WO2020128010A1

WO2020128010A1 - Method and apparatus for radial modulation imaging

Info

Publication number: WO2020128010A1
Application number: PCT/EP2019/086745
Authority: WO
Inventors: Olivier Couture; Mickael Tanter; Pauline MULEKI SEYA
Original assignee: Centre National de la Recherche Scientifique CNRS; Institut National de la Sante et de la Recherche Medicale INSERM; Ecole Superieure de Physique et Chimie Industrielles de Ville de Paris ESPCI; Sorbonne Universite; Universite de Paris
Current assignee: Centre National de la Recherche Scientifique CNRS; Institut National de la Sante et de la Recherche Medicale INSERM; Ecole Superieure de Physique et Chimie Industrielles de Ville de Paris ESPCI; Sorbonne Universite; Universite Paris Cite
Priority date: 2018-12-21
Filing date: 2019-12-20
Publication date: 2020-06-25
Anticipated expiration: 2021-06-21

Abstract

Unfocused ultrasonic pulses having an imaging frequency f₁ are transmitted in a medium containing microbubbles at a number A of propagation directions (P_i) while at least a modulation ultrasonic wave having a smaller modulation frequency f₂ is transmitted in the medium, and raw data from backscattered ultrasonic waves are acquired. For each propagation direction (Pi), a number S of pulses are successively transmitted in the medium (1) respectively at constant phases φ_j of the modulation wave and the corresponding backscattered signals are acquired. Then, a set of basic data corresponding to raw data for each propagation direction (Pi) is filtered over time by a bandpass filter around the modulation frequency f₂, and an image is created from the filtered data.

Description

Method and apparatus for radial modulation imaging.

FIELD

The disclosure relates to methods and apparatus for radial modulation imaging.

BACKGROUND

Most major causes of death are linked to diseases which affect, in some way, the microvasculature. Cancer grows and spreads with the help of angiogenesis, arteriosclerosis often involves the progression of the vaso vasorum and ischemic stroke implies the blockage of a significant number of cerebral vessels of all sizes. In these areas, ultrasound imaging is a key diagnostic tool, mainly because Doppler is sensitive to blood flow [Taylor et al . 1995] .

Recent advances in ultrafast imaging, based on plane- wave emission and parallel beamforming [Tanter et al . 2002, Montaldo et al . 2009], has increased the sensitivity of Doppler by more than an order of magnitude [Bercoff et al . 2011, Tanter and Fink 2014] . For instance, it is now possible to assess the function of the brain with ultrasound through the neuro-vascular coupling [Mace et al . 2011, Deffieux et al . 2018] . However, the finest vessels, smaller than the ultrasound wavelength, are still undetectable since red blood cells are poor scatterers that are moving insufficiently to create a detectable decorrelation .

To improve the sensitivity of ultrasound to the microvasculature, microbubbles were introduced as ultrasound contrast agents [Gramiak and Shah 1968] . The current generation of agents are formed by bubbles of perfluorocarbon gases, a few microns in size, which remain in the vasculature after being intravenously injected [Burns and Wilson 2006] . Their detection is based not only on the impedance mismatch between gas and liquid, but also due to their resonance in the MHz range. Moreover, their oscillation become highly nonlinear in the diagnostic range of ultrasonic pressures and they can even be disrupted with sufficiently high pressure pulses [de Jong et al . 2002, Deng et al . 2002] .

Several detection techniques have been designed to separate microbubbles from tissue. Microbubbles can emit harmonic components, as well as subharmonics and ultraharmonics, which falls at fraction multiple of the fundamental frequency. Harmonic imaging filters the echoes to extract only the content at twice the frequency of the excitation pulse to remove tissue signal [Schrope and al . 1993] . Different harmonic imaging schemes are now common today such as pulse inversion (PI), which implies emitting two successive pulses of inverted polarity [Hwang and al . 1999, Simpson and al . 1999, Burns and al . 2000] . When their respective echoes are combined, linear tissue yields very limited signal while nonlinear microbubbles yield detectable signal. Amplitude modulation (AM) methods, for which echoes from pulses of different amplitudes are combined, are also in use [Brock-Fisher and al . 1996, Caiani and al . 2001] . They can also be combined to PI methods [Eckersley and al . 2005] . To minimize the signal from tissue, another possibility is subharmonic imaging, as the nonlinear signal generated by tissue contains only harmonics components [Forsberg and al . 2000, Goertz and al . 2005] . The generation of very high ultraharmonics can also be implemented to separate microbubbles from tissue, allowing the reception of very high frequencies, resulting in excellent resolution as in ultrasound angiography [Gessner et al . 2013] . These techniques have inherent limitations associated to the fact that they rely on the strong nonlinear oscillations of microbubbles, putting microbubbles at risks of disruption. Moreover, they often lock the emission pulse frequency to the bubble resonance frequency (below 10 MHz) and they consequently limit image resolution .

Other microbubbles detection approaches have been used. The oldest [Gramiak and Shah 1968] is the detection of the microbubble decorrelation through their motion. It was recently exploited for ultrasound localization microscopy [Couture et al . 2011, Errico et al . 2015, Couture et al . 2018] to detect microbubbles with respect to tissue before the localization step which improved the resolution by an order of magnitude as compared to the diffraction limit. For this, spatiotemporal filters were implemented on stacks of ultrafast imaging. Singular-value decomposition of large ensembles of ultrafast images was proved to be superior to nonlinear techniques for flow within the mm/s range, both in-vitro and the clinic [Desailly and al . 2016] . However, we have yet to detect microbubbles moving well below 1 mm/s even at higher imaging frequencies. Spatiotemporal filtering is especially maladjusted for ultrasound molecular imaging, where microbubbles are attached to targets through ligands and, consequently, remain still [Ellegala and al . 2003, Weller and al . 2005, Willmann and al . 2008] .

Another approach to detect microbubbles is to exploit, not their nonlinearities or their motion, but their high compressibility. Radial modulation imaging (RMI), a particular application of a dual frequency band technique is based on the observations of an increase of the microbubbles' amplitude in radio-frequency data in presence of a superimposed ultrasound excitations, say 10 MHz and 0.5 MHz [Deng et al . 2000] . In RMI, a low-frequency pulse (the modulation pulse) induces oscillation of microbubbles and modify its size. The microbubbles are simultaneously interrogated using a high-frequency pulse (the imaging pulse) . One implementation of RMI consists of synchronizing two successive short imaging pulses such that they reach the microbubble when it is in a compressed and an expanded state, induced by the modulation pressure wave. By subtracting successive imaging lines, this approach results in an image of microbubbles, in which tissue scattering is suppressed because it is minimally affected by the modulation pulse. Radial modulation on a single microbubble has an effect on microbubble's radius and echoes' amplitude [Bouakaz et al . 2004] . The predominant effect in RMI was found to be microbubble size and not microbubble resonance or compression-only effect [Emmer and al . 2009] . Simulations suggested that a 10% variation of bubble diameter induced by the modulation excitation is sufficient for radial modulation imaging [Cherin et al . 2008] .

In 2006, RMI showed a 35 dB microbubbles signal enhancement [Shariff et al . 2006] . Angelsen et al . [Angelsen et al . 2007, Hansen et al . 2009] demonstrated numerically that SURF imaging (RMI with a frequency separation between the modulation and the imaging pulses of a factor of 10) with correction of the dispersion of the imaging pulse by the modulation pulse provided higher contrast than PI. This team also pointed out that in presence of the manipulation pulse, the imaging pulse presented a distortion which was different than for conventional harmonic distortion [Hansen et al . 2011] . This technique showed promising results in-vitro with a 20 MHz single-element intravascular ultrasound catheter [Yu et al . 2014] and in-vivo for pig kidney with a bolus injection of Sonovue [Masoy et al . 2008] . Nevertheless, several issues remain to be solved with RMI. For instance, the synchronization between the imaging pulses is non-trivial because microbubbles need to be interrogated in the compression and in the rarefaction phase and not between them (ie, in nodes of the modulation pressure) . Moreover, the time-delay difference corresponding to the dispersion of the imaging pulse by the modulation pulse should be corrected [Angelsen et al . 2007, Hansen et al . 2009] .

Bibliography :

(1) Angelsen BA and Hansen R (2007), "SURF Imaging-A new method for ultrasound contrast agent imaging", In Ultrasonics Symposium, 2007. IEEE. , pp . 531-

541.

