Improving Lateral Resolution in 3-D Imaging With Micro-beamforming Through Adaptive Beamforming by Deep Learning.

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.

Two-Stage Beamforming for Phased Array Imaging Using the Fast Hankel Transform.

An ultrasound scan generates a huge amount of data. To form an image, this data has to be transferred to the imaging system. This is an issue for applications where the data transfer capacity is limited such as hand-held systems, wireless probes, and miniaturized array probes. Two-stage beamforming methods can be used to significantly reduce the data transfer requirements. In the first stage, which is applied in-probe, the amount of data is reduced from channel to scanline data. In the imaging system, the data are then beamformed to obtain images, which are synthetically focused over the entire image. Currently, two approaches exist for the second stage. The first approach is a time-of-flight (TOF) approach called synthetic aperture sequential beamforming (SASB), which has been developed for both linear and phased arrays. SASB does, however, introduce artifacts in the image that can be reduced by tapering the first-stage scanlines at the cost of lateral resolution. The second approach is based on the wave equation, but a computationally efficient method for phased arrays that is producing sector scan data is lacking. Here, we propose an algorithm that uses the fast Hankel transform to obtain a fast algorithm. The imaging performance of this method is evaluated with simulations and experiments. Compared with PSASB, which is an adaption of SASB for phased arrays, our method requires a similar amount of operations to construct the entire image and there is no tradeoff between resolution and artifacts. These results show the advantage of using the wave equation instead of a TOF approach.

Acoustic characterization of a miniature matrix transducer for pediatric 3D transesophageal echocardiography.

This paper presents the design, fabrication and characterization of a miniature PZT-on-CMOS matrix transducer for real-time pediatric 3-dimensional (3D) transesophageal echocardiography (TEE). This 3D TEE probe consists of a 32 × 32 array of PZT elements integrated on top of an Application Specific Integrated Circuit (ASIC). We propose a partitioned transmit/receive array architecture wherein the 8 × 8 transmitter elements, located at the centre of the array, are directly wired out and the remaining receive elements are grouped into 96 sub-arrays of 3 × 3 elements. The echoes received by these sub-groups are locally processed by micro-beamformer circuits in the ASIC that allow pre-steering up to ±37°. The PZT-on-CMOS matrix transducer has been characterized acoustically and has a centre frequency of 5.8 MHz, -6 dB bandwidth of 67%, a transmit efficiency of 6 kPa/V at 30 mm, and a receive dynamic range of 85 dB with minimum and maximum detectable pressures of 5 Pa and 84 kPa respectively. The properties are very suitable for a miniature pediatric real-time 3D TEE probe.

Fast Volumetric Imaging Using a Matrix Transesophageal Echocardiography Probe with Partitioned Transmit-Receive Array.

We describe a 3-D multiline parallel beamforming scheme for real-time volumetric ultrasound imaging using a prototype matrix transesophageal echocardiography probe with diagonally diced elements and separated transmit and receive arrays. The elements in the smaller rectangular transmit array are directly wired to the ultrasound system. The elements of the larger square receive aperture are grouped in 4 × 4-element sub-arrays by micro-beamforming in an application-specific integrated circuit. We propose a beamforming sequence with 85 transmit-receive events that exhibits good performance for a volume sector of 60°â€¯× 60°. The beamforming is validated using Field II simulations, phantom measurements and in vivo imaging. The proposed parallel beamforming achieves volume rates up to 59 Hz and produces good-quality images by angle-weighted combination of overlapping sub-volumes. Point spread function, contrast ratio and contrast-to-noise ratio in the phantom experiment closely match those of the simulation. In vivo 3-D imaging at 22-Hz volume rate in a healthy adult pig clearly visualized the cardiac structures, including valve motion.

Dual stage beamforming in the absence of front-end receive focusing.

Ultrasound front-end receive designs for miniature, wireless, and/or matrix transducers can be simplified considerably by direct-element summation in receive. In this paper we develop a dual-stage beamforming technique that is able to produce a high-quality image from scanlines that are produced with focused transmit, and simple summation in receive (no delays). We call this non-delayed sequential beamforming (NDSB). In the first stage, low-resolution RF scanlines are formed by simple summation of element signals from a running sub-aperture. In the second stage, delay-and-sum beamforming is performed in which the delays are calculated considering the transmit focal points as virtual sources emitting spherical waves, and the sub-apertures as large unfocused receive elements. The NDSB method is validated with simulations in Field II. For experimental validation, RF channel data were acquired with a commercial research scanner using a 5 MHz linear array, and were subsequently processed offline. For NDSB, good average lateral resolution (0.99 mm) and low grating lobe levels (<-40 dB) were achieved by choosing the transmit [Formula: see text] as 0.75 and the transmit focus at 15 mm. NDSB was compared with conventional dynamic receive focusing (DRF) and synthetic aperture sequential beamforming (SASB) with their own respective optimal settings. The full width at half maximum of the NDSB point spread function was on average 20% smaller than that of DRF except for at depths <30 mm and 10% larger than SASB considering all the depths. NDSB showed only a minor degradation in contrast-to-noise ratio and contrast ratio compared to DRF and SASB when measured on an anechoic cyst embedded in a tissue-mimicking phantom. In conclusion, using simple receive electronics front-end, NDSB can attain an image quality better than DRF and slightly inferior to SASB.

Viability of endothelial cells after ultrasound-mediated sonoporation: Influence of targeting, oscillation, and displacement of microbubbles.

Microbubbles (MBs) have been shown to create transient or lethal pores in cell membranes under the influence of ultrasound, known as ultrasound-mediated sonoporation. Several studies have reported enhanced drug delivery or local cell death induced by MBs that are either targeted to a specific biomarker (targeted microbubbles, tMBs) or that are not targeted (non-targeted microbubbles, ntMBs). However, both the exact mechanism and the optimal acoustic settings for sonoporation are still unknown. In this study we used real-time uptake patterns of propidium iodide, a fluorescent cell impermeable model drug, as a measure for sonoporation. Combined with high-speed optical recordings of MB displacement and ultra-high-speed recordings of MB oscillation, we aimed to identify differences in MB behavior responsible for either viable sonoporation or cell death. We compared ntMBs and tMBs with identical shell compositions exposed to long acoustic pulses (500-50,000cycles) at various pressures (150-500kPa). Propidium iodide uptake highly correlated with cell viability; when the fluorescence intensity still increased 120s after opening of the pore, this resulted in cell death. Higher acoustic pressures and longer cycles resulted in more displacing MBs and enhanced sonoporation. Non-displacing MBs were found to be the main contributor to cell death, while displacement of tMBs enhanced reversible sonoporation and preserved cell viability. Consequently, each therapeutic application requires different settings: non-displacing ntMBs or tMBs are advantageous for therapies requiring cell death, especially at 500kPa and 50,000cycles, whereas short acoustic pulses causing limited displacement should be used for drug delivery.