Advertisement

Imaging Methods for Ultrasound Contrast Agents

      Abstract

      Microbubble contrast agents were introduced more than 25 years ago with the objective of enhancing blood echoes and enabling diagnostic ultrasound to image the microcirculation. Cardiology and oncology waited anxiously for the fulfillment of that objective with one clinical application each: myocardial perfusion, tumor perfusion and angiogenesis imaging. What was necessary though at first was the scientific understanding of microbubble behavior in vivo and the development of imaging technology to deliver the original objective. And indeed, for more than 25 years bubble science and imaging technology have evolved methodically to deliver contrast-enhanced ultrasound. Realization of the basic bubbles properties, non-linear response and ultrasound-induced destruction, has led to a plethora of methods; algorithms and techniques for contrast-enhanced ultrasound (CEUS) and imaging modes such as harmonic imaging, harmonic power Doppler, pulse inversion, amplitude modulation, maximum intensity projection and many others were invented, developed and validated. Today, CEUS is used everywhere in the world with clinical indications both in cardiology and in radiology, and it continues to mature and evolve and has become a basic clinical tool that transforms diagnostic ultrasound into a functional imaging modality. In this review article, we present and explain in detail bubble imaging methods and associated artifacts, perfusion quantification approaches, and implementation considerations and regulatory aspects.

      Key Words

      Introduction

      Contemplating the impressive range of techniques for imaging and therapy that exploit the unique interaction of ultrasound with bubbles described in this issue, it is hard to imagine that none of these were considered when ultrasound contrast agents were first developed. Indeed, almost all of the pivotal clinical trials that resulted in the approval of the first agents were made using the same fundamental imaging that was being used routinely in clinical examinations of the time, essentially unchanged from the first bubble injections made by
      • Gramiak R.
      • Shah P.M.
      Echocardiography of the aortic root.
      . It was known, of course, that bubbles driven around their resonant frequency undergo non-linear oscillation and produce specific echoes (
      • Noltigk B.E.
      • Neppiras E.E.
      Cavitation produced by ultrasonics.
      ;
      • Trilling L.
      Collapse and rebound of a gas bubble.
      ;
      • Leighton T.G.
      The acoustic bubble.
      ) and, indeed, that such echoes could be detected from ultrasound contrast agents insonated at diagnostic frequencies (
      • Miller D.L.
      • Neppiras E.A.
      On the oscillation mode of gas-filled micropores.
      ;
      • de Jong N.
      • Hoff L.
      Ultrasound scattering properties of Albunex microspheres.
      ;
      • Leighton T.G.
      The acoustic bubble.
      ;
      • Uhlendorf V.
      • Hoffmann C.
      Nonlinear acoustical response of coated microbubbles in diagnostic ultrasound.
      ), but developers of the new agents were loathe to suggest that new imaging technology was needed to use them, and indeed, it is likely that equipment manufacturers would have been reluctant to contemplate supporting a contrast agent that needed a specialized transducer, for example. Thus the first proposed use of the agents was as an enhancer of the fundamental echo, specifically to “rescue” Doppler signals from blood that were otherwise too weak to be detected, for example, in tumors (
      • Cosgrove D.
      Ultrasound contrast enhancement of tumours.
      ). While such a method proved to be effective in the rare cases in which Doppler failed because of poor signal-to-noise ratio (
      • Fobbe F.
      • Siegert J.
      • Fritzsch T.
      • Koch H.C.
      • Wolf K.J.
      [Color-coded duplex sonography and ultrasound contrast media—Detection of renal perfusion defects in experimental animals].
      ;
      • Ries F.
      • Kaal K.
      • Schultheiss R.
      • Solymosi L.
      • Schlief R.
      Air microbubbles as a contrast medium in transcranial Doppler sonography.
      ;
      • von Bibra H.
      • Sutherland G.
      • Becher H.
      • Neudert J.
      • Nihoyannopoulos P.
      Clinical evaluation of left heart Doppler contrast enhancement by a saccharide-based transpulmonary contrast agent.
      ), in the majority of situations where Doppler fails to detect blood flow, the reason is that the receiver is overwhelmed by the 40-60 dB higher signal from tissue. Doppler methods rely solely on the movement of red blood cells to separate blood flow from this clutter signal. The method therefore fails when the velocity of blood flow approaches that of tissue, at about 1 cm/s, explaining the inability of Doppler to detect tissue perfusion. The addition of microbubbles circumvents this issue by providing a blood flow signal independent of velocity.
      The challenge of imaging perfusion in real time has been met by the development of bubble-specific imaging methods described in this paper. In common, they exploit the non-linear components of the bubble echo to differentiate it from tissue, regardless of the flow velocity of the blood carrying the bubbles, hence suppressing clutter and enabling segmentation of the tissue echo from that of bubbles, whose flow velocity and distribution reflect those of blood (
      • Levine R.A.
      • Teichholz L.E.
      • Goldman M.E.
      • Steinmetz M.Y.
      • Baker M.
      • Meltzer R.S.
      Microbubbles have intracardiac velocities similar to those of red blood cells.
      ). With the advent of bubbles that were able to cross the pulmonary circulation, the first report of second harmonic Doppler to detect flow at the capillary level was by
      • Schrope B.
      • Newhouse V.L.
      • Uhlendorf V.
      Simulated capillary flow measurement using a nonlinear ultrasonic contrast agent.
      in vitro and then in experimental animals (
      • Schrope B.A.
      • Newhouse V.L.
      Second harmonic ultrasonic blood perfusion measurement.
      ), using the agent Levovist. (This and all cited studies that involved human patients and animals in the present review article had received Institutional Review Board/Institutional Animal Care and Use Committee approval.) At the same time, efforts by
      • Burns P.N.
      • Powers J.E.
      • Fritzsch T.
      Harmonic imaging: A new imaging and Doppler method for contrast enhanced ultrasound.
      to modify a clinical scanner with one of the early digital beamformers revealed that existing array transducers had sufficient bandwidth to produce second harmonic gray-scale images in vivo, reversing the contrast between the tissue and the bubbles, at the expense of spatial resolution. Subsequent implementation of harmonic detection in power Doppler imaging with the same agent allowed suppression of tissue clutter and demonstration of flow in moving kidneys down to a vessel diameter identified by histologic examination as 40 µm (
      • Burns P.N.
      • Powers J.E.
      • Hope Simpson D.
      • Brezina A.
      • Kolin A.
      • Chin C.T.
      • Uhlendorf V.
      • Fritzsch T.
      Harmonic power mode Doppler using microbubble contrast agents: An improved method for small vessel flow imaging.
      ). Using harmonic spectral Doppler,
      • Mulvagh S.L.
      • Foley D.A.
      • Aeschbacher B.C.
      • Klarich K.K.
      • Seward J.B.
      Second harmonic imaging of an intravenously administered echocardiographic contrast agent: Visualization of coronary arteries and measurement of coronary blood flow.
      were able to measure coronary flow reserve under vasodilation stress using a transthoracic approach, normally rendered impossible by clutter from cardiac motion. Using the new gray-scale systems, (
      • Porter T.R.
      • Xie F.
      • Kricsfeld D.
      • Armbruster R.W.
      Improved myocardial contrast with second harmonic transient ultrasound response imaging in humans using intravenous perfluorocarbon-exposed sonicated dextrose albumin.
      ;
      • Porter T.R.
      • Xie F.
      • Li S.
      • D'Sa A.
      • Rafter P.
      Increased ultrasound contrast and decreased microbubble destruction rates with triggered ultrasound imaging.
      ) found that by imaging intermittently with a few seconds between acquisitions, harmonic images could be made that revealed myocardial perfusion. A similar intermittent technique revealed perfusion of the liver, which yielded the first reports of enhancement patterns in focal liver lesions (
      • Kono Y.
      • Moriyasu F.
      • Mine Y.
      • Nada T.
      • Kamiyama N.
      • Suginoshita Y.
      • Matsumura T.
      • Kobayashi K.
      • Chiba T.
      Gray-scale second harmonic imaging of the liver with galactose-based microbubbles.
      ;
      • Wilson S.R.
      • Burns P.N.
      • Muradali D.
      • Wilson J.A.
      • Lai X.
      Harmonic hepatic US with microbubble contrast agent: Initial experience showing improved characterization of hemangioma, hepatocellular carcinoma, and metastasis.
      ). It became clear to the scientific community that these signals were associated with disruption of the bubbles by the incident ultrasound, and that the interval was required to allow fresh bubbles to replenish the scan plane (
      • Burns P.N.
      • Wilson S.R.
      • Muradali D.
      • Powers J.E.
      • Greener Y.
      Microbubble destruction is the origin of harmonic signals from FS069.
      ). That it was not possible to reduce the transmit power to a level that allowed real-time perfusion imaging was a reflection of both the lack of sensitivity of the imaging method and the instability of the air-based agents such as Levovist in use at the time. Furthermore, at the relatively high mechanical indexes (MIs) used, the tissue echo was contaminated with propagation harmonics that further reduced the signal-to-clutter ratio.
      As the so-called “second generation” of agents with low-solubility gas cores emerged (
      • Unger E.
      • Shen D.
      • Fritz T.
      • Kulik B.
      • Lund P.
      • Wu G.-L.
      • Yellowhair D.
      • Ramaswami R.
      • Matsunaga T.
      Gas-filled lipid bilayers as ultrasound contrast agents.
      ;
      • Schneider M.
      • Arditi M.
      • Barrau M.B.
      • Brochot J.
      • Broillet A.
      • Ventrone R.
      • Yan F.
      BR1: A new ultrasonographic contrast agent based on sulfur hexafluoride-filled microbubbles.
      ), it was clear that these could be excited at low MI to produce continuous harmonic echoes at real-time frame rates and remain stable for several minutes (

      Averkiou M, Powers JE, Bruce M, Skyba DM. Realtime ultrasonic imaging of perfusion using ultrasonic contrast agents. 2001. U.S. Patent 6,171, 246.

      ). New broadband non-linear imaging methods were quickly developed that formed the basis of the methods used today. These are multipulse techniques, in which successive transmit pulses are modulated by phase (
      • Hope Simpson D.
      • Chin C.T.
      • Burns P.N.
      Pulse inversion Doppler: A new method for detecting nonlinear echoes from microbubble contrast agents.
      ) or amplitude (

      Brock-Fisher GA, Poland MD, Rafter PG. 1996 Means for increasing sensitivity in non-linear ultrasound imaging systems. U.S. Patent US5577505A.

      ) or both (
      • Eckersley R.J.
      • Chin C.T.
      • Burns P.N.
      Optimising phase and amplitude modulation schemes for imaging microbubble contrast agents at low acoustic power.
      ), and the echoes recombined to form the bubble-specific signal.
      The present review article covers bubble imaging methods, presented in historical order with high-MI methods first, followed by the current low-MI methods and reference to tissue harmonic imaging. The principal contrast imaging artifacts are discussed and explained, followed by a short review of the perfusion quantification methods. The final section considers implementation issues including transducers, hardware architecture, transmitters, power supplies and regulatory aspects.

      Bubble imaging methods

      High MI

      The asymmetry of microbubble oscillations leads to non-linear backscattered echoes (
      • Leighton T.G.
      The acoustic bubble.
      ). The recognition of the advantage of these non-linear signals to improve the microbubble signal-to-tissue ratios led to the development of the first dedicated microbubble imaging methods. These first “harmonic imaging” approaches utilized existing/conventional imaging sequences and signal paths with two predominate modifications: (i) a reduction of the transmit frequency to the lower edge of the transducer's bandwidth, and (ii) a shift of the receive filter to twice the transmit frequency.
      The U.S. Food and Drug administration (FDA) regulates the acoustic amplitude of diagnostic ultrasound systems with the MI, which is defined as
      MI=pneg/f
      (1)


      where pneg is the peak negative pressure, and f is the frequency (
      • Apfel R.E.
      • Holland C.K.
      Gauging the likelihood of cavitation from short-pulse, low-duty cycle diagnostic ultrasound.
      ). The FDA requires that MI is ≤1.9; typically, most imaging is done around MI = 1. Imaging of microbubbles began with imaging modes (e.g., B-mode and color Doppler) and MIs readily available on diagnostic ultrasound systems at that time. The role of microbubble disruption by diagnostic ultrasound pressures was not fully appreciated until the mid-1990s (
      • Burns P.N.
      • Wilson S.R.
      • Muradali D.
      • Powers J.E.
      • Greener Y.
      Microbubble destruction is the origin of harmonic signals from FS069.
      ). Real-time imaging (frame rates >15 Hz) of microbubbles at higher MIs destroys them in lower-velocity vessels faster than they can replenish the imaging plane. (
      • Porter T.R.
      • Xie F.
      • Kricsfeld D.
      • Armbruster R.W.
      Improved myocardial contrast with second harmonic transient ultrasound response imaging in humans using intravenous perfluorocarbon-exposed sonicated dextrose albumin.
      ;
      • Porter T.R.
      • Xie F.
      • Li S.
      • D'Sa A.
      • Rafter P.
      Increased ultrasound contrast and decreased microbubble destruction rates with triggered ultrasound imaging.
      ) found that triggering at incremental intervals eventually allowed sample volumes to fully replenish in the heart. Triggering enabled the extended visualization of lower concentrations of agent, solving the problem of large doses and associated attenuation problems (
      • Thapar A.
      • Shalhoub J.
      • Averkiou M.
      • Mannaris C.
      • Davies A.H.
      • Leen E.L.
      Dose-dependent artifact in the far wall of the carotid artery at dynamic contrast-enhanced US.
      ). The two principal high-MI harmonic imaging methods used with triggered acquisitions were harmonic B-mode and harmonic power Doppler imaging, described below. These early harmonic modes were first used in attempts to visualize blood flow in the myocardium and followed by subsequent work in the liver (
      • Main M.L.
      • Grayburn P.A.
      Clinical applications of transpulmonary contrast echocardiography.
      ;
      • Becher H.
      • Burns P.N.
      Handbook of contrast echocardiography: Left ventricular function and myocardial perfusion.
      ;
      • Wermke W.
      • Gaßmann B.
      Tumour diagnostics of the liver with echo enhancers: Colour atlas.
      ).

      Harmonic B-mode imaging

      The lower transmit frequency allows the second harmonic component to fit in the transducer's bandwidth, placing the second harmonic signal in the center of the transducer's bandwidth when possible. Because microbubbles generate larger non-linear signals than tissue, the resulting harmonic signal-to-tissue ratio (10–30 dB) is much better than that of fundamental signals (∼0 dB) (Fig. 1).
      Fig 1
      Fig. 1Fundamental (a) and harmonic (b) images of contrast agent in the left ventricle.
      Relatively narrow transmit bandwidths are used to provide enough separation between the fundamental and second harmonic components, such that there is no overlap in the frequency spectrum. Although developed for contrast, harmonic imaging has many beneficial properties for imaging tissue and is now commonly used in many clinical applications (
      • Averkiou M.A.
      • Roundhill D.N.
      • Powers J.E.
      A new imaging technique based on the nonlinear properties of tissues.
      ). Triggered high-MI acquisitions are used in echocardiography to allow for microbubble replenishment to visualize blood flow in the microcirculation of the myocardium (
      • Porter T.R.
      • Xie F.
      • Kricsfeld D.
      • Armbruster R.W.
      Improved myocardial contrast with second harmonic transient ultrasound response imaging in humans using intravenous perfluorocarbon-exposed sonicated dextrose albumin.
      ;
      • Porter T.R.
      • Xie F.
      • Li S.
      • D'Sa A.
      • Rafter P.
      Increased ultrasound contrast and decreased microbubble destruction rates with triggered ultrasound imaging.
      ;
      • Becher H.
      • Burns P.N.
      Handbook of contrast echocardiography: Left ventricular function and myocardial perfusion.
      ). Real-time acquisitions for left ventricular opacification (where replenishment of the contrast is rapid) visualize the movement of the endocardium in difficult-to-image patients, particularly during stress (
      • Mulvagh S.L.
      • DeMaria A.N.
      • Feinstein S.B.
      • Burns P.N.
      • Kaul S.
      • Miller J.G.
      • Monaghan M.
      • Porter T.R.
      • Shaw L.J.
      • Villanueva F.S.
      Contrast echocardiography: Current and future applications.
      ). While remaining a major clinical application of ultrasound contrast, this is now performed with the low-MI techniques described below.

