- Email: [email protected]
More Than Bubbles: Creating Phase-Shift Droplets from Commercially Available Ultrasound Contrast Agents
Ultrasound in Med. & Biol., Vol. -, No. -, pp. 1–10, 2016 Copyright Ó 2016 World Federation for Ultrasound in Medicine & Biology Printed in the USA. All rights reserved 0301-5629/$ - see front matter
http://dx.doi.org/10.1016/j.ultrasmedbio.2016.09.003
d
Technical Note MORE THAN BUBBLES: CREATING PHASE-SHIFT DROPLETS FROM COMMERCIALLY AVAILABLE ULTRASOUND CONTRAST AGENTS PAUL S. SHEERAN,*y KIMOON YOO,* ROSS WILLIAMS,* MELISSA YIN,* F. STUART FOSTER,*y and PETER N. BURNS*y * Physical Sciences Department, Sunnybrook Research Institute, Toronto, ON, Canada; and y Department of Medical Biophysics, University of Toronto, Toronto, Canada (Received 5 March 2016; revised 30 August 2016; in final form 6 September 2016)
Abstract—Phase-shift perfluorocarbon droplets have been investigated for over 20 years as pre-clinical ultrasound contrast agents with distinctive advantages in imaging and therapy. A number of formulation strategies exist, each with inherent advantages and limitations. In this note, we demonstrate a unique opportunity: that phase-shift droplets can be generated directly from commercially available microbubbles. This may facilitate pre-clinical and translational development by reducing the in-house synthesis expertise and resources required to generate high concentration droplet emulsions. Proof-of-principle in vitro and in vivo is given using droplets created from Definity and MicroMarker. The results demonstrate the role of perfluorocarbon choice in the trade-off between thermal stability and vaporization threshold, and suggest that commercial microbubbles with decafluorobutane cores may be ideal for this approach. (E-mail: [email protected]) Ó 2016 World Federation for Ultrasound in Medicine & Biology. Key Words: Perfluorocarbon, Ultrasound contrast agents, Phase-shift droplets, Acoustic droplet vaporization, Definity, MicroMarker.
cancerous tissues (Kripfgans et al. 2000; Rapoport 2012; Sheeran and Dayton 2012a). Ultrasonically activated phase-shift droplets have made slow progress into the clinic. Many fundamental mechanisms of droplet vaporization and expansion have only recently been revealed (Doinikov et al. 2014; Shpak et al. 2014), and much work remains in characterizing bioeffects, pharmacokinetics and dosing before use in humans can be considered. A variety of formulation strategies exist, including amalgamation/ homogenization/mechanical agitation (Kawabata et al. 2005; Kripfgans et al. 2000), sonication (Fabiilli et al. 2010; Matsuura et al. 2009; Rapoport et al. 2009), microfluidics (Bardin et al. 2011; Seo et al. 2010) and extrusion past porous membranes (Giesecke and Hynynen 2003; Sheeran et al. 2011b). More recently, Sheeran et al. (2011a, 2012b) developed a ‘‘microbubble condensation’’ technique to create phaseshift droplets in which volatile compounds are first encapsulated as microbubbles and are then condensed to the liquid state by decreasing the ambient temperature and/or increasing the ambient pressure. Researchers have demonstrated that microbubble condensation can be coupled with other particle generation techniques such
INTRODUCTION Clinical ultrasound contrast agents take the form of gas-filled microspheres, commonly referred to as microbubbles. In the past two decades, investigations into alternative contrast agents have increased substantially in order to address the fundamental limitations of microbubbles for specific diagnostic and therapeutic applications. Phase-shift perfluorocarbon droplets are perhaps the most extensively researched alternative and are designed to exist metastably in the liquid state under a desired degree of superheat until additional energy in the form of ultrasound or heat nucleates the core and the particles vaporize to the gas state, expanding volumetrically. This approach provides the benefits of increased circulation lifetime, relative ease of generating stable particles at the nanoscale, the ability to generate microbubbles ‘‘on-site’’ through externally applied ultrasound and the possibility of accessing the extravascular space of
Address correspondence to: Paul S. Sheeran, 2075 Bayview Ave, Room S640, Toronto, ON M4N 3M5, Canada. E-mail: pssheeran@ gmail.com 1
2
Ultrasound in Medicine and Biology
as microfluidic generation of microbubbles in order to refine droplet size distributions (Seo and Matsuura 2012; Seo et al. 2015). Each of these formulation strategies has advantages and disadvantages with regard to clinical applicability, resource requirements, and the necessary level of expertise in interfacial chemistry/colloid science (Sheeran et al. 2016b). In this note, we demonstrate that generating phase-shift droplets by microbubble condensation may provide a unique advantage in that it allows for the creation of droplets directly from existing commercial clinical and pre-clinical contrast agents, providing high yield emulsions with minimal resource requirements and without need for in-house bubble synthesis. Here, droplets are created directly from the clinical contrast agent Definity (Lantheus Medical Imaging, Billerica, MA, USA) after formation of bubbles by mechanical agitation, and the pre-clinical contrast agent MicroMarker (Bracco, Geneva, Switzerland and VisualSonics, Toronto, Canada), formed following reconstitution of a lyophilized emulsion. Proof-of-principle is given in vitro and in vivo on clinical and pre-clinical ultrasound machines. MATERIALS AND METHODS Droplet emulsion preparation Phase-shift droplets were generated from two commercially available phospholipid-encapsulated microbubble contrast agents. Definity, the most commonly used clinical ultrasound contrast agent in North America, has a gaseous octafluoropropane core (boiling point 236.7 C), while MicroMarker, a pre-clinical contrast agent designed for small-animal imaging, has a core composed of nitrogen and decafluorobutane (boiling point 22 C). Both agents have a distribution peak diameter below 1 mm and a volume-weighted peak diameter below 5 mm (Helfield et al. 2012; Raymond et al. 2014). Definity microbubbles were formed by activating vials (stored at 4 C) by mechanical agitation according to standard protocols. Octafluoropropane droplets were generated from this suspension with a similar approach to that described in Sheeran et al. (2012b). First, the vial was cooled to 210 C in an isopropanol bath maintained at temperature with dry ice, followed by pressurization of the headspace with room air to approximately 170 kPa above atmospheric pressure over the course of 2 min. MicroMarker microbubbles were prepared by resuspending the lyophilized agent with 0.5 mL of decafluorobutane gas-equilibrated saline. Decafluorobutane droplets were generated by cooling the resulting suspension to 28 C in the isopropanol bath and pressurizing the headspace with room air to approximately 100 kPa above atmospheric pressure over the course of 2 min in order to
Volume -, Number -, 2016
