Characterization of an Array-Based Dual-Frequency Transducer for Superharmonic Contrast Imaging

Abstract

Superharmonic imaging with dual-frequency imaging systems uses conventional low-frequency ultrasound transducers on transmit, and high-frequency transducers on receive to detect higher order harmonic signals from microbubble contrast agents, enabling high-contrast imaging while suppressing clutter from background tissues. Current dual-frequency imaging systems for superharmonic imaging have been used for visualizing tumor microvasculature, with single-element transducers for each of the low- and high-frequency components. However, the useful field of view is limited by the fixed focus of single-element transducers, while image frame rates are limited by the mechanical translation of the transducers. In this article, we introduce an array-based dual-frequency transducer, with low-frequency and high-frequency arrays integrated within the probe head, to overcome the limitations of single-channel dual-frequency probes. The purpose of this study is to evaluate the line-by-line high-frequency imaging and superharmonic imaging capabilities of the array-based dual-frequency probe for acoustic angiography applications in vitro and in vivo. We report center frequencies of 1.86 MHz and 20.3 MHz with −6 dB bandwidths of 1.2 MHz (1.2–2.4 MHz) and 14.5 MHz (13.3–27.8 MHz) for the low- and high-frequency arrays, respectively. With the proposed beamforming schemes, excitation pressure was found to range from 336 to 458 kPa at its azimuthal foci. This was sufficient to induce nonlinear scattering from microbubble contrast agents. Specifically, in vitro contrast channel phantom imaging and in vivo xenograft mouse tumor imaging by this probe with superharmonic imaging showed contrast-to-tissue ratio improvements of 17.7 and 16.2 dB, respectively, compared to line-by-line micro-ultrasound B-mode imaging.

I. INTRODUCTION

TUMOR vasculature lacks the normal blood vessel hierarchy present in healthy tissue, and instead exhibits a chaotic structure with irregular vessel branching [1], [2]. Visualization of abnormal tumor vasculature has the potential to provide prognostic value, and help early detection and monitoring of therapeutic outcomes [3]–[9]. Contrast-enhanced ultrasound (CEUS) using intravascular microbubble (MB) contrast agents has been used to characterize the microcirculation in cancerous lesions in, for example, liver, kidney and abdominal tissues [10]–[12]. CEUS techniques, including amplitude modulation or pulse inversion, detect the nonlinear response of MBs to provide enhanced sensitivity for detecting the blood pool, and can be used to quantitatively measure hemodynamic metrics such as perfusion rate [13], [14]. However, these techniques are hindered by tissue clutter, including tissue harmonics, and have more complex pulsing schemes. For example, in pulse inversion CEUS, two pulses are sent consecutively with the second pulse being the inverse of the first. For image reconstruction, it makes use of signals at the second harmonic of the transmitted pulse frequency [13], but nonlinear signals arising from acoustic wave propagating in tissue limit the contrast-to-tissue ratio (CTR) that can be achieved.

To overcome these challenges, we are investigating superharmonic contrast imaging (SpHI), which takes advantage of the fact that higher order (>third) harmonic components in the nonlinear MB responses significantly predominate in amplitude over corresponding components originating from nonlinear propagation and linearly scattered by tissue. When excited with a single, low-frequency (LF, ~1 to 4 MHz) pulse on transmit, MBs generate a wideband transient response with signal detectible up to 45 MHz [16]. In superharmonic imaging, the higher order harmonics (>third) from MBs are detected, and integrated over the bandwidth of the receiving transducer, to produce the contrast signal. At the same time, the fundamental and lower order harmonic signals from tissue are suppressed, improving the CTR and visualization of vasculature [15].

Dual-frequency (DF) probes are needed for superharmonic imaging because detecting even the second harmonic is near the bandwidth limit of the probes typically used for ultrasound imaging [13]. DF transducers have components that operate in the LF range on transmit to excite MBs and components that operate at higher frequencies (HF, ~5 to 30 MHz) on receive to detect the broadband nonlinear MB signals. DF transducers that effectively separate the LF and HF ranges are therefore essential for superharmonic imaging. Increasing the center frequency and bandwidth of the HF receiving transducer relative to those of the LF transmit transducer should improve image resolution and contrast by capturing more of the wideband response of MBs while efficiently filtering out nonlinear tissue signals.

Several DF transducer designs have been proposed for superharmonic imaging, with LF elements interleaved within the line of HF array elements, positioned beside and parallel to the HF component, or stacked behind the HF array. A summary of published designs and characteristics is shown in Table I. These DF transducers have demonstrated, in various applications, improved CTR and sensitivity to MB contrast agents using higher order harmonics [15], [17]–[20]. However, these array designs also have limitations. The interleaved array design has potentially greater transmit grating lobe levels due to increased LF element pitch [15] and degraded image quality due to the sparse HF array elements. The relatively close LF and HF ranges of the early parallel array configurations [17], [19] limit the detection of MBs signals to the few first harmonics range (≤third), impacting both image resolution and CTR, while the stacked array configuration presented in [20] limits the detection to the fifth harmonic and achieved a moderate imaging resolution with a 15-MHz HF array component.

TABLE I.

SUMMARY OF DF TRANSDUCER DESIGNS AND CHARACTERISTICS REPORTED IN LITERATURE.

Groups	Transducer Type	LF F0(MHz)	LF Bandwidth a	HF F0(MHz)	HF Bandwidth a
Bouakaz et al. [15]	Interleaved LF/HF array	1.1	77% (0.65 – 1.5 MHz)	3.3	82% (1.5 – 4.2 MHz)
Ferin et al. [17]	Parallel LF/HF arrays	3.5	90% (1.9 – 5.1 MHz)	7.5	90% (4.1 – 10.9 MHz)
Van Neer et al. [18]	Interleaved LF/HF array	~1	55% (at −10 dB)	3.7	50% (at −10 dB)
Hu et al. [19]	Parallel LF/HF arrays	1.5	~50%	5.4	73%
Li et al. [20]	Stacked LF/HF arrays	3.4	36.2% (2.78 – 4.0 MHz)	14.8	45.3% (11.45 – 18.15 MHz)
Lukács et al. [23]	Center HF element, outer LF ring	~2	60% (at −3 dB & 3 MHz)	30	100%
Chérin et al. [29]	Parallel LF elements/HF array	1.7	78%	21	52% (13 – 24 MHz)
This paper	Stacked LF/HF arrays	1.86	64.5% (1.2 – 2.4 MHz)	20.3	71% (13.3 – 27.8 MHz)

LF: low-frequency component; HF: high-frequency component; F₀: center frequency cited in references.

^aFractional bandwidth is defined as: (upper – lower cutoff frequency)/F₀; reported at −6 dB unless noted otherwise.