      Harmonic power Doppler

      Contrast-to-tissue ratio is higher for harmonic B-mode than for fundamental B-mode imaging. Interestingly the contrast-to-tissue ratio is constant in the range 0.1 < MI < 2.0 (
      • Keravnou C.P.
      • Mannaris C.
      • Averkiou M.A.
      Accurate measurement of microbubble response to ultrasound with a diagnostic ultrasound scanner.
      ), which suggests that the increase in the microbubble signal caused by increasing the MI is offset by the increase in the tissue harmonic signals generated by non-linear propagation. Applying the same harmonic filters described above to conventional power Doppler further suppressed the background tissue signal because of the wall filter of the Doppler signal path. Harmonic power Doppler used the same conventional Doppler sequencing and again reduced the transmit frequency and increased the receive filtering to twice the transmit frequency. The wall filter provides a second stage of tissue clutter removal. Harmonic power Doppler, in addition to suppressing the tissue, exploits the transient nature (flow and destruction) of microbubble response at high MI (
      • Burns P.N.
      • Powers J.E.
      • Hope Simpson D.
      • Brezina A.
      • Kolin A.
      • Chin C.T.
      • Uhlendorf V.
      • Fritzsch T.
      Harmonic power mode Doppler using microbubble contrast agents: An improved method for small vessel flow imaging.
      ;
      • Tiemann K.
      • Pohl C.
      • Schlosser T.
      • Goenechea J.
      • Bruce M.
      • Veltmann C.
      • Kuntz S.
      • Bangard M.
      • Becher H.
      Stimulated acoustic emission: Pseudo-Doppler shifts seen during the destruction of nonmoving microbubbles.
      ;
      • Becher H.
      • Tiemann K.
      • Kuntz-Hehner S.
      • Omran H.
      • Schlosser T.
      Diagnostic impact of contrast echocardiography for assessment of left ventricular function and myocardial perfusion in patients with coronary artery disease.
      ). Lower pulse repetition frequencies are used in a compromise of balancing transient microbubble responses with the amount tissue motion removal of the wall filter. Because of microbubble disruption, continued insonification of the same plane in real time (frame rate >1 Hz) clears agent faster than sample volumes can be replenished in tissue, so that triggered acquisitions are required for the image plane to refill with microbubbles. Figure 2 is a representative real-time power harmonic Doppler image of the heart acquired at an MI of 1.1 using Optison. For myocardial imaging, triggered acquisitions consist of skipping 2–4 heart cycles (the myocardium replenishes in about 1–4 s), with real-time low-MI imaging to aid the user in maintaining the imaging plane. Although utilized primarily to visualize deficits of blood flow in the myocardium, harmonic power Doppler has also been successfully used to detect and characterize lesions in the liver, where an injection was followed by a single sweep of the liver (
      • Wermke W.
      • Gaßmann B.
      Tumour diagnostics of the liver with echo enhancers: Colour atlas.
      ).
      Fig 2
      Fig. 2Power harmonic image of the heart using Optison acquired at MI=1.4 and while triggering once every four cardiac cycles.

      Additional high-MI applications

      A number of other high-MI approaches have been investigated. Color flow and color power Doppler modes at destructive MI levels, in both fundamental and harmonic modes, can form bubble disruption images. This was especially successful for liver metastasis detection with Levovist (
      • Albrecht T.
      • Blomley M.J.
      • Burns P.N.
      • Wilson S.
      • Harvey C.J.
      • Leen E.
      • Claudon M.
      • Calliada F.
      • Correas J.M.
      • LaFortune M.
      • Campani R.
      • Hoffmann C.W.
      • Cosgrove D.O.
      • LeFevre F.
      Improved detection of hepatic metastases with pulse-inversion US during the liver-specific phase of SHU 508A: Multicenter study.
      ;
      • Wermke W.
      • Gaßmann B.
      Tumour diagnostics of the liver with echo enhancers: Colour atlas.
      ) and other agents (
      • Wilson S.R.
      • Burns P.N.
      • Muradali D.
      • Wilson J.A.
      • Lai X.
      Harmonic hepatic US with microbubble contrast agent: Initial experience showing improved characterization of hemangioma, hepatocellular carcinoma, and metastasis.
      ). Microbubbles collect in the normal liver parenchyma in the late portal phase but not in the metastases. The Doppler sequence causes bubble destruction that is registered as Doppler signal, as illustrated in Figure 3. In a color flow Doppler mode (Fig. 3b), echo decorrelation caused by bubble destruction creates a random “mosaic red–blue” pattern. Various researchers have called these “stimulated acoustic emission” or “agent detection imaging” (
      • Blomley M.J.
      • Albrecht T.
      • Cosgrove D.O.
      • Eckersley R.J.
      • Butler-Barnes J.
      • Jayaram V.
      • Patel N.
      • Heckemann R.A.
      • Bauer A.
      • Schlief R.
      Stimulated acoustic emission to image a late liver and spleen-specific phase of Levovist in normal volunteers and patients with and without liver disease.
      ,
      • Blomley M.J.
      • Albrecht T.
      • Cosgrove D.O.
      • Patel N.
      • Jayaram V.
      • Butler-Barnes J.
      • Eckersley R.J.
      • Bauer A.
      • Schlief R.
      Improved imaging of liver metastases with stimulated acoustic emission in the late phase of enhancement with the US contrast agent SH U 508A: Early experience.
      ;
      • Ichino N.
      • Horiguchi Y.
      • Imai H.
      • Osakabe K.
      • Nishikawa T.
      • Sugita Y.
      • Utsugi H.
      • Togo Y.
      • Sawai T.
      • Mizoguchi Y.
      Contrast-enhanced sonography of pancreatic ductal carcinoma using agent detection imaging.
      ).
      Fig 3
      Fig. 3Bubble destruction imaging with harmonic power Doppler (a) and color flow Doppler (b) at MI=1.4. Microbubbles collect in the normal parenchyma but not in the metastatic lesion. Decorrelation caused by bubble destruction is registered as a Doppler signal and delineates the lesions.
      Harmonic spectral Doppler was used to suppress Doppler clutter in detection of coronary flow (
      • Mulvagh S.L.
      • Foley D.A.
      • Aeschbacher B.C.
      • Klarich K.K.
      • Seward J.B.
      Second harmonic imaging of an intravenously administered echocardiographic contrast agent: Visualization of coronary arteries and measurement of coronary blood flow.
      ). In a more tailored approach, a separate dedicated high-amplitude destruction pulse can be followed by low-amplitude imaging pulses (
      • Frinking P.J.
      • Céspedes E.
      • De Jong N.
      Multi-pulse ultrasound contrast imaging based on a decorrelation detection strategy.
      ;
      • Kirkhorn J.
      • Frinking P.
      • de Jong N.
      • Torp H.
      Improving the sensitivity of power Doppler for ultrasound contrast imaging by using a high power release burst.
      ). Another approach to improve the contrast-to-tissue ratio exploits the subharmonic response of microbubbles, where returning microbubble echoes are imaged at half the transmit frequency, and response from tissue is reduced in comparison to second harmonic imaging (
      • Forsberg F.
      • Shi W.T.
      • Goldberg B.
      Subharmonic imaging of contrast agents.
      ;
      • Eisenbrey J.
      • Dave J.
      • Halldorsdottir V.
      • Merton D.
      • Machado P.
      • Liu J.
      • Miller C.
      • Gonzalez J.
      • Park S.
      • Dianis S.
      Simultaneous grayscale and subharmonic ultrasound imaging on a modified commercial scanner.
      ). Multiple groups have investigated the use of higher harmonics of the microbubble response (superharmonics) in trading off tissue response with increased attenuation for different applications (
      • Bouakaz A.
      • de Jong N.
      Native tissue imaging at superharmonic frequencies.
      ;
      • Shelton S.E.
      • Lindsey B.D.
      • Tsuruta J.K.
      • Foster F.S.
      • Dayton P.A.
      Molecular acoustic angiography: A new technique for high-resolution superharmonic ultrasound molecular imaging.
      ).

      Low MI for real-time imaging

      As the early air-based agents were superseded by those with low-solubility gas cores and hence greater stability, low-MI imaging began to supplant the methods described above. There were two basic reasons for this. First, at low MI (<0.1), the microbubbles are not destroyed and thus they can be repeatedly imaged in real time. Second, at low MI (low acoustic pressure), the propagation of ultrasound in tissue is linear (no harmonic generation takes place at low MIs), whereas the microbubble oscillation and resulting scattering process are non-linear. Therefore, at low MI it is possible to make a “binary,” microbubble-only image based on non-linear backscattered signals while enhancing the contrast-to-tissue ratio (CTR), which is an essential parameter in quantifying the performance of contrast-specific imaging methods.
      The harmonic imaging described in the previous section relies on transmitting at a (fundamental) frequency f0 and forming an image from the second harmonic component 2f0 of the backscattered echoes by the use of filters to remove the fundamental component (
      • Schrope B.A.
      • Newhouse V.L.
      Second harmonic ultrasonic blood perfusion measurement.
      ;
      • Burns P.N.
      Harmonic imaging with ultrasound contrast agents.
      ;
      • Averkiou M.A.
      • Roundhill D.N.
      • Powers J.E.
      A new imaging technique based on the nonlinear properties of tissues.
      ). This technique restricts the bandwidth of the transmitted pulse so that (i) the second harmonic component is (separable from the fundamental, and (b) both components fit within the bandwidth of the transducer. This bandwidth restriction results in reduced axial image resolution (because of the use of lower frequency and longer pulses), requiring the trade-off between CTR and spatial/axial resolution for the implementation of second harmonic imaging. Additional options for contrast-specific imaging, aiming at boosting the CTR, consist of filtering the backscattered signal in either the subharmonic or the ultraharmonic frequency range (i.e., at half of the fundamental harmonic or between the first and second harmonics, respectively)(
      • Forsberg F.
      • Shi W.T.
      • Goldberg B.
      Subharmonic imaging of contrast agents.
      ;
      • Frinking P.J.
      • Bouakaz A.
      • Kirkhorn J.
      • Ten Cate F.J.
      • de Jong N.
      Ultrasound contrast imaging: Current and new potential methods.
      ).

      Pulsing schemes for non-linear signal detection

      As an alternative and improvement to filter-based solutions, multipulse schemes have been developed. They consist of transmitting a series of pulses and then combining their respective scattered signals in a way that all linear components cancel and non-linear components enhance (
      • Averkiou M.A.
      • Mannaris C.
      • Bruce M.
      • Powers J.
      Nonlinear pulsing schemes for the detection of ultrasound contrast agents.
      ). The three most used pulsing schemes on commercial systems today are pulse inversion (PI) (
      • Hope Simpson D.
      • Chin C.T.
      • Burns P.N.
      Pulse inversion Doppler: A new method for detecting nonlinear echoes from microbubble contrast agents.
      ), amplitude modulation (AM) (

      Brock-Fisher GA, Poland MD, Rafter PG. 1996 Means for increasing sensitivity in non-linear ultrasound imaging systems. U.S. Patent US5577505A.

      ;
      • Mor-Avi V.
      • Caiani E.G.
      • Collins K.A.
      • Korcarz C.E.
      • Bednarz J.E.
      • Lang R.M.
      Combined assessment of myocardial perfusion and regional left ventricular function by analysis of contrast-enhanced power modulation images.
      ) and amplitude-modulated pulse inversion (AMPI or contrast pulse sequencing) (
      • Phillips P.J.
      Contrast pulse sequences (CPS): Imaging nonlinear microbubbles.
      ), and will be explained here. Two (or more) pulses (p1, p2) are sent consecutively in the body for imaging (separated by the depth-limited pulse repetition frequency [PRF]), where the second pulse (p2) is the inverse of the first in PI (p2 = –p1), the scaled at half-amplitude in AM (p2 = p1/2) and the inverse scaled at half-amplitude in AMPI (p2 = –p1/2) (see Fig. 4). The backscattered signals (p1s, p2s) from these pulses are collected and summed to form the resulting pulsing schemes according to
      PSPI=12p1s+12p2s,
      (2)


      PSAM=13p1s23p2s
      (3)


      PSAMPI=13p1s+23p2s
      (4)


      If the microbubble oscillation is linear, the backscattered signals resemble the transmitted signals, and the summation of the signals, according to eqns (2)–(4), results in zero. Any non-linear activity results in signals after the summation. Thus, these pulsing schemes cancel the linear components in the backscattered signals while detecting the non-linear ones.
      Fig 4
      Fig. 4Three pulses used for pulsing schemes: normal (a), inverted (b) and half-amplitude (c).
      We illustrate the above basic pulsing schemes for (i) detecting non-linear scattered echoes from microbubbles during contrast-enhanced ultrasound (CEUS) and (ii) detecting non-linear echoes from tissue after non-linear propagation without bubbles.

      Non-linear scattering from microbubbles

      For non-linear scattering from microbubbles we will use a simple bubble dynamics model (the Gilmore equation [
      • Leighton T.G.
      The acoustic bubble.
      ]) to model the scattered echo from a single 2-μm free bubble and generate the pulsing schemes. The three pulses of Figure 4 produced the scattered echoes and spectra in Figure 5 (a–f). The scattered pressure ps is calculated from the radius (R), bubble wall velocity (R˙) and bubble wall acceleration (R¨) with the equation
      ps=ρRr(2R2+R˙R¨)
      (5)


      where ρ is the medium density and r is the initial bubble radius.
      Fig 5
      Fig. 5Simulation of bubble echoes from the ultrasound pulses of . Normal, inverted and half (a–c) and their spectra (d–f). From the bubble echoes, the three pulsing schemes (g–i) are formed with their spectra (j–l).
      The application of eqns (2)–(4) to the echoes in Figure 5 (a–c) produces PI, AM and AMPI, as illustrated in Figure 5 (g–i), respectively, with their spectra shown in Figure 5 (j–l). For convenience and consistency, we use purple to indicate PI, brown for AM and green for AMPI. Despite the very low acoustic pressure to drive these microbubbles, the scattered echoes are quite non-linear and their spectra are rich in harmonic components (2-5). We also observe that PI detects only the even harmonic components whereas AM and AMPI have non-linear signal at all harmonic components including the fundamental. In addition, AM and AMPI have the same amount of fundamental and third harmonic component, with AMPI having a stronger second harmonic component but still a little lower than the second harmonic component of PI.
      This “non-linear fundamental” is something only observed in pulsing schemes and deserves a little more explanation. For a system driven with a single input (a single pulse in our case), the fundamental component in the response is by definition the linear component, and the harmonics the non-linear. However, the “non-linear fundamental” in the spectra of pulsing schemes (AM or AMPI) is the scaled difference in the fundamental component due to the different amounts of non-linear signal in the two pulses used. In AM, the echo of the half-pulse has a certain amount of fundamental and harmonics. The second echo of the full pulse also has a certain amount of fundamental but more harmonics because of the higher amplitude used. One could consider the harmonics as energy that was redistributed from the fundamental. Thus, by scaling the echoes, the result in the fundamental band is not zero, and it is due to the different non-linearity in the response because of the different amplitudes.