condense the decafluorobutane component of the core and expel the nitrogen component (condensation of a pure decafluorobutane core can be accomplished by cooling alone). The droplet suspensions were then centrifuged at 1000 rpm for 2 min at 4 C to separate droplets from microbubbles remaining in the solution, and were held in a centrifuge tube on ice for up to 1 h before testing. Particle sizing Example distributions of droplets generated by condensing Definity and MicroMarker microbubbles were measured using a Nanosight LM10 (Malvern Instruments, Inc., Malvern, UK). The droplet suspensions were diluted 1/200 and 1/100 in room-temperature saline for octafluoropropane droplets (generated from Definity) and decafluorobutane droplets (generated from MicroMarker), respectively. Five 30-s videos of nonoverlapping volumes were collected for each dilution, analyzed and averaged to generate representative distributions. This process was repeated three times by preparing new dilutions from the same vial, and the results averaged for a final representation of the vial size distribution and particle concentration. In vitro experimental design An in vitro vessel flow phantom was constructed by connecting a 6.4-mm diameter (0.24 mm wall thickness) latex tube (Penrose Drain; Medline Industries, Inc., Mundelein, IL, USA) to polyvinyl chloride tubing secured by a custom acrylic holder that was submerged in a large heated water bath. The flow circuit consisted of a 1.5 L sample container held 30–40 cm above the large heated water bath. Temperatures between 25 C and 42 C were maintained in the large water bath by an immersion circulator/heater, and in the upper sample container using a small fish tank heater with an adjustable voltage regulator. Droplet dilutions were imaged and vaporized in the vessel phantom by fixing an ultrasound imaging probe to a mechanical arm and submerging the probe to image the latex tube lengthwise in a set position. Droplets created from the clinical microbubble Definity were tested using a Philips iU22 clinical scanner with an L9-3 linear array probe (Philips Ultrasound, Bothell, WA, USA). Standard B-mode imaging settings were used to capture videos of the samples (mechanical index [MI] 5 0.01, 68 fps). The imaging focus was placed in the center of the tube at 2.2 cm depth. Sample flow in the circuit was stopped and droplets were vaporized by manually triggering a single flash frame—conventionally used to destroy contrast agents for the purposes of volume-flow and molecular imaging (Hudson et al. 2015; Wilson and Burns 2010)— over a range of preset nominal mechanical indices between MI 5 0.05 and MI 5 0.97 (the highest output
Creating Phase-Shift Droplets from Commercial Microbubbles d P. S. SHEERAN et al.
available). Between vaporization captures, an untested sample volume was brought into the imaging plane by restoring forward flow. Output pressures were determined by measuring the center frequency of the flash by membrane hydrophone (Model 804, Sonora Medical Systems, Longmont, CO, USA) and correcting the nominal (derated) value, which assumes an attenuation coefficient of 0.3 dB/(cm $ MHz) set by the US Food and Drug Administration. Here, a calculated free-field peak negative pressure was obtained by negating this attenuation over 2.2 cm at 4.7 MHz, the measured flash frequency. The peak negative pressure inside the tube was estimated by correcting the free-field value to account for attenuation over 0.24 mm of latex according to the frequency dependence characterized by Rickey et al. (1995). Losses due to reflection were neglected due to the similar speed of sound and density between water and latex. Droplets created from the pre-clinical microbubble MicroMarker (designed for small animal imaging) were tested using a VisualSonics Vevo2100 system with an MS400 (30 MHz) imaging probe (VisualSonics) in standard B-mode acquisition. The imaging focus was placed in the center of the tube at 9 mm depth. Due to lack of a manually triggered flash, sample flow in the circuit was stopped and videos of droplets in the imaging plane were captured as a function of increasing B-mode output power between the lowest and highest outputs available (Sheeran et al. 2016a). Here, as above, peak negative pressure inside the tube was estimated by correcting the measured free-field value to account for attenuation over the latex tube wall, with reflection losses neglected. At each output power, 50 frames were collected at 20 fps. Controls were performed by collecting additional videos under the same conditions with only water (i.e., no contrast agents). Dilutions were determined based on preliminary tests with similar in-house droplet formulations, and set to be within approximately one order of magnitude of the concentration used for microbubble characterization in vitro with similar vessel models (Tremblay-Darveau et al. 2016). In each case, a 120-mL bolus of the droplet suspension being tested was drawn from the cooled vial and mixed into 1 L of gas equilibrated deionized water in the sample container heated to the desired temperature. Forward flow was restored until the sample was present in the tube, after which flow was stopped and testing begun immediately. As an exception, a 60-mL bolus of the octafluoropropane droplets was added to the 1 L of deionized water at temperature below 32 C, where preliminary results suggested increased stability and very high contrast production compared to temperatures 32 C and above. Capturing the data over the entire test range per sample typically took 45 s to 1 min. This process was repeated
3
for the same sample an additional two times in the following minutes to characterize in vitro decay of that sample over time. This process was repeated for three to four vials of independently prepared samples of both droplet types. Video analysis Video captures obtained on the Philips iU22 were analyzed offline in QLab quantification software (Philips Ultrasound). Manually drawn regions of interest were placed inside the tube in the region of activation and the change in mean pixel dB after the flash frame was calculated. These results were averaged for each temperature for the three to four vials tested. Decay rates were calculated by fitting an exponential decay curve to the linear intensity values obtained over the three runs for each sample at the two highest flash pressures. These decay rates were averaged for the three to four vials tested at each temperature. Videos obtained from the Vevo2100 were analyzed offline in MATLAB (The MathWorks, Inc., Natick, MA, USA) by first converting the pixel values to linearized intensity (Williams et al. 2013). Manually drawn regions of interest were placed inside the test region and mean pixel values were calculated. These values were converted to dB and adjusted by subtracting the background dB level obtained from the water-only control samples at the same power level. The final values obtained for each sample were averaged for all vials at each temperature. During tests, it was noted that droplets did not decay during the test period, and so fitting with exponential decay was not performed in this case. In vivo vaporization demonstrations All in vivo experiments were approved by the animal care committee of the Sunnybrook Research Institute. Octafluoropropane droplets (generated from Definity) were investigated in the kidney of a New Zealand white rabbit using the Philips iU22 scanner and L9-3 probe attached to a fixed mechanical arm. The rabbit was inducted with dexmedetomidine and maintained under anesthesia with isoflurane according to protocols approved at Sunnybrook Research Institute. The kidney was imaged at 2.5 cm of depth using a side-by-side contrast and B-mode scan, with the focus placed at 2.25 cm. The MI was set to 0.15 for contrast mode and 0.04 for B-mode at a frame rate of 13 Hz. A 200-mL bolus of octafluoropropane droplets was added to 1 mL saline and manually injected over 20 s via ear-vein catheterization, followed by a 2.5-mL saline flush, which falls within previously published doses for microbubbles in this animal model (McDannold et al. 2006; Tremblay-Darveau et al. 2016). Wash-in of the bolus (bubbles produced by spontaneous vaporization) was monitored and a manually