We have extended DF superharmonic imaging to the HF range (20–30 MHz) for greater bandwidth separation and to enable micro-ultrasound B-mode imaging, demonstrating capability in vitro and in vivo for preclinical applications such as molecular imaging and clinical studies [21]–[28]. Although these results are promising, the useful field of view (FOV) is limited by the fixed focus of the single-element LF and HF transducers used, and the 2-D and 3-D frame/volume rates are limited by the necessity of mechanical translation. Now, DF arrays with this larger bandwidth separation are needed to overcome the limitations of using single-element transducers and take advantage of detecting the higher (>fifth) harmonics from MB contrast agents.

Our team has developed and demonstrated a hybrid DF probe with a pair of rectangular transducers (1.7 MHz) parallel to and flanking a commercial HF array (21 MHz) [29]. This probe has also shown the capability of superharmonic imaging for ultrasound localization microscopy applications [30]. However, its fixed plane wave transmit focal zone and the large footprint of the probe limit its applications.

In this article, we report on a vertically integrated, or stacked, coaxial DF transducer consisting of fully populated LF and HF linear arrays. These arrays were designed to operate at 2 and 20 MHz, with the LF array positioned behind the HF array. In Sections II–IV, we present the design and fabrication approaches for the probe followed by the characterization methods and the beamforming schemes needed to focus the arrays on transmit and receive. Acoustic characterization of both LF and HF arrays of the DF probe is reported. Finally, SpHI is demonstrated, both in vitro and in vivo.

II. DF ARRAY DESIGN AND FABRICATION

The DF probe was designed to integrate a LF array, for driving the MBs into nonlinear oscillations, directly behind a HF array. The HF array can be used for both receiving the high harmonic signals from MBs to form a superharmonic contrast image (SpHI-mode) and forming high-resolution B-mode images. With this coaxial design, the LF and HF arrays form coplanar imaging fields focused by a single lens on the front face of the array. The advantages of this approach, compared to the hybrid probe with external LF elements [29], are a more compact footprint, and azimuthal focusing and steering with the LF beam. We targeted a frequency pairing of 2 and 20 MHz for this probe, with the HF array component based on a commercial 15–30 MHz probe (MX250 probe, FUJIFILM-VisualSonics Inc., Toronto, Canada; center frequency: 21 MHz). This frequency pairing is similar to the hybrid probe, and based on prior experience with DF transducers [23], [24], [29]. The HF array was designed with 256 elements, a 90-μm element pitch and 2-mm elevation height. The LF array was designed with 32 elements and wavelength pitch (795 μm), to allow moderate beam steering. It therefore extends slightly beyond the aperture of the HF array. In elevation, the LF element height is 3.7 mm. The DF array configuration is presented in Fig. 1. Relevant material properties and layer thicknesses for the LF array component are listed in Table II.

Fig. 1.

(a) Photo of the vertically integrated DF probe; (b) schematic of transducer layers in the probe. Sections showing (c) LF stack and (d) HF stack with relevant dimensions in millimeters. Schematics are to scale.

TABLE II.

MATERIAL THICKNESSES AND RELEVANT LAYER PROPERTIES IN THE INTEGRATED DF PROBE

	Thickness (mm)	Speed of Sound (m/s)	Density (g/cm3)	Attenuation (dB/mm)
LF Backing	4.2	1960	1.11	17
LF Layer (Tx)	0.75	3370	3.19	-
LF Matching #1	0.25	1989	3.44	16
LF Matching #2	0.33	2670	1.19	-
HF Stack	MX250 (FUJIFILM-VisualSonics Inc.)

Attenuation was measured at 30 MHz.

We have fabricated the HF array using techniques similar to those reported in Lukacs et al. [31]. Matching and backing layers were applied to the piezoelectric substrate, and electrical connections to elements were formed with flexible circuit cabling. For the LF array, we began by fabricating a PZT-5H/epoxy composite using a triangular dicing pattern with 265-μm pitch, 65-μm kerf, and 45° angle [32]. The composite plate was lapped to a thickness of 750 μm then gold electrodes were sputtered onto both sides of the plate. Two matching layers were sequentially cast and machined to quarter-wavelength thickness. The array elements were defined by patterning the electrodes on the back surface of the composite and connecting them to conductive traces on a flexible circuit with matching pitch. A backing layer was bonded to the back surface to complete the LF array preparation. We assembled the DF probe by bonding the LF array front face to the HF array backing, which also acted as an interlayer with an attenuation of 1.2 dB/mm at 2 MHz. An acoustic lens was bonded to the front face of the HF stack and machined to a curvature suitable for HF elevation focusing at 10 mm. Finally, micro-coaxial cables were connected to the flexible circuits of the LF and HF arrays, and the arrays were sealed into a housing. Inductive tuning was used for electrical matching for both LF and HF arrays to compensate for the high impedance related to the relatively small element area and implemented on the flexible circuit. All elements in both LF and HF arrays were found to be connected and functional, as demonstrated by acoustic measurements.

III. ACOUSTIC CHARACTERIZATION AND IMAGING METHODS

A. Instrumentation Setup

The instrumentation required for experimental characterization of the integrated DF probe and for superharmonic imaging is shown in Fig. 2. For superharmonic imaging experiments presented below, two beamforming systems were used to control the LF and HF arrays, respectively. The LF array was connected to a Vantage 128 programmable beamforming system (Verasonics, Inc., Kirkland, WA, USA), and the HF array to a Vevo3100 imaging system (FUJIFILM-VisualSonics, Toronto, ON, Canada), with the two systems synchronized via trigger signals.

Fig. 2.

Schematics of (a) experimental setup using a Vevo3100 system and a Vantage 128 system for controlling each of the HF and LF arrays of the DF transducer for superharmonic imaging; (b) setup for acoustic pressure profile and bandwidth measurements; (c) in vitro contrast channel phantom imaging and (d) xenograft mouse tumor imaging.

In superharmonic imaging mode, the receive channel data from the HF array, which are necessary for the offline beamforming described below, were acquired with the Vevo3100 system. Channel data are accessible using the built-in photoacoustic imaging mode, which enables a receive-only acquisition (i.e. HF transmission is disabled). An external pulse generator (PM5715, Fluke Corporation, Everett, WA, USA) produced a 20-Hz master trigger that is required to initiate imaging in the receive-only acquisition state [29]. Upon reception of a master trigger event, the Vevo3100 generated a secondary trigger that initiated the LF transmit sequence by the Vantage 128 system and the HF receive acquisition sequence.

B. Beamforming Schemes

For this study, we implemented a line-by-line walking-aperture imaging scheme for SpHI, for both transmit (LF) and receive (HF), as illustrated in Fig. 3. The FOV was defined by 256 image lines centered on each HF element. The transmit aperture for SpHI was always set to 19 elements to achieve an f-number of 1.0 for focusing at 10 mm depth in the field, considering the ~5 mm thickness of the acoustic stack. An azimuthal transmit focus was generated for each of the 256 image lines. Therefore, beams focused along the center image lines were unsteered with partially overlapped or identical transmit aperture for adjacent image lines, and the beams near the transducer edge were both focused and effectively steered with the same group of 19 LF elements [Fig. 3(a)].

Fig. 3.