      Non-linear propagation

      Now we consider non-linear propagation of the same three pulses in Figure 4 (produced with a focused transducer), but at a much higher amplitude (source pressures of 200 kPa and pressures arriving at the focus 1–2 MPa) and form the three pulsing schemes of eqns (2)–(4). This scenario describes tissue harmonic imaging (
      • Averkiou M.A.
      Tissue harmonic ultrasonic imaging.
      ) where non-linear signals are generated by the non-linear propagation of ultrasound in a medium and in the absence of microbubbles. The scattering from tissue is linear but the scattered echoes contain non-linear components from non-linear propagation of ultrasound. The distorted pulses that arrive at the focus after non-linear propagation and their spectra are illustrated in Figure 6 (a–c) and (d–f). The application of eqns (2)–(4) to the pulses of Figure 6 (a–c) produces PI, AM and AMPI, as illustrated in Figure 6 (g–i, respectively), with their spectra shown in Figure 6 (j–l). We note that the frequency spectra of the pulsing schemes in Figures 5 and 6 have similar overall trends (see plots j–l). However, for non-linear propagation in tissue the MI must be higher than 0.3. On the other hand, microbubbles are typically destroyed at MIs >0.1. Thus, by scanning at MI <0.1, tissue does not produce harmonics, microbubbles are not destroyed and real-time bubble imaging is possible. This point is illustrated in Figure 7 with the frequency spectra of pulsing schemes from acoustic measurements of propagation of ultrasound in water at both low (Fig. 7a–c) and high (Fig. 7d–f) MI and measured bubble echoes at low MI (0.1). Figure 7 confirms with measurements the same spectral trends of the pulsing schemes discussed in Figures 5 and 6, i.e., that PI detects the even harmonic components, and AM and AMPI detect both even and odd but at slightly lower amplitudes. It is also confirmed that at low MI, no harmonics are generated and detected during propagation with any of the pulsing schemes, and the spectral signatures of bubble echoes and non-linear propagation are qualitatively similar. Finally, Figure 7 illustrates how tissue and microbubbles are separated by simply using low MIs to excite bubbles and avoid tissue harmonics. It is noted that non-linear propagation increases with frequency. When scanning bubbles at higher frequencies, it is sometimes necessary to use slightly higher MIs to make up for using a frequency higher than the resonance frequency, and this may result in a slightly higher amount of second harmonic being generated, thus having incomplete removal of the tissue background (
      • Tang M.X.
      • Kamiyama N.
      • Eckersley R.J.
      Effects of nonlinear propagation in ultrasound contrast agent imaging.
      ).
      Fig 6
      Fig. 6Simulation of non-linear propagation of the three pulses in (at a higher amplitude) from the transducer to the focus (a–c) and their respective spectra (d–f). By addition of the three non-linear pulses according to –(), the resulting pulses for the pulsing schemes are generated (g–i) with their respective spectra (j–l).
      Fig 7
      Fig. 7Measurements of low (0.1)- and high (1.7)-MI ultrasound propagation in water illustrate the effect of non-linear propagation, and measured echoes from microbubbles at low MI illustrate non-linear scattering even at low (0.1) MI.
      It is important to note that in CEUS (and low-MI real-time imaging of microbubbles), the low-MI ultrasound linearly propagates to the area being imaged, the microbubbles non-linearly oscillate and produce non-linear echoes and the bubble echoes linearly propagate back to the transducer because their amplitude is extremely low. For tissue harmonic imaging, high-MI ultrasound propagates non-linearly to the area being imaged, linear scattering from tissue produces echoes that already have harmonics (because of the non-linear propagation) and tissue echoes linearly propagate back to the transducer (just like the microbubble echoes) as their amplitude is extremely low. (see Fig. 8).
      Fig 8
      Fig. 8Schematic explanation of forward propagation, scattering and back-propagation process for contrast-enhanced ultrasound (CEUS) and tissue harmonic imaging (THI).
      With CEUS and the non-linear pulsing schemes discussed above, liver lesions are routinely detected and characterized. In Figure 9, AM at low MI was used to show a tumor (in this case a colorectal metastasis) during the different phases of liver enhancement. The tumor (segmented with a white outline) is enhancing at the peak of the arterial phase (Fig, 9a) at 20 s after contrast injection, but it is already washing out at the peak of the portal phase (Fig. 9b) at 35 s after injection and has washed out completely in the late portal phase (Fig. 9c) at 60 s. To be able to collect this information in real time, microbubble destruction must be completely avoided while tissue signals are suppressed and microbubble non-linear signals are still generated, in this case using the AM technique.
      Fig 9
      Fig. 9Images of a colorectal liver metastasis scanned with amplitude modulation at low mechanical index (0.07) to avoid bubble destruction and reveal tumor enhancement in the different vascular phases. Images at the peak of the arterial phase (a), the peak of the portal venous phase (b) and the late portal phase (c).

      Real-time imaging of stiffer shelled bubbles

      Second-generation ultrasound contrast agents brought improved stability in the circulation and began a migration away from triggered (high-MI) imaging approaches. Agents such as Lumason/Sonovue (Bracco Imaging S.p.A., Milan, Italy), Optison (GE Healthcare, Amersham, UK) and Definity/Luminity (Lantheus Medical Imaging, Billerica, MA, USA) can be imaged continuously at low MI, but the more rigid shell and the size distribution of Sonazoid (GE Healthcare, Amersham, UK) results in differences in both pharmacodynamics and the optimum acoustic conditions needed for imaging in the blood pool (
      • Sontum P.C.
      Physicochemical characteristics of Sonazoid, a new contrast agent for ultrasound imaging.
      ). Once injected, Sonazoid is taken up rapidly by the reticuloendothelial system, which allows visualization of a liver post-vascular phase, known as the Kupffer phase, in addition to the standard liver assessment performed with other agents (
      • Nakano H.
      • Ishida Y.
      • Hatakeyama T.
      • Sakuraba K.
      • Hayashi M.
      • Sakurai O.
      • Hataya K.
      Contrast-enhanced intraoperative ultrasonography equipped with late Kupffer-phase image obtained by sonazoid in patients with colorectal liver metastases.
      ). The more rigid shell means that the agent requires higher acoustic amplitudes both to produce non-linear oscillation and for disruption. Real-time imaging of the agent in this post-vascular phase is optimal with an MI slightly higher MI than that necessary for other agents (0.1 < MI < 0.3). Compared with Definity and Optison, Sonazoid reaches a threshold for shell buckling at a higher amplitude (
      • Tremblay-Darveau C.
      • Sheeran P.S.
      • Vu C.K.
      • Williams R.
      • Zhang Z.
      • Bruce M.
      • Burns P.N.
      The role of microbubble echo phase lag in multipulse contrast-enhanced ultrasound imaging.
      ). As a result, vascular imaging of this agent is performed at a “mid-MI” setting – usually on the order of twofold higher output than used for low-MI agents, which varies by system and imaging strategy. The differences in transmit and receive settings are usually consolidated in an independent “mid-MI” contrast pre-set mode, specific to this agent.

      Image processing—Maximum intensity projection

      Various post-processing algorithms applied in real time can be used to visualize different aspects of the vasculature. One algorithm that has been extensively used is temporal maximum-intensity projection. Real-time CEUS imaging provides critical diagnostic information with high temporal resolution, but it is often useful to consolidate the information captured over many frames to visualize vascular morphology more completely. For example, during periods of the early contrast wash-in, when no perfusion signal is present, accumulating the earliest arriving bubbles from the arterial phase via a maximum-intensity projection (or some other similar approach to accumulate strong and/or non-stationary signals) builds a map of the arterial supply to the site of interest with rich information on the order that vessels fill (
      • Wilson S.R.
      • Jang H.J.
      • Kim T.K.
      • Iijima H.
      • Kamiyama N.
      • Burns P.N.
      Real-time temporal maximum-intensity-projection imaging of hepatic lesions with contrast-enhanced sonography.
      ;
      • Forsberg F.
      • Ro R.J.
      • Fox T.B.
      • Liu J.-B.
      • Chiou S.-Y.
      • Potoczek M.
      • Goldberg B.B.
      Contrast enhanced maximum intensity projection ultrasound imaging for assessing angiogenesis in murine glioma and breast tumor models: A comparative study.
      ). Similarly, applying such a technique in the late phase where bubbles are sparse can be used to compare healthy perfused tissue from a lesion of interest or to accumulate sparse bubbles and build a map of the vasculature in the imaging plane.
      These approaches are often less successful at the peak of the bolus signal, as strong perfusion signals make it difficult to distinguish vessels against a dense capillary network. However, applying one or more high-MI “flash” imaging frames before the accumulation clears the perfusion bed by disrupting bubbles in the scan plane. This flash–accumulation sequence can be triggered at any point in the contrast wash-in to build a map of the vasculature that reperfuses quickly, and follows a logic similar to that of contrast quantification by disruption–replenishment (see Quantification section). Significant tissue motion, such as that from breathing, can degrade the accumulation processing, so some form of motion compensation is usually necessary.

      Contrast imaging artifacts

      A number of imaging artifacts are uniquely associated with the use of microbubble contrast; here we list some of the more important ones.

      Doppler “blooming.”

      The dynamic range of the amplitude of a spectral or color Doppler signal is greater than the range that is displayed, so enhancing the Doppler signal with an ultrasound contrast agent can result in an apparent increase in the maximum Doppler shift frequency in spectral Doppler (resulting in an overestimation of flow velocity) or the size of a region of detected flow in color Doppler (resulting in an overestimation of the extent of a flow-containing region) (
      • Forsberg F.
      • Liu J.B.
      • Burns P.N.
      • Merton D.A.
      • Goldberg B.B.
      Artifacts in ultrasonic contrast agent studies.
      ). This artifact is most apparent when using contrast enhancement to “rescue” weak Doppler signals, one of the first proposed applications for the agents, though less commonly used now.

      Bubble “noise.”

      Spectral Doppler often exposes bubbles to an incident pressure that is sufficient to rupture them, creating popping sounds that can be heard as the bubble echoes disappear from one Doppler pulse to the next (
      • Forsberg F.
      • Liu J.B.
      • Burns P.N.
      • Merton D.A.
      • Goldberg B.B.
      Artifacts in ultrasonic contrast agent studies.
      ). In Doppler, this pulse-to-pulse decorrelation causes a random Doppler shift to be estimated, resulting in a thin streak of frequencies in a spectral display (
      • Tiemann K.
      • Pohl C.
      • Schlosser T.
      • Goenechea J.
      • Bruce M.
      • Veltmann C.
      • Kuntz S.
      • Bangard M.
      • Becher H.
      Stimulated acoustic emission: Pseudo-Doppler shifts seen during the destruction of nonmoving microbubbles.
      ;
      • Fetzer D.T.
      • Rafailidis V.
      • Peterson C.
      • Grant E.G.
      • Sidhu P.
      • Barr R.G.
      Artifacts in contrast-enhanced ultrasound: A pictorial essay.
      ) or a mosaic of colors in color Doppler imaging that can be seen, for example, when sweeping a liver that has bubbles (such as Levovist or Sonazoid) sequestered in Kupffer cells (as seen in Fig. 3). Deliberate disruption of bubbles in this way allows their detection with high sensitivity. The phenomenon, described above, was once incorrectly attributed to a release of energy by the bubble and called “stimulated acoustic emission” (
      • Blomley M.J.
      • Albrecht T.
      • Cosgrove D.O.
      • Eckersley R.J.
      • Butler-Barnes J.
      • Jayaram V.
      • Patel N.
      • Heckemann R.A.
      • Bauer A.
      • Schlief R.
      Stimulated acoustic emission to image a late liver and spleen-specific phase of Levovist in normal volunteers and patients with and without liver disease.
      ).

      Inadequate or failed tissue suppression

      Multipulse CEUS imaging relies on the assumption that the transmitted pulses are precisely formed, that they and their echoes propagate through tissue linearly, and that any phase shifts induced by their path affect each pulse equally. Otherwise, a residual clutter signal from tissue may be present after summation. A common source of tissue clutter is the transmitted waveforms themselves, which may contain distortions in phase or amplitude so that they do not cancel, even in ideal tissue. For example, hardware limitations may limit the precision of the phase of pulses at low output levels, compromising tissue suppression. The result is a “pseudo-enhancement” that can both mask real enhancement caused by bubbles or be incorrectly interpreted as real enhancement. Receiver saturation from strong echoes can also cause non-linearity and failure of cancellation of echoes from specific structures. This is especially prominent where specular reflections occur, such as in subcutaneous fat interfaces, superficial liver capsule, diaphragm and carotid wall. If the overload occurs in the analogue stage before digital conversion, the only mitigation is to reduce analogue gain at the expense of bubble sensitivity. In most systems, these artifacts are almost always evident. A highly echogenic liver with extensive fatty infiltration, for example, may show enhancement without the injection of microbubbles. Lowering the gain reduces such pseudo-enhancement, at the expense of sensitivity to the bubbles. Confirmation that enhancement is due to the bubbles may be made by a high-MI “flash.” This will disrupt the bubbles but not affect the pseudo-enhancement. A final common cause of failed tissue cancellation is motion. Although contrast-specific modes are usually presented on the scanner as gray-scale images, multipulse modes employ a form of Doppler detection (
      • Burns P.N.
      • Wilson S.R.
      • Simpson D.H.
      Pulse inversion imaging of liver blood flow: Improved method for characterizing focal masses with microbubble contrast.
      ). Thus, they are sensitive to motion of tissue or of the transducer, both of which create incompletely cancelled fundamental echoes usually seen as a flash of pseudo-enhancement of tissue. The susceptibility of a given image to tissue/transducer motion depends on the detection algorithm and its characteristics, but fast tissue motion such as respiration, cardiac motion or the snapping of tendons produces artifacts in most non-linear modes (
      • Deslandes M.
      • Guillin R.
      • Cardinal E.
      • Hobden R.
      • Bureau N.J.
      The snapping iliopsoas tendon: New mechanisms using dynamic sonography.
      ).

      Non-linear propagation through a bubble cloud

      A particular form of pseudo-enhancement is produced by the propagation of the ultrasound beam through a non-linear medium, such as a liver or a carotid artery containing bubbles. A distal structure that is echogenic but not enhancing (such as the diaphragm) will appear enhanced because the system detects non-linear backscatter from the object even though it does not contain bubbles (
      • Tang M.X.
      • Eckersley R.J.
      Nonlinear propagation of ultrasound through microbubble contrast agents and implications for imaging.
      ;
      • Gerrit L.
      • Renaud G.G.
      • Akkus Z.
      • van den Oord S.C.
      • Folkert J.
      • Shamdasani V.
      • Entrekin R.R.
      • Sijbrands E.J.
      • de Jong N.
      • Bosch J.G.
      Far-wall pseudoenhancement during contrast-enhanced ultrasound of the carotid arteries: Clinical description and in vitro reproduction.
      ;
      • Thapar A.
      • Shalhoub J.
      • Averkiou M.
      • Mannaris C.
      • Davies A.H.
      • Leen E.L.
      Dose-dependent artifact in the far wall of the carotid artery at dynamic contrast-enhanced US.
      ). Using a high-MI “flash,” as above, will make the pseudo-enhancement disappear, as it destroys the bubbles in the beam's path. In the liver, the time characteristic of the pseudo-enhancement follows that of the organ parenchyma, not of the arterial enhancement of the tumor, thus enabling its identification (
      • Yu H.
      • Jang H.J.
      • Kim T.K.
      • Khalili K.
      • Williams R.
      • Lueck G.
      • Hudson J.
      • Burns P.N.
      Pseudoenhancement within the local ablation zone of hepatic tumors due to a nonlinear artifact on contrast-enhanced ultrasound.
      ). This phenomenon is more prominent at higher frequencies.