4
Ultrasound in Medicine and Biology
triggered flash frame at an MI of 1.31 was triggered at the peak bolus enhancement. The resulting video was analyzed offline in QLab. Decafluorobutane droplets (generated from MicroMarker) were investigated in the liver of a C3H mouse using the Vevo2100 system and MS400 probe. The mouse was anesthetized by isoflurane according to protocols approved at Sunnybrook Research Institute. The center of the liver was positioned at the transducer’s elevational focus, 9 mm from the transducer face. Baseline scans were taken before agent injection in B-mode (20 fps) at 1% and 100% power using a two-focus imaging setting with foci at 8 mm and 9 mm for a larger region of vaporization. Following a previously developed protocol, a 100-mL bolus of undiluted decafluorobutane droplets was administered by syringe pump over 1 min, followed by a saline flush at the same volume and rate
Volume -, Number -, 2016
(Sheeran et al. 2016a). Five min after injection, scans were taken again at 1% and 100% power. RESULTS AND DISCUSSION Droplet generation and sizing Microbubbles were prepared by either mechanical agitation or reconstitution (Fig. 1). After condensation, the appearance of the emulsion changed from a milky solution to a more translucent appearance (Fig. 1). In the case of droplets generated from Definity, a clear bubble layer persisted at the top of the emulsion after condensation and centrifugation while the droplets generated from MicroMarker appeared uniform throughout—highlighting the reduced stability of octafluoropropane droplets compared to decafluorobutane droplets. In both cases, condensation produced primarily sub-micron
Fig. 1. Commercial microbubble agents Definity and MicroMarker shown as packaged in vial. After mechanical agitation (Definity) or reconstitution (MicroMarker), a microbubble condensation technique can be used to convert microbubbles to sub-micron phase-shift droplets.
Creating Phase-Shift Droplets from Commercial Microbubbles d P. S. SHEERAN et al.
droplet sizes. The example sizing of one vial of octafluoropropane droplets on the NanoSight produced a peak diameter, 10th percentile diameter (90% of the distribution above this size) and 90th percentile diameter (90% of the distribution below this size) of 216 nm 6 6 nm, 191 nm 6 8 nm and 476 nm 6 65 nm, respectively, and a concentration of 2.90 3 1011 particles/mL 6 0.93 3 1011 particles/mL. The decafluorobutane droplet sample produced a peak diameter, 10th percentile and 90th percentile diameters of 264 nm 6 12 nm, 230 nm 6 6 nm and 626 nm 6 43 nm, respectively, and a concentration of 1.08 3 1011 particles/mL 6 0.12 3 1011 particles/mL. In vitro results: octafluoropropane droplets Octafluoropropane droplet samples diluted to between 0.006% and 0.012% were exposed to manually triggered flash frames on the Philips iU22 to test droplet vaporization as a function of both peak negative pressure and ambient temperature. At 25 C, the droplet samples provided no significant ultrasound contrast within the tube before the flash frame (Fig. 2a). Once exposed to single flash frames with nominal mechanical indices above 0.5, the sample volume interrogated produced a high level of B-mode contrast, which decayed steadily over the course of 20–30 s. As temperature increased, the ‘‘preflash’’ signal magnitude increased in a generally linear fashion from a mean value of 7.6 dB at 25 C to 15.6 dB at 37 C, indicating a greater level of spontaneous vaporization with increasing temperature. At 32 C and 35 C, though bubbles were clearly present in the pre-flash segment, vaporization could still be observed as a marked increase in brightness after the flash frame. However, at
5
37 C, only a decrease in brightness in the region of interest was observed due to microbubble destruction from the flash frame (Fig. 2a). The change in mean pixel dB as a function of flash pressure and temperature produced a clear picture of regimes in which the flash frame can increase contrast (‘‘flash activation’’ of droplets) or destroy contrast (‘‘flash destruction’’ of microbubbles present by spontaneous vaporization) (Fig. 3a). At 25 C, the flash produced as much as 12 dB of enhancement at peak negative pressures above 1.5 MPa. At 32 C and 35 C, bubbles are clearly present in-plane before the flash frame, and low flash pressures tended to destroy rather than create contrast. Positive contrast was observed at pressures above 0.9 MPa and 0.5 MPa for 32 C and 35 C, respectively. The magnitude of the contrast enhancement decreased with increasing temperature as a result of a greater background signal before the flash. At 37 C, the pre-flash signal appeared to be dominated by microbubbles (Fig. 2a), and the flash frame only produced microbubble destruction, resulting in a decrease in measured signal strength. In vitro results: decafluorobutane droplets Decafluorobutane droplet samples diluted to 0.012% were exposed to increasing output power on the Vevo 2100 to test droplet vaporization as a function of both peak negative pressure and ambient temperature. At 25 C, droplet samples provided no significant contrast within the tube at any output power (Fig. 2b). At temperatures higher than 35 C, droplet vaporization produced individually distinguishable microbubbles that translated axially due to acoustic streaming and radiation force over
Fig. 2. Ultrasound imaging and activation of droplet suspensions in a latex tube. (a) Imaged with a Philips iU22 and L9-3 probe, octafluoropropane droplets generated from Definity show high stability at 25 C and can be vaporized with the manual flash to form contrast-providing microbubbles. At 37 C, however, droplet show a high level of spontaneous vaporization before the flash, and as a result the flash frame diminishes enhancement. (b) Imaged with a VisualSonics Vevo2100 and MS-400 probe, decafluorobutane droplets generated from MicroMarker show high stability at all temperatures, and could not be vaporized at 25 C at the maximum output of the probe. At higher temperatures, vaporization produced echogenic microbubbles within the tube.