Schematics showing (a) the Tx beamforming scheme at two illustrative locations where the same subaperture of 19 LF elements is used to generate a focused beam. Line A is located at the edge of the array and line B is closer to the central part of FOV. The focused beam generated for image line A is effectively steered. The receive apertures for line A and line B are shown in blue and red, with Rx elements common to both image lines shown in black; (b) parameters used in the calculation of the round-trip time-of-flight for beamforming.

The geometry and relevant parameters for transmit and receive beamforming using the DF probe are shown in Fig. 3(b). Assuming a field point positioned at a distance z from the lens surface, and on an image line centered on the HF element i at position $\left(x_i^{HF},0\right)$, the propagation time from the element k of the LF aperture to the field point $\left(x_i^{HF},z\right)$, is given by:

$$\begin{gathered} PT\left(x_i^{HF},z;x_k^{LF}\right) \\ =\left(\frac{Th}{(Th+z)c_l}+\frac{z}{(Th+z)c}\right)\times \sqrt{{\left({\mathit{x}}_k^{LF}-x_i^{HF}\right)}^2+{(Th+z)}^2} \end{gathered}$$ (1)

with $\left(x_k^{LF},-Th\right)$ the position of the LF element k of the Tx aperture, Th the sum of the layer thicknesses from the LF piezoelectric surface to the lens surface, c_l the thickness-weighted speed-of-sound average through the layers, and c the speed of sound in tissue. Thus, to focus the beam at a distance z = F₀ from the lens, the relative transmit delay applied to the LF element k must follow:

$${\tau }_k=\max \left(PT\left(x_i^{HF},F_{\mathbf{0}};x_k^{LF}\right)overk\right)-PT\left(x_i^{HF},F_{\mathbf{0}};x_k^{LF}\right)$$ (2)

Note that τ^max = max(τ_k) and τ^min = min (τ_k) = 0 are the transmit delays applied to the LF elements whose axes are, respectively, the closest to and the furthest away from image line i being beamformed. The acoustic excitation is therefore first transmitted by the element at the edge of the aperture and furthest away from the focal point, and the acquisition trigger signal is transmitted at the same time. Equations (1) and (2) can be used whether the Tx beam was solely focused (in the central part of the FOV) or focused and steered (on the edge of the FOV). Experimentally, transmit delays were calculated with the approximation that the speed of sound c_l was equal to c, which was set to 1540 m/s. Focal depth F₀ was set to 10.6 mm, and the effective layer thickness Th was equal to 4.4 mm.

On receive, dynamic focusing with the HF array was performed using a delay-and-sum (DAS) algorithm. The transmit time-of-flight along each image line i to the field point $\left(x_i^{HF},z\right)$ can be approximated by:

$$TO{F}^{Tx}(z)\approx \raisebox{1ex}{$z$}\!\left/ \!\raisebox{-1ex}{$c$}\right.+{\tau }_{Th}+{\tau }^{max}$$ (3)

with τ_Th = Th/c the propagation time through the multiple layers between the surface of the LF piezoelectric layer and the surface of the lens. The receive time-of-flight from the field point $\left(x_i^{HF},z\right)$ to the receive element centered at $\left(x_j^{HF},0\right)$ is given by:

$$TO{F}^{Rx}\left(x_i^{HF},z;x_j^{HF}\right)=\frac{1}{c}\sqrt{{\left(x_j^{HF}-x_i^{HF}\right)}^2+z^2}$$ (4)

The total round-trip time-of-flight (TOF^Total) from the Tx aperture to the field point $\left(x_i^{HF},z\right)$, to HF element j located at $\left(x_j^{HF},0\right)$:

$$TO{F}^{Total}\left(x_i^{HF},z;x_j^{HF}\right)=TO{F}^{Tx}(\text{z})+TO{F}^{Rx}\left(x_i^{HF},z;x_j^{HF}\right)$$ (5)

A constant delay τ_FOV corresponding to the propagation time to and from the FOV proximal distance $z_{FOV}^{prox\text{.}}$ (i.e. the top of the image) must now be removed from T OF^Total. This delay, which was implemented on the Vevo3100 system for data acquisition, represents the delay between trigger signal and the acquisition time of the first radio-frequency (RF) sample acquired by the system on each channel. It is calculated using the following equation:

$${\tau }_{FOV}=TO{F}^{Total}\left(x_i^{HF},z_{FOV}^{prox\text{.}};x_i^{HF}\right)$$ (6)

Combining (3)-(6), the beamformed RF sample of image line i at distance z can be expressed as:

$$RF\left(x_i^{HF},z\right)=\sum _j{S}_j\left[n_j(z)\right],elementj\in Rxaperture$$ (7)

with S _j the signal received by HF element j, and n _j (z), the index of the sample from S _j to be compounded, given by:

$$\begin{gathered} n_j(z) \\ =round[\frac{f_s}{c}\times \left(z+\sqrt{{\left(x_j^{HF}-x_i^{HF}\right)}^2+z^2}\right)+{\tau }_{Th} \\ +{\tau }^{max}-{\tau }_{FOV}] \end{gathered}$$ (8)

Dynamic focusing was implemented on receive, with a constant f-number of 1.5, so that the number of elements used in the receive aperture increased with depth. However, at the same depth, the number of HF elements in the receive aperture was progressively reduced for image lines toward the edge of the FOV while maintaining the delay profile for this depth.

The Vevo3100 acquisition imaging system has 64 parallel receive channels connected with 1:4 multiplexors to 256 HF array elements, forming four adjacent subapertures. Since the specified receive aperture can be larger than 64 elements, especially at deeper field points, four transmit/receive events were used for each scan line to ensure complete data capture. Beamforming was performed offline. The envelopes of the beamformed RF data were calculated and logarithmically compressed for display. The image dynamic range was selected to suppress the background noise.

C. Pressure Field Simulations

Field II [33] simulation was used to calculate the pressure fields of an ideal and solitary LF array that is set back an effective distance Th from the probe front face. Water was used as the propagation medium, rather than the HF array layers and then water, as a free-field comparison for experimental measurements. The LF array was modeled with planar elements segmented into 5 (azimuth) by 21 (elevation) subelements. The transducer impulse response was set as a 60% bandwidth Gaussian modulated sinusoid and elements were excited with a single cycle sinusoidal 2-MHz pulse.

The pressure field of a single LF element was simulated over a 12-mm wide (in azimuth) and 15-mm deep FOV in front of the array for calculating element directivity. At each given depth, the simulated pressure was normalized to the on-axis value to evaluate the –6 dB directivity angle.

Simulated LF beams were focused in the azimuthal plane for each image line using the transmit beamforming scheme described above. For each transmit focal position, the pressure was simulated over the full FOV (x = ±15 mm from central axis; z = 2.75–25 mm). This set of simulations was used to estimate the variations of the focused beam pressure across the range of image lines. Simulated pressure field profiles were normalized to the peak pressure of the beam focused along a central image line (center of HF element #128), then logarithmically compressed and displayed with a dynamic range of −20 dB.