      Bubble destruction in the near field

      A commonly observed artifact in CEUS is near-field bubble destruction. The high acoustic pressure near the transducer face and the overlap between sequential transmit beams at superficial depths have the effect of disrupting microbubbles over time, even at lower MIs. This results in reduced enhancement at the apex in an apical cardiac image or appearance of a dark “band” in the first 1–3 cm around the capsule in a liver scan. If the bubbles are stationary or near-stationary, rotating the transducer 90° or moving to a new scan plane often reveals the extent of the depletion relative to tissue that has not been imaged (
      • Fetzer D.T.
      • Rafailidis V.
      • Peterson C.
      • Grant E.G.
      • Sidhu P.
      • Barr R.G.
      Artifacts in contrast-enhanced ultrasound: A pictorial essay.
      ). For flowing bubbles (e.g., in a cardiac chamber or arteries of the liver), depending on the bubble velocity and the frame rate of the image, new bubbles may not have time to wash into the imaging plane between insonations. Thus, this artifact is flow velocity dependent. If the flow in the lesion is particularly slow, continuous insonation can selectively destroy those bubbles that dwell for a long period and are thus subject to many pulses of sound. The result is that continuous scanning of a liver lesion such as a hemangioma can lose contrast enhancement and “wash out”; intermittent scanning is therefore preferred (
      • Dietrich C.F.
      • Mertens J.C.
      • Braden B.
      • Schuessler G.
      • Ott M.
      • Ignee A.
      Contrast-enhanced ultrasound of histologically proven liver hemangiomas.
      ). Conversely, raising the MI is sometimes used as a deliberate strategy to highlight larger vessels in the imaging plane. As microbubble disruption occurs as a function of pulse design and PRF, bubble destruction can be mitigated by control of such parameters as PRF, MI or transmit beam pattern.

      Reverberation

      Late-returning echoes produced by previous transmit events that have returned from deep interfaces may produce pseudo-enhancement in contrast imaging (
      • Fetzer D.T.
      • Rafailidis V.
      • Peterson C.
      • Grant E.G.
      • Sidhu P.
      • Barr R.G.
      Artifacts in contrast-enhanced ultrasound: A pictorial essay.
      ). As these echoes may not be created by the same transmit configuration (i.e., differences in polarity, amplitude between transmits), the reverberation signal may not cancel after summation. Many of the multipulse sequences used for CEUS naturally cancel reverberations, but those not cancelled may become evident depending on the actual sequence and timing. Because this artifact is produced by an insufficient delay between pulses to allow for echo decay, increasing the pulse repetition interval sufficiently can usually remove the artifact entirely. However, this comes at the expense of frame rate.

      Dosing artifacts

      An optimum contrast examination requires an optimum contrast agent dose, which is almost always considerably less than that included in the product labeling. Too low a dose not only produces inadequate enhancement of perfused structures, but enhancement that is both depth dependent (with enhancement of deeper-lying structures being lost to attenuation) and time dependent (with inadequate enhancement in the portal or late phases in the liver, for example). Too high a dose, on the other hand, produces attenuation in the parenchyma of perfused organs such as the liver, both shadowing distal structures and exacerbating the non-linear propagation artifact and blooming artifacts described above. Although it is true that a larger body habitus generally requires a slightly higher bolus dose, it should be noted that the optimum dose also changes according to cardiopulmonary function, which itself varies from subject to subject. Using a higher ultrasound frequency and examining poorly perfused structures generally require a slightly higher dose of contrast agent (
      • Dietrich C.F.
      • Ignee A.
      • Greis C.
      • Cui X.W.
      • Schreiber-Dietrich D.G.
      • Hocke M.
      Artifacts and pitfalls in contrast-enhanced ultrasound of the liver.
      ).

      The high-MI “veil.”

      In some situations where the highest sensitivity is required to determine the distribution of bubbles in the liver (e.g., in the post-vascular phase of Sonazoid), a high-MI imaging mode can be used that detects the bubbles by deliberately disrupting them. Some scanners have modes that are specifically designed in this way, displaying the high fundamental signal that results from the decorrelation between two consecutive pulses. However, if the density of bubbles in the liver is sufficient, attenuation by bubbles in superficial tissue will shield distal bubbles, so the first frame only disrupts a layer of superficial bubbles. After these are destroyed, the next frame will disrupt bubbles in a deeper layer, and so on, until all the bubbles in the imaging frame are disrupted (Fig. 10). The effect is a band of enhancement that travels from superficial to deep tissue, disrupting bubbles and appearing like a bright “veil.” The width of the band and the speed of its passage distally depend on the bubble population and ultrasound beam characteristics, as well as the imaging mode and frame rate, but will typically descend through the liver in 0.5–3 s (
      • Wilson S.R.
      • Burns P.N.
      • Muradali D.
      • Wilson J.A.
      • Lai X.
      Harmonic hepatic US with microbubble contrast agent: Initial experience showing improved characterization of hemangioma, hepatocellular carcinoma, and metastasis.
      ).
      Fig 10
      Fig. 10Demonstration of the high MI “veil.” These are consequent images from a loop illustrating bubble destruction in the liver progressing in depth with time. The white arrow on the right side of the images indicates the approximate location of maximum brightness.

      Quantification

      The blood pool nature of ultrasound contrast agents together with the well-established non-linear imaging techniques for their imaging lend themselves to quantification of blood flow in both the macro- and microcirculation. Two methods have been developed for CEUS blood flow quantification: bolus transit and constant infusion with destruction–replenishment.

      Bolus transit method

      Blood flow in tissue can be assessed within the image plane after the passage of a bolus injection of contrast agent administered into a peripheral vein (see Fig. 11a). Approaches have been developed for perfusion quantification of tumors and organs from CEUS video loops based on indicator dilution theory. The indicator dilution technique allows for measurements of flow parameters such as flow rate (F), volume (V) and mean transit time (MTT). If a known amount of indicator is injected in an unknown volume, and its concentration measured, this volume can be determined, under the assumption of good mixing. In CEUS, the indicator is the injected microbubbles, but their concentration in the body as a function of time is not known. Instead, a measure of the backscattered intensity from the microbubbles is used. However, the backscattered intensity is proportional to microbubble concentration (
      • Lampaskis M.
      • Averkiou M.
      Investigation of the relationship of nonlinear backscattered ultrasound intensity with microbubble concentration at low MI.
      ), thus allowing the use of an indicator dilution technique (
      • Williams R.
      • Hudson J.M.
      • Lloyd B.A.
      • Sureshkumar A.R.
      • Lueck G.
      • Milot L.
      • Atri M.
      • Bjarnason G.A.
      • Burns P.N.
      Dynamic microbubble contrast-enhanced US to measure tumor response to targeted therapy: A proposed clinical protocol with results from renal cell carcinoma patients receiving antiangiogenic therapy.
      ;
      • Dietrich C.
      • Averkiou M.
      • Correas J.M.
      • Lassau N.
      • Leen E.
      • Piscaglia F.
      An EFSUMB introduction into dynamic contrast-enhanced ultrasound (DCE-US) for quantification of tumour perfusion.
      ;
      • Leen E.
      • Averkiou M.
      • Arditi M.
      • Burns P.
      • Bokor D.
      • Gauthier T.
      • Kono Y.
      • Lucidarme O.
      Dynamic contrast enhanced ultrasound assessment of the vascular effects of novel therapeutics in early stage trials.
      ). Further details are in the review on CEUS quantitation (S.Turco, P. Frinking, R. Wildeboer, M. Arditi, M. Averkiou, H. Wijkstra, J. Lindner, M. Mischi) published in the current issue.
      Fig 11
      Fig. 11(a) Spread of a bolus injection after passage through the heart and lungs and arrival at a lesion in the liver. (b) Steady-state infusion of microbubbles and the destruction–replenishment protocol.
      Briefly, the technique consists of collecting a CEUS video loop after a bolus injection of microbubbles, placing a region of interest on the image and plotting the linearized image intensity as a function of time, referred to as the time–intensity curve (TIC) (see Fig. 12a). The TIC is fitted with an indicator dilution model to remove noise and extract certain parameters that characterize the passage of the bolus, such as peak intensity (Ip), wash-in time, MTT and area under the curve (
      • Dietrich C.
      • Averkiou M.
      • Correas J.M.
      • Lassau N.
      • Leen E.
      • Piscaglia F.
      An EFSUMB introduction into dynamic contrast-enhanced ultrasound (DCE-US) for quantification of tumour perfusion.
      ;
      • Leen E.
      • Averkiou M.
      • Arditi M.
      • Burns P.
      • Bokor D.
      • Gauthier T.
      • Kono Y.
      • Lucidarme O.
      Dynamic contrast enhanced ultrasound assessment of the vascular effects of novel therapeutics in early stage trials.
      ).
      Fig 12
      Fig. 12Time–intensity curves from a bolus transit method (a) and a destruction–replenishment method (b). A lognormal model is fitted to the data in both cases. MTT = mean transit time; WIT = wash-in time.
      Breathing and transducer motion introduce noise into the TICs and limit the accuracy of the extracted parameters. Respiratory gating with post-processing techniques have been implemented and have limited the issue of motion in perfusion quantification (
      • Averkiou M.
      • Lampaskis M.
      • Kyriakopoulou K.
      • Skarlos D.
      • Klouvas G.
      • Strouthos C.
      • Leen E.
      Quantification of tumor microvascularity with respiratory gated contrast enhanced ultrasound for monitoring therapy.
      ;
      • Christofides D.
      • Leen E.
      • Averkiou M.A.
      Automatic respiratory gating for contrast ultrasound evaluation of liver lesions.
      ,
      • Christofides D.
      • Leen E.L.
      • Averkiou M.A.
      Improvement of the accuracy of liver lesion DCEUS quantification with the use of automatic respiratory gating.
      ). Limitations of the bolus method include the uncertainty of the input function of bubble concentration to the region of interest caused by the spreading of the bolus during its transit to the measurement site. Methods to deconvolve the arterial input function from the true tissue enhancement have been investigated (
      • Hockham N.
      • Coussios C.C.
      • Arora M.
      A real-time controller for sustaining thermally relevant acoustic cavitation during ultrasound therapy.
      ). In practice, the ultrasound image plane must remain in a single position for the entire study, so that studies employing a single bolus are also limited to a single plane. This issue is mitigated with the introduction of 4-D CEUS imaging and further analysis of the passage of a bolus in a volume. The single-bolus method remains a widely used technique owing to its general simplicity of implementation, its high contrast signal-to-noise ratio in most tissues of interest and its relative success as a therapy monitoring tool (
      • Lamuraglia M.
      • Bridal S.L.
      • Santin M.
      • Izzi G.
      • Rixe O.
      • Paradiso A.
      • Lucidarme O.
      Clinical relevance of contrast-enhanced ultrasound in monitoring anti-angiogenic therapy of cancer: Current status and perspectives.
      ).

      Constant infusion with destruction–replenishment

      This method, also known as the “negative bolus” indicator dilution method, relies on the unique ability of ultrasound to disrupt the contrast agent within the imaging plane, creating a “negative” bolus (see Fig. 11b). The wash-in of new agent is then imaged, the rate of which is proportional to flow. The method does not require knowledge of an injected input function and has provided good accuracy compared with invasive flow measurement (Hudson et al. 2011). The contrast agent is introduced into the venous circulation as a continuous, diluted infusion, typically by means of an infusion pump or a drip bag. The systemic concentration of contrast agent will stabilize after approximately 1–2 min and will remain constant for the length of the investigation, typically 10–15 min for a single dose of Definity (
      • Williams R.
      • Hudson J.M.
      • Lloyd B.A.
      • Sureshkumar A.R.
      • Lueck G.
      • Milot L.
      • Atri M.
      • Bjarnason G.A.
      • Burns P.N.
      Dynamic microbubble contrast-enhanced US to measure tumor response to targeted therapy: A proposed clinical protocol with results from renal cell carcinoma patients receiving antiangiogenic therapy.
      ). Perfusion rate and other hemodynamic parameters are then assessed after disruption using a brief burst of high-MI ultrasound called “flash,” followed by an immediate return to low-MI imaging. In the original work by
      • Wei K.
      • Jayaweera A.
      • Firoozan S.
      • Linka A.
      • Skyba D.
      • Kaul S.
      Quantification of myocardial blood flow with ultrasound-induced destruction of microbubbles administered as a constant venous infusion.
      , the subsequent enhancement as the scan plane replenishes was modeled by a mono-exponential function. This model assumes that the vascular tree can be considered a perfect mixing chamber, whereas in reality it comprises a branching conduit structure. Models that more accurately reflect these fractal microvascular geometries (
      • Karshafian R.
      • Burns P.N.
      • Henkelman M.R.
      Transit time kinetics in ordered and disordered vascular trees.
      ;
      • Krix M.
      • Kiessling F.
      • Vosseler S.
      • Farhan N.
      • Mueller M.M.
      • Bohlen P.
      • Fusenig N.E.
      • Delorme S.
      Sensitive noninvasive monitoring of tumor perfusion during antiangiogenic therapy by intermittent bolus-contrast power Doppler sonography.
      ), as well as realistic ultrasound fields (
      • Potdevin T.
      • Fowlkes J.
      • Moskalik A.
      • Carson P.
      Analysis of refill curve shape in ultrasound contrast agent studies.
      ), have evolved into comprehensive models that incorporate microvascular structure and ultrasound beam geometries throughout the scan field (
      • Averkiou M.A.
      Tissue harmonic ultrasonic imaging.
      ;
      • Arditi M.
      • Frinking P.J.
      • Zhou X.
      • Rognin N.G.
      A new formalism for the quantification of tissue perfusion by the destruction–replenishment method in contrast ultrasound imaging.
      ). Each development has led to improved measurement reproducibility (
      • Hudson J.M.
      • Leung K.
      • Burns P.N.
      The lognormal perfusion model for disruption replenishment measurements of blood flow: In vivo validation.
      ), as well as accuracy of flow measurement compared with metered animal experiments (
      • Hudson J.M.
      • Williams R.
      • Lloyd B.
      • Atri M.
      • Kim T.K.
      • Bjarnason G.A.
      • Burns P.N.
      Improved flow measurement using microbubble contrast agents and disruption–replenishment: Clinical application to tumour monitoring.
      ).
      With replenishment times for most tissues typically requiring no more than 10 s (see Fig. 12b), flow measurements can be repeated in quick succession, and multiple imaging planes can be quantified during a single-dose examination (
      • Williams R.
      • Hudson J.M.
      • Lloyd B.A.
      • Sureshkumar A.R.
      • Lueck G.
      • Milot L.
      • Atri M.
      • Bjarnason G.A.
      • Burns P.N.
      Dynamic microbubble contrast-enhanced US to measure tumor response to targeted therapy: A proposed clinical protocol with results from renal cell carcinoma patients receiving antiangiogenic therapy.
      ) with a 1-D transducer. Because the input function to this indicator dilution model is defined not by a bolus but by a brief disruption event in the tissue of interest, transit of the agent through the cardiopulmonary system no longer influences the results. In addition to reflecting blood flow and blood volume, it is important to note that when a volume of tissue is imaged, the change in enhancement is dependent on the vascular geometry as well as the flow (
      • Weisskoff R.M.
      • Chesler D.
      • Boxerman J.L.
      • Rosen B.R.
      Pitfalls in MR measurement of tissue blood flow with intravascular tracers: Which mean transit time?.
      ); more recent models introduce a parameter that relates the shape of the TIC to the fractal geometry of the microvascular network (
      • Arditi M.
      • Frinking P.J.
      • Zhou X.
      • Rognin N.G.
      A new formalism for the quantification of tissue perfusion by the destruction–replenishment method in contrast ultrasound imaging.
      ;
      • Hudson J.M.
      • Karshafian R.
      • Burns P.N.
      Quantification of flow using ultrasound and microbubbles: A disruption replenishment model based on physical principles.
      ), potentially enabling CEUS to distinguish highly disorganized tissues, such as those developed during malignant angiogenesis, from healthy tissues. A significant advantage of the method is that the infusion of a single dose allows acquisition of up to 5–10 planes in a single lesion, whose average value can account for some of the vascular heterogeneity found in large tumors. Disadvantages of the disruption–replenishment technique include a lower contrast signal-to-noise ratio compared with the level of the bolus peak, as well as the need for an infusion pump to administer those contrast agents (such as SonoVue) that are not neutrally buoyant.

