Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

3-D Super-Resolution Ultrasound (SR-US) Imaging with a 2-D Sparse Array

3-D Super-Resolution Ultrasound (SR-US) Imaging with a 2-D Sparse Array 3-D Super-Resolution Ultrasound (SR-US) Imaging with a 2-D Sparse Array Sevan Harput, Kirsten Christensen-Jeffries, Alessandro Ramalli, Jemma Brown, Jiaqi Zhu, Ge Zhang, Chee Hau Leow, Matthieu Toulemonde, Enrico Boni, Piero Tortoli, Robert J. Eckersley , Chris Dunsby , and Meng-Xing Tang Abstract—High frame rate 3-D ultrasound imaging technology pre-clinical studies using microbubbles [5]–[12] and nan- combined with super-resolution processing method can visualize odroplets [13]–[16]. These studies generated super-resolved 3-D microvascular structures by overcoming the diffraction images of 3-D structures using 1-D ultrasound arrays where limited resolution in every spatial direction. However, 3-D super- super-resolution cannot be achieved in the elevational di- resolution ultrasound imaging using a full 2-D array requires a rection. In addition to this, out-of-plane motion cannot be system with large number of independent channels, the design of which might be impractical due to the high cost, complexity, compensated for when data is only acquired in 2-D. However, and volume of data produced. with the implementation of 3-D SR-US imaging using a 2- In this study, a 2-D sparse array was designed and fabricated D array, diffraction limited resolution can be overcome in with 512 elements chosen from a density-tapered 2-D spiral every direction and there is then the potential for 3-D motion layout. High frame rate volumetric imaging was performed using tracking and correction. two synchronized ULA-OP 256 research scanners. Volumetric images were constructed by coherently compounding 9-angle Many studies have contributed to the development of SR- plane waves acquired in 3 milliseconds at a pulse repetition US imaging methods by improving the localization precision, frequency of 3000 Hz. To allow microbubbles sufficient time to reducing the acquisition time, increasing microbubble tracking move between consequent compounded volumetric frames, a 7- accuracy, and extending the super-resolution into the third millisecond delay was introduced after each volume acquisition. dimension. These developments are explained in detail by a This reduced the effective volume acquisition speed to 100 Hz and the total acquired data size by 3.3-fold. Localization-based 3- recent review [17]. Researchers mainly employed two different D super-resolution images of two touching sub-wavelength tubes approaches to generate a super-resolution image of a volume were generated from 6000 volumes acquired in 60 seconds. In by mechanically scanning the volume with a linear probe conclusion, this work demonstrates the feasibility of 3D super- and stacking 2-D SR-US images, or by using arrays that resolution imaging and super-resolved velocity mapping using a can acquire volumetric information electronically. Errico et customized 2D sparse array transducer. al. have taken steps towards 3-D with a coronal scan of an entire rat brain by using 128 elements of a custom-built linear I. I NTRODUCTION array at a frequency of 15 MHz. Motion of the probe was Visualization of the microvasculature beyond the diffraction controlled with a micro-step motor to generate 2-D super- limited resolution has been achieved by localizing spatially resolution images over different imaging planes at a frame isolated microbubbles through multiple frames. In the absence rate of 500 Hz [18]. Lin et al. performed a 3-D mechanical of tissue and probe motion, localization precision determines scan of a rat FSA tumor using a linear array mounted on the maximum achievable resolution, which can be on the a motorized precision motion stage synchronized with the order of several micrometers at clinical ultrasound frequen- imaging system. They generated 3-D super-resolution images cies [1], [2]. If motion is present and subsequently corrected by calculating the maximum intensity projection from all 2- post-acquisition, then the motion correction accuracy can D super-resolution slices, acquired using plane-wave imaging limit the achievable spatial resolution [3], [4]. Researchers with a frame rate of 500 Hz [19]. Although sub-diffraction demonstrated the use of 2-D super-resolution ultrasound (SR- imaging has not been published using a 2-D imaging probe US) imaging in many different controlled experiments and with a high volumetric imaging rate, 3-D super-resolution has been achieved by previous studies. O’Reilly and Hynynen used S. Harput, J. Zhu, G. Zhang, C. H. Leow, M. Toulemonde, and M. X. Tang a subset of 128 elements from a 1372-element hemispherical are with the ULIS Group, Department of Bioengineering, Imperial College London, London, SW7 2AZ, UK. transcranial therapy array at a rate of 10 Hz. They generated 3- K. Christensen-Jeffries, J. Brown, and R. J. Eckersley are with the Biomed- D super-resolution images of a spiral tube phantom through an ical Engineering Department, Division of Imaging Sciences, King’s College ex vivo human skullcap at an imaging center frequency of 612 London, SE1 7EH, London, UK. A. Ramalli is with the Department of Information Engineering, University kHz [20]. Desailly et al. implemented a plane wave ultrafast of Florence, 50139 Florence, IT and Lab. on Cardiovascular Imaging & imaging method using an ultrasound clinical scanner with Dynamics, Dept. of Cardiovascular Sciences, KU Leuven, Leuven, Belgium. 128 fully programmable emission-reception channels. They E. Boni, P. Tortoli are with the Department of Information Engineering, University of Florence, 50139 Florence, IT. placed 2 parallel series of 64 transducers to image microflu- C. Dunsby is with the Department of Physics and the Centre for Pathology, idic channels and obtained 3-D super-localization by fitting Imperial College London, London, SW7 2AZ, UK. parallel parabolas in the elevation direction [21]. Christensen- These authors contributed equally to this work. E-mail: s.harput@imperial.ac.uk and mengxing.tang@imperial.ac.uk Jeffries et al. generated volumetric 3-D super-resolution at the arXiv:1902.01608v1 [physics.med-ph] 5 Feb 2019 2 Fig. 2. Optical image of two 200 m cellulose tubes arranged in a double helix pattern. To create this pattern, two tubes were wrapped around each other which created contact-points that are visible in the optical image. Both tubes had constant microbubble flow in opposite directions. cally for high volumetric rate 3-D super-resolution ultrasound imaging based on a density-tapered spiral layout [37], [38]. The capability of the 2-D sparse array for 3-D SR-US imaging was demonstrated in simulations and experiments. II. MATERIALS AND M ETHODS Fig. 1. Layout of the 2-D sparse array with red and green circles showing the chosen elements. The pitch between consecutive elements in the x and y A. 2-D Sparse Array directions is 300 m. Empty rows (9, 18, and 27) are due to manufacturing A 2-D sparse array was designed by selecting 512 elements limitations and are not related to the density-tapered 2-D spiral method. from a 32  35 gridded layout of a 2D matrix array (Vermon S.A., Tours, France) as shown in Fig. 1. It was fabricated with overlapping imaging region of two orthogonal transducers at an individual element size of 300  300 m, center frequency the focus. They used two identical linear arrays to image sub- of 3.7 MHz and a bandwidth of 60%. In the y direction, diffraction cellulose tubes using amplitude modulated plane- row numbers 9, 18, and 27 were intentionally left blank for wave transmission at 3 MHz with a frame rate of 400 Hz [22]. wiring, hence the total number of available elements is 1024. The development of high-speed programmable ultrasound The method to select the location of sparse array elements systems and 2-D arrays created new opportunities for volumet- is based on the density-tapered 2-D spiral layout [37]. This ric imaging with high spatio-temporal resolution. In parallel method arranges the elements according to seeds generated to these hardware developments, novel 3-D imaging methods from Fermats spiral function with an additional spatial density based on small numbers of transmit-receive pairs enabled modulation to reduce the side lobes of the transmitted beam a more reliable visualization of tissue volumes [23], the profile. This deterministic, aperiodic, and balanced positioning analysis of fast and complex blood flow in 3-D [24]–[27], procedure guarantees uniform performance over a wide range the characterization of mechanical properties of tissue by 4- of imaging angles. D shear-wave imaging [23], [28], the tracking of the pulse It is not possible to connect all 512 elements to a single wave propagation along the arterial wall [29], the estima- ultrasound probe adapter. Therefore, two sparse array layouts, tion of 4-D tissue motion [30], and other in vivo transient hereinafter referred to as Aperture#1 and Aperture#2, were events. These technological advances in 3-D imaging also offer designed as shown with red and green elements in Fig. 1. new opportunities for SR-US. Although volumetric imaging Both sparse arrays were based on an ungridded, 10.4-mm- methods have already shown significant benefits for various wide spiral with 256 seeds [37], whose density tapering was ultrasound imaging applications, 3-D imaging with large 2-D modulated according to a 50%-Tukey window. The elements arrays requires a high number of hardware channels and huge belonging to Aperture#1 were selected among those of the computational power. Vermon 2-D matrix array, by activating the available ele- In this study, we demonstrate the feasibility of 3D super- ments whose positions were closest to the ideal positions resolution imaging and super-resolved flow velocity mapping of the ungridded spiral. Similarly, the elements belonging to using a density-tapered sparse array instead of a full 2-D array Aperture#2 were also selected among those of the Vermon to reduce the number of channels and hence the amount of matrix array, but excluding those that were already assigned data while maintaining the frame rate. A similar approach to Aperture#1. The two layouts were connected to two inde- was in previous non-super-resolution studies on minimally pendent connectors (model DLP 408, ITT Cannon, CA, USA) redundant 2-D arrays [31] and sparse 2-D arrays [32]–[36], but so that an approximation of a 256-element density tapered uses a greater number of elements to improve transmit power spiral array could be driven by an independent ULA-OP 256 and receive sensitivity. Our method significantly differs from system [39], [40]. Moreover, by synchronizing two ULA- row-column addressing and multiplexing approaches since it OP 256 systems to simultaneously control the two layouts, maintains simultaneous access to all probe elements through a 512-element dense array (Aperture#1 + Aperture#2) with independent channels. The sparse array was designed specifi- integrated Tukey apodization could be driven. 4 Fig. 5. Isosurface of the 3-D ultrasound image is plotted in copper at -10 dB level. 2-D maximum intensity projections with a 30 dB dynamic range is overlaid on the volumetric image. 5 mm depth (Fig. 3 (top-left)) is a combination of a plane wave and a dispersed tail, which is a result of missing rows. At the depth of 10 mm, as shown in Fig. 3 (top-right), the tail resembles a superposition of multiple edge waves as a result of discontinuities in the array. At this point, the radiated beam shape is not suitable for generating a good quality image. Around 15 mm depth, as shown in Fig. 3 (bottom- left), the tail becomes less prominent and edge waves diminish below 14 dB; however, it can still produce image artefacts as demonstrated by [47]. Further away from the transducer, the residual waves behind the wavefront disappear and the ultrasound field becomes more uniform, which is suitable for plane wave imaging after 20 mm depth as shown in Fig. 3 (bottom-right). The 3-D simulations displayed in Fig. 4 also support the same conclusion: due to the choice of elements and three unconnected rows, the ultrasound field is not uniform for the first 20 mm. B. 3-D Super-resolution Experimental Results Before performing the experiments on a cellulose microvas- Fig. 6. (Top) 3-D super-resolution image of the two 200 m tubes arranged in a double helix shape. Depth-encoded colorscale is added to improve the culature phantom, the imaging performance of the 2-D sparse visualization. (Bottom) Velocity and direction (positive towards increasing y array was characterized with a point target using the tip of a direction) of tracked microbubbles. 