Pressure fields were simulated in the elevational plane with the LF array positioned at the probe front face (i.e. at z = 0) and delays applied for an elevational focus at 10 mm; no azimuthal focusing was applied. At each given depth, simulated pressure values were normalized to the peak value at that depth. After logarithmic compression, the fields were displayed with a dynamic range of −6 dB. The −6 dB beamwidth was calculated and plotted as a function of depth.

D. Experimental Pressure Profile Measurements

Acoustic pressure measurements were performed using a calibrated 40-μm aperture needle hydrophone (NH0040, Precision Acoustics Ltd, Dorchester, UK) connected to a pre-amplifier (AU1579, L3 Narda-Miteq, Hauppauge, NY, USA), low-pass filter (BLP-70+, Mini-Circuits, Brooklyn, NY, USA) and digitizer with 500-MHz sampling frequency (DP210, Acqiris SA, Geneva, Switzerland). LF array signals were additionally low-pass filtered (5-MHz cutoff). For all measurements, the hydrophone was placed inside a tank filled with degassed and deionized water at room temperature [Fig. 2(b)]. A 3-D micro-positioning system (U511-, Aerotech Inc., Pittsburgh, PA, USA) was used to move the hydrophone along three orthogonal directions. For LF pressure field measurements, the motion step size was set to 0.15 and 0.06 mm in the transducer azimuthal and elevational dimensions. Along the depth axis, the step size was set to 0.2 mm for the LF single-element beam measurement, and 0.25 mm for focused beam measurements. For HF pressure field measurements, the motion step size was set to 0.06 mm along both the azimuthal and elevational axes, and 0.15 mm along the depth axis.

Hydrophone data acquisition was triggered by the Vevo3100 and the Vantage 128, respectively, for HF and LF array characterization. Received signals were averaged to increase the signal-to-noise ratio (SNR), and corrected for the frequency-dependent hydrophone sensitivity and receive electronics transfer functions to obtain absolute pressure amplitude. For acoustic measurements from the LF components, the hydrophone was first aligned with the probe using a stationary HF beam produced by the Vevo3100 system operating in Pulsed Wave Doppler mode.

For the measurement of the pressure fields produced by individual LF elements, the elements were driven with a 67 V (peak positive), one-cycle 1.9841 MHz pulse (the frequency closest to 2 MHz supported by the Vantage 128 system). For LF element directivity estimation, signals were acquired for an element near the center of the array over a FOV 10 mm wide in azimuth and 15 mm in depth. Pressure measurements were also performed as a function of depth on the axis of each individual LF element. The variation in pressure amplitude across all elements was evaluated at a depth of 3.8 mm, where the peak pressure for most of the elements was found. The center frequency and −6 dB bandwidth of individual LF elements were estimated from the spectra of the pressure waveforms acquired at this depth. The center frequency was estimated as the weighted frequency average (centroid) over the −6 dB bandwidth. The mean and the standard deviation (SD) of the center frequency and bandwidth were calculated across all elements.

For focused LF beam measurements, transmit delays were applied as previously described. A one-cycle 1.9841 MHz pulse was used with a peak voltage amplitude of 56.3 V (maximum allowed). The pressure fields for a transmit beam focused along a central image line (centered on HF element #128) and the outer-most image line were measured over a FOV covering ±15 mm laterally from the central axis and to 25 mm depth. Normalized pressure amplitudes were logarithmically compressed and displayed with a dynamic range of −20 dB.

Further pressure measurements were acquired for each transmit focal position: axial pressure profiles were acquired along each image line (i.e. centered on the respective HF element). The pressure values at 8.75 mm were normalized to the value of the central image lines at this depth.

The pressure field of the LF transmit beam focused on the image line centered on HF element #128 was acquired in the elevational plane. Pressure amplitudes at each given depth were normalized to the on-axis peak at that depth, logarithmically compressed, and displayed with a dynamic range of −6 dB. The −6 dB beamwidth was calculated and plotted as a function of depth.

The azimuthal pressure fields of the HF transmit beams focused along a central image line and the outer-most image line were acquired. Pulsed Wave Doppler mode was used to produce a single, stationary beam focused at a depth of 10 mm. The transmit event was set to use a one-cycle, 21-MHz excitation pulse and an f-number of 2.5 at 1 % transmit power. The −6 dB azimuthal beamwidths and depth of field were evaluated. The elevational pressure field was acquired with the transmit beam focused along the central image line, and the −6 dB beamwidth was calculated and plotted as a function of depth.

Pressure waveforms generated by the HF array were acquired with the hydrophone and compared to the waveforms of a commercial HF probe to assess the effects of stacking the LF array behind the HF array on the micro-ultrasound B-mode imaging performance. An MX250 probe (FUJIFILM-VisualSonics Inc., Toronto, Canada) was selected as the reference since the design specification of the HF component of the DF array (DF-HF) is based on this probe. For each array (DF-HF and MX250), the beam was set to focus at 11 mm and the hydrophone was positioned at the focal position. The −6 dB bandwidths of both probes were estimated from the spectra of the pressure waveforms acquired at the focus. Center frequencies were computed as the weighted average over the −6 dB bandwidth.

E. In Vitro Contrast Channel Phantom Imaging

The contrast imaging performance of the DF array was evaluated using a tissue-mimicking channel phantom [Fig. 2(c)]. The phantom contained a 1.4 mm × 1.4 mm square, wall-less channel embedded in tissue-mimicking material made of 2% (w/w) agar (BD Difco™ Agar Technical, Ref. 281210, Fisher Scientific Canada, Ottawa, ON, Canada) with 1% (w/w) silica (S-5505, Sigma-Aldrich, St Louis, MO, USA). The channel was filled with MicroMarker®MBs (FUJIFILM-VisualSonics Inc., Toronto, Canada) prepared according to manufacturer’s protocol and diluted in phosphate-buffered saline (PBS) to a concentration of 2 × 10⁶ MBs/mL. The MB solution was continuously stirred to maintain a uniform concentration and fed into the channel by gravity. The flow was stopped prior to imaging. Freshly prepared MB solutions were used for each set of image acquisitions.

Superharmonic contrast images were acquired using the DF probe and the instrumentation setup shown in Fig. 2(a) and (c). Acquisition of channel data for each SpHI-mode image line was interleaved with acquisition of B-mode image frames using the HF component of the array. This interleaving was necessitated by the settings on the receive-only acquisition state of the Vevo3100 system, but allowed concurrent capture of SpHI-mode and micro-ultrasound images. The built-in B-mode on the Vevo3100 was used for micro-ultrasound imaging.

Contrast images were also acquired with the HF component of the DF probe only, and the MX250 probe. The built-in nonlinear contrast (NLC) mode on the Vevo3100 system was used, which is based on an amplitude-modulation pulse sequence [35]. With both probes, 20 contrast frames interleaved with 20 B-mode frames were acquired, at an overall frame rate of 25 Hz.