      Parametric imaging

      Parametric imaging is an extension of contrast TIC quantification. Rather than quantify the averaged properties of a larger region of interest, calculations are instead performed at the scale of the imaging resolution. With motion compensation (similar to accumulation imaging) and some spatial/temporal smoothing, the local calculations (e.g., time to peak, area under the curve or replenishment rate) are mapped to a color and displayed at the spatial position from which they were calculated, often as an overlay with the gray-scale tissue reference (
      • Ellegala D.B.
      • Leong-Poi H.
      • Carpenter J.E.
      • Klibanov A.L.
      • Kaul S.
      • Shaffrey M.E.
      • Sklenar J.
      • Lindner J.R.
      Imaging tumor angiogenesis with contrast ultrasound and microbubbles targeted to αvβ3.
      ). This provides a map of the quantification results, allowing such features as spatial heterogeneity to be assessed. The use of parametric images can provide insight into functional aspects of the underlying pathology. Both pre-clinical and clinical studies have illustrated the utility of parametric imaging in localizing potentially malignant regions of tissue and predicting response to therapy (
      • Kuenen M.P.
      • Mischi M.
      • Wijkstra H.
      Contrast-ultrasound diffusion imaging for localization of prostate cancer.
      ;
      • Hudson J.M.
      • Williams R.
      • Tremblay-Darveau C.
      • Sheeran P.S.
      • Milot L.
      • Bjarnason G.A.
      • Burns P.N.
      Dynamic contrast enhanced ultrasound for therapy monitoring.
      ). The Vuebox (Bracco Suisse SA, Switzerland) quantification package provides an array of parametric imaging capabilities and an interface to automatically read and linearize DICOM exports from major manufacturers (
      • Greis C.
      Quantitative evaluation of microvascular blood flow by contrast-enhanced ultrasound (CEUS).
      ;
      • Tranquart F.
      • Mercier L.
      • Frinking P.
      • Gaud E.
      • Arditi M.
      Perfusion quantification in contrast-enhanced ultrasound (CEUS)–Ready for research projects and routine clinical use.
      ).

      Implementation considerations

      Numerous implementation considerations result from the compromises that must be made between the performance of a detection method and the design constraints and hardware costs needed to implement it. For example, the ability to “perfectly” invert or scale by half pulses imposes very stringent requirements on transmit waveform generation. Additional implementation considerations derive from clinical needs and regulatory requirements.

      Dual display (fundamental tissue/non-linear contrast)

      Before the arrival of microbubbles after a bolus injection, the ideal contrast image is blank, all tissue echoes having been suppressed. Without a tissue reference image, it is very hard for the clinician to follow lesion enhancement. The image can be provided by either a dual display, with separate contrast and conventional images, or an overlay mode with contrast shown in a color tint over a gray-scale tissue image. Most systems offer either or both of these capabilities. Dual imaging adds demands to the system design that inevitably require trade-offs. Multipulse modes inherently reduce frame rates to less than that of a normal B-mode image. Sending additional pulses for a second simultaneous image reduces the frame rate further and also increases the risk of more bubble destruction. One solution is to extract the B-mode image from the contrast imaging pulse sequence, using different receive processing. While attractive, this limits the choices of transmit parameters to those that will work adequately for both contrast and tissue, potentially compromising both. In practice, the tissue image is almost universally of substandard quality compared with a normal high-MI conventional image.

      Imaging frequency

      A fundamental—and constantly exploited—property of ultrasound images is that they scale with respect to frequency (i.e., a 10-MHz image is identical to a 1-MHz image scaled by 0.1). The addition of bubbles with a fixed dimension sacrifices this property. Their fixed dimension places the resonant frequency of the majority of bubbles from approved agents below 5 MHz. Thus, non-linear imaging methods begin to fail at center frequencies above this. Although some efforts have been made to create bubble populations with a smaller median diameter and hence higher resonant frequency by manufacture (
      • Daeichin V.
      • van Rooij T.
      • Skachkov I.
      • Ergin B.
      • Specht P.A.
      • Lima A.
      • Ince C.
      • Bosch J.G.
      • van der Steen A.F.
      • de Jong N.
      • Kooiman K.
      Microbubble composition and preparation for high-frequency contrast-enhanced ultrasound imaging: In vitro and in vivo evaluation.
      ) or by filtration of existing formulations (
      • Goertz D.E.
      • de Jong N.
      • van der Steen A.F.
      Attenuation and size distribution measurements of Definity and manipulated Definity populations.
      ), in general the Laplace pressure and other factors place a lower limit on the minimum median size of a gas-filled bubble population. The result is that non-linear response decreases with increasing frequency in most settings (
      • Sun C.
      • Sboros V.
      • Butler M.B.
      • Moran C.M.
      In vitro acoustic characterization of three phospholipid ultrasound contrast agents from 12 to 43 MHz.
      ), posing a challenge to the use of microbubble contrast in high-frequency systems dedicated to small animal imaging. Higher doses of smaller-sized (e.g., decanted) bubbles may be used, and sometimes fundamental imaging combined with image subtraction is employed.

      Transducers

      In practice, the contrast imaging performance of an ultrasound scanner is not a simple result of the signal processing to detect non-linear echoes, but is determined by a more complex combination of the performance of many individual components such as the beamformer, the waveform generator and the transducer. The bandwidth and sensitivity requirements placed on the transducer are particularly demanding: it is necessary to transmit at the bottom of the passband of the transducer to have room for at least the second harmonic to be roughly in the middle of its bandwidth (or even higher harmonics) while at the same time using very low amplitudes to avoid bubble destruction. Simply stated, contrast imaging requires large bandwidth and high sensitivity, which are inherent trade-offs in transducer design. Over the years, transducer materials and manufacturing processes have improved dramatically and, with it, contrast ultrasound performance. For example, single-crystal technology (
      • Gururaja T.
      • Panda R.
      • Chen J.
      • Beck H.
      Single crystal transducers for medical imaging applications.
      ;
      • Rehrig P.W.
      • Hackenberger W.S.
      • Jiang X.
      • Shrout T.R.
      • Zhang S.
      • Speyer R.
      2003 Status of piezoelectric single crystal growth for medical transducer applications.
      ) has led to considerable improvements in bandwidth and sensitivity and resulted in greatly improved contrast modes. The use of the broad bandwidth capacitive micromachined ultrasonic transducers for non-linear applications like the detection of microbubbles is challenged by their inherent non-linearity on transmit, but their application to contrast imaging is still an active area of research (
      • Lohfink A.
      • Eccardt P.C.
      Investigation of nonlinear CMUT behavior.
      ;
      • Balantekin M.
      • Degertekin F.L.
      Accurate modeling of capacitive micromachined ultrasonic transducers in pulse-echo operation.
      ;
      • Fouan D.
      • Bouakaz A.
      Investigation of classical pulse sequences for contrast-enhanced ultrasound imaging with a cMUT probe.
      ).

      Practical requirements for multipulse schemes

      System requirements for PI and amplitude modulation

      The majority of tissue cancellation in contrast imaging relies on the design of the transmit pulses. With pulse inversion, perfect inversion of the pulse is required. With amplitude modulation, a perfect replica of the high-amplitude pulse at exactly half the amplitude is required. In this section, we outline some of the difficulties of producing such pulse sequences on commercially available ultrasound machines.
      Cost and performance trade-offs are made at every stage of design of an ultrasound system. The beamformer in an ultrasound system is usually the highest-cost component in the system. Research laboratories publishing experiments with new pulsing schemes often use a ∼$10,000 arbitrary function generator and an ∼$10,000 radiofrequency amplifier. To put this in perspective, the cost budget for the entire transmit section in a commercial ultrasound system is likely to be <$10/channel. Here is a very brief list of some of the demands made on that transmit beamformer:
      • 1.
        Switching between pulse shapes and amplitudes in a few microseconds;
      • 2.
        Variation between pulses in a sequence less than –60 dB;
      • 3.
        Transmitting very long (1 ms), very high amplitude (>1 MPa) pulses for shear wave elastography;
      • 4.
        Transmitting very short (<1 μs), very low amplitude (<0.05 MPa) pulses for contrast imaging;
      • 5.
        Phase and amplitude matching between channels of –40 to –60 dB.
      For 40-dB tissue signal reduction in contrast imaging, all errors in the transmit pulses (amplitude and phase) must be less than 1% between pulses.

      Linear versus switched transmitters

      High-end ultrasound systems employ linear transmitters for control of transmit frequency, bandwidth and sidelobe level. This is usually accomplished using a digitally stored waveform, a digital-to-analogue converter and a power amplifier on every channel. To save cost, some systems use a switched transmitter in which the output switches rapidly between +Vt and –Vt at the transmit frequency. While this allows less control of sidelobes, both axial and lateral, they can be less susceptible to some of the cancellation errors discussed below. Switched transmitters also consume less power as the transistors are either in the on or the off state. In some systems, a unipolar pulse is transmitted, switching between 0 and Vt. Effective contrast imaging with such a unipolar pulse can be very challenging because of the DC component in the output, but is nonetheless possible.

      Matrix transducer challenges

      The recent development of matrix transducers leads to even more challenges for the contrast system designer. Today's matrix transducers can have 20,000 or more elements. Clearly, a transducer cable composed of 20,000 coaxial cables is impractical. During transmit, the application-specific integrator circuits (ASICs) themselves must create the transmit pulses based on programming downloaded from the system. To save size, cost and power consumption, matrix transducers generally use some form of switching transmitters. The receive solution is subarray beamforming in which small patches of elements are beamformed together in ASICs in the transducer itself; then signals from the patches are joined together in the system beamformer.

      Power supply challenges

      Manufacturers build clinical ultrasound systems to be robust across a wide range of imaging modes and power supply configurations, while also remaining cost effective. As such, challenges can arise in contrast imaging related to power supply noise. CEUS imaging often operates in an extreme of both the probe frequency response (lowest frequencies feasible to balance bubble response, pulse fidelity and receive response) and commanded voltages from the hardware, and so patterned noise from the power supply can appear in the final images. As these often reside in the lower edge of the receive bandwidth, they can be mitigated by careful choice of transmit frequency and receive filtering or by using techniques to eliminate the patterned noise from the received echoes. More complex probes that have embedded ASICs (e.g., matrix probes) can also generate substantial patterned noise because of the proximity of the electronics to the transmit/receive pathways. These noise artifacts are often addressed through similar approaches.

      Pulse inversion

      Most common linear amplifiers use a class B or class AB amplifier for linearity and efficiency. Switching transmitters use a similar structure, but instead of being driven by a sinusoidal pulse, they are driven into either the on or off state. Of most interest is the configuration of NPN and PNP transistors. A similar circuit would use N-channel and P-channel field effect transistors. The voltage drop across a diode junction in a transistor is often cited as 0.7 V. In fact, the drop across a PN junction is about 0.7 V, but across an NP junction, it is about 0.6 V, which can lead to slight asymmetry between inverted pulses. In most applications, this is trivial and can be ignored. For performing PI in tissue harmonic imaging, for example, a 0.1-V difference with 150-V transmit voltages is trivial. However, for contrast PI, transmit voltages are about 5 V, placing a –30-dB limit on the amount of tissue cancellation that can be achieved if this error is not taken into account in the transmitter design.

      Amplitude modulation and aperture designs

      Amplitude modulation requires transmitting two identical pulses with an exact amplitude ratio, usually a factor of 2. This requires switching the transmit amplitude virtually instantaneously (1–10 μs), like PI at very low voltages. This, in itself, is challenging because of the stability required of transmitters for color flow and other multipulse sequences. Whatever technique is used to provide a stable transmit voltage acts in opposition to rapid changes. In addition, all amplifiers tend to have a slight distortion at the zero crossing of the waveform when switching from the NPN to the PNP transistor. That distortion is relatively independent of input amplitude, so the magnitude of the zero crossing switching transient is roughly the same between the full- and half-amplitude pulses, which makes it twice as significant for the half-amplitude pulse, leading to reduced tissue cancellation.
      To avoid the aforementioned issue, a new technique, often called “checkerboard apertures,” was developed (

      Brock-Fisher GA, Poland MD, Rafter PG. 1996 Means for increasing sensitivity in non-linear ultrasound imaging systems. U.S. Patent US5577505A.

      ;
      • Phillips P.
      • Gardner E.
      Contrast-agent detection and quantification.
      ;
      • Whittingham T.A.
      Contrast-specific imaging techniques: Technical perspective.
      ;
      • Tremblay-Darveau C.
      • Sheeran P.S.
      • Vu C.K.
      • Williams R.
      • Zhang Z.
      • Bruce M.
      • Burns P.N.
      The role of microbubble echo phase lag in multipulse contrast-enhanced ultrasound imaging.
      ). With this technique the transmit voltage is never changed, but the aperture is. It consists of transmitting first on even-numbered elements, then on all elements and finally on odd-numbered elements (Fig. 13). The scattered echoes from these transmit events are then combined by adding the echoes of the even and odd apertures and subtracting them from those of the full aperture. The beam pattern of the full aperture is somewhat different from those of the even and odd apertures. However, by adding the complementary even and odd apertures and subtracting from the full, all linear signals cancel. This is now a common technique used in many commercial ultrasound systems. Historically, this aperture technique for AM was developed by

      Brock-Fisher GA, Poland MD, Rafter PG. 1996 Means for increasing sensitivity in non-linear ultrasound imaging systems. U.S. Patent US5577505A.

      because their system was not capable of inverting pulses to implement PI. The use of spatially distinct complementary apertures has offered a very straightforward implementation of AM, which, in addition to cancelling linear signals (tissue), also results in additional non-linear signals because of the spatial differences of the beams. Other mechanisms of AM non-linear signal enhancement attributed to small variations in the pressure-dependent phase lag in the microbubble response have been suggested (
      • Emmer M.
      • Vos H.J.
      • Goertz D.E.
      • van Wamel A.
      • Versluis M.
      • de Jong N.
      Pressure-dependent attenuation and scattering of phospholipid-coated microbubbles at low acoustic pressures.
      ;
      • Tremblay-Darveau C.
      • Sheeran P.S.
      • Vu C.K.
      • Williams R.
      • Zhang Z.
      • Bruce M.
      • Burns P.N.
      The role of microbubble echo phase lag in multipulse contrast-enhanced ultrasound imaging.
      ).
      Fig 13
      Fig. 13Even, full and odd apertures used in amplitude modulation (AM). Despite the beam patterns of the different apertures being spatially different, their combination in AM results in zero linear signal.
      These checkerboard apertures have their own limitations, though. There can be electrical cross-talk between the wires going to the on and off elements or acoustic cross-talk between the on and off elements in the transducer itself. Any cross-talk between on and off elements will change the even and odd beam patterns so that when summed, they do not exactly cancel the full beam pattern, thus reducing tissue cancellation.