100 m metal wire. The full-width-half-maximum (FWHM) of the 3-D B-mode point-spread-function (PSF) was measured as 793, 772, and 499 m in the x, y & z directions respectively visualization of the structure displayed in copper color, 2-D maximum-intensity-projection (MIP) slices in three directions by using linear interpolation [48]. The localization precision were plotted. It was not possible to visualize the two separate was measured to be the standard deviation of the localization 200 m tubes in these MIP slices or in the volumetric image. positions over 100 frames. The 3-D super-localization preci- sion of the overall system at 25 mm was found to be 18 m Fig. 6 (top) shows the 3-D super-resolved volume of the in the worst imaging plane (x direction), where the imaging imaged sub-wavelength structures by combining localizations from all acquired frames. A total of 2824 microbubbles were wavelength is 404 m in water at 25 C. The volumetric B-mode image of two cellulose tubes with- localized within the 6000 volumes after compounding. Due out microbubble flow is shown in Fig. 5. In addition to the 3-D to the large number of localizations, the 3-D structure of the 5 Velocity Tracks (mm/s) -200 -100 0 0 -5 -10 -200 -100 0 100 200 Distance ( m) Horizontal Projection Vertical Projection 0.5 -300 -200 -100 0 100 200 300 Distance ( m) Fig. 7. (Top) Figure shows the MIP of the B-mode image belonging to a Fig. 8. (Top) Figure shows the velocity tracks belonging to a 1 mm long 1 mm long section of the tube projected into a 2D plane that is orthogonal to section of the tube projected into a 2D plane that is orthogonal to the direction the direction of the flow. The super-resolution image was projected into the of the flow. Black circle represents the 200 m tube circumference. (Bottom) same 2D plane and overlaid on the B-mode image. Black circle represents The FWHM of the tube is measured as 130 m and 75 m from 1-D the 200 m tube circumference. (Bottom) The B-mode FWHM of the tube projections in the horizontal and vertical directions of the top panel plot is measured as 1380 m and 590 m from 1-D projections in the horizontal respectively. and vertical directions of the top panel plot respectively. The super-resolution FWHM of the tube is measured as 134 m and 108 m from 1-D projections in the horizontal and vertical directions of the top panel plot respectively. and 200 m respectively. In the B-mode image two touching tubes appeared as a single scattering object with a FWHM tubes cannot be clearly visualized in a single 2-D image. To of 1380 m and 590 m in the horizontal and vertical 1-D improve the visualization, 3-D SR-US images are plotted with projections respectively. depth information color-coded in the image. Microbubble tracking effectively worked as another layer of Fig. 6 (bottom) displays the velocity profiles of tracked filtering by removing the potentially erroneous non-traceable microbubbles. Only 1076 microbubble-pairs out of 2824 mi- super-localizations. Around the same section of the tube shown crobubbles were traceable from consecutive frames using a in Fig. 7 (top) and Fig. 8 (top), 96% of the microbubble nearest-neighbor method. Using these microbubble tracks, two velocity tracks were within a diameter of 200 m, which sub-wavelength tubes with opposing flows were easily distin- was 89% for the super-localized microbubbles without velocity guishable by color-coding the direction of their velocity vec- tracking. The FWHM of a single tube appeared as 130 m and tors. The percentage of microbubbles that were followed over 75 m from the projection of velocity tracks to horizontal and two or more volumes of 76.2% was attributed to microbubble vertical directions, as plotted in Fig. 8 (bottom). destruction, which can also be observed in the super-resolution images showing a high number of microbubble localizations Microbubble tracking made the separation between the tubes at the inlet and almost no localizations at the outlet of the more clear when tubes are in contact around the central section tubes, see Fig. 6. of the 3-D SR-US and velocity maps displayed in Fig. 6. The thickness of the imaged tubes was measured at the The velocity profiles of microbubbles at this location with inlet where the tube is clearly isolated in the 3-D SR-US two touching tubes were re-plotted in Fig. 9 (top) for clear image around the [3 mm, -3 mm] coordinates in x and y visualization. From this plotted volume, a 1 mm long section respectively. To perform the thickness measurement, a 1 mm of the tube was chosen and projected into a 2-D plane that long section of the imaged tube was chosen and projected into is shown in Fig. 9 (middle). In this 2-D maximum intensity a 2-D plane that is orthogonal to the direction of the tube as projection, the weighted center locations between positive and shown in Fig. 7 (top) both for B-mode and 3-D SR-US images. negative velocity tracks have a distance of 239 m. The 1- Fig. 7 (bottom) shows the 1-D projections in the horizontal D projection of the velocity tracks had a FWHM of 122 m and vertical directions where the FWHM of the super-resolved and 115 m for positive and negative flows respectively with a tube was measured as 134 m and 108 m and the 20 dB peak-to-peak distance of 190 m between two opposing tracks width of the super-resolved tube was measured as 213 m as plotted in Fig. 9 (bottom). Normalized Amplitude Distance ( m) 6 these systems had 1 of 2 transducer elements multiplexed in re- ception [23], [28]. Many researchers have developed methods to use a large number of active elements with fewer channels (usually between 128 and 256) to reduce the cost and com- plexity of the ultrasound systems and the probes. It has been demonstrated in several studies that row-column addressed matrix arrays [47], [51]–[53], microbeamformers [54]–[56] and channel multiplexing can be an alternative to fully ad- dressed 2-D matrix arrays. However, these methods have less flexibility and limitations due to the elements not being continuously connected to the ultrasound system. In this paper, a 2-D sparse array imaging probe has been developed for 3-D super-resolution imaging. This has ad- dressed the main limitation of the existing 2-D imaging of poor spatial resolution in the elevational plane. In addition to super- resolution imaging, 3-D velocity mapping was implemented to reveal the flow inside the microstructures. Using the sparse Velocity Tracks (mm/s) array approach instead of a full matrix array reduced the -500 10 number of channels to half, and hence the connection issues, cost and data size while still achieving the same volumetric acquisition speed since all elements of 2-D spiral array are always connected to the system. Although this can sacrifice the maximum achievable transmit pressure and receive sensitivity, 0 0 it is not a significant issue with SR-US due to the low pressure -2 required and high sensitivity achievable in microbubble imag- -4 ing. In terms of B-mode image resolution, the axial resolution -6 is comparable, since both arrays have the same bandwidth; -8 while a slightly worse lateral resolution is expected for the -10 sparse array, since the full matrix array has a larger aperture -500 0 500 size. It is hard to distinguish the grating lobes and the side Distance ( m) lobes of a sparse array, but here we consider the unwanted leakage outside the main lobe as grating lobes if it is as a result of element-to-element spacing, and as side lobes if 0.5 it is as a result of finite aperture size. The side lobe and edge wave suppression characteristics of the sparse array will -0.5 outperform an un-apodized full matrix array thanks to the integrated apodization, although the fixed apodization might -1 be a limitation for some applications. Both arrays will have -500 -400 -300 -200 -100 0 100 200 300 400 500 Distance ( m) higher grating lobes in y direction due to the three empty rows. The highest grating lobe of the full matrix array will appear Fig. 9. (Top) 3D velocity profiles of microbubbles re-plotted from Fig. 8 at 8 as high as 17% of the main lobe amplitude, calculated (bottom) to display the details more clearly when two tubes are in contact. using the array factor equation in [57]. A sparse choice of (Middle) Figure shows the velocity tracks belonging to a 1 mm long section of the tube projected into a 2-D plane. (Bottom) The 1-D projection of the middle elements spreads the grating lobes to a wider range due to panel plot towards vertical direction shows that the peak-to-peak distance the irregular placement of elements, where the highest grating between two opposing velocity tracks is 190 m. lobe will appear at 18 as high as 16% of the main lobe amplitude. Using the plane-wave imaging method instead of line-by- IV. D ISCUSSION line scanning increases the temporal resolution of the volumet- A better 3-D image quality may be achieved by using a ric imaging. Faster 3-D image acquisition provides a higher large number of independent array elements with the fastest microbubble localization rate and improves velocity estima- possible volumetric imaging rate; however this requires the tions due to more frequent sampling. However, a microbubble same number of hardware channels as the number of elements travelling with a velocity of 10 mm/s will be exposed to 3000 and the ability to process very large stacks of data. Due to the ultrasound pulses while travelling through the imaging region high cost, full 2-D array imaging using an ultrasound system of 10 mm at a PRF of 3000 Hz. At this insonation rate, even to control very large numbers of independent elements has at a relatively low MI of 0.07 almost all microbubbles were only been used by a few research groups [23], [28], [49], destroyed before reaching the center of the imaging region. [50]. These systems had 1024 channels capable of driving a At this point, a new transmitting strategy was implemented to 32  32 2-D array with at least 4 connectors. Even some of reduce the microbubble destruction rate instead of reducing Normalized Amplitude Distance ( m) 7 the PRF, which may introduce artefacts on 9 compounded REFERENCES volumes due to moving microbubbles, or ultrasound pressure, [1] O. M. Viessmann, R. J. Eckersley, K. Christensen-Jeffries, M.-X. which will decrease the SNR and localization precision. To Tang, and C. Dunsby, “Acoustic super-resolution with ultrasound and microbubbles,” Phys. Med. Biol, vol. 58, pp. 6447–6458, 2013. improve the microbubble circulation time, a 7 millisecond [2] Y. Desailly, J. Pierre, O. Couture, and M. Tanter, “Resolution limits of pause was added between each compounded volume acquisi- ultrafast ultrasound localization microscopy,” Phys. Med. Biol, vol. 60, tion that took 3 milliseconds. This strategy reduced insonation pp. 8723–8740, 2015. [3] S. Harput, K. Christensen-Jeffries, Y. Li, J. Brown, R. J. Eckersley, rate by 3.3-fold and the volumetric acquisition rate to 100 C. Dunsby, and M.-X. Tang, “Two stage sub-wavelength motion correc- Hz without sacrificing the PRF during compounding. The tion in human microvasculature for ceus imaging,” in IEEE International second advantage of this transmit strategy is a reduction in Ultrasonics Symposium (IUS), 2017, pp. 1–4. [4] S. Harput, K. Christensen-Jeffries, J. Brown, Y. Li, K. J. Williams, the redundant microbubble localizations at the same spatial A. H. Davies, R. J. Eckersley, C. Dunsby, and M. Tang, “Two-stage location without reducing the PRF. In order for microbubbles motion correction for super-resolution ultrasound imaging in human to provide new spatial information in each frame, they must lower limb,” IEEE Trans. Ultrason., Ferroelectr., Freq. Control, vol. 65, no. 5, pp. 803–814, 2018. be re-located by the flow. Especially for microvasculature [5] K. Christensen-Jeffries, R. J. Browning, M.-X. Tang, C. Dunsby, and with slow physiological flow, increasing the frame rate will R. J. Eckersley, “In vivo acoustic super-resolution and super-resolved no longer increase the obtained spatial information but result velocity mapping using microbubbles,” IEEE Trans. Med. Imag., vol. 34, no. 2, pp. 433–440, 2015. in redundant location information. Nevertheless, using high [6] D. Ackermann and G. Schmitz, “Detection and tracking of multiple frame rates may still be valuable for improving the SNR and microbubbles in ultrasound b-mode images,” IEEE Trans. Ultrason., velocity tracking. The proposed transmit strategy improved the Ferroelectr., Freq. Control, vol. 63, no. 1, pp. 72–82, 2016. [7] A. Bar-Zion, C. Tremblay-Darveau, O. Solomon, D. Adam, and Y. C. image quality for this study, but it could not totally prevent the Eldar, “Fast vascular ultrasound imaging with enhanced spatial resolu- microbubble destruction that caused a lower number of local- tion and background rejection,” IEEE Trans. Med. Imag., vol. 36, pp. izations at outlet of the tubes visible in the 3-D SR-US image. 169–180, 2017. [8] J. Foiret, H. Zhang, T. Ilovitsh, L. Mahakian, S. Tam, and K. W. Ferrara, In the future, the relationship between PRF, pause interval, “Ultrasound localization microscopy to image and assess microvascula- microbubble flow velocity, and compounding strategies should ture in a rat kidney,” Scientific Reports, vol. 7, no. 13662, pp. 1–12, be investigated for different applications and physiological [9] S. Harput, K. Christensen-Jeffries, J. Brown, R. J. Eckersley, C. Dunsby, flow rates. and M.-X. Tang, “Localisation of multiple non-isolated microbubbles with frequency decomposition in super-resolution imaging,” in IEEE International Ultrasonics Symposium (IUS), 2017, pp. 1–4. [10] P. Song, J. D. Trzasko, A. Manduca, R. Huang, R. Kadirvel, D. F. V. CONCLUSION Kallmes, and S. Chen, “Improved super-resolution ultrasound microves- sel imaging with spatiotemporal nonlocal means filtering and bipartite The main limitation of localization-based SR-US imaging graph-based microbubble tracking,” IEEE Trans. Ultrason., Ferroelectr., Freq. Control, vol. 65, no. 2, pp. 149–167, 2018. performed in 2-D is the lack of super-resolution in the el- [11] T. Opacic, S. Dencks, B. Theek, M. Piepenbrock, D. Ackermann, A. Rix, evation direction. In this study, this issue was addressed by T. Lammers, E. Stickeler, S. Delorme, G. Schmitz, and F. Kiessling, using a bespoke 2-D sparse array that achieved an estimated “Motion model ultrasound localization microscopy for preclinical and localization precision of 18 m in the worst imaging plane, clinical multiparametric tumor characterization,” Nature Communica- tions, vol. 9, no. 1527, pp. 1–13, 2018. which is approximately 22 times smaller than the wavelength. [12] T. Ilovitsh, A. Ilovitsh, J. Foiret, B. Z. Fite, and K. W. Ferrara, “Acous- Compounded plane wave imaging with a PRF of 3000 Hz tical structured illumination for super-resolution ultrasound imaging,” enabled super-resolution imaging in all spatial directions with Communications Biology, vol. 1, no. 3, 2018. [13] G. P. Luke, A. S. Hannah, and S. Y. Emelianov, “Super-resolution an image acquisition time of 60 seconds. The structure of two ultrasound imaging in vivo with transient laser-activated nanodroplets,” 200 m, smaller than half wavelength, tubes arranged in a Nano Letters, vol. 16, pp. 2556–2559, 2016. double helix shape were super resolved and flow velocities [14] H. Yoon, K. A. Hallam, C. Yoon, and S. Y. Emelianov, “Super-resolution imaging with ultrafast ultrasound imaging of optically triggered per- within these tubes were estimated. 3-D sub-diffraction imaging fluorohexane nanodroplets,” IEEE Trans. Ultrason., Ferroelectr., Freq. was achieved in vitro using the 2D sparse array probe. Control, vol. 65, no. pre-print, pp. 1–1, 2018. [15] G. Zhang, S. Harput, S. Lin, K. Christensen-Jeffries, C. H. Leow, J. Brown, C. Dunsby, R. J. Eckersley, and M.-X. Tang, “Acoustic wave sparsely activated localization microscopy (awsalm): Super-resolution ACKNOW LEDGMENTS ultrasound imaging using acoustic activation and deactivation of nan- odroplets,” Applied Physics Letters, vol. 113, no. 1, p. 014101, 2018. [16] G. Zhang, S. Harput, H. Hu, K. Christensen-Jeffries, J. Zhu, J. Brown, This work was supported mainly by the EPSRC under C. H. Leow, C. Dunsby, R. J. Eckersley, and M.-X. Tang, “Fast acoustic Grant EP/N015487/1 and EP/N014855/1, in part by the wave sparsely activated localization microscopy (fast-awsalm) using octafluoropropane nanodroplets,” in IEEE International Ultrasonics King’s College London (KCL) and Imperial College London Symposium (IUS), 2018, pp. 1–4. EPSRC Centre for Doctoral Training in Medical Imaging [17] O. Couture, V. Hingot, B. Heiles, P. Muleki-Seya, and M. Tanter, (EP/L015226/1), in part by the Wellcome EPSRC Centre for “Ultrasound localization microscopy and super-resolution: A state of the art,” IEEE Trans. Ultrason., Ferroelectr., Freq. Control, vol. 65, no. 8, Medical Engineering at KCL (WT 203148/Z/16/Z), in part by pp. 1304–1320, 2018. the Department of Health through the National Institute for [18] C. Errico, J. Pierre, S. Pezet, Y. Desailly, Z. Lenkei, O. Couture, and Health Research comprehensive Biomedical Research Center M. Tanter, “Ultrafast ultrasound localization microscopy for deep super- resolution vascular imaging,” Nature, vol. 527, pp. 499–507, 2015. Award to Guy’s and St Thomas’ NHS Foundation Trust in [19] S. Lin, G. Zhang, C. H. Leow, and M.