F. In Vivo Superharmonic Imaging of Xenograft Tumor

Breast cancer cells (LM2–4) were injected subcutaneously into the hindlimb of a four-week-old female mouse (SCID hairless outbred – SHO, Charles River Laboratory, Wilmington, MA, USA). The mouse was anesthetized using gaseous isofluorane (4% initiation, 2% maintenance) during the imaging process. The mouse body was maintained at a constant temperature using a heating pad. Ultrasound gel was applied to couple the transducer to the mouse skin. For in vivo mouse imaging, the MB solution was prepared by diluting the stock MB solution with 0.9% saline to a final concentration of 10⁸ MBs/mL. A 100 μL volume of this MB solution was injected through a tail vein catheter using an infusion pump (New Era Pump Systems, Inc., Farmingdale, NY, USA) at 500 μL/min. Pre- and post-injection SpHI data (interleaved SpHI-mode and B-mode frames) were acquired with the setup shown in Fig. 2(a) and (d). Post-injection acquisition was initiated at the end of MB solution injection. After imaging, the animal was euthanized in a chamber with CO₂. Animal experimental procedures were approved by the Animal Care Committee at Sunnybrook Research Institute.

G. Signal Processing and CTR Computation

Raw unbeamformed RF data for SpHI images and beamformed IQ data for B-mode and NLC images were exported from the Vevo3100 system for offline signal processing and image reconstruction. Unbeamformed SpHI RF data were bandpass filtered (6–35 MHz) using a finite-impulse-response (FIR) filter designed from the analysis of their spectra. The 6-MHz cutoff frequency was chosen so that the higher order harmonic (>third) signals, originating predominantly from MB oscillations, could be used for contrast image reconstruction. The 35-MHz cutoff frequency was chosen to remove the electronic noise beyond the HF array bandwidth.

For each imaging mode (SpHI, micro-ultrasound B-mode and NLC), the CTR in the reconstructed flow channel image was computed as:

$$CT{R}^{mode}=20log10\left(\frac{\overline{E_{MB}}}{\overline{E_t}}\right)$$ (9)

where, $\overline{E_{MB}}$ and $\overline{E_t}$ are the means of the signal envelope from regions of interest (ROIs) corresponding to the channel filled with MBs and tissue-mimicking phantom matrix, respectively. $\overline{E_t}$ was measured in an imaging plane 4–5 mm away from the channel plane but at the same depth. For in vivo imaging, $\overline{E_{MB}}$ and $\overline{E_t}$ were calculated in two manually selected ROIs within the tumor, one with strong MB contrast signals, and the second with signals at the noise floor level. $\overline{E_{MB}}$ and $\overline{E_t}$ are mean signal envelope levels within these ROIs. For micro-ultrasound B-mode and NLC-mode, CTR was reported as the average over all acquired frames.

IV. RESULTS

A. Acoustic Characterization

The simulated beam in the azimuthal plane and acoustic measurements for single LF elements are shown in Fig. 4. The measured and simulated −6 dB directivity angles were found to be 17° and 34°, respectively. The measured −6 dB beamwidth at 2 mm depth was 7 mm. The waveform generated at a depth of 3.6 mm by an element adjacent to the center of the array has a center frequency at 1.84 MHz with a −6 dB bandwidth of 1.2 MHz (1.3–2.5 MHz) [Fig. 4(c)]; no harmonics were observed in the measured signal. Across the LF array elements, the mean center frequency was 1.86 MHz with a SD of 0.02 MHz, with mean (SD) upper and lower −6 dB ranges of 2.4 (0.03) MHz and 1.2 (0.03) MHz, respectively [Fig. 4(d)]. Driven at 67 V (peak positive), the individual LF elements generated an average pressure amplitude of 56.4 kPa at 3.8 mm below the probe surface. Less than 4 dB variation was observed among all 32 elements, although edge elements were found to generate higher pressure amplitudes compared to central ones [Fig. 4(e)].

Fig. 4.

Single LF element transmit pressure field (a) simulated by Field II using the physical width of the LF element (0.73 mm), and (b) measured output from element 16 of the LF array. Pressure is normalized to the on-axis peak value at each depth. The red continuous line is the −6 dB contour. The red dashed line is the axis of the array element; profiles are displayed at a dynamic range of −6 dB. (c) The LF waveform and its spectrum acquired on axis, at a depth of 3.6 mm below the probe surface, for a single cycle transmit pulse at 1.9841 MHz. (d) Center frequency of all elements with their respective upper and lower bound −6 dB frequencies; and (e) pressure measured on axis of individually excited LF elements at a depth of 3.8 mm below the probe surface.

Simulated and measured pressure beams for focused transmits from the LF array are presented in Fig. 5. For beams focused along the central and edge image lines, the predicted azimuthal beamwidths (−6 dB) were 1.2 and 1.35 mm, and the peak pressure of the two beams was found at 10.25 and 9.5 mm in depth, respectively. The −6 dB depth of field, where the pressure drops to half of the axial peak, was 12.75 mm (2.75–15.5 mm) for the simulated beam focused along the central image line. The predicted elevational focus was found at 4.1 mm in depth with a −6 dB beamwidth of 1.5 mm.

Fig. 5.

Acoustic pressure maps for the DF array. (a-d) Simulated and (e-h) measured pressure field for the LF array, with the overlay showing measured pressure field for the HF array. Azimuthal beam plots for transmit focus at ~10 mm below the LF array at (a, e) −11.5 mm and (b, f) −0.045 mm in azimuthal positions; pressure amplitudes normalized to the peak value when focused at x =−0.045 mm. The LF beam plots are displayed with −20 dB dynamic range; the HF beam plots are shown with −6 dB, −12 dB and −20 dB contours. (c,g) Elevational pressure fields at 0 mm in azimuth; pressure fields normalized to the on-axis value at each depth. The LF beam plots are displayed with −6 dB dynamic range; the HF beam is shown with −6 dB contour. (d,h) Elevational −6 dB beamwidth over depth for LF (blue) and HF (orange) arrays; effective focal depths are indicated by arrows. The depth-axis of the plots is relative to the probe surface. Image axes are in millimeters.

The measured azimuthal −6 dB beamwidths were 1.35 and 1.5 mm for beams focused along the central and edge image lines, respectively. The peak pressure amplitudes obtained with focused transmits (56.3 V peak-amplitude excitation) along the central and edge image lines were found to be 458 kPa at 8 mm depth and 336.6 kPa at 6.5 mm depth, respectively. The −6 dB depth of field was found to be 11.75 mm (3–14.75 mm) for the beam focused along the central image line. The elevational focus was found to be located at 4.5 mm below the probe surface with a −6 dB beamwidth of 1.98 mm, broadening to 4.56 mm at 20 mm depth.

Lateral beam profiles for both measurements and simulations at the central image line and the image line at the edge of FOV are shown in Fig. 6(a), with pressures normalized to the on-axis value obtained for the beam focused along the central image line. Measured and simulated −6 and −10 dB beamwidths at both lateral locations were comparable, although the grating lobes were substantially higher in the simulated beams than in the measured beams. Fig. 6(b) shows the simulated and measured pressure amplitudes at focal depths for each focused transmit beam. The values have been normalized to their respective peak values found at this depth. For both simulations and measurements, the pressure at the central scan lines showed variations of less than 1 dB, but pressure dropped in the edge image lines by 3.75 and 3.0 dB in simulations and measurements, respectively.