(2) Bercoff J, Montaldo G, Loupas T, Savery D, Meziere

F, Fink M and Tanter M (2011), "Ultrafast compound Doppler imaging: Providing full blood flow characterization", IEEE Transactions on

Ultrasonics, Ferroelectrics and Frequency Control. Vol . 58(1), pp. 134-147.

(3) Bouakaz A and de Jong N (2004), "New contrast imaging method using double frequency exposure", In Ultrasonics Symposium, 2004 IEEE. Vol. 1, pp . 339-

342.

(4) Brock-Fisher GA, Poland MD and Rafter PG (1996),

"Means for increasing sensitivity in non-linear ultrasound imaging systems". November 26, 1996.

(5) Burns PN and Wilson SR (2006), "Microbubble contrast for radiological imaging: 1. Principles", Ultrasound quarterly. Vol. 22(1), pp . 5-13. LWW.

(6) Burns PN, Wilson SR and Simpson DH (2000), "Pulse inversion imaging of liver blood flow: improved method for characterizing focal masses with microbubble contrast", Investigative radiology. Vol. 35(1), pp . 58. LWW.

(7) Caiani EG, Collins KA, Korcarz CE, Bednarz JE, Lang RM and others (2001), "Combined assessment of myocardial perfusion and regional left ventricular function by analysis of contrast-enhanced power modulation images", Circulation. Am Heart Assoc.

(8) Cherin E, Brown J, Masoy S-E, Shariff H, Karshafian R, Williams R, Burns PN and Foster FS (2008), "Radial modulation imaging of microbubble contrast agents at high frequency", Ultrasound in medicine & biology. Vol. 34(6), pp . 949-962. Elsevier.

(9) Couture 0, Besson B, Montaldo G, Fink M and Tanter

M (2011), "Microbubble ultrasound super localization imaging (MUSLI)", In Ultrasonics

Symposium (IUS), 2011 IEEE International. , pp . 1285-1287.

(10) Couture 0, Hingot V, Heiles B, Muleki-Seya P and

Tanter M (2018), "Ultrasound localization microscopy and super-resolution: A state of the art", IEEE transactions on ultrasonics, ferroelectrics , and frequency control. Vol. 65(8), pp. 1304-1320. IEEE

(11) De Jong N, Bouakaz A and Frinking P (2002), "Basic acoustic properties of microbubbles",

Echocardiography. Vol. 19(3), pp . 229-240. Wiley

Online Library.

(12) Deffieux T, Demene C, Pernot M and Tanter M (2018),

"Functional ultrasound neuroimaging: a review of the preclinical and clinical state of the art", Current opinion in neurobiology. Vol. 50, pp . 128-

135. Elsevier.

(13) Demene C, Deffieux T, Pernot M, Osmanski B-F, Biran

V, Gennisson J-L, Sieu L-A, Bergel A, Franqui S, Correas J-M and others (2015), " Spatiotemporal clutter filtering of ultrafast ultrasound data highly increases Doppler and fUltrasound sensitivity", IEEE transactions on medical imaging. Vol. 34(11), pp. 2271-2285. IEEE.

(14) Deng CX and Lizzi FL (2002), "A review of physical phenomena associated with ultrasonic contrast agents and illustrative clinical applications", Ultrasound in medicine & biology. Vol. 28(3), pp . 277-286. Elsevier. (15) Deng CX, Lizzi FL, Kalisz A, Rosado A, Silverman RH and Coleman DJ (2000), "Study of ultrasonic contrast agents using a dual-frequency band technique", Ultrasound in medicine & biology. Vol. 26(5), pp . 819-831. Elsevier.

(16) Desailly Y, Tissier A-M, Correas J-M, Wintzenrieth

F, Tanter M and Couture 0 (2016), "Contrast enhanced ultrasound by real-time spatiotemporal filtering of ultrafast images", Physics in Medicine & Biology. Vol. 62(1), pp . 31. IOP Publishing.

(17) Eckersley RJ, Chin CT and Burns PN (2005), "Optimising phase and amplitude modulation schemes for imaging microbubble contrast agents at low acoustic power", Ultrasound in medicine & biology. Vol. 31(2), pp . 213-219. Elsevier.

(18) Ellegala DB, Carpenter JE, Klibanov AL, Shaffrey ME, Lindner JR and others (2003), "Imaging tumor angiogenesis with contrast ultrasound and microbubbles targeted to nb3", Circulation. Am Heart Assoc.

(19) Emmer M, Vos HJ, Versluis M and De Jong N (2009), "Radial modulation of single microbubbles", IEEE transactions on ultrasonics, ferroelectrics , and frequency control. Vol. 56(11) IEEE.

(20) Errico C, Pierre J, Pezet S, Desailly Y, Lenkei Z, Couture 0 and Tanter M (2015), "Ultrafast ultrasound localization microscopy for deep super resolution vascular imaging", Nature. Vol. 527(7579), pp . 499. Nature Publishing Group.

(21) Forsberg F, Shi WT and Goldberg BB (2000), "Subharmonic imaging of contrast agents", Ultrasonics. Vol. 38(1-8), pp . 93-98. Elsevier.

(22) Gessner RC, Frederick CB, Foster FS and Dayton PA

(2013), "Acoustic angiography: a new imaging modality for assessing microvasculature architecture", Journal of Biomedical Imaging. Vol. 2013, pp . 14. Hindawi Publishing Corp.

(23) Goertz DE, Cherin E, Needles A, Karshafian R, Brown AS, Burns PN and Foster FS (2005), "High frequency nonlinear B-scan imaging of microbubble contrast agents", IEEE transactions on ultrasonics, ferroelectrics , and frequency control. Vol. 52(1), pp. 65-79. IEEE.

(24) Gramiak R and Shah PM (1968), "Echocardiography of the aortic root", Investigative Radiology. Vol. 3, pp . 356-366. Elsevier.

(25) Hansen R and Angelsen BA (2009), "SURF imaging for contrast agent detection", IEEE transactions on ultrasonics, ferroelectrics , and frequency control. Vol. 56(2), pp. 280-290. IEEE.

(26) Hansen R, Masoy S-E, Tangen TA and Angelsen BA (2011), "Nonlinear propagation delay and pulse distortion resulting from dual frequency band transmit pulse complexes", The Journal of the Acoustical Society of America. Vol. 129(2), pp .

1117-1127. ASA.

(27) Hoff L (2001), "Acoustic characterization of contrast agents for medical ultrasound imaging" Springer Science & Business Media.

(28) Hwang J-J and Simpson DH (1999), "Two pulse technique for ultrasonic harmonic imaging" . September 14, 1999.

(29) Mace E, Montaldo G, Cohen I, Baulac M, Fink M and

Tanter M (2011), "Functional ultrasound imaging of the brain", Nature methods. Vol. 8(8), pp . 662.

Nature Publishing Group.

(30) Masoy S-E, Standal 0, Nasholm P, Johansen TF, Angelsen B and Hansen R (2008), "SURF imaging: In vivo demonstration of an ultrasound contrast agent detection technique", IEEE transactions on ultrasonics, ferroelectrics , and frequency control. Vol . 55(5), pp. 1112-1121. IEEE.

(31) Montaldo G, Tanter M, Bercoff J, Benech N and Fink M (2009), "Coherent plane-wave compounding for very high frame rate ultrasonography and transient elastography" , IEEE transactions on ultrasonics, ferroelectrics , and frequency control. Vol. 56(3), pp. 489-506. IEEE.

(32) Needles A, Arditi M, Rognin NG, Mehi J, Coulthard

T, Bilan-Tracey C, Gaud E, Frinking P, Hirson D and Foster FS (2010), "Nonlinear contrast imaging with an array-based micro-ultrasound system", Ultrasound in medicine & biology. Vol. 36(12), pp . 2097-2106.

Elsevier .

(33) Schneider M (1999), "Characteristics of SonoVuetrade mark", Echocardiography. Vol. 16(7), pp. 743-746.

(34) Schrope BA and Newhouse VL (1993), "Second harmonic ultrasonic blood perfusion measurement", Ultrasound in Medicine and Biology. Vol. 19(7), pp . 567-579.

Elsevier .

(35) Shariff HH, Bevan PD, Karshafian R, Williams R and Burns PN (2006), "Radial Modulation Imaging: Raising the Frequency for Contrast Imaging", In Ultrasonics Symposium, 2006. IEEE. , pp . 104-107.