-X. Tang, “Effects of microchan- partnership with KCL and King’s College Hospital NHS nel confinement on acoustic vaporisation of ultrasound phase change Foundation Trust, in part by the Graham-Dixon Foundation contrast agents,” Physics in Medicine & Biology, vol. 62, no. 17, pp. and in part by NVIDIA GPU grant. 6884–6898, 2017. 8 [20] M. A. O’Reilly and K. Hynynen, “A super-resolution ultrasound method system for continuous real-time high frame rate imaging,” IEEE Trans. for brain vascular mapping,” Medical Physics, vol. 40, no. 110701, 2013. Ultrason., Ferroelectr., Freq. Control, vol. 64, no. 9, pp. 1276–1284, [21] Y. Desailly, O. Couture, M. Fink, and M. Tanter, “Sono-activated 2017. ultrasound localization microscopy,” Applied Physics Letters, vol. 103, [41] M. Toulemonde, A. Stanziola, Y. Li, S. Lin, M. Butler, W. C. Duncan, no. 174107, 2013. V. Sboros, R. Eckersley, and M.-X. Tang, “Effects of motion on high [22] K. Christensen-Jeffries, J. Brown, P. Aljabar, M.-X. Tang, C. Dunsby, frame rate contrast enhanced echocardiography and its correction,” in and R. J. Eckersley, “3-d in vitro acoustic super-resolution and super- IEEE International Ultrasonics Symposium (IUS), 2017. resolved velocity mapping using microbubbles,” IEEE Trans. Ultrason., [42] C. Demene, T. Deffieux, M. Pernot, B.-F. Osmanski, V. Biran, J.-L. Ferroelectr., Freq. Control, vol. 64, no. 10, pp. 1478–1486, 2017. Gennisson, L.-A. Sieu, A. Bergel, S. Franqui, J.-M. Correas, I. Cohen, [23] J. Provost, C. Papadacci, J. E. Arango, M. Imbault, M. Fink, J.-L. O. Baud, and M. Tanter, “Spatiotemporal clutter filtering of ultrafast Gennisson, M. Tanter, and M. Pernot, “3d ultrafast ultrasound imaging ultrasound data highly increases doppler and fultrasound sensitivity,” in vivo,” Phys. Med. Biol, vol. 59, no. 19, pp. L1–L13, 2014. IEEE Trans Med Imaging, vol. 34, no. 11, pp. 2271–2285, 2015. [24] M. J. Pihl, M. B. Stuart, B. G. Tomov, M. F. Rasmussen, and J. A. [43] A. Stanziola, C. H. Leow, E. Bazigou, P. D. Weinberg, and M. Tang, Jensen, “A transverse oscillation approach for estimation of three- “Asap: Super-contrast vasculature imaging using coherence analysis and dimensional velocity vectors, part ii: experimental validation,” IEEE high frame-rate contrast enhanced ultrasound,” IEEE Transactions on Trans. Ultrason., Ferroelectr., Freq. Control, vol. 61, no. 10, pp. 1608– Medical Imaging, vol. 37, no. 8, pp. 1847–1856, 2018. 1618, 2014. [44] K. Christensen-Jeffries, S. Harput, J. Brown, P. N. T. Wells, P. Aljabar, [25] J. Provost, C. Papadacci, C. Demene, J. Gennisson, M. Tanter, and C. Dunsby, M.-X. Tang, and R. J. Eckersley, “Microbubble axial M. Pernot, “3-d ultrafast doppler imaging applied to the noninvasive localization errors in ultrasonic super-resolution imaging,” IEEE Trans. mapping of blood vessels in vivo,” IEEE Trans. Ultrason., Ferroelectr., Ultrason., Ferroelectr., Freq. Control, vol. 64, no. 11, pp. 1644–1654, Freq. Control, vol. 62, no. 8, pp. 1467–1472, 2015. 2017. [26] M. Correia, J. Provost, M. Tanter, and M. Pernot, “4d ultrafast ultrasound [45] J. Jensen and N. B. Svendsen, “Calculation of pressure fields from flow imaging: in vivo quantification of arterial volumetric flow rate in arbitrarily shaped, apodized, and excited ultrasound transducers,” IEEE a single heartbeat,” Phys. Med. Biol, vol. 61, no. 23, p. L48, 2016. Trans. Ultrason., Ferroelectr., Freq. Control, vol. 39, pp. 262–267, 1992. [27] S. Holbek, C. Ewertsen, H. Bouzari, M. J. Pihl, K. L. Hansen, M. B. [46] J. Jensen, “Field: A program for simulating ultrasound systems,” in Stuart, C. Thomsen, M. B. Nielsen, and J. A. Jensen, “Ultrasonic 3-d Medical & Biological Engineering & Computing, vol. 34, no. 1, 1996, vector flow method for quantitative in vivo peak velocity and flow rate pp. 351–353. estimation,” IEEE Trans. Ultrason., Ferroelectr., Freq. Control, vol. 64, [47] M. F. Rasmussen, T. L. Christiansen, E. V. Thomsen, and J. A. no. 3, pp. 544–554, 2017. Jensen, “3-d imaging using rowcolumn-addressed arrays with integrated [28] J. Gennisson, J. Provost, T. Deffieux, C. Papadacci, M. Imbault, M. Per- apodization part i: Apodization design and line element beamforming,” not, and M. Tanter, “4-d ultrafast shear-wave imaging,” IEEE Trans. IEEE Trans. Ultrason., Ferroelectr., Freq. Control, vol. 62, no. 5, pp. Ultrason., Ferroelectr., Freq. Control, vol. 62, no. 6, pp. 1059–1065, 947–958, 2015. 2015. [48] S. Harput, J. McLaughlan, D. M. Cowell, and S. Freear, “New per- [29] I. Apostolakis, P. Nauleau, C. Papadacci, M. D. McGarry, and E. E. formance metrics for ultrasound pulse compression systems,” in IEEE Konofagou, “Feasibility and validation of 4-d pulse wave imaging International Ultrasonics Symposium (IUS), 2014, pp. 440–443. in phantoms and in vivo,” IEEE Trans. Ultrason., Ferroelectr., Freq. [49] J. A. Jensen, M. F. Holten-Lund, R. T. Nilsson, M. Hansen, U. D. Control, vol. 64, no. 9, pp. 1305–1317, 2017. Larsen, R. P. Domsten, B. G. Tomov, M. B. Stuart, S. I. Nikolov, [30] S. Salles, H. Liebgott, D. Garcia, and D. Vray, “Full 3-d transverse oscil- M. J. Pihl, Y. Du, J. H. Rasmussen, and M. F. Rasmussen, “Sarus: lations: a method for tissue motion estimation,” IEEE Trans. Ultrason., A synthetic aperture real-time ultrasound system,” IEEE Transactions Ferroelectr., Freq. Control, vol. 62, no. 8, pp. 1473–1485, 2015. on Ultrasonics, Ferroelectrics, and Frequency Control, vol. 60, no. 9, [31] M. Karaman, I. O. Wygant, O. Oralkan, and B. T. Khuri-Yakub, pp. 1838–1852, 2013. “Minimally redundant 2-d array designs for 3-d medical ultrasound [50] L. Petrusca, F. Varray, R. Souchon, A. Bernard, J.-Y. Chapelon, H. Lieb- imaging,” IEEE Trans. Med. Imag., vol. 28, no. 7, pp. 1051–1061, 2009. gott, W. A. NDjin, and M. Viallon, “Fast volumetric ultrasound b-mode [32] A. Austeng and S. Holm, “Sparse 2-d arrays for 3-d phased array and doppler imaging with a new high-channels density platform for imaging - design methods,” IEEE Trans. Ultrason., Ferroelectr., Freq. advanced 4d cardiac imaging/therapy,” Applied Sciences, vol. 8, no. 2, Control, vol. 49, no. 8, pp. 1073–1086, 2002. 2018. [33] B. Diarra, M. Robini, P. Tortoli, C. Cachard, and H. Liebgott, “Design of [51] A. Savoia, V. Bavaro, G. Caliano, A. Caronti, R. Carotenuto, P. Gatta, optimal 2-d nongrid sparse arrays for medical ultrasound,” IEEE Trans. C. Longo, and M. Pappalardo, “Crisscross 2d cmut array: Beamform- Biomed. Eng, vol. 60, no. 11, pp. 3093–3102, 2013. ing strategy and synthetic 3d imaging results,” in IEEE International [34] E. Roux, A. Ramalli, P. Tortoli, C. Cachard, M. Robini, and H. Liebgott, Ultrasonics Symposium (IUS), 2007, pp. 1514–1517. “2-d ultrasound sparse arrays multidepth radiation optimization using [52] S. Holbek, T. L. Christiansen, M. B. Stuart, C. Beers, E. V. Thomsen, simulated annealing and spiral-array inspired energy functions,” IEEE and J. A. Jensen, “3-d vector flow estimation with rowcolumn-addressed Trans. Ultrason., Ferroelectr., Freq. Control, vol. 63, no. 12, pp. 2138– arrays,” IEEE Trans. Ultrason., Ferroelectr., Freq. Control, vol. 63, 2149, 2016. no. 11, pp. 1799–1814, 2016. [35] E. Roux, A. Ramalli, H. Liebgott, C. Cachard, M. Robini, and P. Tortoli, [53] M. Flesch, M. Pernot, J. Provost, G. Ferin, A. Nguyen-Dinh, M. Tanter, “Wideband 2-d array design optimization with fabrication constraints and T. Deffieux, “4d in vivo ultrafast ultrasound imaging using a row- for 3-d us imaging,” IEEE Trans. Ultrason., Ferroelectr., Freq. Control, column addressed matrix and coherently-compounded orthogonal plane vol. 64, no. 1, pp. 108–125, 2017. waves,” Phys. Med. Biol, vol. 62, no. 11, p. 4571, 2017. [36] E. Roux, F. Varray, L. Petrusca, C. Cachard, P. Tortoli, and H. Liebgott, [54] B. J. Savord, “Beamforming methods and apparatus for three- “Experimental 3-d ultrasound imaging with 2-d sparse arrays using dimensional ultrasound imaging using two-dimensional transducer ar- focused and diverging waves,” Scientific Reports, vol. 8, no. 9108, pp. ray,” USA Patent US6 013 032A, 2000. 1–12, 2018. [55] G. Matrone, A. S. Savoia, M. Terenzi, G. Caliano, F. Quaglia, and [37] A. Ramalli, E. Boni, A. S. Savoia, and P. Tortoli, “Density-tapered spiral G. Magenes, “A volumetric cmut-based ultrasound imaging system sim- arrays for ultrasound 3-d imaging,” IEEE Trans. Ultrason., Ferroelectr., ulator with integrated reception and -beamforming electronics models,” Freq. Control, vol. 62, no. 8, pp. 1580–1588, 2015. IEEE Trans. Ultrason., Ferroelectr., Freq. Control, vol. 61, no. 5, pp. [38] S. Harput, K. Christensen-Jeffries, J. Brown, J. Zhu, G. Zhang, C. H. 792–804, 2014. Leow, M. Toulemonde, A. Ramalli, E. Boni, P. Tortoli, R. J. Eckersley, [56] P. Santos, G. U. Haugen, L. Lvstakken, E. Samset, and J. Dhooge, C. Dunsby, and M.-X. Tang, “3-d super-resolution ultrasound imaging “Diverging wave volumetric imaging using subaperture beamforming,” using a 2-d sparse array with high volumetric imaging rate,” in IEEE IEEE Trans. Ultrason., Ferroelectr., Freq. Control, vol. 63, no. 12, pp. International Ultrasonics Symposium (IUS), 2018, pp. 1–4. 2114–2124, 2016. [39] E. Boni, L. Bassi, A. Dallai, F. Guidi, V. Meacci, A. Ramalli, S. Ricci, [57] S. Harput and A. Bozkurt, “Ultrasonic phased array device for acoustic and P. Tortoli, “Ula-op 256: A 256-channel open scanner for develop- imaging in air,” Sensors Journal, IEEE, vol. 8, no. 11, pp. 1755–1762, ment and real-time implementation of new ultrasound methods,” IEEE 2008. Trans. Ultrason., Ferroelectr., Freq. Control, vol. 63, no. 10, pp. 1488– 1495, 2016. [40] E. Boni, L. Bassi, A. Dallai, V. Meacci, A. Ramalli, M. Scaringella, F. Guidi, S. Ricci, and P. Tortoli, “Architecture of an ultrasound http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Physics arXiv (Cornell University)

3-D Super-Resolution Ultrasound (SR-US) Imaging with a 2-D Sparse Array

3-D Super-Resolution Ultrasound (SR-US) Imaging with a 2-D Sparse Array

3-D Super-Resolution Ultrasound (SR-US) Imaging with a 2-D Sparse Array Sevan Harput, Kirsten Christensen-Jeffries, Alessandro Ramalli, Jemma Brown, Jiaqi Zhu, Ge Zhang, Chee Hau Leow, Matthieu Toulemonde, Enrico Boni, Piero Tortoli, Robert J. Eckersley , Chris Dunsby , and Meng-Xing Tang Abstract—High frame rate 3-D ultrasound imaging technology pre-clinical studies using microbubbles [5]–[12] and nan- combined with super-resolution processing method can visualize odroplets [13]–[16]. These studies generated super-resolved 3-D microvascular structures by overcoming the diffraction images of 3-D structures using 1-D ultrasound arrays where limited resolution in every spatial direction. However, 3-D super- super-resolution cannot be achieved in the elevational di- resolution ultrasound imaging using a full 2-D array requires a rection. In addition to this, out-of-plane motion cannot be system with large number of independent channels, the design of which might be impractical due to the high cost, complexity, compensated for when data is only acquired in 2-D. However, and volume of data produced. with the implementation of 3-D SR-US imaging using a 2- In this study, a 2-D sparse array was designed and fabricated D array, diffraction limited resolution can be overcome in with 512 elements chosen from a density-tapered 2-D spiral every direction and there is then the potential for 3-D motion layout. High frame rate volumetric imaging was performed using tracking and correction. two synchronized ULA-OP 256 research scanners. Volumetric images were constructed by coherently compounding 9-angle Many studies have contributed to the development of SR- plane waves acquired in 3 milliseconds at a pulse repetition US imaging methods by improving the localization precision, frequency of 3000 Hz. To allow microbubbles sufficient time to reducing the acquisition time, increasing microbubble tracking move between consequent compounded volumetric frames, a 7- accuracy, and extending the super-resolution into the third millisecond delay was introduced after each volume acquisition. dimension. These developments are explained in detail by a This reduced the effective volume acquisition speed to 100 Hz and the total acquired data size by 3.3-fold. Localization-based 3- recent review [17]. Researchers mainly employed two different D super-resolution images of two touching sub-wavelength tubes approaches to generate a super-resolution image of a volume were generated from 6000 volumes acquired in 60 seconds. In by mechanically scanning the volume with a linear probe conclusion, this work demonstrates the feasibility of 3D super- and stacking 2-D SR-US images, or by using arrays that resolution imaging and super-resolved velocity mapping using a can acquire volumetric information electronically. Errico et customized 2D sparse array transducer. al. have taken steps towards 3-D with a coronal scan of an entire rat brain by using 128 elements of a custom-built linear I. I NTRODUCTION array at a frequency of 15 MHz. Motion of the probe was Visualization of the microvasculature beyond the diffraction controlled with a micro-step motor to generate 2-D super- limited resolution has been achieved by localizing spatially resolution images over different imaging planes at a frame isolated microbubbles through multiple frames. In the absence rate of 500 Hz [18]. Lin et al. performed a 3-D mechanical of tissue and probe motion, localization precision determines scan of a rat FSA tumor using a linear array mounted on the maximum achievable resolution, which can be on the a motorized precision motion stage synchronized with the order of several micrometers at clinical ultrasound frequen- imaging system. They generated 3-D super-resolution images cies [1], [2]. If motion is present and subsequently corrected by calculating the maximum intensity projection from all 2- post-acquisition, then the motion correction accuracy can D super-resolution slices, acquired using plane-wave imaging limit the achievable spatial resolution [3], [4]. Researchers with a frame rate of 500 Hz [19]. Although sub-diffraction demonstrated the use of 2-D super-resolution ultrasound (SR- imaging has not been published using a 2-D imaging probe US) imaging in many different controlled experiments and with a high volumetric imaging rate, 3-D super-resolution has been achieved by previous studies. O’Reilly and Hynynen used S. Harput, J. Zhu, G. Zhang, C. H. Leow, M. Toulemonde, and M. X. Tang a subset of 128 elements from a 1372-element hemispherical are with the ULIS Group, Department of Bioengineering, Imperial College London, London, SW7 2AZ, UK. transcranial therapy array at a rate of 10 Hz. They generated 3- K. Christensen-Jeffries, J. Brown, and R. J. Eckersley are with the Biomed- D super-resolution images of a spiral tube phantom through an ical Engineering Department, Division of Imaging Sciences, King’s College ex vivo human skullcap at an imaging center frequency of 612 London, SE1 7EH, London, UK. A. Ramalli is with the Department of Information Engineering, University kHz [20]. Desailly et al. implemented a plane wave ultrafast of Florence, 50139 Florence, IT and Lab. on Cardiovascular Imaging & imaging method using an ultrasound clinical scanner with Dynamics, Dept. of Cardiovascular Sciences, KU Leuven, Leuven, Belgium. 