Fig. 6.

(a) Measured and simulated pressure plots at 8.75 mm (within LF focal region of central image line) below the probe surface as a function of azimuthal position for LF beams focused along a central image line (x =−0.045 mm) and the edge image line (x =−11.5 mm). (b) Normalized measured and simulated pressures at their azimuthal focal depths for the beam focused along a central image line (measured: 8.75 mm; simulated: 10 mm) as a function of image line azimuthal position. Normalization is relative to the respective measured and simulated maximum peak pressure amplitudes.

The measured azimuthal −6 dB beamwidths of the transmitted HF pressure beams were 0.3 and 0.42 mm for beams focused along the central and edge image lines [Fig. 5(e) and (f)]. The corresponding peak pressure amplitudes were found to be 1.4 MPa and 794.4 kPa, at depths of 10.35 and 9 mm, respectively. The −6 dB depth of field was 5.1 mm (7.95–13.05 mm) for the central image line, falling within the focal region of the LF transmit beam. The elevational focal depth was found to locate between 10.8 and 11.7 mm with a −6 dB beamwidth of 0.66 mm, which was also within the LF beam but deeper than the LF beam focus [Fig. 5(g) and (h)].

Pressure waveforms measured at an azimuthal focus of 11 mm with both the HF component of the DF probe and the standard MX250 probe are shown in Fig. 7. The DF-HF array had a center frequency of 20.3 MHz and a −6 dB bandwidth of 14.5 MHz (13.3–27.8 MHz). The MX250 probe showed a similar spectrum, with a center frequency of 20.5 MHz and a −6 dB bandwidth of 15.6 MHz (12.6–28.2 MHz). The DF-HF also produced −6 and −20 dB pulse lengths of 54 and 100 ns, respectively, that were comparable to those recorded with the MX250 probe (58 ns; 100 ns).

Fig. 7.

HF pressure waveforms, and their spectra, acquired with a hydrophone from (a) the DF-HF and (b) a standard 21 MHz probe (MX250). Both transducers were driven by a Vevo3100 system using a single cycle voltage pulse at 21 MHz, and measurement performs on Tx aperture axis, at an azimuthal focus set to 11 mm below the probe surface.

B. In Vitro Contrast Imaging of Channel Phantom

Acquired B-mode and contrast mode images of the tissue-mimicking phantom are shown in Fig. 8, with CTR values listed in Table III. With both probes, contrast enhancement was observed in B-mode and amplitude-modulated NLC-mode, as indicated by the elevated signal level within the contrast-filled channel compared to that of the matrix. Further quantification showed that the DF-HF array presented a CTR improvement of 5.2 dB in NLC imaging mode compared to B-mode imaging (CT R^NLC: 11.3 dB; CT R^Bmode: 6.1 dB), while the MX250 probe had a CTR improvement of 6.3 dB (CT R^NLC: 13 dB; CT R^Bmode: 6.7 dB).

Fig. 8.

Contrast images of a flow channel embedded in a tissue-mimicking phantom with and without MBs. Amplitude-modulated NLC images acquired with (left) a standard 21 MHz transducer and (middle) the HF array of the DF probe (DF-HF). (right) Images acquired in SpHI-mode using the DF probe. Color-rendered contrast images are overlaid on interleaved gray-scale B-mode images. Gray-scale micro-ultrasound B-mode images and the color-rendered images are displayed in −50 dB and −35 dB dynamic ranges, respectively. The ROIs used to compute CTR are indicated by the white box on the image. The depth axis is relative to the probe surfaces and identical for all images. Image axes are in millimeters.

TABLE III.

CTR MEASUREMENTS OF In Vitro AND In Vivo IMAGING

	Flow Channel Imaging			Tumor Imaging
	B-mode	NLC	SpHI	B-mode	SpHI
MX250	6.7	13	-	-
DF-HF	6.1	11.3	-	−0.8
DF	-	-	23.8	-	15.4

The reported CTR are in dBs

The CTR observed in images acquired in SpHI-mode was significantly higher than that obtained in B-mode and NLC-mode, with both the DF-HF array and MX250. CT R^SpHI was found to be 23.8 dB, and therefore 17.7 and 12.5 dB higher than the CT R^Bmode and CT R^NLC obtained with the DF-HF array, respectively. It was similarly 17.1 and 10.8 dB higher than the CT R^Bmode and CT R^NLC obtained with the MX250, respectively. This CTR improvement is clearly visible in Fig. 8, where clutter signal from the tissue-mimicking matrix was more efficiently suppressed in SpHI-mode compared to the NLC-mode with the HF arrays (either the MX250 or the DF-HF array).

C. In Vivo Superharmonic Imaging of Xenograft Mouse Tumor

Superharmonic images of the mouse tumor were acquired and CTR measured (Fig. 9, Table III). B-mode imaging, using the HF array component, shows no contrast enhancement following contrast injection (CT R^Bmode: −0.8 dB). The SpHI image shows significant contrast signal enhancement with a CT R^SpHI of 15.4 dB. The overall 16.2 dB CTR improvement is consistent with that found from in vitro contrast channel imaging (17.7 dB). Additionally, the SpHI image following contrast injection was able to identify a necrotic core within the tumor, which appeared as a region lacking contrast signals. The necrotic core was not detectable on micro-ultrasound B-mode images prior to or after contrast injection.

Fig. 9.

Superharmonic imaging of a xenograft mouse tumor showing B-mode images acquired with the DF-HF array, SpHI contrast images, and the color-rendered SpHI images overlaid on the gray-scale micro-ultrasound B-mode images. Panels show pre- and post-injection images. Image artifacts due to trigger crosstalk are indicated by arrowheads on the SpHI image. The two ROIs used to compute CTR in SpHI and B-mode images are indicated by the white boxes on the SpHI image. All gray-scale images are displayed with −50 dB dynamic range. The color-rendered SpHI images are displayed in a −35 dB dynamic range in the overlay panel. The depth axis is relative to the probe surface and identical for all images. Image axes are in millimeters.

V. DISCUSSION

A. Multilayer Array Structure

The design of the DF array, with a LF array positioned directly behind a HF array, can be adapted for different frequency combinations or array sizes. The benefit of adding a LF array to an existing HF array design is that the most challenging aspects of fabrication (i.e., defining and connecting to a fine-pitch array) have been overcome, as evidenced by the 100% yield in connecting to array elements. Alignment between LF and HF arrays is important to ensure overlapping fields of view, but can be achieved with precise tooling to position and hold components during assembly.