6
Ultrasound in Medicine and Biology
Volume -, Number -, 2016
a
b
Fig. 3. Contrast enhancement of droplets as a function of temperature and estimated peak negative pressure inside the latex tube. Markers represent average value across samples, while bars represent the range of the individual measurements at each pressure. (a) Octafluoropropane droplets were vaporized at 4.7 MHz using a manually triggered flash frame on the iU22, with contrast enhancement calculated relative to the signal just before the flash. For clarity, the 25 C and 35 C series have been shifted left and right, respectively, by 20 kPa. (b) Decafluorobutane droplets were vaporized at 30 MHz by increasing the imaging output power on the Vevo2100 and the contrast enhancement calculated relative to agent-free controls. For clarity, the 42 C and 37 C series have been shifted left and right, respectively, by 25 kPa.
the 50 frames. No significant change was observed in the signal for the lowest output powers at each temperature, indicating the decafluorobutane droplets were stable against spontaneous vaporization over the entire temperature range—consistent with other reports at these temperatures (Mountford et al. 2015b; Sheeran et al. 2012b, 2016a).
Calculating the increase in mean pixel dB over the no-agent controls illustrates the relationship of temperature and pressure to vaporization of decafluorobutane droplets (Fig. 3b). At temperatures of 35 C and higher, vaporization produced maximum B-mode contrast on the order of 10–12 dB within the tube. The peak negative pressure required to vaporize droplets at 37 C occurred
Creating Phase-Shift Droplets from Commercial Microbubbles d P. S. SHEERAN et al.
Fig. 4. In vivo demonstrations of droplet performance. (a) Octafluoropropane droplets imaged with the iU22 showed substantial spontaneous vaporization upon intravenous injection and imaging in the rabbit kidney (contrast mode shown with kidney region manually outlined), but the manually triggered flash produced a modest increase in contrast as a result of droplet vaporization, calculated in the plot below. (b) Vaporization of decafluorobutane droplets in the mouse liver with the Vevo2100 produced a bright band of enhancement above the focal depth compared to the same region before injection.
7
8
Ultrasound in Medicine and Biology
near 3 MPa, which is in agreement with a recent study testing similar in-house preparations (Sheeran et al. 2016a). In vivo results Upon intravenous injection of octafluoropropane droplets (generated from Definity), a bolus wash-in of microbubbles was observed in the kidney as a result of spontaneous vaporization. Unlike the in vitro results at physiologic temperatures, the manually triggered flash at peak wash-in produced an increase in contrast on the order of 5 dB within the kidney region, which dissipated over the course of several seconds (Fig. 4a). The difference between in vitro performance (Fig. 2a, high initial microbubble signal diminished by flash) and in vivo performance (enhancement over microbubble signal upon flash) is likely a result of clearance of microbubbles from circulation and filtration by the lungs en route to the kidney, allowing the un-vaporized droplets to be activated and visualized. This is consistent with the results for inhouse preparations of octafluoropropane droplets, which show that some portion of the original population remains unvaporized at body temperature within the test period (Mountford et al. 2015b; Sheeran et al. 2012b, 2014, 2015). In this study, attempts to produce additional vaporization in the minutes following administration were less successful, indicating a relatively short circulation profile consistent with previous reports (Porter et al. 2016; Sheeran et al. 2015). For decafluorobutane droplets, imaging was performed 5 min after injection of the bolus, similar to the protocol in Sheeran et al. (2016a). The liver was first imaged in B-mode at 1% power, and no droplet vaporization was observed. Immediately following this, output power was increased to 100%, producing a clear, bright band within the liver superficial to the depth of focus as a result of droplet vaporization (Fig. 4b). Enhancement at later time points was not tested. In vitro results: temperature dependence and instability We expect droplets to decay due to a combination of droplet dissolution and spontaneous vaporization, which both depend on ambient temperature. The observed spontaneous vaporization as a function of temperature in this study matches closely with that recently reported by Mountford et al. (2015b) for octafluoropropane droplets, with general stability against spontaneous vaporization at temperatures below 30 C and a high degree of spontaneous vaporization at temperatures near 37 C. The similarity in the stability properties between this study (observed acoustically) and that of Mountford et al. (measured optically) despite differences in size and encapsulation supports the notion that the spontaneous vaporization of encapsulated perfluorocarbon droplets
Volume -, Number -, 2016
primarily depends on the nearness of the compound’s critical temperature to the ambient temperature (Eberhart 1976). At 25 C, the octafluoropropane droplets generated from Definity produced an exponential decay rate of 0.05 min21 6 0.06 min21 (half-life of approximately 14.7 min). Upon exposure to higher temperatures, the decay rate increased more than an order of magnitude to 1.1 min21 6 0.13 min21 (half-life of approximately 0.6 min) at 35 C. In contrast, the decafluorobutane droplets generated from MicroMarker did not produce a discernible decay over the test period. Future testing of this sort may be useful in order to experimentally derive the dissolution rate of each droplet core compound as a function of temperature, encapsulation and concentration, which has not been previously characterized over a temperature range relevant to medical ultrasound. Limitations and future directions As a proof-of-principle, this study gave only preliminary evidence of droplet thresholds and size, which may be refined in future investigations. In order to demonstrate activation of the two droplet types, which had substantially different thermal stability and activation thresholds, two different commercial imaging systems were required. The system-dependent imaging properties and the wide separation in center frequency used to assess activation