(36) Simpson DH, Chin CT and Burns PN (1999), "Pulse inversion Doppler: a new method for detecting nonlinear echoes from microbubble contrast agents", In IEEE Transactions on Ultrasonics, Ferroelectrics and Frequency Control. Vol. 46(2), pp . 372-382.

(37) Tanter M, Bercoff J, Sandrin L and Fink M (2002), "

Ultrafast compound imaging for 2-D motion vector estimation: Application to transient elastography

", IEEE transactions on ultrasonics, ferroelectrics , and frequency control. Vol. 49(10), pp. 1363-1374. IEEE.

(38) Tanter M and Fink M (2014), "Ultrafast imaging in biomedical ultrasound", IEEE transactions on ultrasonics, ferroelectrics , and frequency control. Vol. 61(1), pp. 102-119. IEEE.

(39) Taylor KJW, Burns PN and Wells PNT ( 1995 ), "Clinical applications of Doppler ultrasound", Raven Press, New York, NY.

(40) Tremblay-Darveau C, Sheeran PS, Vu CK, Williams R, Zhang Z, Bruce M and Burns PN (2018) , "The role of microbubble echo phase lag in multi-pulse contrast- enhanced ultrasound imaging", IEEE Transactions on Ultrasonics, Ferroelectrics , and Frequency Control. IEEE .

(41) Weller GER, Wong MKK, Modzelewski RA, Lu E, Klibanov AL, Wagner WR and Villanueva FS (2005), "Ultrasonic imaging of tumor angiogenesis using contrast microbubbles targeted via the tumor binding peptide arginine-arginine-leucine", Cancer research. Vol. 65(2), pp . 533-539. AACR.

(42) Willmann JK, Paulmurugan R, Chen K, Gheysens 0,

Rodriguez-Porcel M, Lutz AM, Chen IY, Chen X and Gambhir SS (2008), "US imaging of tumor angiogenesis with microbubbles targeted to vascular endothelial growth factor receptor type 2 in mice", Radiology. Vol. 246(2), pp . 508-518. Radiological

Society of North America.

(43) Yu FTH, Villanueva FS and Chen X (2014), "Radial modulation contrast imaging using a 20-MHz single element intravascular ultrasound catheter", IEEE transactions on ultrasonics, ferroelectrics , and frequency control. Vol. 61(5), pp . 779-791. IEEE.

(44) 0. Couture, S. Bannouf, G. Montaldo, J-F . Aubry, M.

Fink and M. Tanter. "Ultrafast imaging of ultrasound contrast agents". Ultrasound in medicine & biology, 35(11), pp .1908-1916, 2009.

(45) 0. Couture, M. Fink and M. Tanter. "Ultrasound contrast plane wave imaging." IEEE transactions on ultrasonics, ferroelectrics , and frequency control, Vol . 59(12): 2676-2683, 2012.

(46) C. Tremblay-Darveau, R. Williams, L. Milot, M.

Bruce and P. N. Burns. "Combined perfusion and doppler imaging using plane-wave nonlinear detection and microbubble contrast agents". IEEE transactions on ultrasonics, ferroelectrics , and frequency control, 61(12), pp .1988-2000 , 2014.

(47) C. Barriere and D. Royer, "Diffraction effects in the parametric interaction of acoustic waves: application to measurements of the nonlinearity parameter B/A in liquids". IEEE transactions on ultrasonics, ferroelectrics , and frequency control, 48(6), pp.1706-1715, 2001.

SUMMARY

Embodiments described therein provide for enhanced methods and apparatus for radial modulation imaging.

To this end, a method is provided for radial modulation imaging using an array ultrasonic probe having a plurality of ultrasonic transducers and a modulation ultrasonic probe, said array ultrasonic probe and said modulation ultrasonic probe being both controlled by a control system, said method including at least the following steps:

(a) an acquisition step wherein unfocused ultrasonic pulses having an imaging frequency fi are transmitted in a medium containing microbubbles by said array ultrasonic probe at a number A of propagation directions while at least a modulation ultrasonic wave having a modulation frequency f å is transmitted in the medium by said modulation ultrasonic probe, and raw data from backscattered ultrasonic waves are acquired by said array ultrasonic probe, wherein said number A is an integer equal to 1 or larger, wherein said imaging frequency fi is at least 4 times said modulation frequency 2, wherein for the or each propagation direction, a number S of unfocused ultrasonic pulses are successively transmitted in the medium respectively at instants corresponding to constant phases cp_j of the modulation wave and the corresponding backscattered signals are acquired, j being an index comprised between 1 and S, and the S phases cp_j being distinct from one another, wherein a number N of data frames are acquired during said acquisition step, each data frame including acquiring A.S successive backscattered signals, N being at least 20;

(c) a filtering step wherein in which a set of basic data corresponding to raw data for the or each propagation direction is filtered over time by a bandpass filter having a bandpass around the modulation frequency 2, thus obtaining filtered data;

(d) an imaging step wherein at least one image is created on the basis of said filtered data.

Thanks to these dispositions, the radial modulation imaging method is particularly efficient and also simpler to implement compared to classical radial modulation imaging. As a matter of fact, the efficiency and accuracy of the method is obtained in particular because at least three different phases of the modulation of the microbubbles are interrogated by the unfocused ultrasonic pulses, thus avoiding the need to synchronize the unfocused ultrasonic pulses with a predetermined state of the microbubbles as in classical radial modulation imaging (notably compression and rarefaction phases and not between them) . Besides, no time-delay difference corresponding to the dispersion of the imaging pulse by the modulation wave needs to be corrected as in classical RMI, since the timing of the unfocused ultrasonic pulses in the present disclosure is entirely controlled by the phases cp_j of the modulation wave corresponding to the instants of the unfocused ultrasonic pulses.

Additionally, it should be noted that the method of the present disclosure enables to obtain images in a very short time.

The method may further include one and / or other of the following features:

- said bandpass filter has a bandpass of x.f_å, x being comprised between 0.01 and 0.3;

at step (c) , after being filtered by said bandpass filter, said set of basic data is multiplied by a sinusoidal signal at said modulation frequency

and filtered by a lowpass filter having a cutoff frequency lower than the modulation frequency f2, thus obtaining said filtered data;

- said cutoff frequency is comprised between 0.05 and 0.2 times the modulation frequency f2;

- said modulation ultrasonic wave has a mechanical index MI comprised between 0.005 and 0.3, wherein:

P_neg being the maximum negative pressure of the modulation ultrasonic wave in MPa and f2 being expressed in MHz;

- said modulation ultrasonic wave is a sinusoidal wave having a known phase;

- said sinusoidal wave is continuous at least during transmission of the unfocused ultrasonic pulses in the medium;

- said constant phases cp_j are spread over the range 0-2n and a maximum difference between said constant phases cp being at least n/2;

- said constant phases cp_j are substantially equally spread over the range 0-2n;

- said microbubbles have a resonance frequency and said modulation frequency f is comprised between 0.3 and 3 times said resonance frequency;

N is at least 500;

- S is not more than 30, for instance not more than 10; in a specific example, S = 4;

- A is not more than 25, for instance not more than 7; in a specific example, A = 3;

- said medium is a living body having a vasculature and said microbubbles are contained in said vasculature;

- the method further comprises, before the filtering step (c) :

(b) a beamforming step in which said raw data corresponding to the or each propagation direction (Pi) is beamformed, thus obtaining said set of basic data over time;

- said number A is 1 and said at least one image corresponds to said filtered basic data;

- said number A is more than 1 and said at least one image is obtained by coherently summing said filtered basic data corresponding respectively to the propagation directions ;

- said basic data are said raw data and said imaging step (d) includes computing said at least one image from said filtered data by synthetic imaging;

- at step d) , a set of images is obtained over time and said set of images is filtered by a bandpass filter around m.f2, wherein m is an integer equal to or larger than 2 such that m.f2 is lower than PRF/ (2.A.S), where PRF is a frequency of repetition of the unfocused ultrasonic pulses .