128 fully programmable emission-reception channels. They E. Boni, P. Tortoli are with the Department of Information Engineering, University of Florence, 50139 Florence, IT. placed 2 parallel series of 64 transducers to image microflu- C. Dunsby is with the Department of Physics and the Centre for Pathology, idic channels and obtained 3-D super-localization by fitting Imperial College London, London, SW7 2AZ, UK. parallel parabolas in the elevation direction [21]. Christensen- These authors contributed equally to this work. E-mail: s.harput@imperial.ac.uk and mengxing.tang@imperial.ac.uk Jeffries et al. generated volumetric 3-D super-resolution at the arXiv:1902.01608v1 [physics.med-ph] 5 Feb 2019 2 Fig. 2. Optical image of two 200 m cellulose tubes arranged in a double helix pattern. To create this pattern, two tubes were wrapped around each other which created contact-points that are visible in the optical image. Both tubes had constant microbubble flow in opposite directions. cally for high volumetric rate 3-D super-resolution ultrasound imaging based on a density-tapered spiral layout [37], [38]. The capability of the 2-D sparse array for 3-D SR-US imaging was demonstrated in simulations and experiments. II. MATERIALS AND M ETHODS Fig. 1. Layout of the 2-D sparse array with red and green circles showing the chosen elements. The pitch between consecutive elements in the x and y A. 2-D Sparse Array directions is 300 m. Empty rows (9, 18, and 27) are due to manufacturing A 2-D sparse array was designed by selecting 512 elements limitations and are not related to the density-tapered 2-D spiral method. from a 32  35 gridded layout of a 2D matrix array (Vermon S.A., Tours, France) as shown in Fig. 1. It was fabricated with overlapping imaging region of two orthogonal transducers at an individual element size of 300  300 m, center frequency the focus. They used two identical linear arrays to image sub- of 3.7 MHz and a bandwidth of 60%. In the y direction, diffraction cellulose tubes using amplitude modulated plane- row numbers 9, 18, and 27 were intentionally left blank for wave transmission at 3 MHz with a frame rate of 400 Hz [22]. wiring, hence the total number of available elements is 1024. The development of high-speed programmable ultrasound The method to select the location of sparse array elements systems and 2-D arrays created new opportunities for volumet- is based on the density-tapered 2-D spiral layout [37]. This ric imaging with high spatio-temporal resolution. In parallel method arranges the elements according to seeds generated to these hardware developments, novel 3-D imaging methods from Fermats spiral function with an additional spatial density based on small numbers of transmit-receive pairs enabled modulation to reduce the side lobes of the transmitted beam a more reliable visualization of tissue volumes [23], the profile. This deterministic, aperiodic, and balanced positioning analysis of fast and complex blood flow in 3-D [24]–[27], procedure guarantees uniform performance over a wide range the characterization of mechanical properties of tissue by 4- of imaging angles. D shear-wave imaging [23], [28], the tracking of the pulse It is not possible to connect all 512 elements to a single wave propagation along the arterial wall [29], the estima- ultrasound probe adapter. Therefore, two sparse array layouts, tion of 4-D tissue motion [30], and other in vivo transient hereinafter referred to as Aperture#1 and Aperture#2, were events. These technological advances in 3-D imaging also offer designed as shown with red and green elements in Fig. 1. new opportunities for SR-US. Although volumetric imaging Both sparse arrays were based on an ungridded, 10.4-mm- methods have already shown significant benefits for various wide spiral with 256 seeds [37], whose density tapering was ultrasound imaging applications, 3-D imaging with large 2-D modulated according to a 50%-Tukey window. The elements arrays requires a high number of hardware channels and huge belonging to Aperture#1 were selected among those of the computational power. Vermon 2-D matrix array, by activating the available ele- In this study, we demonstrate the feasibility of 3D super- ments whose positions were closest to the ideal positions resolution imaging and super-resolved flow velocity mapping of the ungridded spiral. Similarly, the elements belonging to using a density-tapered sparse array instead of a full 2-D array Aperture#2 were also selected among those of the Vermon to reduce the number of channels and hence the amount of matrix array, but excluding those that were already assigned data while maintaining the frame rate. A similar approach to Aperture#1. The two layouts were connected to two inde- was in previous non-super-resolution studies on minimally pendent connectors (model DLP 408, ITT Cannon, CA, USA) redundant 2-D arrays [31] and sparse 2-D arrays [32]–[36], but so that an approximation of a 256-element density tapered uses a greater number of elements to improve transmit power spiral array could be driven by an independent ULA-OP 256 and receive sensitivity. Our method significantly differs from system [39], [40]. Moreover, by synchronizing two ULA- row-column addressing and multiplexing approaches since it OP 256 systems to simultaneously control the two layouts, maintains simultaneous access to all probe elements through a 512-element dense array (Aperture#1 + Aperture#2) with independent channels. The sparse array was designed specifi- integrated Tukey apodization could be driven. 4 Fig. 5. Isosurface of the 3-D ultrasound image is plotted in copper at -10 dB level. 2-D maximum intensity projections with a 30 dB dynamic range is overlaid on the volumetric image. 5 mm depth (Fig. 3 (top-left)) is a combination of a plane wave and a dispersed tail, which is a result of missing rows. At the depth of 10 mm, as shown in Fig. 3 (top-right), the tail resembles a superposition of multiple edge waves as a result of discontinuities in the array. At this point, the radiated beam shape is not suitable for generating a good quality image. Around 15 mm depth, as shown in Fig. 3 (bottom- left), the tail becomes less prominent and edge waves diminish below 14 dB; however, it can still produce image artefacts as demonstrated by [47]. Further away from the transducer, the residual waves behind the wavefront disappear and the ultrasound field becomes more uniform, which is suitable for plane wave imaging after 20 mm depth as shown in Fig. 3 (bottom-right). The 3-D simulations displayed in Fig. 4 also support the same conclusion: due to the choice of elements and three unconnected rows, the ultrasound field is not uniform for the first 20 mm. B. 3-D Super-resolution Experimental Results Before performing the experiments on a cellulose microvas- Fig. 6. (Top) 3-D super-resolution image of the two 200 m tubes arranged in a double helix shape. Depth-encoded colorscale is added to improve the culature phantom, the imaging performance of the 2-D sparse visualization. (Bottom) Velocity and direction (positive towards increasing y array was characterized with a point target using the tip of a direction) of tracked microbubbles. 100 m metal wire. The full-width-half-maximum (FWHM) of the 3-D B-mode point-spread-function (PSF) was measured as 793, 772, and 499 m in the x, y & z directions respectively visualization of the structure displayed in copper color, 2-D maximum-intensity-projection (MIP) slices in three directions by using linear interpolation [48]. The localization precision were plotted. It was not possible to visualize the two separate was measured to be the standard deviation of the localization 200 m tubes in these MIP slices or in the volumetric image. positions over 100 frames. The 3-D super-localization preci- sion of the overall system at 25 mm was found to be 18 m Fig. 6 (top) shows the 3-D super-resolved volume of the in the worst imaging plane (x direction), where the imaging imaged sub-wavelength structures by combining localizations from all acquired frames. A total of 2824 microbubbles were wavelength is 404 m in water at 25 C. The volumetric B-mode image of two cellulose tubes with- localized within the 6000 volumes after compounding. Due out microbubble flow is shown in Fig. 5. In addition to the 3-D to the large number of localizations, the 3-D structure of the 5 Velocity Tracks (mm/s) -200 -100 0 0 -5 -10 -200 -100 0 100 200 Distance ( m) Horizontal Projection Vertical Projection 0.5 -300 -200 -100 0 100 200 300 Distance ( m) Fig. 7. (Top) Figure shows the MIP of the B-mode image belonging to a Fig. 8. (Top) Figure shows the velocity tracks belonging to a 1 mm long 1 mm long section of the tube projected into a 2D plane that is orthogonal to section of the tube projected into a 2D plane that is orthogonal to the direction the direction of the flow. The super-resolution image was projected into the of the flow. Black circle represents the 200 m tube circumference. (Bottom) same 2D plane and overlaid on the B-mode image. Black circle represents The FWHM of the tube is measured as 130 m and 75 m from 1-D the 200 m tube circumference. (Bottom) The B-mode FWHM of the tube projections in the horizontal and vertical directions of the top panel plot is measured as 1380 m and 590 m from 1-D projections in the horizontal respectively. and vertical directions of the top panel plot respectively. The super-resolution FWHM of the tube is measured as 134 m and 108 m from 1-D projections in the horizontal and vertical directions of the top panel plot respectively. and 200 m respectively. In the B-mode image two touching tubes appeared as a single scattering object with a FWHM tubes cannot be clearly visualized in a single 2-D image. To of 1380 m and 590 m in the horizontal and vertical 1-D improve the visualization, 3-D SR-US images are plotted with projections respectively. depth information color-coded in the image. Microbubble tracking effectively worked as another layer of Fig. 6 (bottom) displays the velocity profiles of tracked filtering by removing the potentially erroneous non-traceable microbubbles. Only 1076 microbubble-pairs out of 2824 mi- super-localizations. Around the same section of the tube shown crobubbles were traceable from consecutive frames using a in Fig. 7 (top) and Fig. 8 (top), 96% of the microbubble nearest-neighbor method. Using these microbubble tracks, two velocity tracks were within a diameter of 200 m, which sub-wavelength tubes with opposing flows were easily distin- was 89% for the super-localized microbubbles without velocity guishable by color-coding the direction of their velocity vec- tracking. The FWHM of a single tube appeared as 130 m and tors. The percentage of microbubbles that were followed over 75 m from the projection of velocity tracks to horizontal and two or more volumes of 76.2% was attributed to microbubble vertical directions, as plotted in Fig. 8 (bottom). destruction, which can also be observed in the super-resolution images showing a high number of microbubble localizations Microbubble tracking made the separation between the tubes at the inlet and almost no localizations at the outlet of the more clear when tubes are in contact around the central section tubes, see Fig. 6. of the 3-D SR-US and velocity maps displayed in Fig. 6. The thickness of the imaged tubes was measured at the The velocity profiles of microbubbles at this location with inlet where the tube is clearly isolated in the 3-D SR-US two touching tubes were re-plotted in Fig. 9 (top) for clear image around the [3 mm, -3 mm] coordinates in x and y visualization. From this plotted volume, a 1 mm long section respectively. To perform the thickness measurement, a 1 mm of the tube was chosen and projected into a 2-D plane that long section of the imaged tube was chosen and projected into is shown in Fig. 9 (middle). In this 2-D maximum intensity a 2-D plane that is orthogonal to the direction of the tube as projection, the weighted center locations between positive and shown in Fig. 7 (top) both for B-mode and 3-D SR-US images. negative velocity tracks have a distance of 239 m. The 1- Fig. 7 (bottom) shows the 1-D projections in the horizontal D projection of the velocity tracks had a FWHM of 122 m and vertical directions where the FWHM of the super-resolved and 115 m for positive and negative flows respectively with a tube was measured as 134 m and 108 m and the 20 dB peak-to-peak distance of 190 m between two opposing tracks width of the super-resolved tube was measured as 213 m as plotted in Fig. 9 (bottom). Normalized Amplitude Distance ( m) 6 these systems had 1 of 2 transducer elements multiplexed in re- ception [23], [28]. Many researchers have developed methods to use a large number of active elements with fewer channels (usually between 128 and 256) to reduce the cost and com- plexity of the ultrasound systems and the probes. It has been demonstrated in several studies that row-column addressed matrix arrays [47], [51]–[53], microbeamformers [54]–[56] and channel multiplexing can be an alternative to fully ad- dressed 2-D matrix arrays. However, these methods have less flexibility and limitations due to the elements not being continuously connected to the ultrasound system. In this paper, a 2-D sparse array imaging probe has been developed for 3-D super-resolution imaging. This has ad- dressed the main limitation of the existing 2-D imaging of poor spatial resolution in the elevational plane. In addition to super- resolution imaging, 3-D velocity mapping was implemented to reveal the flow inside the microstructures. Using the sparse Velocity Tracks (mm/s) array approach instead of a full matrix array reduced the -500 10 number of channels to half, and hence the connection issues, cost and data size while still achieving the same volumetric acquisition speed since all elements of 2-D spiral array are always connected to the system. Although this can sacrifice the maximum achievable transmit pressure and receive sensitivity, 0 0 it is not a significant issue with SR-US due to the low pressure -2 required and high sensitivity achievable in microbubble imag- -4 ing. In terms of B-mode image resolution, the axial resolution -6 is comparable, since both arrays have the same bandwidth; -8 while a slightly worse lateral resolution is expected for the -10 sparse array, since the full matrix array has a larger aperture -500 0 500 