      Plane wave transmits/ultrafast

      Researchers have investigated the use of plane wave high-frame-rate acquisitions to improve both sensitivity and visualization of microbubble flow (
      • Couture O.
      • Fink M.
      • Tanter M.
      Ultrasound contrast plane wave imaging.
      ;
      • Tanter M.
      • Fink M.
      Ultrafast imaging in biomedical ultrasound.
      ;
      • Tremblay-Darveau C.
      • Williams R.
      • Milot L.
      • Bruce M.
      • Burns P.
      Visualizing the tumour microvasculature with a nonlinear plane-wave Doppler imaging scheme based on amplitude modulation.
      ,
      • Tremblay-Darveau C.
      • Williams R.
      • Sheeran P.S.
      • Milot L.
      • Bruce M.
      • Burns P.N.
      Concepts and tradeoffs in velocity estimation with plane-wave contrast-enhanced Doppler.
      ). This work has also led to remarkable images of vasculature in the azimuthal plane through super-resolution optical approaches, similar to fluorescence photo-activated localization microscopy (
      • Betzig E.
      • Patterson G.H.
      • Sougrat R.
      • Lindwasser O.W.
      • Olenych S.
      • Bonifacino J.S.
      • Davidson M.W.
      • Lippincott-Schwartz J.
      • Hess H.F.
      Imaging intracellular fluorescent proteins at nanometer resolution.
      ), to circumvent diffraction limits in spatial resolution (
      • Errico C.
      • Pierre J.
      • Pezet S.
      • Desailly Y.
      • Lenkei Z.
      • Couture O.
      • Tanter M.
      Ultrafast ultrasound localization microscopy for deep super-resolution vascular imaging.
      ). Although multiple ultrasound research platforms support plane-wave acquisitions, a limited number of commercial ultrasound system architectures support plane-wave acquisitions, and to date, major ultrasound manufacturers have yet to release a plane-wave microbubble imaging mode. An additional challenge faced by plane wave-based approaches using 1-D arrays is the fixed elevational focus, which limits the penetration of unfocused beams (
      • Lai T.Y.
      • Bruce M.
      • Averkiou M.A.
      Modeling of the acoustic field produced by diagnostic ultrasound arrays in plane and diverging wave modes.
      ).

      Regulatory aspects

      The medical devices are heavily regulated, and diagnostic ultrasound is no exception. First and foremost is the acoustic output. There are three primary measures of acoustic output mandated by FDA and European Medicines Evaluation Agency (EMEA); mechanical index (MI), spatial peak temporal average (SPTA) and surface temperature (ST). The first is an instantaneous peak pressure measurement, and the second two are average measurements.

      Mechanical index

      The MI (eqn [1]) is a measure of the highest instantaneous peak negative pressure that exists anywhere in an ultrasound image plane or volume. It was defined as a regulatory measure to prevent cavitation in non-bubble-containing tissue. The onscreen MI is valid only at the location in the image with the highest derated acoustic pressure. Away from that region the acoustic pressure is lower, but is not generally documented. During optimization, most companies find an optimum default power level to minimize microbubble destruction, but maximize the returned signal. Suboptimal optimization may cause near-field bubble destruction or even a band near the focal depth where the acoustic pressure may be the highest.
      The regulatory bodies do not specify exactly how the MI should be measured, only that it is measured and it never exceeds the limit of 1.9. This leads to differences in stated MI and levels of bubble destruction between different equipment manufacturers. For example there can be slight differences in the output of different transducers of the same model. To ensure that no transducer ever puts out more than the allowed acoustic power, the company will likely use an assortment of manufactured transducers to measure the spread of output. Then the company can decide to label the onscreen MI as either what the transducer with the highest output delivered or what the average transducer delivered. In the first case, the MI may go up very close to the limit of 1.9, because the hottest transducer was used for the testing. In the second case, the onscreen MI will show what the average transducer delivered, but would then have to be lower to ensure that the hottest transducer never exceeded the limit. Therefore, if a random transducer from different companies is measured in a lab, the company that uses the former method will appear to be at a higher MI than the latter, even if the measured output is the same. During contrast imaging, MI is rarely a limiting factor from a regulatory point of view except in high-MI triggered modes. It is, however, often the limit for flash.

      Spatial peak temporal average

      Spatial peak temporal average (SPTA) is used as an estimate of the likelihood of heating tissue internally; the mandated limit is 750 mW/cm2. Like MI, it is defined only for the location in the image with the highest average in situ power, again derated for the assumed tissue attenuation. SPTA is rarely a limiting factor for contrast imaging, except during flash.

      Surface temperature

      Surface temperature (ST) is also a reflection of average power, but is a measure of how much the surface of the transducer heats up during imaging to prevent burning the skin. FDA limits this value to 41°C. Again, this is rarely an issue for contrast imaging.

      Power considerations for flash

      flash for contrast still needs to meet the above limits, but how this is calculated is a little more complex. MI is an instantaneous measure, and flash must of course meet the limit of 1.9. However, because flash is typically used intermittently, during its use both SPTA and ST are usually much lower than they would be if the system were transmitting continuously at the flash amplitude. For example, if we consider the case of 1 flash frame every second and a real-time imaging frame rate of 30 Hz, the SPTA and ST, which are calculated assuming real-time scanning, are overestimated by a factor of 30.
      Although SPTA and ST are measures of average power, they do not necessarily track each other. In a scanned mode, the power deposited at any given location within tissue (SPTA) is proportional to how often the beam hits that location. In a phased array, all or nearly all transducer elements are used at all times regardless of where the beam is pointing. In a linear or curvilinear transducer, the active aperture is scanned across the face of the transducer, spreading out the power across the face of the transducer, so transducer heating tends to be less of an issue, unless the sector angle is reduced. All of this needs to be considered by the acoustic output control software and affects how flash is handled.

      The future

      Although refinement of these methods over the past decade has been steady but incremental, the fundamental change in acquisition of image data brought about by synthetic aperture and related techniques has had a large impact on contrast-specific imaging, allowing much longer ensemble lengths in multipulse schemes and commensurate improvement in imaging performance (
      • Tremblay-Darveau C.
      • Williams R.
      • Milot L.
      • Bruce M.
      • Burns P.N.
      Combined perfusion and Doppler imaging using plane-wave nonlinear detection and microbubble contrast agents.
      ). Similarly, multifrequency approaches have opened the way to contrast-specific imaging at higher frequencies that are many multiples of bubble resonance (
      • Cherin E.
      • Brown J.
      • Masoy S.E.
      • Shariff H.
      • Karshafian R.
      • Williams R.
      • Burns P.N.
      • Foster F.S.
      Radial modulation imaging of microbubble contrast agents at high frequency.
      ). Higher spatial resolution with sparse bubbles has also been achieved using super-resolution methods (
      • Christensen-Jeffries K.
      • Browning R.J.
      • Tang M.X.
      • Dunsby C.
      • Eckersley R.J.
      In vivo acoustic super-resolution and super-resolved velocity mapping using microbubbles.
      ;
      • Opacic T.
      • Dencks S.
      • Theek B.
      • Piepenbrock M.
      • Ackermann D.
      • Rix A.
      • Lammers T.
      • Stickeler E.
      • Delorme S.
      • Schmitz G.
      Motion model ultrasound localization microscopy for preclinical and clinical multiparametric tumor characterization.
      ).
      Three-dimensional contrast acquisitions remain an active area of research and industry effort. Several studies have reported the potential benefit of 3-D characterization in visualizing the vascular activity within an entire volume of interest and reducing heterogeneity present in single scan planes. Especially for applications involving quantification, such as for assessing response to therapy, the use of 3-D ultrasound may provide significant benefits over conventional 2-D approaches (
      • Wang H.
      • Kaneko O.F.
      • Tian L.
      • Hristov D.
      • Willmann J.K.
      Three-dimensional ultrasound molecular imaging of angiogenesis in colon cancer using a clinical matrix array ultrasound transducer.
      ;
      • El Kaffas A.
      • Sigrist R.M.S.
      • Fisher G.
      • Bachawal S.
      • Liau J.
      • Wang H.
      • Karanany A.
      • Durot I.
      • Rosenberg J.
      • Hristov D.
      Quantitative three-dimensional dynamic contrast-enhanced ultrasound imaging: First-in-human pilot study in patients with liver metastases.
      ). Several approaches to volumetric CEUS have been evaluated for anatomic evaluation and quantification, including single-volume freehand “panoramic” sweeps using 1-D arrays and multivolume acquisitions using mechanically swept 1-D arrays and matrix array probes. Challenges remain in 3-D ultrasound to provide resolution and frame rate comparable to those of 2-D CEUS scanning, which are often critical to diagnosis. Plane wave techniques may play a role in maximizing volumetric frame rate with a limited number of transmit events. While for 2-D imaging, plane wave imaging provides higher frame rates, it also allows longer ensemble lengths in multipulse contrast modes. Thus, an image composed of three pulses per line can be replaced by one with 100 or more pulses per image with no loss of frame rate. These long ensemble lengths allow for Doppler velocity estimation and, when combined with phase and/or amplitude modulation, create separate spectra at every point in the image corresponding to the linear and non-linear components (
      • Hope Simpson D.
      • Burns P.N.
      Pulse Inversion Doppler: A new method for detecting nonlinear echoes from microbubble contrast agents.
      ). From the resulting data, perfusion images can be overlaid with velocity-resolved non-linear color Doppler, formed at frame rates of 50–100 Hz, with spectra available at each point (
      • Tremblay-Darveau C.
      • Williams R.
      • Milot L.
      • Bruce M.
      • Burns P.N.
      Combined perfusion and Doppler imaging using plane-wave nonlinear detection and microbubble contrast agents.
      ,
      • Tremblay-Darveau C.
      • Williams R.
      • Milot L.
      • Bruce M.
      • Burns P.
      Visualizing the tumour microvasculature with a nonlinear plane-wave Doppler imaging scheme based on amplitude modulation.
      ,
      • Tremblay-Darveau C.
      • Williams R.
      • Sheeran P.S.
      • Milot L.
      • Bruce M.
      • Burns P.N.
      Concepts and tradeoffs in velocity estimation with plane-wave contrast-enhanced Doppler.
      ;
      • Khaing Z.Z.
      • Cates L.N.
      • DeWees D.M.
      • Hannah A.
      • Mourad P.
      • Bruce M.
      • Hofstetter C.P.
      Contrast-enhanced ultrasound to visualize hemodynamic changes after rodent spinal cord injury.
      ). Such techniques may offer the advantage of a “one-stop” CEUS examination, obviating the need for pre-contrast imaging and Doppler examination.
      Finally, it should be noted that 25 years later, we are still using contrast agents that were developed with no regard to optimizing imaging methods; to a great extent it is a coincidence of fortune that the bubble resonance for a transpulmonary agent lies within the diagnostic frequency range; that lipid buckling and low-amplitude thresholds amplify non-linear behavior; and that the disruption threshold of commonly used agents is sufficiently high to allow continuous imaging over a range of depths. Yet because of their polydisperse size distribution, it is likely that only a small percentage of bubbles contribute significantly to the received echo. Monodisperse bubble populations offer the potential to improve the efficiency and specificity of bubble-specific imaging (
      • Segers T.
      • Kruizinga P.
      • Kok M.P.
      • Lajoinie G.
      • de Jong N.
      • Versluis M.
      Monodisperse versus polydisperse ultrasound contrast agents: Non-linear response, sensitivity, and deep tissue imaging potential.
      ), as well as non-invasive pressure measurement (
      • Tremblay-Darveau C.
      • Williams R.
      • Milot L.
      • Bruce M.
      • Burns P.N.
      Combined perfusion and Doppler imaging using plane-wave nonlinear detection and microbubble contrast agents.
      ). Efforts using bubble sizing (
      • Segers T.
      • de Jong N.
      • Versluis M.
      Uniform scattering and attenuation of acoustically sorted ultrasound contrast agents: Modeling and experiments.
      ) or microfluidic synthesis (
      • Segers T.
      • Lohse D.
      • Versluis M.
      • Frinking P.
      Universal equations for the coalescence probability and long-term size stability of phospholipid-coated monodisperse microbubbles formed by flow focusing.
      ) appear promising: it seems likely that in the future, functionalized bubbles designed in conjunction with specific imaging methods will propel the field in entirely new directions.

      References

        • Albrecht T.
        • Blomley M.J.
        • Burns P.N.
        • Wilson S.
        • Harvey C.J.
        • Leen E.
        • Claudon M.
        • Calliada F.
        • Correas J.M.
        • LaFortune M.
        • Campani R.
        • Hoffmann C.W.
        • Cosgrove D.O.
        • LeFevre F.
        Improved detection of hepatic metastases with pulse-inversion US during the liver-specific phase of SHU 508A: Multicenter study.
        Radiology. 2003; 227: 361-370
        • Apfel R.E.
        • Holland C.K.
        Gauging the likelihood of cavitation from short-pulse, low-duty cycle diagnostic ultrasound.
        Ultrasound Med Biol. 1991; 17: 179-185
        • Arditi M.
        • Frinking P.J.
        • Zhou X.
        • Rognin N.G.
        A new formalism for the quantification of tissue perfusion by the destruction–replenishment method in contrast ultrasound imaging.
        IEEE Trans Ultrason Ferroelectr Freq Control. 2006; 53: 1118-1129
        • Averkiou M.A.
        Tissue harmonic ultrasonic imaging.
        C R Acad Sci Ser IV Phys. 2001; 2: 1139-1151
        • Averkiou M.A.
        • Roundhill D.N.
        • Powers J.E.
        A new imaging technique based on the nonlinear properties of tissues.
        Proc IEEE Int Ultrason Symp. 1997; 1/2: 1561-1566
      1. Averkiou M, Powers JE, Bruce M, Skyba DM. Realtime ultrasonic imaging of perfusion using ultrasonic contrast agents. 2001. U.S. Patent 6,171, 246.