The coaxial structure of the DF probe requires that the LF wave travels through the multiple layers of the HF array stack. Despite this fact, the LF array generates short pulses, with a 64.5% fractional bandwidth (1.2–2.4 MHz), and no significant undesirable ring down [Fig. 4(c)]. With respect to the HF array, the design has not been significantly altered from the standard 21 MHz probe. Since the probe has a backing layer to prevent reflections from back-propagating waves, the presence of the LF array should not significantly affect its performance. Indeed, its estimated center frequency (20.3 MHz) and fractional bandwidth (71%, 13.3–27.8 MHz) were marginally lower than those of the MX250 (20.5 MHz; 76%, 12.6–28.2 MHz), and resulted in a similar pulse shape and amplitude when driven with the same excitation pulse (Fig. 7). NLC imaging was also demonstrated with the HF component of the DF probe, providing a CTR 1.7 dB lower than that obtained with the commercial HF probe. This is likely due to the narrower elevational dimension (DF-HF: 2 mm; MX250: 2.75 mm) of the DF-HF, and a smaller radius of curvature of the DF-HF lens, restricting its penetration depth and leading to smaller acoustic pressure at the focal depth.

The propagation of the diverging wavefront emitted from the wavelength-width LF elements is made complex by refraction at interfaces between the different material layers of the transducer array. This is evidenced by the narrow directivity measured for a single LF element compared to the directivity expected from simulation for an element of the same size transmitting directly in water. Further effects are the suppression of grating lobes predicted by the simulation for beams focused along image lines at the center and edge of the FOV [Figs. 5(a) and (e), and 6(a)], the slight shift in focal position, and a slight difference in effective steering angle for beams focused along the outer image lines compared to the simulation. These shifts in focal position have minimal effects on the resultant image since the LF beams still cover the HF receive beam [Fig. 5(e)–(g)]. And with dynamic receive beamforming implemented for SpHI-mode, the depth of field of the HF array would be elongated along the image line.

The outer LF elements produced higher pressure in the field than the inner elements. This may be in part due to the LF array extending beyond the piezoelectric layer of the HF array, affecting reflections at transducer layer interfaces, and to boundary conditions at the edge of the array, which would affect the pressure wave amplitude at these locations. With the focused beams, this variation in single-element performance worked in our favor and a smaller pressure variation (−3 dB) was found across the full FOV for the focused transmit beams than predicted by simulations (−3.75 dB).

Despite the use of a focusing lens, the measured elevational focus of the LF array was at approximately 5 mm, almost 5 mm shallower than the targeted focus. This shallow focus is attributable to the small LF elevation aperture of 4.8λ at 2 MHz, and small lens aperture. The simulations also predicted a short focus in the elevation plane, though shallower still than the experimental measurement. Although this beam divergence beyond the elevational focus limits the efficiency, the LF array was capable to deliver sufficient acoustic pressure level in the imaging FOV (Fig. 9). In future iterations of the design, the LF elevation dimension and lens aperture can be increased, with finite element modeling to guide the design, in order to improve the focusing of the LF array.

B. Superharmonic Imaging with Array-Based DF Probe

For the present work, we have used focused transmit beams to excite the MBs for SpHI and a simple walking-aperture transmit and receive beamforming approach. With the equipment used for LF transmission and to overcome the lack of focusing in the elevation plane, line-by-line focused transmit beams in the azimuthal plane were necessary to generate the transmit pressure levels required to elicit high harmonic responses from MBs. The experiments demonstrated superharmonic imaging capability with DF coaxial arrays and showed some of the advantages of using an array, including electronic beam scanning over the FOV governed by the array aperture. In related work, the improvement in the depth of field of this DF array design compared to the confocal, single-channel DF probes has been demonstrated [36]. Our overall frame rate was limited due to the current instrumentation setup, which required multiple transmit and receive events to reconstruct image lines in SpHI-mode. Further development of the array control instrumentation and investigation of beamforming approaches, including plane wave imaging [29], [37], [38], will be undertaken to enable real-time and ultrafast superharmonic imaging [30].

Overall, the DF probe has shown sufficient bandwidth separation between LF and HF components of the array. The broad receiving frequency range allows the detection of high-order MB harmonic signals. Although the lower bound of the −6 dB bandwidth of the HF array is around 13 MHz, a bandpass filter with 6-MHz cutoff was applied to further suppress any LF signals received from tissue.

CTR measurements mentioned in the literature vary due to differences in the type of tissue being imaged and contrast agents used, frequency and excitation pressure, as well as signal processing algorithms, and computation methods, making a direct comparison to other probes challenging. Nonetheless, compared to micro-ultrasound B-mode and NLC-mode, substantial improvements in CTR were observed with superharmonic imaging both for in vitro phantom imaging and in vivo tumor imaging (Table III), demonstrating the advantage of imaging with higher harmonics for suppressing nonlinear signal from tissue. Particularly, since necrosis is a hallmark for aggressive tumors and extensive tumor necrosis is associated with poor prognosis [39], the ability of the DF probe in SpHI-mode to identify the necrotic core in tumor imaging marks its potential in other preclinical or clinical applications such as the assessment of treatment outcomes.

VI. CONCLUSION

We have developed and characterized an integrated DF coaxial array transducer that enables both SpHI and micro-ultrasound B-mode imaging for acoustic angiography applications. The LF array component is positioned directly behind the micro-ultrasound imaging array such that the transmitted LF waves used to excite the MBs must pass through the HF array. We have demonstrated that, compared to an equivalent standard HF array, the HF array of the DF probe is not adversely affected by the presence of the LF array in terms of output pulse and amplitude-modulated contrast imaging performance.

The center frequencies (−6 dB bandwidths) for the LF and HF array were found to be 1.86 MHz (1.2–2.4 MHz) and 20.3 MHz (13.3–27.8 MHz), respectively, showing good bandwidth separation between the LF and HF arrays. We have also presented a focused beamforming scheme with which the LF array was able to generate pressure amplitudes sufficient for exciting MB contrast agents, and the transmit beams had less than 3 dB pressure variations laterally across the FOV at a depth of 10 mm. We have presented in vitro imaging of a contrast-filled channel phantom and in vivo imaging of a xenograft mouse tumor model with breast cancer cells. As expected, due to efficient clutter signal suppression, the CT R^SpHI in vitro (23.8 dB) and in vivo (15.4 dB) were found to be significantly higher than those obtained in micro-ultrasound B-mode imaging, with an improvement of 17.7 and 16.2 dB, respectively.

The length of the array and beamforming control over both the LF and HF arrays yield advantages of extended FOV and flexible beam focusing schemes compared to single-channel DF transducers. For future iterations of the DF probe, we will use finite element modeling to improve and tailor the design for specific preclinical or clinical imaging applications, and investigate and implement high frame rate and plane wave beamforming approaches.

Biography

Jing Yang received the B.A.Sc degree in Engineering Science from University of Toronto, Toronto, ON, Canada in 2016. She is currently pursuing the Ph.D degree with the Department of Medical Biophysics, Faculty of Medicine, University of Toronto. Her research focuses on superharmonic ultrasound imaging with ultrafast imaging techniques.