size. It is hard to distinguish the grating lobes and the side Distance ( m) lobes of a sparse array, but here we consider the unwanted leakage outside the main lobe as grating lobes if it is as a result of element-to-element spacing, and as side lobes if 0.5 it is as a result of finite aperture size. The side lobe and edge wave suppression characteristics of the sparse array will -0.5 outperform an un-apodized full matrix array thanks to the integrated apodization, although the fixed apodization might -1 be a limitation for some applications. Both arrays will have -500 -400 -300 -200 -100 0 100 200 300 400 500 Distance ( m) higher grating lobes in y direction due to the three empty rows. The highest grating lobe of the full matrix array will appear Fig. 9. (Top) 3D velocity profiles of microbubbles re-plotted from Fig. 8 at 8 as high as 17% of the main lobe amplitude, calculated (bottom) to display the details more clearly when two tubes are in contact. using the array factor equation in [57]. A sparse choice of (Middle) Figure shows the velocity tracks belonging to a 1 mm long section of the tube projected into a 2-D plane. (Bottom) The 1-D projection of the middle elements spreads the grating lobes to a wider range due to panel plot towards vertical direction shows that the peak-to-peak distance the irregular placement of elements, where the highest grating between two opposing velocity tracks is 190 m. lobe will appear at 18 as high as 16% of the main lobe amplitude. Using the plane-wave imaging method instead of line-by- IV. D ISCUSSION line scanning increases the temporal resolution of the volumet- A better 3-D image quality may be achieved by using a ric imaging. Faster 3-D image acquisition provides a higher large number of independent array elements with the fastest microbubble localization rate and improves velocity estima- possible volumetric imaging rate; however this requires the tions due to more frequent sampling. However, a microbubble same number of hardware channels as the number of elements travelling with a velocity of 10 mm/s will be exposed to 3000 and the ability to process very large stacks of data. Due to the ultrasound pulses while travelling through the imaging region high cost, full 2-D array imaging using an ultrasound system of 10 mm at a PRF of 3000 Hz. At this insonation rate, even to control very large numbers of independent elements has at a relatively low MI of 0.07 almost all microbubbles were only been used by a few research groups [23], [28], [49], destroyed before reaching the center of the imaging region. [50]. These systems had 1024 channels capable of driving a At this point, a new transmitting strategy was implemented to 32  32 2-D array with at least 4 connectors. Even some of reduce the microbubble destruction rate instead of reducing Normalized Amplitude Distance ( m) 7 the PRF, which may introduce artefacts on 9 compounded REFERENCES volumes due to moving microbubbles, or ultrasound pressure, [1] O. M. Viessmann, R. J. Eckersley, K. Christensen-Jeffries, M.-X. which will decrease the SNR and localization precision. To Tang, and C. Dunsby, “Acoustic super-resolution with ultrasound and microbubbles,” Phys. Med. Biol, vol. 58, pp. 6447–6458, 2013. improve the microbubble circulation time, a 7 millisecond [2] Y. Desailly, J. Pierre, O. Couture, and M. Tanter, “Resolution limits of pause was added between each compounded volume acquisi- ultrafast ultrasound localization microscopy,” Phys. Med. Biol, vol. 60, tion that took 3 milliseconds. This strategy reduced insonation pp. 8723–8740, 2015. [3] S. Harput, K. Christensen-Jeffries, Y. Li, J. Brown, R. J. Eckersley, rate by 3.3-fold and the volumetric acquisition rate to 100 C. Dunsby, and M.-X. Tang, “Two stage sub-wavelength motion correc- Hz without sacrificing the PRF during compounding. The tion in human microvasculature for ceus imaging,” in IEEE International second advantage of this transmit strategy is a reduction in Ultrasonics Symposium (IUS), 2017, pp. 1–4. [4] S. Harput, K. Christensen-Jeffries, J. Brown, Y. Li, K. J. Williams, the redundant microbubble localizations at the same spatial A. H. Davies, R. J. Eckersley, C. Dunsby, and M. Tang, “Two-stage location without reducing the PRF. In order for microbubbles motion correction for super-resolution ultrasound imaging in human to provide new spatial information in each frame, they must lower limb,” IEEE Trans. Ultrason., Ferroelectr., Freq. Control, vol. 65, no. 5, pp. 803–814, 2018. be re-located by the flow. Especially for microvasculature [5] K. Christensen-Jeffries, R. J. Browning, M.-X. Tang, C. Dunsby, and with slow physiological flow, increasing the frame rate will R. J. Eckersley, “In vivo acoustic super-resolution and super-resolved no longer increase the obtained spatial information but result velocity mapping using microbubbles,” IEEE Trans. Med. Imag., vol. 34, no. 2, pp. 433–440, 2015. in redundant location information. Nevertheless, using high [6] D. Ackermann and G. Schmitz, “Detection and tracking of multiple frame rates may still be valuable for improving the SNR and microbubbles in ultrasound b-mode images,” IEEE Trans. Ultrason., velocity tracking. The proposed transmit strategy improved the Ferroelectr., Freq. Control, vol. 63, no. 1, pp. 72–82, 2016. [7] A. Bar-Zion, C. Tremblay-Darveau, O. Solomon, D. Adam, and Y. C. image quality for this study, but it could not totally prevent the Eldar, “Fast vascular ultrasound imaging with enhanced spatial resolu- microbubble destruction that caused a lower number of local- tion and background rejection,” IEEE Trans. Med. Imag., vol. 36, pp. izations at outlet of the tubes visible in the 3-D SR-US image. 169–180, 2017. [8] J. Foiret, H. Zhang, T. Ilovitsh, L. Mahakian, S. Tam, and K. W. Ferrara, In the future, the relationship between PRF, pause interval, “Ultrasound localization microscopy to image and assess microvascula- microbubble flow velocity, and compounding strategies should ture in a rat kidney,” Scientific Reports, vol. 7, no. 13662, pp. 1–12, be investigated for different applications and physiological [9] S. Harput, K. Christensen-Jeffries, J. Brown, R. J. Eckersley, C. Dunsby, flow rates. and M.-X. Tang, “Localisation of multiple non-isolated microbubbles with frequency decomposition in super-resolution imaging,” in IEEE International Ultrasonics Symposium (IUS), 2017, pp. 1–4. [10] P. Song, J. D. Trzasko, A. Manduca, R. Huang, R. Kadirvel, D. F. V. CONCLUSION Kallmes, and S. Chen, “Improved super-resolution ultrasound microves- sel imaging with spatiotemporal nonlocal means filtering and bipartite The main limitation of localization-based SR-US imaging graph-based microbubble tracking,” IEEE Trans. Ultrason., Ferroelectr., Freq. Control, vol. 65, no. 2, pp. 149–167, 2018. performed in 2-D is the lack of super-resolution in the el- [11] T. Opacic, S. Dencks, B. Theek, M. Piepenbrock, D. Ackermann, A. Rix, evation direction. In this study, this issue was addressed by T. Lammers, E. Stickeler, S. Delorme, G. Schmitz, and F. Kiessling, using a bespoke 2-D sparse array that achieved an estimated “Motion model ultrasound localization microscopy for preclinical and localization precision of 18 m in the worst imaging plane, clinical multiparametric tumor characterization,” Nature Communica- tions, vol. 9, no. 1527, pp. 1–13, 2018. which is approximately 22 times smaller than the wavelength. [12] T. Ilovitsh, A. Ilovitsh, J. Foiret, B. Z. Fite, and K. W. Ferrara, “Acous- Compounded plane wave imaging with a PRF of 3000 Hz tical structured illumination for super-resolution ultrasound imaging,” enabled super-resolution imaging in all spatial directions with Communications Biology, vol. 1, no. 3, 2018. [13] G. P. Luke, A. S. Hannah, and S. Y. Emelianov, “Super-resolution an image acquisition time of 60 seconds. The structure of two ultrasound imaging in vivo with transient laser-activated nanodroplets,” 200 m, smaller than half wavelength, tubes arranged in a Nano Letters, vol. 16, pp. 2556–2559, 2016. double helix shape were super resolved and flow velocities [14] H. Yoon, K. A. Hallam, C. Yoon, and S. Y. Emelianov, “Super-resolution imaging with ultrafast ultrasound imaging of optically triggered per- within these tubes were estimated. 3-D sub-diffraction imaging fluorohexane nanodroplets,” IEEE Trans. Ultrason., Ferroelectr., Freq. was achieved in vitro using the 2D sparse array probe. Control, vol. 65, no. pre-print, pp. 1–1, 2018. [15] G. Zhang, S. Harput, S. Lin, K. Christensen-Jeffries, C. H. Leow, J. Brown, C. Dunsby, R. J. Eckersley, and M.-X. Tang, “Acoustic wave sparsely activated localization microscopy (awsalm): Super-resolution ACKNOW LEDGMENTS ultrasound imaging using acoustic activation and deactivation of nan- odroplets,” Applied Physics Letters, vol. 113, no. 1, p. 014101, 2018. [16] G. Zhang, S. Harput, H. Hu, K. Christensen-Jeffries, J. Zhu, J. Brown, This work was supported mainly by the EPSRC under C. H. Leow, C. Dunsby, R. J. Eckersley, and M.-X. Tang, “Fast acoustic Grant EP/N015487/1 and EP/N014855/1, in part by the wave sparsely activated localization microscopy (fast-awsalm) using octafluoropropane nanodroplets,” in IEEE International Ultrasonics King’s College London (KCL) and Imperial College London Symposium (IUS), 2018, pp. 1–4. EPSRC Centre for Doctoral Training in Medical Imaging [17] O. Couture, V. Hingot, B. Heiles, P. Muleki-Seya, and M. Tanter, (EP/L015226/1), in part by the Wellcome EPSRC Centre for “Ultrasound localization microscopy and super-resolution: A state of the art,” IEEE Trans. Ultrason., Ferroelectr., Freq. Control, vol. 65, no. 8, Medical Engineering at KCL (WT 203148/Z/16/Z), in part by pp. 1304–1320, 2018. the Department of Health through the National Institute for [18] C. Errico, J. Pierre, S. Pezet, Y. Desailly, Z. Lenkei, O. Couture, and Health Research comprehensive Biomedical Research Center M. Tanter, “Ultrafast ultrasound localization microscopy for deep super- resolution vascular imaging,” Nature, vol. 527, pp. 499–507, 2015. Award to Guy’s and St Thomas’ NHS Foundation Trust in [19] S. Lin, G. Zhang, C. H. Leow, and M.-X. Tang, “Effects of microchan- partnership with KCL and King’s College Hospital NHS nel confinement on acoustic vaporisation of ultrasound phase change Foundation Trust, in part by the Graham-Dixon Foundation contrast agents,” Physics in Medicine & Biology, vol. 62, no. 17, pp. and in part by NVIDIA GPU grant. 6884–6898, 2017. 8 [20] M. A. O’Reilly and K. Hynynen, “A super-resolution ultrasound method system for continuous real-time high frame rate imaging,” IEEE Trans. for brain vascular mapping,” Medical Physics, vol. 40, no. 110701, 2013. Ultrason., Ferroelectr., Freq. Control, vol. 64, no. 9, pp. 1276–1284, [21] Y. Desailly, O. Couture, M. Fink, and M. Tanter, “Sono-activated 2017. ultrasound localization microscopy,” Applied Physics Letters, vol. 103, [41] M. Toulemonde, A. Stanziola, Y. Li, S. Lin, M. Butler, W. C. Duncan, no. 174107, 2013. V. Sboros, R. Eckersley, and M.-X. Tang, “Effects of motion on high [22] K. Christensen-Jeffries, J. Brown, P. Aljabar, M.-X. Tang, C. Dunsby, frame rate contrast enhanced echocardiography and its correction,” in and R. J. Eckersley, “3-d in vitro acoustic super-resolution and super- IEEE International Ultrasonics Symposium (IUS), 2017. resolved velocity mapping using microbubbles,” IEEE Trans. Ultrason., [42] C. Demene, T. Deffieux, M. Pernot, B.-F. Osmanski, V. Biran, J.-L. Ferroelectr., Freq. Control, vol. 64, no. 10, pp. 1478–1486, 2017. Gennisson, L.-A. Sieu, A. Bergel, S. Franqui, J.-M. Correas, I. Cohen, [23] J. Provost, C. Papadacci, J. E. Arango, M. Imbault, M. Fink, J.-L. O. Baud, and M. Tanter, “Spatiotemporal clutter filtering of ultrafast Gennisson, M. Tanter, and M. Pernot, “3d ultrafast ultrasound imaging ultrasound data highly increases doppler and fultrasound sensitivity,” in vivo,” Phys. Med. Biol, vol. 59, no. 19, pp. L1–L13, 2014. IEEE Trans Med Imaging, vol. 34, no. 11, pp. 2271–2285, 2015. [24] M. J. Pihl, M. B. Stuart, B. G. Tomov, M. F. Rasmussen, and J. A. [43] A. Stanziola, C. H. Leow, E. Bazigou, P. D. Weinberg, and M. Tang, Jensen, “A transverse oscillation approach for estimation of three- “Asap: Super-contrast vasculature imaging using coherence analysis and dimensional velocity vectors, part ii: experimental validation,” IEEE high frame-rate contrast enhanced ultrasound,” IEEE Transactions on Trans. Ultrason., Ferroelectr., Freq. Control, vol. 61, no. 10, pp. 1608– Medical Imaging, vol. 37, no. 8, pp. 1847–1856, 2018. 1618, 2014. [44] K. Christensen-Jeffries, S. Harput, J. Brown, P. N. T. Wells, P. Aljabar, [25] J. Provost, C. Papadacci, C. Demene, J. Gennisson, M. Tanter, and C. Dunsby, M.-X. Tang, and R. J. Eckersley, “Microbubble axial M. Pernot, “3-d ultrafast doppler imaging applied to the noninvasive localization errors in ultrasonic super-resolution imaging,” IEEE Trans. mapping of blood vessels in vivo,” IEEE Trans. Ultrason., Ferroelectr., Ultrason., Ferroelectr., Freq. Control, vol. 64, no. 11, pp. 1644–1654, Freq. Control, vol. 62, no. 8, pp. 1467–1472, 2015. 2017. [26] M. Correia, J. Provost, M. Tanter, and M. Pernot, “4d ultrafast ultrasound [45] J. Jensen and N. B. Svendsen, “Calculation of pressure fields from flow imaging: in vivo quantification of arterial volumetric flow rate in arbitrarily shaped, apodized, and excited ultrasound transducers,” IEEE a single heartbeat,” Phys. Med. Biol, vol. 61, no. 23, p. L48, 2016. Trans. Ultrason., Ferroelectr., Freq. Control, vol. 39, pp. 262–267, 1992. [27] S. Holbek, C. Ewertsen, H. Bouzari, M. J. Pihl, K. L. Hansen, M. B. [46] J. Jensen, “Field: A program for simulating ultrasound systems,” in Stuart, C. Thomsen, M. B. Nielsen, and J. A. Jensen, “Ultrasonic 3-d Medical & Biological Engineering & Computing, vol. 34, no. 1, 1996, vector flow method for quantitative in vivo peak velocity and flow rate pp. 351–353. estimation,” IEEE Trans. Ultrason., Ferroelectr., Freq. Control, vol. 64, [47] M. F. Rasmussen, T. L. Christiansen, E. V. Thomsen, and J. A. no. 3, pp. 544–554, 2017. Jensen, “3-d imaging using rowcolumn-addressed arrays with integrated [28] J. Gennisson, J. Provost, T. Deffieux, C. Papadacci, M. Imbault, M. Per- apodization part i: Apodization design and line element beamforming,” not, and M. Tanter, “4-d ultrafast shear-wave imaging,” IEEE Trans. IEEE Trans. Ultrason., Ferroelectr., Freq. Control, vol. 62, no. 5, pp. Ultrason., Ferroelectr., Freq. Control, vol. 62, no. 6, pp. 1059–1065, 947–958, 2015. 