made it impossible to directly compare some details, such as relative contrast enhancement between the two droplet types and the pressures required for conversion—for which frequency dependence is a complex phenomenon (Kripfgans et al. 2000; Sheeran et al. 2016a; Shpak et al. 2014; Vlaisavljevich et al. 2015; Williams et al. 2013). Regardless, some general conclusions can be drawn. Sub-micron droplets with decafluorobutane cores were not able to be activated at the peak output of the L9-3 probe at 37 C (nominal peak MI of 0.97 in B-mode), although they could be activated at the frequencies available on small-animal imaging systems (Sheeran et al. 2016a). Conversely, droplets with octafluoropropane cores could be vaporized using the clinical system, but exhibited poor thermal stability at the ideal temperatures. The low vaporization threshold of these droplets prevented characterization on the small-animal imaging system, as the lowest imaging pressures available exceeded the vaporization threshold at temperatures of 32 C or higher, leaving no option for imaging without activation. The fact that the more stable decafluorobutane droplets could not be activated on the iU22 with the L9-3 probe should not be interpreted to mean that they require activation pressures beyond the diagnostic limit. In fact, most experimental characterizations of these droplets indicate activation at 5–8 MHz is possible with
Creating Phase-Shift Droplets from Commercial Microbubbles d P. S. SHEERAN et al.
mechanical indices on the order of 1.3 (Sheeran et al. 2013, 2015). However, this does not mean a clinical machine and transducer will necessarily be able to produce these absolute pressures/MIs within tissue. For example, most soft tissue imaging settings assume a lower than normal tissue attenuation of 0.3 dB/(cm $ MHz), as opposed to values near 0.5 dB/(cm $ MHz) or higher (Goss et al. 1979) in order to ensure the actual pressures are not higher than expected. So a nominal MI of 1.3 at 5 MHz in 2 cm of tissue may produce an actual MI on the order of 1.0 or less, below the activation threshold. Here, the estimated peak negative pressure inside the tube reached a maximum of 2.67 MPa, or MI 5 1.23—just below where activation likely begins for the decafluorobutane droplets at this frequency. Future studies are necessary to identify clinical systems with ideal vaporization frequencies and output pressures to efficiently convert sub-micron droplets in vivo. Thus, for use with low clinical frequencies (1–5 MHz) within the regulatory output limits, this study highlights that there is motivation for development of phase-shift droplets with properties in between the octafluoropropane and decafluorobutane droplets tested here—ideally with reduced vaporization thresholds while maintaining high thermal stability. This might be accomplished by choosing an alternative compound that has similar properties (hydrophobic, inert, non-toxic, low solubility in aqueous media) but with a more ideal boiling point/critical temperature, by incorporating nucleation seeds to reduce activation thresholds (Lee et al. 2015; Matsuura et al. 2009), or through mixtures (Kawabata et al. 2005; Sheeran et al. 2012b). However, recent studies suggest that mixtures of such highly superheated compounds will not behave ideally in vivo (Mountford et al. 2015a). It may also be possible to reduce the vaporization pressure simply by choosing a decafluorobutane microbubble with a larger mean diameter than MicroMarker, as the vaporization threshold is inversely related to droplet diameter (Kripfgans et al. 2004; Sheeran et al. 2012b). This study provides proof-of-principle that one can develop phase-shift droplets from existing commercial agents through simple adjustment of ambient conditions, converting them into a new type of agent with advantages compared to microbubbles for specific ultrasoundmediated applications. The ability to produce phaseshift droplets without the necessity of in-house bubble or droplet synthesis expertise may facilitate preclinical development and exploration of new applications. That these droplets can be generated from precursors already approved for the clinic and with existing good manufacturing protocols may also provide new opportunities for commercialization and translation, although assessment of pharmacokinetics and safety would be
9
necessary. The two different commercial microbubbles were selected in order to demonstrate the applicability of the condensation technique for different perfluorocarbon cores as well as differences in microbubble preparation (mechanical agitation vs. reconstitution). Although MicroMarker is intended for pre-clinical use only, it is similar in encapsulation, size distribution, concentration and method of reconstitution to currently available clinical microbubbles. Thus, it can be reasonably expected that the condensation technique can be applied to those formulations as well. Other perfluorocarbon-based commercial microbubbles that may potentially be converted into droplets include Optison (GE Healthcare, Milwaukee, WI, USA), Sonazoid (GE Healthcare/Daiichi Sankyo, Tokyo, Japan), BR38 and the VEGFR2-targeted agent BR55 (Bracco). In particular, we propose that those with decafluorobutane cores, such as Sonazoid, may provide the most promise moving forward. CONCLUSIONS In this study, we have proposed that ultrasonically vaporized phase-shift droplets can be generated directly from commercially available microbubble contrast agents by utilizing a microbubble condensation technique. This presents a unique strategy to create new imaging and therapy agents from existing ones. The results demonstrate the role of perfluorocarbon choice in the trade-off between thermal stability and vaporization threshold, and suggest that commercial microbubbles with decafluorobutane cores may be ideal for this approach. Acknowledgments—This work has been supported by a grant from the Canadian Institutes of Health Research (MOP119346 and MOP125994) and the Ontario Institute for Cancer Research. P .S. appreciates the generous support of the Banting Postdoctoral Fellowship, administered by the Canadian Institutes of Health Research. We thank Megan Thompson and Linda Nghiem for assistance with animal studies, Charles Tremblay-Darveau and Zachary Zajac for assistance with animal experiments and data processing and Chelsea Munding for technical discussions.