Besides, the disclosure proposes an apparatus for radial modulation imaging including an array ultrasonic probe having a plurality of ultrasonic transducers (2a) and a modulation ultrasonic probe, said array ultrasonic probe and said modulation ultrasonic probe being both controlled by a control system which is adapted to perform:

(a) an acquisition step wherein unfocused ultrasonic pulses having an imaging frequency fi are transmitted by said array ultrasonic probe at a number A of propagation directions while at least a modulation ultrasonic wave having a modulation frequency f å is transmitted in the medium by said modulation ultrasonic probe (11), and raw data from backscattered ultrasonic waves are acquired by said array ultrasonic probe (2), wherein said number A is an integer equal to 1 or larger, wherein said imaging frequency fi is at least 4 times said modulation frequency f2, wherein for the or each propagation direction (Pi), a number S of unfocused ultrasonic pulses are successively transmitted in the medium (1) respectively at instants corresponding to constant phases cp of the modulation wave and the corresponding backscattered signals are acquired, j being an index comprised between 1 and S, and the S phases cp being distinct from one another, wherein a number N of data frames are acquired during said acquisition step, each data frame including acquiring A.S successive backscattered signals, N being at least 20;

(c) a filtering step wherein in which a set of basic data corresponding to raw data for the or each propagation direction (Pi) is filtered over time by a bandpass filter having a bandpass around the modulation frequency f2, thus obtaining filtered data;

The apparatus may further include one and / or other of the following features:

- said bandpass filter has a bandpass of x.f2, x being comprised between 0.01 and 0.3;

said control system is adapted to, at step (c) , after filtering by said bandpass filter, multiplying said set of basic data by a sinusoidal signal at said modulation frequency f and filtering it by a lowpass filter having a cutoff frequency lower than the modulation frequency f₂, thus obtaining said filtered data;

- said cutoff frequency is comprised between 0.05 and 0.2 times the modulation frequency f ;

said modulation ultrasonic wave has a mechanical index MI comprised between 0.005 and 0.3, wherein:

- said modulation ultrasonic wave is a sinusoidal wave having a known phase;

N is at least 500;

- the control system is adapted to, before the filtering step (c) :

(b) perform a beamforming step in which said raw data corresponding to the or each propagation direction (Pi) is beamformed, thus obtaining said set of basic data over time;

- said number A is more than 1 and said control system is adapted to obtained said at least one image by coherently summing said filtered basic data corresponding respectively to the propagation directions;

- said basic data are said raw data and at said imaging step (d) , the control system is adapted to compute said at least one image from said filtered data by synthetic imaging;

at step d) , the control system is adapted to obtain a set of images over time and to filter said set of images by a bandpass filter around m.f2, wherein m is an integer equal to or larger than 2 such that m.f2 is lower than PRF/ (2.A.S), where PRF is a frequency of repetition of the unfocused ultrasonic pulses.

BRIEF DESCRIPTION OF THE DRAWINGS

Other features and advantages of the disclosure appear from the following detailed description of one non limiting example thereof, with reference to the accompanying drawings. In the following, the method of the present disclosure will be called uRMI, for "ultrafast RMI" .

In the drawings:

- Figure 1 is a schematic drawing showing an apparatus for radial modulation imaging;

- Figure 2 is a block diagram showing part of the apparatus of Figure 1;

- Figure 3 is a schematic drawing illustrating operation of the apparatus of Figures 1-2;

- Figure 4 is graph showing scattered pressure from simulations of a microbubble in MA (modulation of amplitude) imaging, microbubble disruption and RMI as a function of the mechanical index MI;

- Figure 5 shows an example of normalized frequency spectrums (divided by the sampling frequency fs, i.e. the frame rate=PRF/A . S ) for S = 3, 4, 5 and 10;

- Figure 6 shows examples of images for A=3, S= 4a and 100 kPa acquisition without flow of microbubbles: (a) one Bmode image with microbubbles, (b) one uRMI image with microbubbles and with modulation ultrasonic wave, (c) sum of 1000 uRMI images with microbubbles and with modulation ultrasonic wave, (d) one uRMI image without microbubbles without modulation ultrasonic wave, (e) one uRMI image without microbubbles with modulation ultrasonic wave, and (f) one uRMI image with microbubbles without modulation ultrasonic wave. The black and white rectangles correspond to the region of interest (ROI) of the vessel phantom and the agar, respectively. The scale bar represents 1 mm and the colorbar is in dB;

- Figure 7 shows: contrast-to-tissue ratio between microbubbles in the vessel and the phantom as a function of the amplitude of the 1 MHz modulation wave A = 3, S = 4 without flow of microbubbles and for a 1 and 3 mL/min flow

(a) and examples of images for 100 (b) , 125 (c) , 200 (d) and 250 (e) kPa without flow. The black and white rectangles correspond to the ROI of the vessel phantom and the agar, respectively. The scale bar represents 1 mm and the colorbar is in dB . Corresponding normalized frequency spectrums (divided by the sampling frequency fs) of images data-set at the vessel phantom location (f) zoomed around the modulation frequency (g) showing a split peak. The dotted lines show the limits of the bandpass filter;

- Figure 8 shows: contrast-to-tissue ratio between microbubbles in the vessel and the phantom as a function of the flow speed of microbubbles for A = 3, S = 4, 100 kPa low-frequency acquisition (a) and examples of images for 1

(b) , 5 (c) , 10 (d) and 20 (e) mL/min. The plain black and white rectangles correspond to the ROI of the vessel phantom and the agar, respectively. The dotted black and white rectangles correspond to the ROI in the center of the vessel phantom and the agar, respectively. The scale bar represents 1 mm and the colorbar is in dB . Corresponding normalized frequency spectrums (divided by the sampling frequency fs) of images data-set at the vessel phantom location of ROI1 (f) zoomed around the modulation frequency (g) . The dotted lines show the limits of the bandpass filter;

Figure 9 shows: (a) Contrast-to-tissue ratio between microbubbles in the vessel and the phantom as a function of microbubbles modulation-states number for A = 3, 100 kPa low-frequency acquisition without flow of microbubbles and for a 3 mL/min flow speed. Examples of uRMI images for a 3mL/min flow speed for 5 modulation- states (b) and for S = 10 (c) and examples of multi-band uRMI images for S = 5 (d) and for S = 10 (e) . The black and white rectangles correspond to the ROI of the vessel phantom and the agar, respectively. The scale bar represents 1 mm and the colorbar is in dB . Corresponding normalized frequency spectrums (divided by the sampling frequency fs) of images data-set at the vessel phantom location (f) zoomed around the modulation frequency. The dotted lines show the limits of the bandpass filter;

- Figure 10 shows: Contrast-to-tissue ratio between microbubbles in the vessel and the phantom as a function of the number of angles for S = 4, 100 kPa low- frequency acquisition without flow of microbubbles and for a 3 mL/min flow (a) and examples of images for A = 1 (b) , A

= 3 (c) , A = 5 (d) and A = 7 (e) with a 3 mL/min flow speed. The black and white rectangles correspond to the ROI of the vessel phantom and the agar, respectively. The scale bar represents 1 mm and the colorbar is in dB . Corresponding normalized frequency spectrums (divided by the sampling frequency fs) of images data-set at the vessel phantom location (f) zoomed around the modulation frequency (g) . The dotted lines show the limits of the bandpass filter;

- Figure 11 shows: Contrast-to-tissue ratio between microbubbles in the vessel and the phantom for images-set with and without ultrasonic modulation wave minus its median value over time as a function of the number of angles for S = 4, 100 kPa low-frequency acquisition without flow of microbubbles (a) . Amplitude of microbubbles in the vessel and amplitude in the phantom without flow as a function of the number of angles for S = 4, 100 kPa low-frequency acquisition from uRMI (b) ;

— Figure 12 shows: Contrast-to-tissue ratio between microbubbles in the vessel and the phantom as a function of the flow speed for A = 3, S = 4, 100 kPa low-frequency acquisition and for amplitude modulation, SVD filtering and microbubbles disruption (a) and examples of images for uRMI (b) , SVD filter (c) , amplitude modulation (d) Subtraction of an image with microbubbles to an image without microbubbles (e) . The black and white rectangles correspond to the ROI of the vessel phantom and the agar, respectively. The scale bar represents 1 mm and the colorbar is in dB;

- Figure 13 shows contrast-to-tissue ratio between microbubbles in the vessel and the phantom obtained with uRMI and the lock-in-amplifier and with a bandpass filter, as a function of the amplitude of the 1 MHz modulation pulse (a) (A = 3, S = 4, acquisition without flow) , of the flow speed (b) (A = 3, S = 4, 15 kPa low-frequency acquisition) , of the number of microbubble modulation- states (c) (A = 3, 15 kPa low-frequency acquisition without flow) and of the number of compounding angles (d) (S = 4, 15 kPa low-frequency acquisition, 3 mL/min flow speed);

- Figure 14 shows microbubble contrast as a function of the number of images in the filtering process of the lock-in-amplifier (A = 3, S = 4, without flow) : (a) effect of the number of images on the filtering by evaluating microbubble contrast on the sum of the first 150 images only and (b) effect of the number of images on microbubble contrast by evaluating microbubble contrast on the sum of all images.