2015. [48] S. Harput, J. McLaughlan, D. M. Cowell, and S. Freear, “New per- [29] I. Apostolakis, P. Nauleau, C. Papadacci, M. D. McGarry, and E. E. formance metrics for ultrasound pulse compression systems,” in IEEE Konofagou, “Feasibility and validation of 4-d pulse wave imaging International Ultrasonics Symposium (IUS), 2014, pp. 440–443. in phantoms and in vivo,” IEEE Trans. Ultrason., Ferroelectr., Freq. [49] J. A. Jensen, M. F. Holten-Lund, R. T. Nilsson, M. Hansen, U. D. Control, vol. 64, no. 9, pp. 1305–1317, 2017. Larsen, R. P. Domsten, B. G. Tomov, M. B. Stuart, S. I. Nikolov, [30] S. Salles, H. Liebgott, D. Garcia, and D. Vray, “Full 3-d transverse oscil- M. J. Pihl, Y. Du, J. H. Rasmussen, and M. F. Rasmussen, “Sarus: lations: a method for tissue motion estimation,” IEEE Trans. Ultrason., A synthetic aperture real-time ultrasound system,” IEEE Transactions Ferroelectr., Freq. Control, vol. 62, no. 8, pp. 1473–1485, 2015. on Ultrasonics, Ferroelectrics, and Frequency Control, vol. 60, no. 9, [31] M. Karaman, I. O. Wygant, O. Oralkan, and B. T. Khuri-Yakub, pp. 1838–1852, 2013. “Minimally redundant 2-d array designs for 3-d medical ultrasound [50] L. Petrusca, F. Varray, R. Souchon, A. Bernard, J.-Y. Chapelon, H. Lieb- imaging,” IEEE Trans. Med. Imag., vol. 28, no. 7, pp. 1051–1061, 2009. gott, W. A. NDjin, and M. Viallon, “Fast volumetric ultrasound b-mode [32] A. Austeng and S. Holm, “Sparse 2-d arrays for 3-d phased array and doppler imaging with a new high-channels density platform for imaging - design methods,” IEEE Trans. Ultrason., Ferroelectr., Freq. advanced 4d cardiac imaging/therapy,” Applied Sciences, vol. 8, no. 2, Control, vol. 49, no. 8, pp. 1073–1086, 2002. 2018. [33] B. Diarra, M. Robini, P. Tortoli, C. Cachard, and H. Liebgott, “Design of [51] A. Savoia, V. Bavaro, G. Caliano, A. Caronti, R. Carotenuto, P. Gatta, optimal 2-d nongrid sparse arrays for medical ultrasound,” IEEE Trans. C. Longo, and M. Pappalardo, “Crisscross 2d cmut array: Beamform- Biomed. Eng, vol. 60, no. 11, pp. 3093–3102, 2013. ing strategy and synthetic 3d imaging results,” in IEEE International [34] E. Roux, A. Ramalli, P. Tortoli, C. Cachard, M. Robini, and H. Liebgott, Ultrasonics Symposium (IUS), 2007, pp. 1514–1517. “2-d ultrasound sparse arrays multidepth radiation optimization using [52] S. Holbek, T. L. Christiansen, M. B. Stuart, C. Beers, E. V. Thomsen, simulated annealing and spiral-array inspired energy functions,” IEEE and J. A. Jensen, “3-d vector flow estimation with rowcolumn-addressed Trans. Ultrason., Ferroelectr., Freq. Control, vol. 63, no. 12, pp. 2138– arrays,” IEEE Trans. Ultrason., Ferroelectr., Freq. Control, vol. 63, 2149, 2016. no. 11, pp. 1799–1814, 2016. [35] E. Roux, A. Ramalli, H. Liebgott, C. Cachard, M. Robini, and P. Tortoli, [53] M. Flesch, M. Pernot, J. Provost, G. Ferin, A. Nguyen-Dinh, M. Tanter, “Wideband 2-d array design optimization with fabrication constraints and T. Deffieux, “4d in vivo ultrafast ultrasound imaging using a row- for 3-d us imaging,” IEEE Trans. Ultrason., Ferroelectr., Freq. Control, column addressed matrix and coherently-compounded orthogonal plane vol. 64, no. 1, pp. 108–125, 2017. waves,” Phys. Med. Biol, vol. 62, no. 11, p. 4571, 2017. [36] E. Roux, F. Varray, L. Petrusca, C. Cachard, P. Tortoli, and H. Liebgott, [54] B. J. Savord, “Beamforming methods and apparatus for three- “Experimental 3-d ultrasound imaging with 2-d sparse arrays using dimensional ultrasound imaging using two-dimensional transducer ar- focused and diverging waves,” Scientific Reports, vol. 8, no. 9108, pp. ray,” USA Patent US6 013 032A, 2000. 1–12, 2018. [55] G. Matrone, A. S. Savoia, M. Terenzi, G. Caliano, F. Quaglia, and [37] A. Ramalli, E. Boni, A. S. Savoia, and P. Tortoli, “Density-tapered spiral G. Magenes, “A volumetric cmut-based ultrasound imaging system sim- arrays for ultrasound 3-d imaging,” IEEE Trans. Ultrason., Ferroelectr., ulator with integrated reception and -beamforming electronics models,” Freq. Control, vol. 62, no. 8, pp. 1580–1588, 2015. IEEE Trans. Ultrason., Ferroelectr., Freq. Control, vol. 61, no. 5, pp. [38] S. Harput, K. Christensen-Jeffries, J. Brown, J. Zhu, G. Zhang, C. H. 792–804, 2014. Leow, M. Toulemonde, A. Ramalli, E. Boni, P. Tortoli, R. J. Eckersley, [56] P. Santos, G. U. Haugen, L. Lvstakken, E. Samset, and J. Dhooge, C. Dunsby, and M.-X. Tang, “3-d super-resolution ultrasound imaging “Diverging wave volumetric imaging using subaperture beamforming,” using a 2-d sparse array with high volumetric imaging rate,” in IEEE IEEE Trans. Ultrason., Ferroelectr., Freq. Control, vol. 63, no. 12, pp. International Ultrasonics Symposium (IUS), 2018, pp. 1–4. 2114–2124, 2016. [39] E. Boni, L. Bassi, A. Dallai, F. Guidi, V. Meacci, A. Ramalli, S. Ricci, [57] S. Harput and A. Bozkurt, “Ultrasonic phased array device for acoustic and P. Tortoli, “Ula-op 256: A 256-channel open scanner for develop- imaging in air,” Sensors Journal, IEEE, vol. 8, no. 11, pp. 1755–1762, ment and real-time implementation of new ultrasound methods,” IEEE 2008. Trans. Ultrason., Ferroelectr., Freq. Control, vol. 63, no. 10, pp. 1488– 1495, 2016. [40] E. Boni, L. Bassi, A. Dallai, V. Meacci, A. Ramalli, M. Scaringella, F. Guidi, S. Ricci, and P. Tortoli, “Architecture of an ultrasound
Loading next page...
 
/lp/arxiv-cornell-university/3-d-super-resolution-ultrasound-sr-us-imaging-with-a-2-d-sparse-array-Kpl6jbBiyr

References (71)

ISSN
0885-3010
eISSN
ARCH-3341
DOI
10.1109/TUFFC.2019.2943646
Publisher site
See Article on Publisher Site

Abstract

3-D Super-Resolution Ultrasound (SR-US) Imaging with a 2-D Sparse Array Sevan Harput, Kirsten Christensen-Jeffries, Alessandro Ramalli, Jemma Brown, Jiaqi Zhu, Ge Zhang, Chee Hau Leow, Matthieu Toulemonde, Enrico Boni, Piero Tortoli, Robert J. Eckersley , Chris Dunsby , and Meng-Xing Tang Abstract—High frame rate 3-D ultrasound imaging technology pre-clinical studies using microbubbles [5]–[12] and nan- combined with super-resolution processing method can visualize odroplets [13]–[16]. These studies generated super-resolved 3-D microvascular structures by overcoming the diffraction images of 3-D structures using 1-D ultrasound arrays where limited resolution in every spatial direction. However, 3-D super- super-resolution cannot be achieved in the elevational di- resolution ultrasound imaging using a full 2-D array requires a rection. In addition to this, out-of-plane motion cannot be system with large number of independent channels, the design of which might be impractical due to the high cost, complexity, compensated for when data is only acquired in 2-D. However, and volume of data produced. with the implementation of 3-D SR-US imaging using a 2- In this study, a 2-D sparse array was designed and fabricated D array, diffraction limited resolution can be overcome in with 512 elements chosen from a density-tapered 2-D spiral every direction and there is then the potential for 3-D motion layout. High frame rate volumetric imaging was performed using tracking and correction. two synchronized ULA-OP 256 research scanners. Volumetric images were constructed by coherently compounding 9-angle Many studies have contributed to the development of SR- plane waves acquired in 3 milliseconds at a pulse repetition US imaging methods by improving the localization precision, frequency of 3000 Hz. To allow microbubbles sufficient time to reducing the acquisition time, increasing microbubble tracking move between consequent compounded volumetric frames, a 7- accuracy, and extending the super-resolution into the third millisecond delay was introduced after each volume acquisition. dimension. These developments are explained in detail by a This reduced the effective volume acquisition speed to 100 Hz and the total acquired data size by 3.3-fold. Localization-based 3- recent review [17]. Researchers mainly employed two different D super-resolution images of two touching sub-wavelength tubes approaches to generate a super-resolution image of a volume were generated from 6000 volumes acquired in 60 seconds. In by mechanically scanning the volume with a linear probe conclusion, this work demonstrates the feasibility of 3D super- and stacking 2-D SR-US images, or by using arrays that resolution imaging and super-resolved velocity mapping using a can acquire volumetric information electronically. Errico et customized 2D sparse array transducer. al. have taken steps towards 3-D with a coronal scan of an entire rat brain by using 128 elements of a custom-built linear I. I NTRODUCTION array at a frequency of 15 MHz. Motion of the probe was Visualization of the microvasculature beyond the diffraction controlled with a micro-step motor to generate 2-D super- limited resolution has been achieved by localizing spatially resolution images over different imaging planes at a frame isolated microbubbles through multiple frames. In the absence rate of 500 Hz [18]. Lin et al. performed a 3-D mechanical of tissue and probe motion, localization precision determines scan of a rat FSA tumor using a linear array mounted on the maximum achievable resolution, which can be on the a motorized precision motion stage synchronized with the order of several micrometers at clinical ultrasound frequen- imaging system. They generated 3-D super-resolution images cies [1], [2]. If motion is present and subsequently corrected by calculating the maximum intensity projection from all 2- post-acquisition, then the motion correction accuracy can D super-resolution slices, acquired using plane-wave imaging limit the achievable spatial resolution [3], [4]. Researchers with a frame rate of 500 Hz [19]. Although sub-diffraction demonstrated the use of 2-D super-resolution ultrasound (SR- imaging has not been published using a 2-D imaging probe US) imaging in many different controlled experiments and with a high volumetric imaging rate, 3-D super-resolution has been achieved by previous studies. O’Reilly and Hynynen used S. Harput, J. Zhu, G. Zhang, C. H. Leow, M. Toulemonde, and M. X. Tang a subset of 128 elements from a 1372-element hemispherical are with the ULIS Group, Department of Bioengineering, Imperial College London, London, SW7 2AZ, UK. transcranial therapy array at a rate of 10 Hz. They generated 3- K. Christensen-Jeffries, J. Brown, and R. J. Eckersley are with the Biomed- D super-resolution images of a spiral tube phantom through an ical Engineering Department, Division of Imaging Sciences, King’s College ex vivo human skullcap at an imaging center frequency of 612 London, SE1 7EH, London, UK. A. Ramalli is with the Department of Information Engineering, University kHz [20]. Desailly et al. implemented a plane wave ultrafast of Florence, 50139 Florence, IT and Lab. on Cardiovascular Imaging & imaging method using an ultrasound clinical scanner with Dynamics, Dept. of Cardiovascular Sciences, KU Leuven, Leuven, Belgium. 128 fully programmable emission-reception channels. They E. Boni, P. Tortoli are with the Department of Information Engineering, University of Florence, 50139 Florence, IT. placed 2 parallel series of 64 transducers to image microflu- C. Dunsby is with the Department of Physics and the Centre for Pathology, idic channels and obtained 3-D super-localization by fitting Imperial College London, London, SW7 2AZ, UK. parallel parabolas in the elevation direction [21]. Christensen- These authors contributed equally to this work. E-mail: s.harput@imperial.ac.uk and mengxing.tang@imperial.ac.uk Jeffries et al. generated volumetric 3-D super-resolution at the arXiv:1902.01608v1 [physics.med-ph] 5 Feb 2019 2 Fig. 2. Optical image of two 200 m cellulose tubes arranged in a double helix pattern. To create this pattern, two tubes were wrapped around each other which created contact-points that are visible in the optical image. Both tubes had constant microbubble flow in opposite directions. cally for high volumetric rate 3-D super-resolution ultrasound imaging based on a density-tapered spiral layout [37], [38]. The capability of the 2-D sparse array for 3-D SR-US imaging was demonstrated in simulations and experiments. II. MATERIALS AND M ETHODS Fig. 1. Layout of the 2-D sparse array with red and green circles showing the chosen elements. The pitch between consecutive elements in the x and y A. 2-D Sparse Array directions is 300 m. Empty rows (9, 18, and 27) are due to manufacturing A 2-D sparse array was designed by selecting 512 elements limitations and are not related to the density-tapered 2-D spiral method. from a 32  35 gridded layout of a 2D matrix array (Vermon S.A., Tours, France) as shown in Fig. 1. It was fabricated with overlapping imaging region of two orthogonal transducers at an individual element size of 300  300 m, center frequency the focus. They used two identical linear arrays to image sub- of 3.7 MHz and a bandwidth of 60%. In the y direction, diffraction cellulose tubes using amplitude modulated plane- row numbers 9, 18, and 27 were intentionally left blank for wave transmission at 3 MHz with a frame rate of 400 Hz [22]. wiring, hence the total number of available elements is 1024. The development of high-speed programmable ultrasound The method to select the location of sparse array elements systems and 2-D arrays created new opportunities for volumet- is based on the density-tapered 2-D spiral layout [37]. This ric imaging with high spatio-temporal resolution. In parallel method arranges the elements according to seeds generated to these hardware developments, novel 3-D imaging methods from Fermats spiral function with an additional spatial density based on small numbers of transmit-receive pairs enabled modulation to reduce the side lobes of the transmitted beam a more reliable visualization of tissue volumes [23], the profile. This deterministic, aperiodic, and balanced positioning analysis of fast and complex blood flow in 3-D [24]–[27], procedure guarantees uniform performance over a wide range the characterization of mechanical properties of tissue by 4- of imaging angles. D shear-wave imaging [23], [28], the tracking of the pulse It is not possible to connect all 512 elements to a single wave propagation along the arterial wall [29], the estima- ultrasound probe adapter. Therefore, two sparse array layouts, tion of 4-D tissue motion [30], and other in vivo transient hereinafter referred to as Aperture#1 and Aperture#2, were events. These technological advances in 3-D imaging also offer designed as shown with red and green elements in Fig. 1. new opportunities for SR-US. Although volumetric imaging Both sparse arrays were based on an ungridded, 10.4-mm- methods have already shown significant benefits for various wide spiral with 256 seeds [37], whose density tapering was ultrasound imaging applications, 3-D imaging with large 2-D modulated according to a 50%-Tukey window. The elements arrays requires a high number of hardware channels and huge belonging to Aperture#1 were selected among those of the computational power. Vermon 2-D matrix array, by activating the available ele- In this study, we demonstrate the feasibility of 3D super- ments whose positions were closest to the ideal positions resolution imaging and super-resolved flow velocity mapping of the ungridded spiral. Similarly, the elements belonging to using a density-tapered sparse array instead of a full 2-D array Aperture#2 were also selected among those of the Vermon to reduce the number of channels and hence the amount of matrix array, but excluding those that were already assigned data while maintaining the frame rate. A similar approach to Aperture#1. The two layouts were connected to two inde- was in previous non-super-resolution studies on minimally pendent connectors (model DLP 408, ITT Cannon, CA, USA) redundant 2-D arrays [31] and sparse 2-D arrays [32]–[36], but so that an approximation of a 256-element density tapered uses a greater number of elements to improve transmit power spiral array could be driven by an independent ULA-OP 256 and receive sensitivity. Our method significantly differs from system [39], [40]. Moreover, by synchronizing two ULA- row-column addressing and multiplexing approaches since it OP 256 systems to simultaneously control the two layouts, maintains simultaneous access to all probe elements through a 512-element dense array (Aperture#1 + Aperture#2) with independent channels. The sparse array was designed specifi- integrated Tukey apodization could be driven. 