        • Averkiou M.A.
        • Mannaris C.
        • Bruce M.
        • Powers J.
        Nonlinear pulsing schemes for the detection of ultrasound contrast agents.
        in: Proceedings, 155th Meeting of the Acoustical Society of America (Acoustic ’08), Paris, France2008: 915-920 (June 29–July 4)
        • Averkiou M.
        • Lampaskis M.
        • Kyriakopoulou K.
        • Skarlos D.
        • Klouvas G.
        • Strouthos C.
        • Leen E.
        Quantification of tumor microvascularity with respiratory gated contrast enhanced ultrasound for monitoring therapy.
        Ultrasound Med Biol. 2010; 36: 68-77
        • Balantekin M.
        • Degertekin F.L.
        Accurate modeling of capacitive micromachined ultrasonic transducers in pulse-echo operation.
        Proc IEEE Int Ultrason Symp. 2008; : 2107-2110
        • Becher H.
        • Burns P.N.
        Handbook of contrast echocardiography: Left ventricular function and myocardial perfusion.
        Springer Science & Business Media, 2012
        • Becher H.
        • Tiemann K.
        • Kuntz-Hehner S.
        • Omran H.
        • Schlosser T.
        Diagnostic impact of contrast echocardiography for assessment of left ventricular function and myocardial perfusion in patients with coronary artery disease.
        Eur Heart J Suppl. 2002; 4: C12-C21
        • Betzig E.
        • Patterson G.H.
        • Sougrat R.
        • Lindwasser O.W.
        • Olenych S.
        • Bonifacino J.S.
        • Davidson M.W.
        • Lippincott-Schwartz J.
        • Hess H.F.
        Imaging intracellular fluorescent proteins at nanometer resolution.
        Science. 2006; 313: 1642-1645
        • Blomley M.J.
        • Albrecht T.
        • Cosgrove D.O.
        • Eckersley R.J.
        • Butler-Barnes J.
        • Jayaram V.
        • Patel N.
        • Heckemann R.A.
        • Bauer A.
        • Schlief R.
        Stimulated acoustic emission to image a late liver and spleen-specific phase of Levovist in normal volunteers and patients with and without liver disease.
        Ultrasound Med Biol. 1999; 25: 1341-1352
        • Blomley M.J.
        • Albrecht T.
        • Cosgrove D.O.
        • Patel N.
        • Jayaram V.
        • Butler-Barnes J.
        • Eckersley R.J.
        • Bauer A.
        • Schlief R.
        Improved imaging of liver metastases with stimulated acoustic emission in the late phase of enhancement with the US contrast agent SH U 508A: Early experience.
        Radiology. 1999; 210: 409-416
        • Bouakaz A.
        • de Jong N.
        Native tissue imaging at superharmonic frequencies.
        IEEE Trans Ultrason Ferroelectr Freq Control. 2003; 50: 496-506
      2. Brock-Fisher GA, Poland MD, Rafter PG. 1996 Means for increasing sensitivity in non-linear ultrasound imaging systems. U.S. Patent US5577505A.