MORE DETAILED DESCRIPTION

In the Figures, the same references denote identical or similar elements.

The apparatus shown on Figures 1 and 2 is adapted to ultrafast radial modulation imaging of a medium 1. The medium 1 can be tissues of a living being, for instance a mammal and in particular a human.

The apparatus may include for instance at least a ID or 2D array ultrasonic probe 2 having a plurality of ultrasonic transducers 2a and a modulation ultrasonic probe 11, said array ultrasonic probe 2 and said modulation ultrasonic probe 11 being both controlled by a control system 3 , 4 , 12.

Although a 2D array probe 2 is shown on Figure 1, the disclosure is applicable as well to the case of a ID array probe 2.

The array ultrasonic probe 2 may have for instance a few tens to a few thousands transducer elements 2a (T_ki, or T_k in the case of a ID probe) , with a pitch which can be for instance lower than 1mm. The transducer elements 2a of the array ultrasonic probe 2 may transmit for instance at a central imaging frequency fi comprised between 1 and 40 MHz, for instance of 15 MHz or close to 15 MHz.

One example of usable array ultrasonic probe 2 is a 15 MHz linear (ID) transducer array from Vermon, France (128 elements, pitch: 0.11mm and bandwidth: 9-24 MHz at -10 dB) .

The modulation ultrasonic probe 11 may further include at least one ultrasonic transducer T, or more, transmitting for instance at a central modulating frequency f2 comprised between 0.1 and 5 MHz, for instance 1 MHz or close to 1 MHz. Such ultrasonic modulation probe 11 may be either unitary with or separate from the array ultrasonic probe 2.

The imaging frequency fi is at least 4 times said modulation frequency f₂. For instance, the imaging frequency fi is at least 10 times said modulation frequency f₂.

The control system may for instance include a specific control unit 3, a function generator 12 (FG) and a computer 4.

In this example, the control unit 3 is used for controlling the array ultrasonic probe 2 and acquiring signals therefrom, the function generator 12 is used for controlling modulation ultrasonic probe 11 while the computer 4 is used for controlling control unit 3 and possibly function generator 11 (function generator 11 may also be autonomous and thus controlled independently of the control unit 3, provided the phase thereof remains constant), and generating image sequences from the signals acquired by control unit 3.

These functionalities could be fulfilled by any variant of the control system described above. For instance, in a variant, a single electronic device could fulfill all the functionalities of control unit 3 and computer 4. In another variant, the control unit 3 or the above single electronic device could include the function generator 12.

As shown on Figure 2, control unit 3 may include for instance (in the case of a n*n 2D array - the description would be the same, mutatis mutandis, in the case of a n ID array) :

- n*n analog/digital converters 5 (ADi_j) individually connected to the n transducers Tj_j of 2D array ultrasonic probe 2;

- n*n buffer memories 6 (Bj_j) respectively connected to the n*n analog/digital converters 5;

- a function generator 12 (FG) controlling the modulation transducer 11;

- a central processing unit 7 (CPU) communicating with the buffer memories 6 and the computer

4;

- a memory 8 (MEM) connected to the central processing unit 7 ;

- a digital signal processor 9 (DSP) connected to the central processing unit 7.

An example of control unit 3 is 256 the Vantage ® system (Verasonics, Kirkland, WA) .

As shown on Figure 3, array ultrasonic probe 2 may be controlled to transmit unfocused ultrasonic pulses at said imaging frequency fi in the medium 1, at a number A of propagation directions Pi. A is an integer which can be 1 or more, for instance not more than 25, in particular not more than 7. Each unfocused ultrasonic pulse can be in particular a plane ultrasonic wave. The propagation directions Pi may be defined by their respective angles eg with the normal N to the array ultrasonic probe 2. In the case of plane waves, the angles eg are the respective inclinations of the plane waves with respect to the array ultrasonic probe 2, as shown on Figure 3. In a particular exemple, A=3 and the angles eg may be -10°, 0°, +10°.

The array ultrasonic probe 2 is able to image a certain field Ci in the medium 1, in order to image blood vessels 21 in the medium 1 (including the smallest ones of very small diameter d) by detecting microbubbles 22 in the blood 23 of the blood vessels 21.

The modulation ultrasonic probe 11 may be disposed to transmit the modulation ultrasonic wave, for instance a plane or diverging wave, in the medium 1, in a certain field C2 covering the field Ci at least in part.

The modulation ultrasonic probe 11 may be controlled by the function generator 12 to transmit the modulation ultrasonic wave for instance as a sinusoidal wave of known phase, which is either continuous during all the acquisition or continuous at least during transmission of the unfocused ultrasonic pulses in the medium 1 by array ultrasonic probe 2. An example of function generator 12 is the DG1022A, RIGOL, Beaverton, OR, USA.

It may usually be required that the modulation ultrasonic probe 11 be inclined on the surface of the medium 1 so that the field C2 covers the field Ci, in which case this inclination can be obtained by any support 11a made in a material transparent to ultrasonic waves.

The modulation ultrasonic wave has a mechanical index MI comprised between 0.005 and 0.3, wherein:

MI = P_mglJ (1) ,

P_neg being the maximum negative pressure of the modulation ultrasonic wave in MPa and f2 being expressed in MHz.

The microbubbles may be for instance Sonovue ® microbubbles (Bracco) . The microbubbles may have a resonance frequency and said modulation frequency f2 may be comprised between 0.3 and 3 times said resonance frequency. Operation - first mode :

The apparatus may operate as follows.

(a) Acquisition :

The array ultrasonic probe 2 and modulation ultrasonic probe 11 are placed on the surface of the medium 1. While the modulation ultrasonic probe 11 is controlled by the function generator 12 to transmit the modulation ultrasonic wave having the modulation frequency f2 in the medium 1, the array ultrasonic probe 2 is controlled by the control unit 3 to transmit the unfocused ultrasonic pulses having said imaging frequency fi in the medium 1.

A series of unfocused (e.g. planar) ultrasonic pulses are transmitted at said number A of successive propagation directions Pi, and for the or each propagation direction, a number S of unfocused ultrasonic pulses are successively transmitted in the medium respectively at instants corresponding to constant phases cp of the modulation wave, j being an index comprised between 1 and

S .

The unfocused ultrasonic pulses may be transmitted for instance every 60ps, so that the pulse repetition frequency PRF is in that case 1/60 MHz = 0.0166 MHz.

The number S is an integer equal to 3 or larger. S may be not more than 30, for instance not more than 10; in a specific example, S = 4.

The phases cp_j may be spread over the range 0-2n. A maximum difference between said constant phases cpj may be at least n/2. In particular, the phases cpj may substantially equally spread over the range 0-2n.

Successive raw data sets (called RF data sets in the technical field) of backscattered ultrasonic waves are acquired by the array ultrasonic probe 2 and the control unit 3 from the successive unfocused ultrasonic pulses over time. Each raw data set corresponds to one propagation direction Pi and one phase cp_j .

The acquired signals sensed by ultrasonic transducers 2a may be sampled for instance at 31.25 MHz.

As it will be explained below, the A.S data sets corresponding to the A propagation directions and for each propagation direction, to the S constant phases, constitute a data frame which enables to reconstruct an image of the medium 1 and in particular of the microbubbles 22 in the medium. A number N of successive data frames is acquired in the acquisition step, corresponding to N images. N may be at least 20, for instance at least 500. A specific example of N is 1000. The frame rate, also called sampling frequency f_s, is equal to: fs = PRF/ (A.S) . With PRF = 1/60 MHz, A = 3 and S = 4, fs = PRF/12 = 1,38 kHz.