4 Fig. 5. Isosurface of the 3-D ultrasound image is plotted in copper at -10 dB level. 2-D maximum intensity projections with a 30 dB dynamic range is overlaid on the volumetric image. 5 mm depth (Fig. 3 (top-left)) is a combination of a plane wave and a dispersed tail, which is a result of missing rows. At the depth of 10 mm, as shown in Fig. 3 (top-right), the tail resembles a superposition of multiple edge waves as a result of discontinuities in the array. At this point, the radiated beam shape is not suitable for generating a good quality image. Around 15 mm depth, as shown in Fig. 3 (bottom- left), the tail becomes less prominent and edge waves diminish below 14 dB; however, it can still produce image artefacts as demonstrated by [47]. Further away from the transducer, the residual waves behind the wavefront disappear and the ultrasound field becomes more uniform, which is suitable for plane wave imaging after 20 mm depth as shown in Fig. 3 (bottom-right). The 3-D simulations displayed in Fig. 4 also support the same conclusion: due to the choice of elements and three unconnected rows, the ultrasound field is not uniform for the first 20 mm. B. 3-D Super-resolution Experimental Results Before performing the experiments on a cellulose microvas- Fig. 6. (Top) 3-D super-resolution image of the two 200 m tubes arranged in a double helix shape. Depth-encoded colorscale is added to improve the culature phantom, the imaging performance of the 2-D sparse visualization. (Bottom) Velocity and direction (positive towards increasing y array was characterized with a point target using the tip of a direction) of tracked microbubbles. 100 m metal wire. The full-width-half-maximum (FWHM) of the 3-D B-mode point-spread-function (PSF) was measured as 793, 772, and 499 m in the x, y & z directions respectively visualization of the structure displayed in copper color, 2-D maximum-intensity-projection (MIP) slices in three directions by using linear interpolation [48]. The localization precision were plotted. It was not possible to visualize the two separate was measured to be the standard deviation of the localization 200 m tubes in these MIP slices or in the volumetric image. positions over 100 frames. The 3-D super-localization preci- sion of the overall system at 25 mm was found to be 18 m Fig. 6 (top) shows the 3-D super-resolved volume of the in the worst imaging plane (x direction), where the imaging imaged sub-wavelength structures by combining localizations from all acquired frames. A total of 2824 microbubbles were wavelength is 404 m in water at 25 C. The volumetric B-mode image of two cellulose tubes with- localized within the 6000 volumes after compounding. Due out microbubble flow is shown in Fig. 5. In addition to the 3-D to the large number of localizations, the 3-D structure of the 5 Velocity Tracks (mm/s) -200 -100 0 0 -5 -10 -200 -100 0 100 200 Distance ( m) Horizontal Projection Vertical Projection 0.5 -300 -200 -100 0 100 200 300 Distance ( m) Fig. 7. (Top) Figure shows the MIP of the B-mode image belonging to a Fig. 8. (Top) Figure shows the velocity tracks belonging to a 1 mm long 1 mm long section of the tube projected into a 2D plane that is orthogonal to section of the tube projected into a 2D plane that is orthogonal to the direction the direction of the flow. The super-resolution image was projected into the of the flow. Black circle represents the 200 m tube circumference. (Bottom) same 2D plane and overlaid on the B-mode image. Black circle represents The FWHM of the tube is measured as 130 m and 75 m from 1-D the 200 m tube circumference. (Bottom) The B-mode FWHM of the tube projections in the horizontal and vertical directions of the top panel plot is measured as 1380 m and 590 m from 1-D projections in the horizontal respectively. and vertical directions of the top panel plot respectively. The super-resolution FWHM of the tube is measured as 134 m and 108 m from 1-D projections in the horizontal and vertical directions of the top panel plot respectively. and 200 m respectively. In the B-mode image two touching tubes appeared as a single scattering object with a FWHM tubes cannot be clearly visualized in a single 2-D image. To of 1380 m and 590 m in the horizontal and vertical 1-D improve the visualization, 3-D SR-US images are plotted with projections respectively. depth information color-coded in the image. Microbubble tracking effectively worked as another layer of Fig. 6 (bottom) displays the velocity profiles of tracked filtering by removing the potentially erroneous non-traceable microbubbles. Only 1076 microbubble-pairs out of 2824 mi- super-localizations. Around the same section of the tube shown crobubbles were traceable from consecutive frames using a in Fig. 7 (top) and Fig. 8 (top), 96% of the microbubble nearest-neighbor method. Using these microbubble tracks, two velocity tracks were within a diameter of 200 m, which sub-wavelength tubes with opposing flows were easily distin- was 89% for the super-localized microbubbles without velocity guishable by color-coding the direction of their velocity vec- tracking. The FWHM of a single tube appeared as 130 m and tors. The percentage of microbubbles that were followed over 75 m from the projection of velocity tracks to horizontal and two or more volumes of 76.2% was attributed to microbubble vertical directions, as plotted in Fig. 8 (bottom). destruction, which can also be observed in the super-resolution images showing a high number of microbubble localizations Microbubble tracking made the separation between the tubes at the inlet and almost no localizations at the outlet of the more clear when tubes are in contact around the central section tubes, see Fig. 6. of the 3-D SR-US and velocity maps displayed in Fig. 6. The thickness of the imaged tubes was measured at the The velocity profiles of microbubbles at this location with inlet where the tube is clearly isolated in the 3-D SR-US two touching tubes were re-plotted in Fig. 9 (top) for clear image around the [3 mm, -3 mm] coordinates in x and y visualization. From this plotted volume, a 1 mm long section respectively. To perform the thickness measurement, a 1 mm of the tube was chosen and projected into a 2-D plane that long section of the imaged tube was chosen and projected into is shown in Fig. 9 (middle). In this 2-D maximum intensity a 2-D plane that is orthogonal to the direction of the tube as projection, the weighted center locations between positive and shown in Fig. 7 (top) both for B-mode and 3-D SR-US images. negative velocity tracks have a distance of 239 m. The 1- Fig. 7 (bottom) shows the 1-D projections in the horizontal D projection of the velocity tracks had a FWHM of 122 m and vertical directions where the FWHM of the super-resolved and 115 m for positive and negative flows respectively with a tube was measured as 134 m and 108 m and the 20 dB peak-to-peak distance of 190 m between two opposing tracks width of the super-resolved tube was measured as 213 m as plotted in Fig. 9 (bottom). Normalized Amplitude Distance ( m) 6 these systems had 1 of 2 transducer elements multiplexed in re- ception [23], [28]. Many researchers have developed methods to use a large number of active elements with fewer channels (usually between 128 and 256) to reduce the cost and com- plexity of the ultrasound systems and the probes. It has been demonstrated in several studies that row-column addressed matrix arrays [47], [51]–[53], microbeamformers [54]–[56] and channel multiplexing can be an alternative to fully ad- dressed 2-D matrix arrays. However, these methods have less flexibility and limitations due to the elements not being continuously connected to the ultrasound system. In this paper, a 2-D sparse array imaging probe has been developed for 3-D super-resolution imaging. This has ad- dressed the main limitation of the existing 2-D imaging of poor spatial resolution in the elevational plane. In addition to super- resolution imaging, 3-D velocity mapping was implemented to reveal the flow inside the microstructures. Using the sparse Velocity Tracks (mm/s) array approach instead of a full matrix array reduced the -500 10 number of channels to half, and hence the connection issues, cost and data size while still achieving the same volumetric acquisition speed since all elements of 2-D spiral array are always connected to the system. Although this can sacrifice the maximum achievable transmit pressure and receive sensitivity, 0 0 it is not a significant issue with SR-US due to the low pressure -2 required and high sensitivity achievable in microbubble imag- -4 ing. In terms of B-mode image resolution, the axial resolution -6 is comparable, since both arrays have the same bandwidth; -8 while a slightly worse lateral resolution is expected for the -10 sparse array, since the full matrix array has a larger aperture -500 0 500 size. It is hard to distinguish the grating lobes and the side Distance ( m) lobes of a sparse array, but here we consider the unwanted leakage outside the main lobe as grating lobes if it is as a result of element-to-element spacing, and as side lobes if 0.5 it is as a result of finite aperture size. The side lobe and edge wave suppression characteristics of the sparse array will -0.5 outperform an un-apodized full matrix array thanks to the integrated apodization, although the fixed apodization might -1 be a limitation for some applications. Both arrays will have -500 -400 -300 -200 -100 0 100 200 300 400 500 Distance ( m) higher grating lobes in y direction due to the three empty rows. The highest grating lobe of the full matrix array will appear Fig. 9. (Top) 3D velocity profiles of microbubbles re-plotted from Fig. 8 at 8 as high as 17% of the main lobe amplitude, calculated (bottom) to display the details more clearly when two tubes are in contact. using the array factor equation in [57]. A sparse choice of (Middle) Figure shows the velocity tracks belonging to a 1 mm long section of the tube projected into a 2-D plane. (Bottom) The 1-D projection of the middle elements spreads the grating lobes to a wider range due to panel plot towards vertical direction shows that the peak-to-peak distance the irregular placement of elements, where the highest grating between two opposing velocity tracks is 190 m. lobe will appear at 18 as high as 16% of the main lobe amplitude. Using the plane-wave imaging method instead of line-by- IV. D ISCUSSION line scanning increases the temporal resolution of the volumet- A better 3-D image quality may be achieved by using a ric imaging. Faster 3-D image acquisition provides a higher large number of independent array elements with the fastest microbubble localization rate and improves velocity estima- possible volumetric imaging rate; however this requires the tions due to more frequent sampling. However, a microbubble same number of hardware channels as the number of elements travelling with a velocity of 10 mm/s will be exposed to 3000 and the ability to process very large stacks of data. Due to the ultrasound pulses while travelling through the imaging region high cost, full 2-D array imaging using an ultrasound system of 10 mm at a PRF of 3000 Hz. At this insonation rate, even to control very large numbers of independent elements has at a relatively low MI of 0.07 almost all microbubbles were only been used by a few research groups [23], [28], [49], destroyed before reaching the center of the imaging region. [50]. These systems had 1024 channels capable of driving a At this point, a new transmitting strategy was implemented to 32  32 2-D array with at least 4 connectors. Even some of reduce the microbubble destruction rate instead of reducing Normalized Amplitude Distance ( m) 7 the PRF, which may introduce artefacts on 9 compounded REFERENCES volumes due to moving microbubbles, or ultrasound pressure, [1] O. M. Viessmann, R. J. Eckersley, K. Christensen-Jeffries, M.-X. which will decrease the SNR and localization precision. To Tang, and C. Dunsby, “Acoustic super-resolution with ultrasound and microbubbles,” Phys. Med. Biol, vol. 58, pp. 6447–6458, 2013. improve the microbubble circulation time, a 7 millisecond [2] Y. Desailly, J. Pierre, O. Couture, and M. Tanter, “Resolution limits of pause was added between each compounded volume acquisi- ultrafast ultrasound localization microscopy,” Phys. Med. Biol, vol. 60, tion that took 3 milliseconds. This strategy reduced insonation pp. 8723–8740, 2015. [3] S. Harput, K. Christensen-Jeffries, Y. Li, J. Brown, R. J. Eckersley, rate by 3.3-fold and the volumetric acquisition rate to 100 C. Dunsby, and M.-X. Tang, “Two stage sub-wavelength motion correc- Hz without sacrificing the PRF during compounding. The tion in human microvasculature for ceus imaging,” in IEEE International second advantage of this transmit strategy is a reduction in Ultrasonics Symposium (IUS), 2017, pp. 1–4. [4] S. Harput, K. Christensen-Jeffries, J. Brown, Y. Li, K. J. Williams, the redundant microbubble localizations at the same spatial A. H. Davies, R. J. Eckersley, C. Dunsby, and M. Tang, “Two-stage location without reducing the PRF. In order for microbubbles motion correction for super-resolution ultrasound imaging in human to provide new spatial information in each frame, they must lower limb,” IEEE Trans. Ultrason., Ferroelectr., Freq. Control, vol. 65, no. 5, pp. 803–814, 2018. be re-located by the flow. Especially for microvasculature [5] K. Christensen-Jeffries, R. J. Browning, M.