        • Burns P.N.
        Harmonic imaging with ultrasound contrast agents.
        Clin Radiol. 1996; 51: 50-55
        • Burns P.N.
        • Powers J.E.
        • Fritzsch T.
        Harmonic imaging: A new imaging and Doppler method for contrast enhanced ultrasound.
        Radiology. 1992; 185 (Abstract): 142
        • Burns P.N.
        • Powers J.E.
        • Hope Simpson D.
        • Brezina A.
        • Kolin A.
        • Chin C.T.
        • Uhlendorf V.
        • Fritzsch T.
        Harmonic power mode Doppler using microbubble contrast agents: An improved method for small vessel flow imaging.
        IEEE Trans Ultrason Ferroelectr Freq Control. 1994; : 1547-1550
        • Burns P.N.
        • Wilson S.R.
        • Muradali D.
        • Powers J.E.
        • Greener Y.
        Microbubble destruction is the origin of harmonic signals from FS069.
        Radiology. 1996; 201: 158
        • Burns P.N.
        • Wilson S.R.
        • Simpson D.H.
        Pulse inversion imaging of liver blood flow: Improved method for characterizing focal masses with microbubble contrast.
        Invest Radiol. 2000; 35: 58-71
        • Cherin E.
        • Brown J.
        • Masoy S.E.
        • Shariff H.
        • Karshafian R.
        • Williams R.
        • Burns P.N.
        • Foster F.S.
        Radial modulation imaging of microbubble contrast agents at high frequency.
        Ultrasound Med Biol. 2008; 34: 949-962
        • Christensen-Jeffries K.
        • Browning R.J.
        • Tang M.X.
        • Dunsby C.
        • Eckersley R.J.
        In vivo acoustic super-resolution and super-resolved velocity mapping using microbubbles.
        IEEE Trans Med Imaging. 2014; 34: 433-440
        • Christofides D.
        • Leen E.
        • Averkiou M.A.
        Automatic respiratory gating for contrast ultrasound evaluation of liver lesions.
        IEEE Trans Ultrason Ferroelectr Freq Control. 2014; 61: 25-32
        • Christofides D.
        • Leen E.L.
        • Averkiou M.A.
        Improvement of the accuracy of liver lesion DCEUS quantification with the use of automatic respiratory gating.
        Eur Radiol. 2016; 26: 417-424
        • Cosgrove D.
        Ultrasound contrast enhancement of tumours.
        Clin Radiol. 1996; 51: 44-49
        • Couture O.
        • Fink M.
        • Tanter M.
        Ultrasound contrast plane wave imaging.
        IEEE Trans Ultrason Ferroelectr Freq Control. 2012; 59: 2676-2683
        • Daeichin V.
        • van Rooij T.
        • Skachkov I.
        • Ergin B.
        • Specht P.A.
        • Lima A.
        • Ince C.
        • Bosch J.G.
        • van der Steen A.F.
        • de Jong N.
        • Kooiman K.
        Microbubble composition and preparation for high-frequency contrast-enhanced ultrasound imaging: In vitro and in vivo evaluation.
        IEEE Trans Ultrason Ferroelectr Freq Control. 2017; 64: 555-567
        • de Jong N.
        • Hoff L.
        Ultrasound scattering properties of Albunex microspheres.
        Ultrasonics. 1993; 31: 175-181
        • Deslandes M.
        • Guillin R.
        • Cardinal E.
        • Hobden R.
        • Bureau N.J.
        The snapping iliopsoas tendon: New mechanisms using dynamic sonography.
        AJR Am J Roentgenol. 2008; 190: 556-581
        • Dietrich C.F.
        • Mertens J.C.
        • Braden B.
        • Schuessler G.
        • Ott M.
        • Ignee A.
        Contrast-enhanced ultrasound of histologically proven liver hemangiomas.
        Hepatology. 2007; 45: 1139-1145
        • Dietrich C.
        • Averkiou M.
        • Correas J.M.
        • Lassau N.
        • Leen E.
        • Piscaglia F.
        An EFSUMB introduction into dynamic contrast-enhanced ultrasound (DCE-US) for quantification of tumour perfusion.
        Ultraschall Med. 2012; 33: 344-351
        • Dietrich C.F.
        • Ignee A.
        • Greis C.
        • Cui X.W.
        • Schreiber-Dietrich D.G.
        • Hocke M.
        Artifacts and pitfalls in contrast-enhanced ultrasound of the liver.
        Ultraschall Med. 2014; 35 (quiz 26–27): 108-125
        • Eckersley R.J.
        • Chin C.T.
        • Burns P.N.
        Optimising phase and amplitude modulation schemes for imaging microbubble contrast agents at low acoustic power.
        Ultrasound Med Biol. 2005; 31: 213-219
        • Eisenbrey J.
        • Dave J.
        • Halldorsdottir V.
        • Merton D.
        • Machado P.
        • Liu J.
        • Miller C.
        • Gonzalez J.
        • Park S.
        • Dianis S.
        Simultaneous grayscale and subharmonic ultrasound imaging on a modified commercial scanner.
        Ultrasonics. 2011; 51: 890-897
        • El Kaffas A.
        • Sigrist R.M.S.
        • Fisher G.
        • Bachawal S.
        • Liau J.
        • Wang H.
        • Karanany A.
        • Durot I.
        • Rosenberg J.
        • Hristov D.
        Quantitative three-dimensional dynamic contrast-enhanced ultrasound imaging: First-in-human pilot study in patients with liver metastases.
        Theranostics. 2017; 7: 3745
        • Ellegala D.B.
        • Leong-Poi H.
        • Carpenter J.E.
        • Klibanov A.L.
        • Kaul S.
        • Shaffrey M.E.
        • Sklenar J.
        • Lindner J.R.
        Imaging tumor angiogenesis with contrast ultrasound and microbubbles targeted to αvβ3.
        Circulation. 2003; 108: 336-341
        • Emmer M.
        • Vos H.J.
        • Goertz D.E.
        • van Wamel A.
        • Versluis M.
        • de Jong N.
        Pressure-dependent attenuation and scattering of phospholipid-coated microbubbles at low acoustic pressures.
        Ultrasound Med Biol. 2009; 35: 102-111
        • Errico C.
        • Pierre J.
        • Pezet S.
        • Desailly Y.
        • Lenkei Z.
        • Couture O.
        • Tanter M.
        Ultrafast ultrasound localization microscopy for deep super-resolution vascular imaging.
        Nature. 2015; 527: 499-502
        • Fetzer D.T.
        • Rafailidis V.
        • Peterson C.
        • Grant E.G.
        • Sidhu P.
        • Barr R.G.
        Artifacts in contrast-enhanced ultrasound: A pictorial essay.
        Abdom Radiol. 2018; 43: 977-997
        • Fobbe F.
        • Siegert J.
        • Fritzsch T.
        • Koch H.C.
        • Wolf K.J.
        [Color-coded duplex sonography and ultrasound contrast media—Detection of renal perfusion defects in experimental animals].
        Rofo Fortschr Geb Rontgenstr Neuen Bildgeb Verfahr. 1991; 154: 242-245
        • Forsberg F.
        • Liu J.B.
        • Burns P.N.
        • Merton D.A.
        • Goldberg B.B.
        Artifacts in ultrasonic contrast agent studies.
        J Ultrasound Med. 1994; 13: 357-365
        • Forsberg F.
        • Shi W.T.
        • Goldberg B.
        Subharmonic imaging of contrast agents.
        Ultrasonics. 2000; 38: 93-98
        • Forsberg F.
        • Ro R.J.
        • Fox T.B.
        • Liu J.-B.
        • Chiou S.-Y.
        • Potoczek M.
        • Goldberg B.B.
        Contrast enhanced maximum intensity projection ultrasound imaging for assessing angiogenesis in murine glioma and breast tumor models: A comparative study.
        Ultrasonics. 2011; 51: 382-389
        • Fouan D.
        • Bouakaz A.
        Investigation of classical pulse sequences for contrast-enhanced ultrasound imaging with a cMUT probe.
        IEEE Trans Ultrason Ferroelectr Freq Control. 2016; 63: 1496-1504
        • Frinking P.J.
        • Céspedes E.
        • De Jong N.
        Multi-pulse ultrasound contrast imaging based on a decorrelation detection strategy.
        Proc IEEE Int Ultrason Symp. 1998; : 1787-1790
        • Frinking P.J.
        • Bouakaz A.
        • Kirkhorn J.
        • Ten Cate F.J.
        • de Jong N.
        Ultrasound contrast imaging: Current and new potential methods.
        Ultrasound Med Biol. 2000; 26: 965-975
        • Gerrit L.
        • Renaud G.G.
        • Akkus Z.
        • van den Oord S.C.
        • Folkert J.
        • Shamdasani V.
        • Entrekin R.R.
        • Sijbrands E.J.
        • de Jong N.
        • Bosch J.G.
        Far-wall pseudoenhancement during contrast-enhanced ultrasound of the carotid arteries: Clinical description and in vitro reproduction.
        Ultrasound Med Biol. 2012; 38: 593-600
        • Goertz D.E.
        • de Jong N.
        • van der Steen A.F.
        Attenuation and size distribution measurements of Definity and manipulated Definity populations.
        Ultrasound Med Biol. 2007; 33: 1376-1388
        • Gramiak R.
        • Shah P.M.
        Echocardiography of the aortic root.
        Invest Radiol. 1968; 3: 356-366
        • Greis C.
        Quantitative evaluation of microvascular blood flow by contrast-enhanced ultrasound (CEUS).
        Clin Hemorheol Microcirc. 2011; 49: 137-149
        • Gururaja T.
        • Panda R.
        • Chen J.
        • Beck H.
        Single crystal transducers for medical imaging applications.
        In: Proc IEEE Int Ultrason Symp. 1999; : 969-972
        • Hockham N.
        • Coussios C.C.
        • Arora M.
        A real-time controller for sustaining thermally relevant acoustic cavitation during ultrasound therapy.
        IEEE Trans Ultrason Ferroelectr Freq Control. 2010; 57: 2685-2694
        • Hope Simpson D.
        • Burns P.N.
        Pulse Inversion Doppler: A new method for detecting nonlinear echoes from microbubble contrast agents.
        Proc IEEE Int Ultrason Symp. 1997; : 1597-1600
        • Hope Simpson D.
        • Chin C.T.
        • Burns P.N.
        Pulse inversion Doppler: A new method for detecting nonlinear echoes from microbubble contrast agents.
        IEEE Trans Ultrason Ferroelectr Freq Control. 1999; 46: 372-382
        • Hudson J.M.
        • Karshafian R.
        • Burns P.N.
        Quantification of flow using ultrasound and microbubbles: A disruption replenishment model based on physical principles.
        Ultrasound Med Biol. 2009; 35: 2007-2020
        • Hudson J.M.
        • Leung K.
        • Burns P.N.
        The lognormal perfusion model for disruption replenishment measurements of blood flow: In vivo validation.
        Ultrasound Med Biol. 2011; 37: 1571-1578
        • Hudson J.M.
        • Williams R.
        • Lloyd B.
        • Atri M.
        • Kim T.K.
        • Bjarnason G.A.
        • Burns P.N.
        Improved flow measurement using microbubble contrast agents and disruption–replenishment: Clinical application to tumour monitoring.
        Ultrasound Med Biol. 2011; 37: 1210-1221
        • Hudson J.M.
        • Williams R.
        • Tremblay-Darveau C.
        • Sheeran P.S.
        • Milot L.
        • Bjarnason G.A.
        • Burns P.N.
        Dynamic contrast enhanced ultrasound for therapy monitoring.
        Eur J Radiol. 2015; 84: 1650-1657
        • Ichino N.
        • Horiguchi Y.
        • Imai H.
        • Osakabe K.
        • Nishikawa T.
        • Sugita Y.
        • Utsugi H.
        • Togo Y.
        • Sawai T.
        • Mizoguchi Y.
        Contrast-enhanced sonography of pancreatic ductal carcinoma using agent detection imaging.
        J Med Ultrason. 2006; 33: 29-35
        • Karshafian R.
        • Burns P.N.
        • Henkelman M.R.
        Transit time kinetics in ordered and disordered vascular trees.
        Phys Med Biol. 2003; 48: 3225-3237
        • Keravnou C.P.
        • Mannaris C.
        • Averkiou M.A.
        Accurate measurement of microbubble response to ultrasound with a diagnostic ultrasound scanner.
        IEEE Trans Ultrason Ferroelectr Freq Control. 2015; 62: 176-184
        • Khaing Z.Z.
        • Cates L.N.
        • DeWees D.M.
        • Hannah A.
        • Mourad P.
        • Bruce M.
        • Hofstetter C.P.
        Contrast-enhanced ultrasound to visualize hemodynamic changes after rodent spinal cord injury.
        J Neurosurg Spine. 2018; 29: 306-313
        • Kirkhorn J.
        • Frinking P.
        • de Jong N.
        • Torp H.
        Improving the sensitivity of power Doppler for ultrasound contrast imaging by using a high power release burst.
        in: Proceedings, 4th Heart Centre European Symposium on Ultrasound Contrast Imaging, Rotterdam, The Netherlands1999: 56-60
        • Kono Y.
        • Moriyasu F.
        • Mine Y.
        • Nada T.
        • Kamiyama N.
        • Suginoshita Y.
        • Matsumura T.
        • Kobayashi K.
        • Chiba T.
        Gray-scale second harmonic imaging of the liver with galactose-based microbubbles.
        Invest Radiol. 1997; 32: 120-125
        • Krix M.
        • Kiessling F.
        • Vosseler S.
        • Farhan N.
        • Mueller M.M.
        • Bohlen P.
        • Fusenig N.E.
        • Delorme S.
        Sensitive noninvasive monitoring of tumor perfusion during antiangiogenic therapy by intermittent bolus-contrast power Doppler sonography.
        Cancer Res. 2003; 63: 8264-8270
        • Kuenen M.P.
        • Mischi M.
        • Wijkstra H.
        Contrast-ultrasound diffusion imaging for localization of prostate cancer.
        IEEE Trans Med Imaging. 2011; 30: 1493-1502
        • Lai T.Y.
        • Bruce M.
        • Averkiou M.A.
        Modeling of the acoustic field produced by diagnostic ultrasound arrays in plane and diverging wave modes.
        IEEE Trans Ultrason Ferroelectr Freq Control. 2019; 66: 1158-1169
        • Lampaskis M.
        • Averkiou M.
        Investigation of the relationship of nonlinear backscattered ultrasound intensity with microbubble concentration at low MI.
        Ultrasound Med Biol. 2010; 36: 306-312
        • Lamuraglia M.
        • Bridal S.L.
        • Santin M.
        • Izzi G.
        • Rixe O.
        • Paradiso A.
        • Lucidarme O.
        Clinical relevance of contrast-enhanced ultrasound in monitoring anti-angiogenic therapy of cancer: Current status and perspectives.
        Crit Rev Oncol Hematol. 2010; 73: 202-212
        • Leen E.
        • Averkiou M.
        • Arditi M.
        • Burns P.
        • Bokor D.
        • Gauthier T.
        • Kono Y.
        • Lucidarme O.
        Dynamic contrast enhanced ultrasound assessment of the vascular effects of novel therapeutics in early stage trials.
        Eur Radiol. 2012; 22: 1442-1450
        • Leighton T.G.
        The acoustic bubble.
        Academic Press, London1994
        • Levine R.A.
        • Teichholz L.E.
        • Goldman M.E.
        • Steinmetz M.Y.
        • Baker M.
        • Meltzer R.S.
        Microbubbles have intracardiac velocities similar to those of red blood cells.
        J Am Coll Cardiol. 1984; 3: 28-33
        • Lohfink A.
        • Eccardt P.C.
        Investigation of nonlinear CMUT behavior.
        Proc IEEE Int Ultrason Symp. 2005; : 585-588
        • Main M.L.
        • Grayburn P.A.
        Clinical applications of transpulmonary contrast echocardiography.
        Am Heart J. 1999; 137: 144-153
        • Miller D.L.
        • Neppiras E.A.
        On the oscillation mode of gas-filled micropores.
        J Acoust Soc Am. 1985; 77: 946-953
        • Mor-Avi V.
        • Caiani E.G.
        • Collins K.A.
        • Korcarz C.E.
        • Bednarz J.E.
        • Lang R.M.
        Combined assessment of myocardial perfusion and regional left ventricular function by analysis of contrast-enhanced power modulation images.
        Circulation. 2001; 104: 352-357
        • Mulvagh S.L.
        • Foley D.A.
        • Aeschbacher B.C.
        • Klarich K.K.
        • Seward J.B.
        Second harmonic imaging of an intravenously administered echocardiographic contrast agent: Visualization of coronary arteries and measurement of coronary blood flow.
        J Am Coll Cardiol. 1996; 27: 1519-1525
        • Mulvagh S.L.
        • DeMaria A.N.
        • Feinstein S.B.
        • Burns P.N.
        • Kaul S.
        • Miller J.G.
        • Monaghan M.
        • Porter T.R.
        • Shaw L.J.
        • Villanueva F.S.
        Contrast echocardiography: Current and future applications.
        J Am Soc Echocardiogr. 2000; 13: 331-342
        • Nakano H.
        • Ishida Y.
        • Hatakeyama T.
        • Sakuraba K.
        • Hayashi M.
        • Sakurai O.
        • Hataya K.
        Contrast-enhanced intraoperative ultrasonography equipped with late Kupffer-phase image obtained by sonazoid in patients with colorectal liver metastases.
        World J Gastroenterol. 2008; 14: 3207-3211
        • Noltigk B.E.
        • Neppiras E.E.
        Cavitation produced by ultrasonics.
        Proc Phys Soc B. 1950; 63: 674-685
        • Opacic T.
        • Dencks S.
        • Theek B.
        • Piepenbrock M.
        • Ackermann D.
        • Rix A.
        • Lammers T.
        • Stickeler E.
        • Delorme S.
        • Schmitz G.
        Motion model ultrasound localization microscopy for preclinical and clinical multiparametric tumor characterization.
        Nat Commun. 2018; 9: 1527
        • Phillips P.J.
        Contrast pulse sequences (CPS): Imaging nonlinear microbubbles.
        Proc IEEE Int Ultrason Symp. 2001; : 1739-1745
        • Phillips P.
        • Gardner E.
        Contrast-agent detection and quantification.
        Eur Radiol. 2004; 14: P4-P10
        • Porter T.R.
        • Xie F.
        • Kricsfeld D.
        • Armbruster R.W.
        Improved myocardial contrast with second harmonic transient ultrasound response imaging in humans using intravenous perfluorocarbon-exposed sonicated dextrose albumin.
        J Am Coll Cardiol. 1996; 27: 1497-1501
        • Porter T.R.
        • Xie F.
        • Li S.
        • D'Sa A.
        • Rafter P.
        Increased ultrasound contrast and decreased microbubble destruction rates with triggered ultrasound imaging.
        J Am Soc Echocardiogr. 1996; 9: 599-605
        • Potdevin T.
        • Fowlkes J.
        • Moskalik A.
        • Carson P.
        Analysis of refill curve shape in ultrasound contrast agent studies.
        Med Phys. 2004; 31: 623-632
        • Rehrig P.W.
        • Hackenberger W.S.
        • Jiang X.
        • Shrout T.R.
        • Zhang S.
        • Speyer R.
        2003 Status of piezoelectric single crystal growth for medical transducer applications.
        Proc IEEE Int Ultrason Symp. 2003; : 766-769
        • Ries F.
        • Kaal K.
        • Schultheiss R.
        • Solymosi L.
        • Schlief R.
        Air microbubbles as a contrast medium in transcranial Doppler sonography.
        J Neuroimaging. 1991; 4: 173-187
        • Schneider M.
        • Arditi M.
        • Barrau M.B.
        • Brochot J.
        • Broillet A.
        • Ventrone R.
        • Yan F.
        BR1: A new ultrasonographic contrast agent based on sulfur hexafluoride-filled microbubbles.
        Invest Radiol. 1995; 30: 451-457
        • Schrope B.A.
        • Newhouse V.L.
        Second harmonic ultrasonic blood perfusion measurement.
        Ultrasound Med Biol. 1993; 19: 567-579
        • Schrope B.
        • Newhouse V.L.
        • Uhlendorf V.
        Simulated capillary flow measurement using a nonlinear ultrasonic contrast agent.
        Ultrason Imaging. 1992; 14: 134-158
        • Segers T.
        • de Jong N.
        • Versluis M.
        Uniform scattering and attenuation of acoustically sorted ultrasound contrast agents: Modeling and experiments.
        J Acoust Soc Am. 2016; 140: 2506
        • Segers T.
        • Lohse D.
        • Versluis M.
        • Frinking P.
        Universal equations for the coalescence probability and long-term size stability of phospholipid-coated monodisperse microbubbles formed by flow focusing.
        Langmuir. 2017; 33: 10329-10339
        • Segers T.
        • Kruizinga P.
        • Kok M.P.
        • Lajoinie G.
        • de Jong N.
        • Versluis M.
        Monodisperse versus polydisperse ultrasound contrast agents: Non-linear response, sensitivity, and deep tissue imaging potential.
        Ultrasound Med Biol. 2018; 44: 1482-1492
        • Shelton S.E.
        • Lindsey B.D.
        • Tsuruta J.K.
        • Foster F.S.
        • Dayton P.A.
        Molecular acoustic angiography: A new technique for high-resolution superharmonic ultrasound molecular imaging.
        Ultrasound Med Biol. 2016; 42: 769-781
        • Sontum P.C.
        Physicochemical characteristics of Sonazoid, a new contrast agent for ultrasound imaging.
        Ultrasound Med Biol. 2008; 34: 824-833
        • Sun C.
        • Sboros V.
        • Butler M.B.
        • Moran C.M.
        In vitro acoustic characterization of three phospholipid ultrasound contrast agents from 12 to 43 MHz.
        Ultrasound Med Biol. 2014; 40: 541-550
        • Tang M.X.
        • Eckersley R.J.
        Nonlinear propagation of ultrasound through microbubble contrast agents and implications for imaging.
        IEEE Trans Ultrason Ferroelectr Freq Control. 2006; 53: 2406-2415
        • Tang M.X.
        • Kamiyama N.
        • Eckersley R.J.
        Effects of nonlinear propagation in ultrasound contrast agent imaging.
        Ultrasound Med Biol. 2010; 36: 459-466
        • Tanter M.
        • Fink M.
        Ultrafast imaging in biomedical ultrasound.
        IEEE Trans Ultrason Ferroelectr Freq Control. 2014; 61: 102-119
        • Thapar A.
        • Shalhoub J.
        • Averkiou M.
        • Mannaris C.
        • Davies A.H.
        • Leen E.L.
        Dose-dependent artifact in the far wall of the carotid artery at dynamic contrast-enhanced US.
        Radiology. 2012; 262: 672-679
        • Tiemann K.
        • Pohl C.
        • Schlosser T.
        • Goenechea J.
        • Bruce M.
        • Veltmann C.
        • Kuntz S.
        • Bangard M.
        • Becher H.
        Stimulated acoustic emission: Pseudo-Doppler shifts seen during the destruction of nonmoving microbubbles.
        Ultrasound Med Biol. 2000; 26: 1161-1167
        • Tranquart F.
        • Mercier L.
        • Frinking P.
        • Gaud E.
        • Arditi M.
        Perfusion quantification in contrast-enhanced ultrasound (CEUS)–Ready for research projects and routine clinical use.
        Ultraschall Med. 2012; 33: S31-S38
        • Tremblay-Darveau C.
        • Williams R.
        • Milot L.
        • Bruce M.
        • Burns P.N.
        Combined perfusion and Doppler imaging using plane-wave nonlinear detection and microbubble contrast agents.
        IEEE Trans Ultrason Ferroelectr Freq Control. 2014; 61: 1988-2000
        • Tremblay-Darveau C.
        • Williams R.
        • Milot L.
        • Bruce M.
        • Burns P.
        Visualizing the tumour microvasculature with a nonlinear plane-wave Doppler imaging scheme based on amplitude modulation.
        IEEE Trans Med Imaging. 2016; 35: 699-709
        • Tremblay-Darveau C.
        • Williams R.
        • Sheeran P.S.
        • Milot L.
        • Bruce M.
        • Burns P.N.
        Concepts and tradeoffs in velocity estimation with plane-wave contrast-enhanced Doppler.
        IEEE Trans Ultrason Ferroelectr Freq Control. 2016; 63: 1890-1905
        • Tremblay-Darveau C.
        • Sheeran P.S.
        • Vu C.K.
        • Williams R.
        • Zhang Z.
        • Bruce M.
        • Burns P.N.
        The role of microbubble echo phase lag in multipulse contrast-enhanced ultrasound imaging.
        IEEE Trans Ultrason Ferroelectr Freq Control. 2018; 65: 1389-1401
        • Trilling L.
        Collapse and rebound of a gas bubble.
        J Appl Phys. 1952; 23: 14-17
        • Uhlendorf V.
        • Hoffmann C.
        Nonlinear acoustical response of coated microbubbles in diagnostic ultrasound.
        Proc IEEE Ultrasonics Symp. 1994; : 1559-1562
        • Unger E.
        • Shen D.
        • Fritz T.
        • Kulik B.
        • Lund P.
        • Wu G.-L.
        • Yellowhair D.
        • Ramaswami R.
        • Matsunaga T.
        Gas-filled lipid bilayers as ultrasound contrast agents.
        Invest Radiol. 1994; 29: 134-136
        • von Bibra H.
        • Sutherland G.
        • Becher H.
        • Neudert J.
        • Nihoyannopoulos P.
        Clinical evaluation of left heart Doppler contrast enhancement by a saccharide-based transpulmonary contrast agent.
        The Levovist Cardiac Working Group. J Am Coll Cardiol. 1995; 25: 500-508
        • Wang H.
        • Kaneko O.F.
        • Tian L.
        • Hristov D.
        • Willmann J.K.
        Three-dimensional ultrasound molecular imaging of angiogenesis in colon cancer using a clinical matrix array ultrasound transducer.
        Invest Radiol. 2015; 50: 322
        • Wei K.
        • Jayaweera A.
        • Firoozan S.
        • Linka A.
        • Skyba D.
        • Kaul S.
        Quantification of myocardial blood flow with ultrasound-induced destruction of microbubbles administered as a constant venous infusion.
        Circulation. 1998; 97: 473-483
        • Weisskoff R.M.
        • Chesler D.
        • Boxerman J.L.
        • Rosen B.R.
        Pitfalls in MR measurement of tissue blood flow with intravascular tracers: Which mean transit time?.
        Magn Reson Med. 1993; 29: 553-559
        • Wermke W.
        • Gaßmann B.
        Tumour diagnostics of the liver with echo enhancers: Colour atlas.
        Springer Science & Business Media. 2012;
        • Whittingham T.A.
        Contrast-specific imaging techniques: Technical perspective.
        in: Quaia E. Contrast media in ultrasonography: Basic principles and clinical applications. Springer, New York2005: 43-70
        • Williams R.
        • Hudson J.M.
        • Lloyd B.A.
        • Sureshkumar A.R.
        • Lueck G.
        • Milot L.
        • Atri M.
        • Bjarnason G.A.
        • Burns P.N.
        Dynamic microbubble contrast-enhanced US to measure tumor response to targeted therapy: A proposed clinical protocol with results from renal cell carcinoma patients receiving antiangiogenic therapy.
        Radiology. 2011; 260: 581-590
        • Wilson S.R.
        • Burns P.N.
        • Muradali D.
        • Wilson J.A.
        • Lai X.
        Harmonic hepatic US with microbubble contrast agent: Initial experience showing improved characterization of hemangioma, hepatocellular carcinoma, and metastasis.
        Radiology. 2000; 215: 153-161
        • Wilson S.R.
        • Jang H.J.
        • Kim T.K.
        • Iijima H.
        • Kamiyama N.
        • Burns P.N.
        Real-time temporal maximum-intensity-projection imaging of hepatic lesions with contrast-enhanced sonography.
        AJR Am J Roentgenol. 2008; 190: 691-695
        • Yu H.
        • Jang H.J.
        • Kim T.K.
        • Khalili K.
        • Williams R.
        • Lueck G.
        • Hudson J.
        • Burns P.N.
        Pseudoenhancement within the local ablation zone of hepatic tumors due to a nonlinear artifact on contrast-enhanced ultrasound.
        AJR Am J Roentgenol. 2010; 194: 653-659