(b) Beamforming step

The raw data sets corresponding to the or each propagation direction Pi are then beamformed by the control system (e.g. by computer 4) as well known in the art

(beamforming in reception) , thus obtaining a set of beamformed data over time for the or each propagation direction Pi (this set of beamformed data is in fact already a set of successive images) . This set of beamformed data will be called hereafter "set of basic data".

( c) Filtering step

Each of the A sets of basic data (corresponding respectively to the A propagation directions Pi) is then filtered over time by the control system (e.g. by computer 4), by a bandpass filter having a bandpass around the modulation frequency 2, thus obtaining filtered basic data .

The bandpass filter may have a bandpass of x.f2, x being comprised between 0.01 and 0.3. For instance, the bandpass may be between 0.9 f2 and 1.1 f2. An example of bandpass filter usable here is a butterworth filter with cut-off frequencies of 0.9 f2 and 1.1 f2 with an order of 48.

In a variant, after being filtered by said bandpass filter, each set of basic data may be multiplied by a sinusoidal signal at said modulation frequency f2 and filtered by a lowpass filter having a cutoff frequency lower than the modulation frequency f2, thus obtaining a set of filtered basic data for each propagation direction Pi (the whole filtering process then forms a lock-in amplifier) .

Said cutoff frequency may be comprised between 0.05 f2 and 0.2 f , for instance 0.1 f2_. An example of lowpass filter usable here is a butterworth filter with cut-off frequency of 0.1 f2 with an order of 48.

(d) Imaging step

At least one image is then created by the control system (e.g. by computer 4) on the basis of said filtered basic data. When A = 1, said at least one image corresponds to said filtered basic data.

When said number A is more than 1, said at least one image (in practice a set of N images over time) is 5 obtained by coherently summing said sets of filtered basic data corresponding respectively to the A different propagation directions.

Such coherent summing is known in the art of synthetic imaging (also called compound imaging) , as 10 taught, inter alia, in the following publications:

Montaldo, G., Tanter, M., Bercoff, J., Benech, N., Fink, M., 2009. Coherent plane-wave compounding for very high frame rate ultrasonography and transient elastography. IEEE Trans. Ultrason. Ferroelectr. Freg. Control 56, 489- lb 506. doi : 10.1109/TUFFC.2009.1067

Nikolov, S.I., 2001. Synthetic aperture tissue and flow ultrasound imaging. Orsted-DTU, Technical University of Denmark, Lyngby, Denmark.

Nikolov, S.I., Kortbek, J., Jensen, J.A., 2010.

20 Practical applications of synthetic aperture imaging, in:

2010 IEEE Ultrasonics Symposium (IUS) . Presented at the 2010 IEEE Ultrasonics Symposium (IUS), pp . 350-358. doi : 10.1109/ULTSYM.2010.5935627

Lockwood, G.R., Talman, J.R., Brunke, S.S., 1998.

25 Real-time 3-D ultrasound imaging using sparse synthetic aperture beamforming. IEEE Trans. Ultrason. Ferroelectr. Freg. Control 45, 980-988. doi : 10.1109/58.710573

Papadacci, C., Pernot, M., Couade, M., Fink, M. & Tanter, M. High-contrast ultrafast imaging of the heart. 30 IEEE transactions on ultrasonics, ferroelectrics, and freguency control 61, 288-301 , doi : 10.1109/tuffc.2014.6722614 (2014) .

The set of images may also be filtered over time by a bandpass filter around m.f2, wherein m is an integer 35 equal to or larger than 2, thus obtaining multiband uRMI . The number m may be:

- m = 2 when A = 5 or 10;

- m = 3 when A = 10;

- m = 4 when A = 10.

The number m is such that m.f2 is lower than

PRF/ (2.A.S), where PRF is a pulse frequency repetition of the unfocused ultrasonic pulses.

Operation - second mode :

In a second mode of operation, the filtering step could be applied to the successive sets of raw data without beamforming. In that case, the method could be as follows:

(a) Acquisition :

The acquisition step of the second mode may be the same as in the first mode.

(b) Beamforming step

In the second mode, the beamforming step is omitted before filtering. In that case, the A sets of raw data corresponding respectively to the A propagation directions Pi constitute the A sets of basic data which are filtered in step (c) .

(c) Filtering step

Each set of basic data is then filtered over time by the control system (e.g. by computer 4), as described for the first mode.

(d) Imaging step

At least one image is then created by the control system (e.g. by computer 4) on the basis of said filtered basic data.

When A = 1, said at least one image is obtained by beamforming as well known in the art (beamforming in reception) .

When said number A is more than 1, said at least one image is obtained by synthetic imaging (also called compound imaging) as taught for instance in the publications cited in step (d) of the first mode.

The set of images may also be filtered over time by a bandpass filter around m.f2, wherein m is an integer equal to or larger than 2, thus obtaining multiband uRMI . The number m may be:

- m = 2 when A = 5 or 10;

- m = 3 when A = 10;

- m = 4 when A = 10.

The number m is such that m.f2 is lower than PRF/ (2.A.S), where PRF is a pulse frequency repetition of the unfocused ultrasonic pulses.

Experimental conditions :

Tests were made with an experimental setup in which Sonovue microbubbles (Bracco) , diluted at 1/3000 (with an approximate concentration of 30,000 to 16.104 microbubbles/mL) circulated at a flow rate between 0 and 20 mL/min in a 2 mm diameter vessel phantom by syringe pump or by gravity. The flow rate was determined by measuring the volume of outflow over 1 minute of flow. To guarantee the absence of flow when required, the flow was stopped 30 seconds before acquisition. This microbubble wall-less vessel phantom was included in a 4 cm thick 5% (w/w) agar phantom comprising 1% (w/w) sigma-cell (20 pm particles, Sigma-Aldrich, Saint Louis, MO) to mimic tissue scattering. A 15 MHz linear transducer array (Vermon, France, 128 elements, pitch: 0.11mm and bandwidth: 9-24 MHz at -10 dB) was placed above the agar phantom and a 1-MHz modulation transducer was positioned next to it, in a way that their pressure fields overlapped at the vessel phantom location.

High-frequency imaging was performed with the 15- MHz transducer driven by a 256 Vantage system (Verasonics, Kirkland, WA) . Plane waves with 1 cycle pulse at 15 MHz with A = 1 to 7 angles (between -10° and 10 °) were generated. The pulse-repetition-frequency (PRF) was of 1/ (60 ps) . The imaging voltage was 300 kPa peak-negative pressure. Ultrafast radial-modulation imaging consisted in addressing the microbubbles at different stage of their oscillations. Several numbers of microbubble modulation- states were tested, each fitted within the same imaging angle. For example, the sequence for a S = 4 modulation- states and A = 3 angles excitation was [state 1 at -10° , state 2 at -10°, state 3 at -10°, state 4 at -10°, state 1, 0°, etc], adding to a 12 pulses acquisition, with a frame rate of 1/ (12*60ps) . The more modulation-states or the more angles used, the lower was the frame rate. The low- frequency was generated, via the 1-MHz transducer, by a function generator (DG1022A, RIGOL, Beaverton, OR, USA) . The low-frequency was chosen in a way that high-frequency pulses every 60 ps reached microbubble in 3, 4, 5 or 10 modulation-states and the corresponded frequencies were 0.988889, 1.004167, 1.003333 and 1.001667 MHz, respectively. The amplitude pressures of the low-frequency were between 50 and 250 kPa.

For each measurement, 1000 images were acquired with the imaging transducer at a sampling frequency of 31.25 MHz. The radio-frequency data for each angle were separated and beamformed.

To increase the sensitivity of the system to the narrowband radial modulation, a lock-in amplifier was implemented on the channel data in the slow time. The demodulation process consisted to filter each channel data around the modulation frequency using a bandpass filter (using a butterworth filter with cut-off frequencies of 0.9 and 1.1 time the modulation frequency with an order of 48) . The modulation frequency is the sampling frequency in the slow time divided by the number of microbubble modulation- states. Then, the filtered signals were multiplied with a sinusoidal signal at the modulation frequency before to be filtered by a lowpass filter in the slow time (butterworth filter with a cut-off frequency of 0.1 time the sampling frequency with an order of 48) .

