenging. Keeping the cable count and data rate of the transducer device at a realistic level without sacrificing image quality to an undesirable extend means that a dedicated design with carefully chosen trade-offs is required and powerful processing of the received signals is desired. This thesis describes the development of a high-frame-rate 3D ultrasound transducer through dedicated transducer design and explores the use of deep learning-based beamforming to achieve high-quality 3D imaging. Specifically, the first part of this thesis focuses on the development of an imaging scheme and the realization and testing of two prototype transducers for high-frame-rate 3D intracardiac echography (3D-ICE). The second part of the thesis implements deep learning in the image reconstruction process to improve the image quality of volumetric ultrasound. Deep learning-based beamforming is implemented and evaluated first for a miniature matrix array, which similar to the 3D-ICE design applies micro-beamforming to achieve cable count reduction and finally for a spiral array which uses a sparse distribution of transducer channels. @en

Ultrasound imaging is an attractive imaging modality due to its low-cost and real-time feedback, although it often falls short in image quality compared to MRI and CT imaging. Conventional ultrasound image reconstruction, such as Delay-and-Sum beamforming, is derived from maximum-likelihood estimation. As such, no prior information is exploited in the image formation process, which limits potential image quality. Maximum-a-posteriori (MAP) beamforming aims to overcome this issue, but often relies on rough approximations of the underlying signal statistics. Deep learning based reconstruction methods have demonstrated impressive results over the past years, but often lack interpretability and require vast amounts of data.In this work we present a neural MAP beamforming technique, which efficiently combines deep learning in the MAP beamforming framework. We show that this model-based deep learning approach can achieve high-quality imaging, improving over the state-of-the-art, without compromising the real-time abilities of ultrasound imaging.@en

There is an increased desire for miniature ultrasound probes with small apertures to provide volumetric images at high frame rates for in-body applications. Satisfying these increased requirements makes simultaneous achievement of a good lateral resolution a challenge. As micro-beamforming is often employed to reduce data rate and cable count to acceptable levels, receive processing methods that try to improve spatial resolution will have to compensate the introduced reduction in focusing. Existing beamformers do not realize sufficient improvement and/or have a computational cost that prohibits their use. Here we propose the use of adaptive beamforming by deep learning (ABLE) in combination with training targets generated by a large aperture array, which inherently has better lateral resolution. In addition, we modify ABLE to extend its receptive field across multiple voxels. We illustrate that this method improves lateral resolution both quantitatively and qualitatively, such that image quality is improved compared with that achieved by existing delay-and-sum, coherence factor, filtered-delay-multiplication-and-sum and Eigen-based minimum variance beamformers. We found that only in silica data are required to train the network, making the method easily implementable in practice.

@en

Atrial fibrillation (AF) is the most common cardiac arrhythmia and is normally treated by RF ablation. Intracardiac echography (ICE) is widely employed during RF ablation procedures to guide the electrophysiologist in navigating the ablation catheter, although only 2-D probes are currently clinically used. A 3-D ICE catheter would not only improve visualization of the atrium and ablation catheter, but it might also provide the 3-D mapping of the electromechanical wave (EW) propagation pattern, which represents the mechanical response of cardiac tissue to electrical activity. The detection of this EW needs 3-D high-frame-rate imaging, which is generally only realizable in tradeoff with channel count and image quality. In this simulation-based study, we propose a high volume rate imaging scheme for a 3-D ICE probe design that employs 1-D micro-beamforming in the elevation direction. Such a probe can achieve a high frame rate while reducing the channel count sufficiently for realization in a 10-Fr catheter. To suppress the grating-lobe (GL) artifacts associated with micro-beamforming in the elevation direction, a limited number of fan-shaped beams with a wide azimuthal and narrow elevational opening angle are sequentially steered to insonify slices of the region of interest. An angular weighted averaging of reconstructed subvolumes further reduces the GL artifacts. We optimize the transmit beam divergence and central frequency based on the required image quality for EW imaging (EWI). Numerical simulation results show that a set of seven fan-shaped transmission beams can provide a frame rate of 1000 Hz and a sufficient spatial resolution to visualize the EW propagation on a large 3-D surface.

@en

3-D contrast enhanced ultrasound enables better visualization of inherently 3-D vascular geometries compared to an intersecting plane. Additionally, it would allow the application of motion correction techniques for all directions. Both contrast detection and motion correction work better on high-frame rate data. However high-frame rate 3-D ultrasound imaging with dense matrix arrays is challenging to realize. Sparse arrays alleviate some of the limitations in cable count and data rate that fully populated arrays encounter, but their increased level of secondary lobes negatively impacts image contrast. Meanwhile the use of unfocused transmit beams needed to achieve high-frame rates negatively impacts resolution. Here we propose to use adaptive beamforming by deep learning (ABLE) to improve the image quality of contrast enhanced ultrasound images acquired with a sparse spiral array. We train the neural network on simulated data and evaluate simulated images and in vivo images of an ex ovo chicken embryo. ABLE improved resolution compared to delay-and-sum (DAS) and spatial coherence (SC) beamforming on the simulated and in vivo data. The qualitative improvements persist after histogram matching, indicating that the image quality improvement of the ABLE images was not purely due to dynamic range stretching.