Emmanuel Chérin received the Ph.D. degree in physical acoustics from Université Pierre et Marie Curie (Paris VI), Paris, France, in 1998.

He joined Imaging Research Group, Sunnybrook Health Sciences Centre, Toronto, ON, Canada, as a Postdoctoral Fellow. Since 2001, he has been a Research Associate with Sunnybrook Research Institute, Toronto, where he is involved in research projects related to high-frequency ultrasound imaging, photoacoustic imaging, ultrasound contrast agents, and transducer characterization.

Jianhua Yin received the Ph.D. degree in physics from Nanjing University, Nanjing, China, in 1990.

His previous research involved surface acoustic wave devices and their applications in electronics, Nanjing University, ultrasonic nondestructive evaluation, Tokai University, Shimizu, Japan, and domain study and characterization of piezoelectric crystals, Pennsylvania State University, State College, PA, USA. He is a Research Physicist with Sunnybrook Research institute, Toronto, ON, Canada, where he is currently working on finite-element modeling, ultrasound transducers, and arrays design and fabrication.

Isabel G. Newsome received the B.S. degree in physics from Clarkson University, Potsdam, NY, USA, in 2016. She is currently pursuing the Ph.D. degree with the Joint Department of Biomedical Engineering, The University of North Carolina at Chapel Hill and North Carolina State University, Chapel Hill, NC, USA, where she focuses on advancements in superharmonic ultrasound imaging technology and the clinical translation of acoustic angiography.

She is currently a Trainee at the UNC’s Integrative Vascular Biology Program.

Mrs. Newsome is a recipient of the Ruth L. Kirschstein Predoctoral Individual National Research Service Award from the National Cancer Institute.

Thomas M. Kiersk received the B.S. degree in biomedical engineering from The University of North Carolina at Chapel Hill, Chapel Hill, NC, USA, in 2017. He is currently pursuing the Ph.D. degree with the Joint Department of Biomedical Engineering, The University of North Carolina at Chapel Hill and the North Carolina State University, where he is researching superharmonic imaging and novel approaches to ultrasound localization microscopy.

Guofeng Pang received the Ph.D. degree in physics from Queen’s University, Kingston, Ontario, Canada in 2003.

Starting in 2004, he worked as a Research Scientist at the Sunnybrook Health Sciences Center, Toronto, ON, Canada, where he focused on research and development of ultrasound linear arrays for biomedical applications. In 2006, he joined FUJIFILM VisualSonics, Inc., Toronto, ON, Canada. At VisualSonics, he has worked on research, development, and fabrication of ultra-high frequency ultrasound array transducers for pre-clinical and clinical micro-ultrasound imaging as a Production Coordinator and Transducer Developer.

Currently, Dr. Pang is a Senior Transducer Developer, where his research involves photoacoustic imaging, dual-frequency imaging, functional imaging, and creating novel ultrasound transducer arrays and medical devices.

Claudia A. Carnevale received the B.S. degree in physics from The University of Guelph, Guelph, ON, Canada, in 2009. She completed post-graduate studies in advanced lasers at Niagara College and joined FUJIFILM VisualSonics, Inc. in 2011.

Mrs. Carnevale is currently the Manager of Transducer Development at VisualSonics., where her team conducts research, development, and fabrication of ultra-high frequency ultrasound array transducers for pre-clinical and clinical micro-ultrasound imaging. Her team is developing products incorporating dual-frequency, photoacoustic, and phased array transducer technology.

Paul A. Dayton received the B.S. degree in physics from Villanova University, Villanova, PA, USA, in 1995 and the M.E. degree in electrical engineering and the Ph.D. degree in biomedical engineering from the University of Virginia, Charlottesville, VA, USA, in 1998 and 2001, respectively.

He pursued postdoctoral research and was later research faculty at the University of California at Davis, Davis, CA, USA. Much of his training was under the mentorship of Dr. Ferrara, where his initial studies involved high-speed optical and acoustical analysis of individual contrast agent microbubbles. In 2007, he moved to the Joint Department of Biomedical Engineering, The University of North Carolina at Chapel Hill, Chapel Hill, NC, USA, and North Carolina University, Raleigh, NC, USA, where he is currently a Professor and the Interim Department Chair. He is also the Associate Director of Education for the Biomedical Research Imaging Center. His research interests involve ultrasound contrast imaging, ultrasound-mediated therapies, and medical devices.

Dr. Dayton is a member of the Technical Program Committee of IEEE UFFC and editorial boards of the IEEE ULTRASONICS, FERROELECTRICS, AND FREQUENCY CONTROL and Molecular Imaging.

F. Stuart Foster is currently a Senior Scientist with Sunnybrook Research Institute, Toronto, ON, Canada, and a Professor with the Department of Medical Biophysics, University of Toronto, Toronto. His current research centers on the development of high-frequency clinical and preclinical imaging systems, array technology, intravascular imaging, photoacoustics, and molecular imaging.

Dr. Foster is a fellow of the American Institute of Ultrasound in Medicine, the Royal Society of Canada, and the Canadian Academy of Engineering. He was elected as a foreign member of the U.S. National Academy of Engineering in 2017. He was a recipient of the Eadie Medal for his major contributions to engineering in Canada, the Queen’s Golden Jubilee Medal, the Manning Award of Distinction for Canadian Innovation, the Ontario Premier’s Discovery Award, the 2010 Rayleigh Award, and the 2020 Biomedical Engineering Award from the IEEE. He is also the Founder and the Former Chairman of FUJIFILM VisualSonics Inc., a company dedicated to preclinical and clinical micro-ultrasound. He co-founded Mouse Imaging Centre (MICe) now at the Toronto Centre for Phenogenomics. He has served on the Board of Directors for the National Cancer Institute of Canada and as the Chairman of its Committee on Research (ACOR). He serves on numerous advisory bodies and is currently an Associate Editor of Ultrasound in Medicine and Biology.

Christine E. M. Démoré received the B.Sc. degree in engineering physics and the Ph.D. degree in physics from Queen’s University, Kingston, ON, Canada, in 2000 and 2006, respectively.

From 2007 to 2015, she was based at Institute for Medical Science and Technology, University of Dundee, Dundee, U.K., and the School of Engineering, University of Glasgow, Glasgow, U.K., in 2016. She is currently a Scientist with Sunnybrook Research Institute, Toronto, ON, Canada, and an Assistant Professor with the Department of Medical Biophysics, University of Toronto, Toronto. Her research includes creating novel ultrasound transducer arrays for biomedical applications, with a specialization in probes for micro-ultrasound imaging. She is also exploring acoustic and photoacoustic contrast-enhanced imaging with new transducer arrays and novel contrast agents.

Dr. Démoré has been an Elected Member of the IEEE UFFC-S Administrative Committee and continues to volunteer with the UFFC-S. She was a recipient of the Royal Society of Edinburgh/Caledonian Research Fund Biomedical Personal Research Fellowship and the IEEE Ultrasonics Young Investigator Award in 2015 for her development of ultrasonic arrays for particle manipulation.