REFERENCES Bardin D, Martz TD, Sheeran PS, Shih R, Dayton PA, Lee AP. Highspeed, clinical-scale microfluidic generation of stable phase-change droplets for gas embolotherapy. Lab Chip 2011;11: 3990–3998. Doinikov AA, Sheeran PS, Bouakaz A, Dayton PA. Vaporization dynamics of volatile perfluorocarbon droplets: A theoretical model and in vitro validation. Med Phys 2014;41:102901. Eberhart JG. The thermodynamic and the kinetic limits of superheat of a liquid. J Colloid Interface Sci 1976;56:262–269. Fabiilli M, Lee J, Kripfgans O, Carson P, Fowlkes J. Delivery of watersoluble drugs using acoustically triggered perfluorocarbon double emulsions. Pharm Res 2010;27:2753–2765. Giesecke T, Hynynen K. Ultrasound-mediated cavitation thresholds of liquid perfluorocarbon droplets in vitro. Ultrasound Med Biol 2003;29:1359–1365.
10
Ultrasound in Medicine and Biology
Goss SA, Frizzell LA, Dunn F. Ultrasonic absorption and attenuation in mammalian tissues. Ultrasound Med Biol 1979;5:181–186. Helfield BL, Huo X, Williams R, Goertz DE. The effect of preactivation vial temperature on the acoustic properties of Definity. Ultrasound Med Biol 2012;38:1298–1305. Hudson JM, Williams R, Tremblay-Darveau C, Sheeran PS, Milot L, Bjarnason GA, Burns PN. Dynamic contrast enhanced ultrasound for therapy monitoring. Eur J Radiol 2015;84:1650–1657. Kawabata K, Sugita N, Yoshikawa H, Azuma T, Umemura S. Nanoparticles with multiple perfluorocarbons for controllable ultrasonically induced phase shifting. Jpn J Appl Phys 2005;44:5. Kripfgans OD, Fabiilli ML, Carson PL, Fowlkes JB. On the acoustic vaporization of micrometer-sized droplets. J Acoust Soc Am 2004; 116:272–281. Kripfgans OD, Fowlkes JB, Miller DL, Eldevik OP, Carson PL. Acoustic droplet vaporization for therapeutic and diagnostic applications. Ultrasound Med Biol 2000;26:1177–1189. Lee JY, Carugo D, Crake C, Owen J, De Saint Victor M, Seth A, Coussios C, Stride E. Nanoparticle-loaded protein–polymer nanodroplets for improved stability and conversion efficiency in ultrasound imaging and drug delivery. Adv Mater 2015;27: 5484–5492. Matsuura N, Williams R, Gorelikov I, Chaudhuri J, Rowlands J, Hynynen K, Foster S, Burns P, Resnik N. Nanoparticle-loaded perfluorocarbon droplets for imaging and therapy. In: IEEE International Ultrasonics Symposium (IUS) 2009:5–8. McDannold N, Vykhodtseva N, Hynynen K. Targeted disruption of the blood–brain barrier with focused ultrasound: Association with cavitation activity. Phys Med Biol 2006;51:793. Mountford PA, Smith WS, Borden MA. Fluorocarbon nanodrops as acoustic temperature probes. Langmuir 2015a;31:10656–10663. Mountford PA, Thomas AN, Borden MA. Thermal activation of superheated lipid-coated perfluorocarbon drops. Langmuir 2015b;31: 4627–4634. Porter TR, Arena C, Sayyed S, Lof J, High RR, Xie F, Dayton PA. Targeted transthoracic acoustic activation of systemically administered nanodroplets to detect myocardial perfusion abnormalities. Circ Cardiovasc Imaging 2016;9:e003770. Rapoport N. Phase-shift, stimuli-responsive perfluorocarbon nanodroplets for drug delivery to cancer. Wiley Interdiscip Rev Nanomed Nanobiotechnol 2012;4:492–510. Rapoport NY, Kennedy AM, Shea JE, Scaife CL, Nam KH. Controlled and targeted tumor chemotherapy by ultrasound-activated nanoemulsions/microbubbles. J Control Release 2009;138:268–276. Raymond JL, Haworth KJ, Bader KB, Radhakrishnan K, Griffin JK, Huang SL, McPherson DD, Holland CK. Broadband attenuation measurements of phospholipid-shelled ultrasound contrast agents. Ultrasound Med Biol 2014;40:410–421. Rickey DW, Picot PA, Christopher DA, Fenster A. A wall-less vessel phantom for Doppler ultrasound studies. Ultrasound Med Biol 1995;21:1163–1176. Seo M, Gorelikov I, Williams R, Matsuura N. Microfluidic assembly of monodisperse, nanoparticle-incorporated perfluorocarbon microbubbles for medical imaging and therapy. Langmuir 2010;26: 13855–13860.