@en

In this letter, a compact high-voltage (HV) transmit circuit for dense 2-D transducer arrays used in 3-D ultrasonic imaging systems is presented. Stringent area requirements are addressed by a unipolar pulser with embedded transmit/receive switch. Combined with a capacitive HV level shifter, it forms the ultrasonic HV transmit circuit with the lowest reported HV transistor count and area without any static power consumption. The balanced latched-based level shifter implementation makes the design insensitive to transients on the HV supply caused by pulsing, facilitating application in probes with limited local supply decoupling, such as imaging catheters. Favorable scaling through resource sharing benefits massively arrayed architectures while preserving full individual functionality. A prototype of 8 × 9 elements was fabricated in the TSMC 0.18 μm HV BCD technology and a 160μm×160μm PZT transducer matrix is manufactured on the chip. The system is designed to drive 65-V peak-to-peak pulses on 2-pF transducer capacitance and hardware sharing of six elements allows for an area of only 0.008 mm2 per element. Electrical characterization as well as acoustic results obtained with the 6-MHz central frequency transducer are demonstrated.@en

Intra-cardiac echography (ICE) probes (Fig. 32.2.1) are widely used in electrophysiology for their good procedure guidance and relatively safe application. ASICs are increasingly employed in these miniature probes to enhance signal quality and reduce the number of connections needed in mm-diameter catheters [1]-[5]. 3D visualization in real-time is additionally enabled by 2D transducer arrays with, for each transducer element, a high-voltage (HV) transmit (TX) part, to generate acoustic pulses of sufficient pressure, and a receive (RX) path, to process the resulting echoes. To achieve the required reduction in RX channels, micro-beamforming (BF), which merges the signals from a subarray using a delay-and-sum operation, has been shown to be an effective solution [3], [4]. However, due to the frame-rate reduction that is associated with BF, these designs cannot serve emerging high-frame-rate imaging modes (1000 volumes/s) like 3D blood-flow and elastography imaging. In-probe digitization has recently been investigated to provide further channel-count reduction, make data transmission more robust, and enable pre-processing in the probe [1]-[3]. However, these earlier designs have either no TX functionality [2], [3] or only low-voltage (LV) TX [1] integrated. Combining BF and digitization with area-hungry HV transmitters in a pitch-matched scalable fashion while supporting high-frame-rate imaging remains an unmet challenge. The work presented in this paper meets this target, enabled by a hybrid ADC, the small die size of which allows for co-integration with 65V element-level pulsers.

@en

This work describes an ASIC design for high-frame-rate 3D intracardiac echocardiography probes. The chip is the first to combine element-level high-voltage pulsers and time-gain-compensation analog frontends as well as subarray beamformers and in-probe digitization in a pitch-matched fashion. The integration challenge is met by a shared hybrid beamforming ADC with the highest reported area and power efficiency. The achieved beamformer size of three elements enables acquisition at 1000 volumes/s while, in combination with a custom datalink, still providing sufficient channel-count reduction for catheter integration. @en

The image quality of ultrasound localization microscopy (ULM) images is driven by the ability to accurately detect and track the location of microbubbles (MBs) in vascular networks. This task becomes increasingly challenging in imaging environments with high MB concentrations and low signal-to-noise ratios, making it difficult to differentiate and localize individual MBs. Recent developments in deep learning (DL) have demonstrated significant improvements over conventional methods but depend on vast amounts of realistic training data with the corresponding ground truth labels, which are difficult to obtain. The alternative, simulated data, in turn, poses challenges in generalizability of the method. In this work, we present a hybrid pipeline for ULM that comprises data generation, localization, and tracking. It combines the current state-of-the-art, utilizing both conventional and DL techniques. We show that using this approach, we can create high-quality velocity maps while being able to generalize well across different domains.

@en

Effective treatment of movement disorders requires thorough understanding of human limb control. Joint dynamics can be assessed using robotic manipulators and system identification. Due to tendon compliance, joint angle and muscle length are not proportional. This study uses plane-wave ultrasound imaging to investigate the dynamic relation between ankle joint angle and muscle fiber stretch. The first goal is to determine the feasibility of using ultrasound imaging with system identification; the second goal is to assess the relation between ankle angle, muscle stretch, and reflex size. Soleus and gastrocnemius muscle stretches were assessed with ultrasound imaging and image tracking. For small (1° SD) continuous motions, muscle stretch was proportional to ankle angle during a relax task, but images were too noisy to make that assessment during an active position task. For transient perturbations with high velocity (> 90°/s) the muscle length showed oscillations that were not present in the ankle angle, demonstrating a non-proportional relationship and muscle-tendon interaction. The gastrocnemius velocity predicted the size of the short-latency reflex better than the ankle angle velocity. Concluding, plane-wave ultrasound muscle imaging is feasible for system identification experiments and shows that muscle length and ankle angle are proportional during a relax task with small continuous perturbations.@en