Reconstruction of images was performed by combining coherently the summing angles. In the case of 5 and 10 microbubble modulation-states, the beamformed images were also filtered at the modulation frequency multiplied by 2 (for 5 and 10 microbubbles modulation-states), 3 (for 10 microbubbles modulation-states) and 4 (for 10 microbubbles modulation-states) . For these two microbubbles modulation- states number, the reconstructed set of images correspond to the product of the filtered images for the modulation frequency and its multiples. These uRMI images filtered at several frequencies were named multi-band uRMI .

The amplitudes of the signal in the agar and in the vessel phantom were estimated for the sum in intensity of these 1000 processed images. Acquisitions were repeated 3 times .

Other techniques

The uRMI results were compared with other techniques to detect microbubbles: amplitude modulation, microbubbles disruption and SVD spatiotemporal filter.

1) Amplitude modulation (AM)

A Verasonic plane-waves amplitude modulation sequence with the 15-MHz transducer was implemented. In this sequence for each angles of each image with an amplitude of 300 kPa, 3 pulses were defined: a first 1 cycle 15 MHz pulses with all elements, a second with only odd elements and a third with only even elements. The amplitude modulation image was estimated by subtracting the RF data from odd and even elements to the image with all elements prior to beamforming these different images. The PRF was 1/ (3*60) ps, the same as PRF from uRMI imaging for 3 angles. One thousand images were acquired and the amplitude of the signal in the agar and in the vessel phantom were estimated for the sum in intensity of these 1000 images. Acquisitions were repeated 3 times. 2) Microbubble disruption

The contrast-to-tissue ratio obtained by microbubbles disruption should be the difference in contrast in presence to in absence of microbubbles. To evaluate the amplitude of the signal due to microbubbles, 1000 images were acquired before and after microbubbles injection in the vessel phantom. The amplitude of the signal in the agar and in the vessel phantom were estimated on the sum in intensity of the images before microbubbles injection and on the sum in intensity of images after microbubble injection. The microbubble contrast corresponding to the subtraction of the contrast-to-tissue ratio after injection to the microbubble contrast before injection. The ROI used were two times smaller than for amplitude modulation, uRMI or SVD filterting in order to limit the presence of artefacts in the vessel ROI. In comparison, the other techniques naturally limit the amplitude of the phantom.

3) SVD filter

The same sequence as the ultrafast RMI sequence without the 1 MHz modulation pulse was used. After the images acquisition, these images were beamformed and filtered with a SVD filter to remove eigenvalues corresponding to tissue and noise (eigenvalues between 3/1000 and 120/1000 were kept) . The amplitude of the signal in the agar and in the vessel phantom were estimated for the sum in intensity of these 1000 images. Acquisitions were repeated 3 times.

Experimental results

Simulation results are presented in Figure 4. For the same MI, the scattered pressure for RMI is higher than AM imaging but almost always lower than microbubbles disruption. The scattered pressure for microbubbles disruption increases linearly with the MI. However, the scattered pressure for AM and RMI increases in a quadratic manner with the MI. The scattered pressure for AM is very low compared to RMI and microbubbles disruption.

After an uRMI acquisition, a set of 1000 ultrafast Bmode images were obtained. From these Bmode images, the vessel phantom and a phantom location were determined. The selected vessel and phantom region-of-interest had the same dimensions and the same depth. The temporal fast Fourier transform of the Bmode images set, averaged at the vessel phantom location, was calculated and examples of frequency spectrum in slow time (normalized by the sampling frequency of the acquisition) are presented in Figure 5 for S = 3, 4, 5 and 10 microbubbles modulation-states. For a 3 microbubble modulation-states, peaks are present in the spectrum due to the modulation at f/fs=0.33 and 0.66. For a 4 microbubble modulation-states, peaks are present in the spectrum due to the modulation at f/fs=0.25, 0.5 and 0.75. For a 5 microbubble modulation-states, peaks are present in the spectrum due to the modulation at f/fs=0.2, 0.4, 0.6 and 0.8. For a 10 microbubble modulation-states, peaks are present in the spectrum due to the modulation at f/fs=0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8 and 0.9. Due to the number of microbubble modulation-states, a peak due to modulation was expected at f/fs=0.33 for 3 microbubble modulation-states, f/fs=0.25 for 4 microbubble modulation- states, f/fs=0.2 for 5 microbubble modulation-states and f/fs=0.1 for 10 microbubble modulation-states.

The uRMI technique was applied in presence and absence of microbubbles, with and without low-frequency excitation to evaluate their effect on the modulation and demodulation process. Example of microbubbles images acquired are presented in Figure 6. In presence of microbubbles and low- frequency excitation, microbubbles in the vessel phantom are clearly observed in the image (Figure 6 (b) ) . When the sum of the 1000 images is used, the entire vessel phantom is well defined in the uRMI image (Figure 6 (c) ) . However, without microbubbles (Figure 6d) and (e) ) or low-frequency excitation (Figure 6(d) and (f) ) , microbubbles cannot be detected. The corresponded contrast-to-tissue ratios are listed in Table 1. In presence of microbubbles, the contrast-to-tissue ratio in the Bmode image shows an increase around 3 dB . In absence of low-frequency excitation, the contrast-to-tissue ratio for uRMI is close to 0 dB . In presence of low-frequency excitation, the contrast-to-tissue ratio for uRMI is 2.41 dB without microbubbles and 12.65 dB with microbubbles.

Table 1: Contrast-to-tissue ratio between microbubbles in the vessel and the phantom for a 3 angles high-frequency excitation to evaluate the modulation/demodulation process: with and without microbubbles and with and without a low- frequency excitation (4 microbubble modulation-states, 100 kPa) .

The effect of the low-frequency excitation pressure on the contrast-to-tissue ratio is presented in Figure 7 for S =4 microbubbles modulation-states and a A = 3 angles high- frequency excitation without flow and with a 1 and 3 mL/min flow speed. The contrast-to-tissue ratio in the vessel phantom as compared to agar increases with low-frequency excitation pressure from 50 to 150 kPa and then is stable from 150 to 250 kPa without flow. This arises also for 1 and 3 mL/min flow speed (Figure 7 (a) ) . White artefacts were observed only in uRMI images for a low-frequency pressure equal or superior at 150 kPa (Figure 7 (b-e) ) . On the frequency spectrum at the vessel phantom location (Figure 7 (f-g) ) , the amplitude of the peak at the modulation frequency (f/fs=0.25) increases with the amplitude of the 1 MHz modulation. The main peak appears split for amplitude higher or equal to 150 kPa.

The effect of flow speed on the contrast-to-tissue ratio is presented in Figure 8 for 4 microbubbles modulation-states, a 3 angles high-frequency excitation and a 100 kPa low-frequency excitation. The contrast-to-tissue ratio decreases slightly with the flow speed (Figure 8 (a) ) . On the uRMI images, the signal in the centre of the vessel phantom decreases when flow speed increases (Figure 8 (b-e) ) but there is always signal in the border of the vessel phantom for the different flow speed used. It may correspond to microbubbles in the boundary layer and with a lower speed than microbubbles in the centre of the vessel phantom. When using a ROI at the centre of the vessel phantom only (dotted ROI), contrast-to-tissue ratio is slightly higher for flow speed until 5 mL/min, and the decrease of contrast-to-tissue ratio with flow speed is more pronounced. The decrease of contrast-to-tissue ratio with flow speed is related to an evolution of the frequency spectrum at the vessel phantom location (Figure 8 (f-g) ) : the amplitude of the peak at the modulation frequency (f/fs=0.25) decreases when flow speed increase and another peak with a higher frequency appears from 5 mL/min with an amplitude and a centre frequency increasing with the flow speed (from f/fs=0.26 at 5 mL/min to f/fs=0.32 at 20 mL/min) .

The effect of the number of microbubbles modulation- states on the microbubble contrast is presented in Figure 9 for 3 angles high-frequency excitation and a 100 kPa low- frequency excitation without flow and with a 3 mL/min flow speed. Without flow or with 3 mL/min flow speed, contrast- to-tissue ratios are similar (Figure 9 (a) ) : the contrast- to-tissue ratio increases slightly with the number of microbubbles modulation-states (from 10 dB to 12 dB for 3 and 10 microbubbles modulation-states, respectively) and the use of the multiple of the modulation frequency allow to increase this contrast until 17 dB for 5 microbubbles modulation-states and up to 24 dB for 10 microbubbles modulation-states. This increase of contrast-to-tissue ratio is observed on the uRMI and multi-band uRMI images (Figure 9 (b-e) ) .