Volume -, Number -, 2016 Seo M, Matsuura N. Monodisperse, Submicrometer droplets via condensation of microfluidic-generated gas bubbles. Small 2012;8: 2704–2714. Seo M, Williams R, Matsuura N. Size reduction of cosolvent-infused microbubbles to form acoustically responsive monodisperse perfluorocarbon nanodroplets. Lab Chip 2015;15:3581–3590. Sheeran PS, Daghighi Y, Yoo K, Williams R, Cherin E, Foster FS, Burns PN. Image-guided ultrasound characterization of volatile sub-micron phase-shift droplets in the 20-40 MHz frequency range. Ultrasound Med Biol 2016a;42:795–807. Sheeran PS, Dayton PA. Phase-change contrast agents for imaging and therapy. Curr Pharm Des 2012a;18:2152–2165. Sheeran PS, Luois S, Dayton PA, Matsunaga TO. Formulation and acoustic studies of a new phase-shift agent for diagnostic and therapeutic ultrasound. Langmuir 2011a;27:10412–10420. Sheeran PS, Luois SH, Mullin LB, Matsunaga TO, Dayton PA. Design of ultrasonically-activatable nanoparticles using low boiling point perfluorocarbons. Biomaterials 2012b;33:3262–3269. Sheeran PS, Matsunaga TO, Dayton PA. Phase-transition thresholds and vaporization phenomena for ultrasound phase-change nanoemulsions assessed via high-speed optical microscopy. Phys Med Biol 2013;58:4513. Sheeran PS, Matsunaga TO, Dayton PA. Phase change events of volatile liquid perfluorocarbon contrast agents produce unique acoustic signatures. Phys Med Biol 2014;59:379. Sheeran PS, Matsuura N, Borden MA, Williams R, Matsunaga TO, Burns PN, Dayton PA. Methods of generating sub-micron phaseshift perfluorocarbon droplets for applications in medical ultrasonography. IEEE Trans Ultrason Ferroelectr Freq Control 2016b. in review. Sheeran PS, Rojas JD, Puett C, Hjelmquist J, Arena CB, Dayton PA. Contrast-enhanced ultrasound imaging and in vivo circulation kinetics with low boiling point nanoscale phase-change perfluorocarbon agents. Ultrasound Med Biol 2015;41:814–831. Sheeran PS, Wong VP, Luois S, McFarland RJ, Ross WD, Feingold S, Matsunaga TO, Dayton PA. Decafluorobutane as a phase-change contrast agent for low-energy extravascular ultrasonic imaging. Ultrasound Med Biol 2011b;37:1518–1530. Shpak O, Verweij M, Vos HJ, de Jong N, Lohse D, Versluis M. Acoustic droplet vaporization is initiated by superharmonic focusing. Proc Natl Acad Sci 2014;111:1697–1702. Tremblay-Darveau C, Williams R, Milot L, Bruce M, Burns PN. Visualizing the tumor microvasculature with a nonlinear plane-wave doppler imaging scheme based on amplitude modulation. IEEE Trans Med Imaging 2016;35:699–709. Vlaisavljevich E, Aydin O, Durmaz YY, Lin KW, Fowlkes B, ElSayed M, Xu Z. Effects of ultrasound frequency on nanodroplet-mediated histotripsy. Ultrasound Med Biol 2015;41: 2135–2147. Williams R, Wright C, Cherin E, Reznik N, Lee M, Gorelikov I, Foster FS, Matsuura N, Burns PN. Characterization of submicron phase-change perfluorocarbon droplets for extravascular ultrasound imaging of cancer. Ultrasound Med Biol 2013;39:475–489. Wilson SR, Burns PN. Microbubble-enhanced US in body imaging: What role? Radiology 2010;257:24–39.
http://dx.doi.org/10.1016/j.ultrasmedbio.2016.09.003
d
Technical Note MORE THAN BUBBLES: CREATING PHASE-SHIFT DROPLETS FROM COMMERCIALLY AVAILABLE ULTRASOUND CONTRAST AGENTS PAUL S. SHEERAN,*y KIMOON YOO,* ROSS WILLIAMS,* MELISSA YIN,* F. STUART FOSTER,*y and PETER N. BURNS*y * Physical Sciences Department, Sunnybrook Research Institute, Toronto, ON, Canada; and y Department of Medical Biophysics, University of Toronto, Toronto, Canada (Received 5 March 2016; revised 30 August 2016; in final form 6 September 2016)
Abstract—Phase-shift perfluorocarbon droplets have been investigated for over 20 years as pre-clinical ultrasound contrast agents with distinctive advantages in imaging and therapy. A number of formulation strategies exist, each with inherent advantages and limitations. In this note, we demonstrate a unique opportunity: that phase-shift droplets can be generated directly from commercially available microbubbles. This may facilitate pre-clinical and translational development by reducing the in-house synthesis expertise and resources required to generate high concentration droplet emulsions. Proof-of-principle in vitro and in vivo is given using droplets created from Definity and MicroMarker. The results demonstrate the role of perfluorocarbon choice in the trade-off between thermal stability and vaporization threshold, and suggest that commercial microbubbles with decafluorobutane cores may be ideal for this approach. (E-mail: [email protected]) Ó 2016 World Federation for Ultrasound in Medicine & Biology. Key Words: Perfluorocarbon, Ultrasound contrast agents, Phase-shift droplets, Acoustic droplet vaporization, Definity, MicroMarker.