-X. Tang, C. Dunsby, and with slow physiological flow, increasing the frame rate will R. J. Eckersley, “In vivo acoustic super-resolution and super-resolved no longer increase the obtained spatial information but result velocity mapping using microbubbles,” IEEE Trans. Med. Imag., vol. 34, no. 2, pp. 433–440, 2015. in redundant location information. Nevertheless, using high [6] D. Ackermann and G. Schmitz, “Detection and tracking of multiple frame rates may still be valuable for improving the SNR and microbubbles in ultrasound b-mode images,” IEEE Trans. Ultrason., velocity tracking. The proposed transmit strategy improved the Ferroelectr., Freq. Control, vol. 63, no. 1, pp. 72–82, 2016. [7] A. Bar-Zion, C. Tremblay-Darveau, O. Solomon, D. Adam, and Y. C. image quality for this study, but it could not totally prevent the Eldar, “Fast vascular ultrasound imaging with enhanced spatial resolu- microbubble destruction that caused a lower number of local- tion and background rejection,” IEEE Trans. Med. Imag., vol. 36, pp. izations at outlet of the tubes visible in the 3-D SR-US image. 169–180, 2017. [8] J. Foiret, H. Zhang, T. Ilovitsh, L. Mahakian, S. Tam, and K. W. Ferrara, In the future, the relationship between PRF, pause interval, “Ultrasound localization microscopy to image and assess microvascula- microbubble flow velocity, and compounding strategies should ture in a rat kidney,” Scientific Reports, vol. 7, no. 13662, pp. 1–12, be investigated for different applications and physiological [9] S. Harput, K. Christensen-Jeffries, J. Brown, R. J. Eckersley, C. Dunsby, flow rates. and M.-X. Tang, “Localisation of multiple non-isolated microbubbles with frequency decomposition in super-resolution imaging,” in IEEE International Ultrasonics Symposium (IUS), 2017, pp. 1–4. [10] P. Song, J. D. Trzasko, A. Manduca, R. Huang, R. Kadirvel, D. F. V. CONCLUSION Kallmes, and S. Chen, “Improved super-resolution ultrasound microves- sel imaging with spatiotemporal nonlocal means filtering and bipartite The main limitation of localization-based SR-US imaging graph-based microbubble tracking,” IEEE Trans. Ultrason., Ferroelectr., Freq. Control, vol. 65, no. 2, pp. 149–167, 2018. performed in 2-D is the lack of super-resolution in the el- [11] T. Opacic, S. Dencks, B. Theek, M. Piepenbrock, D. Ackermann, A. Rix, evation direction. In this study, this issue was addressed by T. Lammers, E. Stickeler, S. Delorme, G. Schmitz, and F. Kiessling, using a bespoke 2-D sparse array that achieved an estimated “Motion model ultrasound localization microscopy for preclinical and localization precision of 18 m in the worst imaging plane, clinical multiparametric tumor characterization,” Nature Communica- tions, vol. 9, no. 1527, pp. 1–13, 2018. which is approximately 22 times smaller than the wavelength. [12] T. Ilovitsh, A. Ilovitsh, J. Foiret, B. Z. Fite, and K. W. Ferrara, “Acous- Compounded plane wave imaging with a PRF of 3000 Hz tical structured illumination for super-resolution ultrasound imaging,” enabled super-resolution imaging in all spatial directions with Communications Biology, vol. 1, no. 3, 2018. [13] G. P. Luke, A. S. Hannah, and S. Y. Emelianov, “Super-resolution an image acquisition time of 60 seconds. The structure of two ultrasound imaging in vivo with transient laser-activated nanodroplets,” 200 m, smaller than half wavelength, tubes arranged in a Nano Letters, vol. 16, pp. 2556–2559, 2016. double helix shape were super resolved and flow velocities [14] H. Yoon, K. A. Hallam, C. Yoon, and S. Y. Emelianov, “Super-resolution imaging with ultrafast ultrasound imaging of optically triggered per- within these tubes were estimated. 3-D sub-diffraction imaging fluorohexane nanodroplets,” IEEE Trans. Ultrason., Ferroelectr., Freq. was achieved in vitro using the 2D sparse array probe. Control, vol. 65, no. pre-print, pp. 1–1, 2018. [15] G. Zhang, S. Harput, S. Lin, K. Christensen-Jeffries, C. H. Leow, J. Brown, C. Dunsby, R. J. Eckersley, and M.-X. Tang, “Acoustic wave sparsely activated localization microscopy (awsalm): Super-resolution ACKNOW LEDGMENTS ultrasound imaging using acoustic activation and deactivation of nan- odroplets,” Applied Physics Letters, vol. 113, no. 1, p. 014101, 2018. [16] G. Zhang, S. Harput, H. Hu, K. Christensen-Jeffries, J. Zhu, J. Brown, This work was supported mainly by the EPSRC under C. H. Leow, C. Dunsby, R. J. Eckersley, and M.-X. Tang, “Fast acoustic Grant EP/N015487/1 and EP/N014855/1, in part by the wave sparsely activated localization microscopy (fast-awsalm) using octafluoropropane nanodroplets,” in IEEE International Ultrasonics King’s College London (KCL) and Imperial College London Symposium (IUS), 2018, pp. 1–4. EPSRC Centre for Doctoral Training in Medical Imaging [17] O. Couture, V. Hingot, B. Heiles, P. Muleki-Seya, and M. Tanter, (EP/L015226/1), in part by the Wellcome EPSRC Centre for “Ultrasound localization microscopy and super-resolution: A state of the art,” IEEE Trans. Ultrason., Ferroelectr., Freq. Control, vol. 65, no. 8, Medical Engineering at KCL (WT 203148/Z/16/Z), in part by pp. 1304–1320, 2018. the Department of Health through the National Institute for [18] C. Errico, J. Pierre, S. Pezet, Y. Desailly, Z. Lenkei, O. Couture, and Health Research comprehensive Biomedical Research Center M. Tanter, “Ultrafast ultrasound localization microscopy for deep super- resolution vascular imaging,” Nature, vol. 527, pp. 499–507, 2015. Award to Guy’s and St Thomas’ NHS Foundation Trust in [19] S. Lin, G. Zhang, C. H. Leow, and M.-X. Tang, “Effects of microchan- partnership with KCL and King’s College Hospital NHS nel confinement on acoustic vaporisation of ultrasound phase change Foundation Trust, in part by the Graham-Dixon Foundation contrast agents,” Physics in Medicine & Biology, vol. 62, no. 17, pp. and in part by NVIDIA GPU grant. 6884–6898, 2017. 8 [20] M. A. O’Reilly and K. Hynynen, “A super-resolution ultrasound method system for continuous real-time high frame rate imaging,” IEEE Trans. for brain vascular mapping,” Medical Physics, vol. 40, no. 110701, 2013. Ultrason., Ferroelectr., Freq. Control, vol. 64, no. 9, pp. 1276–1284, [21] Y. Desailly, O. Couture, M. Fink, and M. Tanter, “Sono-activated 2017. ultrasound localization microscopy,” Applied Physics Letters, vol. 103, [41] M. Toulemonde, A. Stanziola, Y. Li, S. Lin, M. Butler, W. C. Duncan, no. 174107, 2013. V. Sboros, R. Eckersley, and M.-X. Tang, “Effects of motion on high [22] K. Christensen-Jeffries, J. Brown, P. Aljabar, M.-X. Tang, C. Dunsby, frame rate contrast enhanced echocardiography and its correction,” in and R. J. Eckersley, “3-d in vitro acoustic super-resolution and super- IEEE International Ultrasonics Symposium (IUS), 2017. resolved velocity mapping using microbubbles,” IEEE Trans. Ultrason., [42] C. Demene, T. Deffieux, M. Pernot, B.-F. Osmanski, V. Biran, J.-L. Ferroelectr., Freq. Control, vol. 64, no. 10, pp. 1478–1486, 2017. Gennisson, L.-A. Sieu, A. Bergel, S. Franqui, J.-M. Correas, I. Cohen, [23] J. Provost, C. Papadacci, J. E. Arango, M. Imbault, M. Fink, J.-L. O. Baud, and M. Tanter, “Spatiotemporal clutter filtering of ultrafast Gennisson, M. Tanter, and M. Pernot, “3d ultrafast ultrasound imaging ultrasound data highly increases doppler and fultrasound sensitivity,” in vivo,” Phys. Med. Biol, vol. 59, no. 19, pp. L1–L13, 2014. IEEE Trans Med Imaging, vol. 34, no. 11, pp. 2271–2285, 2015. [24] M. J. Pihl, M. B. Stuart, B. G. Tomov, M. F. Rasmussen, and J. A. [43] A. Stanziola, C. H. Leow, E. Bazigou, P. D. Weinberg, and M. Tang, Jensen, “A transverse oscillation approach for estimation of three- “Asap: Super-contrast vasculature imaging using coherence analysis and dimensional velocity vectors, part ii: experimental validation,” IEEE high frame-rate contrast enhanced ultrasound,” IEEE Transactions on Trans. Ultrason., Ferroelectr., Freq. Control, vol. 61, no. 10, pp. 1608– Medical Imaging, vol. 37, no. 8, pp. 1847–1856, 2018. 1618, 2014. [44] K. Christensen-Jeffries, S. Harput, J. Brown, P. N. T. Wells, P. Aljabar, [25] J. Provost, C. Papadacci, C. Demene, J. Gennisson, M. Tanter, and C. Dunsby, M.-X. Tang, and R. J. Eckersley, “Microbubble axial M. Pernot, “3-d ultrafast doppler imaging applied to the noninvasive localization errors in ultrasonic super-resolution imaging,” IEEE Trans. mapping of blood vessels in vivo,” IEEE Trans. Ultrason., Ferroelectr., Ultrason., Ferroelectr., Freq. Control, vol. 64, no. 11, pp. 1644–1654, Freq. Control, vol. 62, no. 8, pp. 1467–1472, 2015. 2017. [26] M. Correia, J. Provost, M. Tanter, and M. Pernot, “4d ultrafast ultrasound [45] J. Jensen and N. B. Svendsen, “Calculation of pressure fields from flow imaging: in vivo quantification of arterial volumetric flow rate in arbitrarily shaped, apodized, and excited ultrasound transducers,” IEEE a single heartbeat,” Phys. Med. Biol, vol. 61, no. 23, p. L48, 2016. Trans. Ultrason., Ferroelectr., Freq. Control, vol. 39, pp. 262–267, 1992. [27] S. Holbek, C. Ewertsen, H. Bouzari, M. J. Pihl, K. L. Hansen, M. B. [46] J. Jensen, “Field: A program for simulating ultrasound systems,” in Stuart, C. Thomsen, M. B. Nielsen, and J. A. Jensen, “Ultrasonic 3-d Medical & Biological Engineering & Computing, vol. 34, no. 1, 1996, vector flow method for quantitative in vivo peak velocity and flow rate pp. 351–353. estimation,” IEEE Trans. Ultrason., Ferroelectr., Freq. Control, vol. 64, [47] M. F. Rasmussen, T. L. Christiansen, E. V. Thomsen, and J. A. no. 3, pp. 544–554, 2017. Jensen, “3-d imaging using rowcolumn-addressed arrays with integrated [28] J. Gennisson, J. Provost, T. Deffieux, C. Papadacci, M. Imbault, M. Per- apodization part i: Apodization design and line element beamforming,” not, and M. Tanter, “4-d ultrafast shear-wave imaging,” IEEE Trans. IEEE Trans. Ultrason., Ferroelectr., Freq. Control, vol. 62, no. 5, pp. Ultrason., Ferroelectr., Freq. Control, vol. 62, no. 6, pp. 1059–1065, 947–958, 2015. 2015. [48] S. Harput, J. McLaughlan, D. M. Cowell, and S. Freear, “New per- [29] I. Apostolakis, P. Nauleau, C. Papadacci, M. D. McGarry, and E. E. formance metrics for ultrasound pulse compression systems,” in IEEE Konofagou, “Feasibility and validation of 4-d pulse wave imaging International Ultrasonics Symposium (IUS), 2014, pp. 440–443. in phantoms and in vivo,” IEEE Trans. Ultrason., Ferroelectr., Freq. [49] J. A. Jensen, M. F. Holten-Lund, R. T. Nilsson, M. Hansen, U. D. Control, vol. 64, no. 9, pp. 1305–1317, 2017. Larsen, R. P. Domsten, B. G. Tomov, M. B. Stuart, S. I. Nikolov, [30] S. Salles, H. Liebgott, D. Garcia, and D. Vray, “Full 3-d transverse oscil- M. J. Pihl, Y. Du, J. H. Rasmussen, and M. F. Rasmussen, “Sarus: lations: a method for tissue motion estimation,” IEEE Trans. Ultrason., A synthetic aperture real-time ultrasound system,” IEEE Transactions Ferroelectr., Freq. Control, vol. 62, no. 8, pp. 1473–1485, 2015. on Ultrasonics, Ferroelectrics, and Frequency Control, vol. 60, no. 9, [31] M. Karaman, I. O. Wygant, O. Oralkan, and B. T. Khuri-Yakub, pp. 1838–1852, 2013. “Minimally redundant 2-d array designs for 3-d medical ultrasound [50] L. Petrusca, F. Varray, R. Souchon, A. Bernard, J.-Y. Chapelon, H. Lieb- imaging,” IEEE Trans. Med. Imag., vol. 28, no. 7, pp. 1051–1061, 2009. gott, W. A. NDjin, and M. Viallon, “Fast volumetric ultrasound b-mode [32] A. Austeng and S. Holm, “Sparse 2-d arrays for 3-d phased array and doppler imaging with a new high-channels density platform for imaging - design methods,” IEEE Trans. Ultrason., Ferroelectr., Freq. advanced 4d cardiac imaging/therapy,” Applied Sciences, vol. 8, no. 2, Control, vol. 49, no. 8, pp. 1073–1086, 2002. 2018. [33] B. Diarra, M. Robini, P. Tortoli, C. Cachard, and H. Liebgott, “Design of [51] A. Savoia, V. Bavaro, G. Caliano, A. Caronti, R. Carotenuto, P. Gatta, optimal 2-d nongrid sparse arrays for medical ultrasound,” IEEE Trans. C. Longo, and M. Pappalardo, “Crisscross 2d cmut array: Beamform- Biomed. Eng, vol. 60, no. 11, pp. 3093–3102, 2013. ing strategy and synthetic 3d imaging results,” in IEEE International [34] E. Roux, A. Ramalli, P. Tortoli, C. Cachard, M. Robini, and H. Liebgott, Ultrasonics Symposium (IUS), 2007, pp. 1514–1517. “2-d ultrasound sparse arrays multidepth radiation optimization using [52] S. Holbek, T. L. Christiansen, M. B. Stuart, C. Beers, E. V. Thomsen, simulated annealing and spiral-array inspired energy functions,” IEEE and J. A. Jensen, “3-d vector flow estimation with rowcolumn-addressed Trans. Ultrason., Ferroelectr., Freq. Control, vol. 63, no. 12, pp. 2138– arrays,” IEEE Trans. Ultrason., Ferroelectr., Freq. Control, vol. 63, 2149, 2016. no. 11, pp. 1799–1814, 2016. [35] E. Roux, A. Ramalli, H. Liebgott, C. Cachard, M. Robini, and P. Tortoli, [53] M. Flesch, M. Pernot, J. Provost, G. Ferin, A. Nguyen-Dinh, M. Tanter, “Wideband 2-d array design optimization with fabrication constraints and T. Deffieux, “4d in vivo ultrafast ultrasound imaging using a row- for 3-d us imaging,” IEEE Trans. Ultrason., Ferroelectr., Freq. Control, column addressed matrix and coherently-compounded orthogonal plane vol. 64, no. 1, pp. 108–125, 2017. waves,” Phys. Med. Biol, vol. 62, no. 11, p. 4571, 2017. [36] E. Roux, F. Varray, L. Petrusca, C. Cachard, P. Tortoli, and H. Liebgott, [54] B. J. Savord, “Beamforming methods and apparatus for three- “Experimental 3-d ultrasound imaging with 2-d sparse arrays using dimensional ultrasound imaging using two-dimensional transducer ar- focused and diverging waves,” Scientific Reports, vol. 8, no. 9108, pp. ray,” USA Patent US6 013 032A, 2000. 1–12, 2018. [55] G. Matrone, A. S. Savoia, M. Terenzi, G. Caliano, F. Quaglia, and [37] A. Ramalli, E. Boni, A. S. Savoia, and P. Tortoli, “Density-tapered spiral G. Magenes, “A volumetric cmut-based ultrasound imaging system sim- arrays for ultrasound 3-d imaging,” IEEE Trans. Ultrason., Ferroelectr., ulator with integrated reception and -beamforming electronics models,” Freq. Control, vol. 62, no. 8, pp. 1580–1588, 2015. IEEE Trans. Ultrason., Ferroelectr., Freq. Control, vol. 61, no. 5, pp. [38] S. Harput, K. Christensen-Jeffries, J. Brown, J. Zhu, G. Zhang, C. H. 792–804, 2014. Leow, M. Toulemonde, A. Ramalli, E. Boni, P. Tortoli, R. J. Eckersley, [56] P. Santos, G. U. Haugen, L. Lvstakken, E. Samset, and J. Dhooge, C. Dunsby, and M.-X. Tang, “3-d super-resolution ultrasound imaging “Diverging wave volumetric imaging using subaperture beamforming,” using a 2-d sparse array with high volumetric imaging rate,” in IEEE IEEE Trans. Ultrason., Ferroelectr., Freq. Control, vol. 63, no. 12, pp. International Ultrasonics Symposium (IUS), 2018, pp. 1–4. 2114–2124, 2016. [39] E. Boni, L. Bassi, A. Dallai, F. Guidi, V. Meacci, A. Ramalli, S. Ricci, [57] S. Harput and A. Bozkurt, “Ultrasonic phased array device for acoustic and P. Tortoli, “Ula-op 256: A 256-channel open scanner for develop- imaging in air,” Sensors Journal, IEEE, vol. 8, no. 11, pp. 1755–1762, ment and real-time implementation of new ultrasound methods,” IEEE 2008. Trans. Ultrason., Ferroelectr., Freq. Control, vol. 63, no. 10, pp. 1488– 1495, 2016. [40] E. Boni, L. Bassi, A. Dallai, V. Meacci, A. Ramalli, M. Scaringella, F. Guidi, S. Ricci, and P. Tortoli, “Architecture of an ultrasound

Journal

PhysicsarXiv (Cornell University)

Published: Feb 5, 2019

There are no references for this article.