The effect of the number of angles in high-frequency acquisition on the contrast-to-tissue ratio is presented in Figure 10 for 4 microbubbles modulation-states and a 100 kPa low-frequency excitation without flow and with a 3 mL/min flow speed. Without flow, there was no effect of the number of angles in high-frequency acquisition but for a 1 or 3 mL/min flow speed, the contrast-to-tissue ratio decreases with the number of angles (Figure 10(a)) . Individually the contrast of microbubbles in the vessel phantom and agar increases with the number of angles (Figure 11(b)) without flow. This increase of agar amplitude is observed on the uRMI images (Figure 10 (b-e) ) . We wished to determine if the lack of increase of microbubbles contrast with the number of angles was related to an issue of angles compounding due to microbubbles displacement with buoyancy or to an effect of RMI filtering. In this way, acquisitions with and without low- frequency excitation were recorded. The contrast-to-tissue ratio from these acquisitions was evaluated by subtracting the median image of each images set over time (Figure 11(a)) . In these conditions, there was an increase of contrast-to-tissue ratio with the number of angles (Figure 10(a)) with and without low-frequency excitation. Examples of frequency spectrum at the vessel phantom location are presented in Figure lO(f-g) . In these spectrum, the amplitude of the peak at the modulation frequency (f/fs=0.25) increases with the number of angles and another peak with a higher frequency appears until 3 angles with an amplitude and a center frequency increasing with the number of angles (from f/fs=0.26 for 3 angles to f/fs=0.275 for 7 angles) .

The contrast-to-tissue ratio was estimated with other techniques and compared to the contrast-to-tissue ratio obtained with uRMI for 4 modulation-states, 3 angles high- frequency excitation and a 100 kPa low-frequency excitation. Amplitude modulation was applied for the same amplitude of excitation of the high-frequency than the one used for uRMI, 300 kPa. The SVD filter was applied on acquisitions with the same conditions as for uRMI but without low-frequency excitation. Microbubbles maximum contrast after their disruption was evaluated by acquiring an image without microbubbles and with microbubbles, without flow only. Results are presented in Figure 12 as well as example of images obtained with these techniques. Except in the without flow condition, SVD filter provided higher contrast-to-tissue ratios than other techniques (around 16 to 17 dB in our conditions) . Without flow, uRMI provided the best contrast-to-tissue ratio (around 10 dB) . Contrast-to-tissue ratios obtained with amplitude modulation were close to 4 dB and close to 6 dB for microbubbles disruption.

The use of a lock-in amplifier (selective bandpass filter combined with a demodulation stage) instead of a selective bandpass filter without demodulation stage allowed to increase the relative contrast-to-tissue ratio from 26 to 48 % on average (examples are presented on Fig. 13), and up to 65%. The lock-in amplifier was particularly interesting for the low amplitudes of the modulation excitation, a limited number of compounding angles and modulation states, along with the slow flow. Moreover, the microbubbles videos obtained after demodulation with the lock-in amplifier presented fewer artifacts as compared to a simple bandpass filter. The minimum number of images for the lock-in amplifier was 144 (defined by the Matlab® functions used) . There was no effect on microbubbles contrast of the number of images in the lock-in amplifier filter process (Fig. 14 (a) ) . In in vivo conditions, the number of images may be optimized as a function of the number of modulation states, angles and imaging depth, to attain sufficient image repetition frequency. The increase in the number of images led to an increase of microbubbles contrast (Fig. 14 (b) ) .

Claims

CLAIMS :

1. Method for radial modulation imaging using an array ultrasonic probe (2) having a plurality of ultrasonic transducers (2a) and a modulation ultrasonic probe (11), said array ultrasonic probe (2) and said modulation ultrasonic probe (11) being both controlled by a control system (3, 4, 12), said method including at least the following steps:

(a) an acquisition step wherein unfocused ultrasonic pulses having an imaging frequency fi are transmitted in a medium (1) containing microbubbles (22) by said array ultrasonic probe (2) at a number A of propagation directions (Pi) while at least a modulation ultrasonic wave having a modulation frequency f å is transmitted in the medium by said modulation ultrasonic probe (11), and raw data from backscattered ultrasonic waves are acquired by said array ultrasonic probe (2), wherein said number A is an integer equal to 1 or larger, wherein said imaging frequency fi is at least 4 times said modulation frequency f2, wherein for the or each propagation direction (Pi), a number S of unfocused ultrasonic pulses are successively transmitted in the medium (1) respectively at instants corresponding to constant phases cp of the modulation wave and the corresponding backscattered signals are acquired, j being an index comprised between 1 and S, and the S phases cp being distinct from one another, wherein a number N of data frames are acquired during said acquisition step, each data frame including acquiring A.S successive backscattered signals, N being at least 20;

2. Method according to claim 1, wherein said bandpass filter has a bandpass of x.f , x being comprised between 0.01 and 0.3.

3. Method according to claim 1 or claim 2, wherein at step (c) , after being filtered by said bandpass filter, said set of basic data is multiplied by a sinusoidal signal at said modulation frequency f₂ and filtered by a lowpass filter having a cutoff frequency lower than the modulation frequency f₂, thus obtaining said filtered data.

4. Method according to claim 3, wherein said cutoff frequency is comprised between 0.05 and 0.2 times the modulation frequency f₂.

5. Method according to any one of the preceding claims, wherein said modulation ultrasonic wave has a mechanical index MI comprised between 0.005 and 0.3, wherein :

P_neg being the maximum negative pressure of the modulation ultrasonic wave in MPa and f₂ being expressed in MHz.

6. Method according to any one of the preceding claims, wherein said modulation ultrasonic wave is a sinusoidal wave having a known phase.

7. Method according to any one of the preceding claims, wherein said constant phases cp_j are spread over the range 0-2n and a maximum difference between said constant phases cp_j being at least n/2.

8. Method according to any one of the preceding claims, wherein S is not more than 30 and A is not more than 25.

9. Method according to any one of the preceding claims, wherein said medium (1) is a living body having a vasculature (21) and said microbubbles (22) are contained in said vasculature (21) .

10. Method according to any one of the preceding claims, further comprising, before the filtering step (c) :

(b) a beamforming step in which said raw data corresponding to the or each propagation direction (Pi) is beamformed, thus obtaining said set of basic data over time .

11. Method according to claim 10, wherein said number A is 1 and said at least one image corresponds to said filtered basic data.

12. Method according to claim 10, wherein said number A is more than 1 and said at least one image is obtained by coherently summing said filtered basic data corresponding respectively to the propagation directions (Pi) .

13. Method according to any one of claims 1-9, wherein said basic data are said raw data and said imaging step (d) includes computing said at least one image from said filtered data by synthetic imaging.

14. Method according to claim 12 or claim 13, wherein at step d) , a set of images is obtained over time and said set of images is filtered by a bandpass filter around m.f2, wherein m is an integer equal to or larger than 2.

15. Apparatus for radial modulation imaging including an array ultrasonic probe (2) having a plurality of ultrasonic transducers (2a) and a modulation ultrasonic probe (11), said array ultrasonic probe (2) and said modulation ultrasonic probe (11) being both controlled by a control system (3, 4, 12) which is adapted to perform:

(a) an acquisition step wherein unfocused ultrasonic pulses having an imaging frequency fi are transmitted by said array ultrasonic probe (2) at a number A of propagation directions (Pi) while at least a modulation ultrasonic wave having a modulation frequency f is transmitted in the medium by said modulation ultrasonic probe (11), and raw data from backscattered ultrasonic waves are acquired by said array ultrasonic probe (2), wherein said number A is an integer equal to 1 or larger, wherein said imaging frequency fi is at least 4 times said modulation frequency 2, wherein for the or each propagation direction (Pi), a number S of unfocused ultrasonic pulses are successively transmitted in the medium (1) respectively at instants corresponding to constant phases cp_j of the modulation wave and the corresponding backscattered signals are acquired, j being an index comprised between 1 and S, and the S phases cp_j being distinct from one another, wherein a number N of data frames are acquired during said acquisition step, each data frame including acquiring A.S successive backscattered signals, N being at least 20;

(c) a filtering step wherein in which a set of basic data corresponding to raw data for the or each propagation direction (Pi) is filtered over time by a bandpass filter having a bandpass around the modulation frequency 2, thus obtaining filtered data;