cancerous tissues (Kripfgans et al. 2000; Rapoport 2012; Sheeran and Dayton 2012a). Ultrasonically activated phase-shift droplets have made slow progress into the clinic. Many fundamental mechanisms of droplet vaporization and expansion have only recently been revealed (Doinikov et al. 2014; Shpak et al. 2014), and much work remains in characterizing bioeffects, pharmacokinetics and dosing before use in humans can be considered. A variety of formulation strategies exist, including amalgamation/ homogenization/mechanical agitation (Kawabata et al. 2005; Kripfgans et al. 2000), sonication (Fabiilli et al. 2010; Matsuura et al. 2009; Rapoport et al. 2009), microfluidics (Bardin et al. 2011; Seo et al. 2010) and extrusion past porous membranes (Giesecke and Hynynen 2003; Sheeran et al. 2011b). More recently, Sheeran et al. (2011a, 2012b) developed a ‘‘microbubble condensation’’ technique to create phaseshift droplets in which volatile compounds are first encapsulated as microbubbles and are then condensed to the liquid state by decreasing the ambient temperature and/or increasing the ambient pressure. Researchers have demonstrated that microbubble condensation can be coupled with other particle generation techniques such
INTRODUCTION Clinical ultrasound contrast agents take the form of gas-filled microspheres, commonly referred to as microbubbles. In the past two decades, investigations into alternative contrast agents have increased substantially in order to address the fundamental limitations of microbubbles for specific diagnostic and therapeutic applications. Phase-shift perfluorocarbon droplets are perhaps the most extensively researched alternative and are designed to exist metastably in the liquid state under a desired degree of superheat until additional energy in the form of ultrasound or heat nucleates the core and the particles vaporize to the gas state, expanding volumetrically. This approach provides the benefits of increased circulation lifetime, relative ease of generating stable particles at the nanoscale, the ability to generate microbubbles ‘‘on-site’’ through externally applied ultrasound and the possibility of accessing the extravascular space of
Address correspondence to: Paul S. Sheeran, 2075 Bayview Ave, Room S640, Toronto, ON M4N 3M5, Canada. E-mail: pssheeran@ gmail.com 1
2
Ultrasound in Medicine and Biology
as microfluidic generation of microbubbles in order to refine droplet size distributions (Seo and Matsuura 2012; Seo et al. 2015). Each of these formulation strategies has advantages and disadvantages with regard to clinical applicability, resource requirements, and the necessary level of expertise in interfacial chemistry/colloid science (Sheeran et al. 2016b). In this note, we demonstrate that generating phase-shift droplets by microbubble condensation may provide a unique advantage in that it allows for the creation of droplets directly from existing commercial clinical and pre-clinical contrast agents, providing high yield emulsions with minimal resource requirements and without need for in-house bubble synthesis. Here, droplets are created directly from the clinical contrast agent Definity (Lantheus Medical Imaging, Billerica, MA, USA) after formation of bubbles by mechanical agitation, and the pre-clinical contrast agent MicroMarker (Bracco, Geneva, Switzerland and VisualSonics, Toronto, Canada), formed following reconstitution of a lyophilized emulsion. Proof-of-principle is given in vitro and in vivo on clinical and pre-clinical ultrasound machines. MATERIALS AND METHODS Droplet emulsion preparation Phase-shift droplets were generated from two commercially available phospholipid-encapsulated microbubble contrast agents. Definity, the most commonly used clinical ultrasound contrast agent in North America, has a gaseous octafluoropropane core (boiling point 236.7 C), while MicroMarker, a pre-clinical contrast agent designed for small-animal imaging, has a core composed of nitrogen and decafluorobutane (boiling point 22 C). Both agents have a distribution peak diameter below 1 mm and a volume-weighted peak diameter below 5 mm (Helfield et al. 2012; Raymond et al. 2014). Definity microbubbles were formed by activating vials (stored at 4 C) by mechanical agitation according to standard protocols. Octafluoropropane droplets were generated from this suspension with a similar approach to that described in Sheeran et al. (2012b). First, the vial was cooled to 210 C in an isopropanol bath maintained at temperature with dry ice, followed by pressurization of the headspace with room air to approximately 170 kPa above atmospheric pressure over the course of 2 min. MicroMarker microbubbles were prepared by resuspending the lyophilized agent with 0.5 mL of decafluorobutane gas-equilibrated saline. Decafluorobutane droplets were generated by cooling the resulting suspension to 28 C in the isopropanol bath and pressurizing the headspace with room air to approximately 100 kPa above atmospheric pressure over the course of 2 min in order to
Volume -, Number -, 2016
condense the decafluorobutane component of the core and expel the nitrogen component (condensation of a pure decafluorobutane core can be accomplished by cooling alone). The droplet suspensions were then centrifuged at 1000 rpm for 2 min at 4 C to separate droplets from microbubbles remaining in the solution, and were held in a centrifuge tube on ice for up to 1 h before testing. Particle sizing Example distributions of droplets generated by condensing Definity and MicroMarker microbubbles were measured using a Nanosight LM10 (Malvern Instruments, Inc., Malvern, UK). The droplet suspensions were diluted 1/200 and 1/100 in room-temperature saline for octafluoropropane droplets (generated from Definity) and decafluorobutane droplets (generated from MicroMarker), respectively. Five 30-s videos of nonoverlapping volumes were collected for each dilution, analyzed and averaged to generate representative distributions. This process was repeated three times by preparing new dilutions from the same vial, and the results averaged for a final representation of the vial size distribution and particle concentration. In vitro experimental design An in vitro vessel flow phantom was constructed by connecting a 6.4-mm diameter (0.24 mm wall thickness) latex tube (Penrose Drain; Medline Industries, Inc., Mundelein, IL, USA) to polyvinyl chloride tubing secured by a custom acrylic holder that was submerged in a large heated water bath. The flow circuit consisted of a 1.5 L sample container held 30–40 cm above the large heated water bath. Temperatures between 25 C and 42 C were maintained in the large water bath by an immersion circulator/heater, and in the upper sample container using a small fish tank heater with an adjustable voltage regulator. Droplet dilutions were imaged and vaporized in the vessel phantom by fixing an ultrasound imaging probe to a mechanical arm and submerging the probe to image the latex tube lengthwise in a set position. Droplets created from the clinical microbubble Definity were tested using a Philips iU22 clinical scanner with an L9-3 linear array probe (Philips Ultrasound, Bothell, WA, USA). Standard B-mode imaging settings were used to capture videos of the samples (mechanical index [MI] 5 0.01, 68 fps). The imaging focus was placed in the center of the tube at 2.2 cm depth. Sample flow in the circuit was stopped and droplets were vaporized by manually triggering a single flash frame—conventionally used to destroy contrast agents for the purposes of volume-flow and molecular imaging (Hudson et al. 2015; Wilson and Burns 2010)— over a range of preset nominal mechanical indices between MI 5 0.05 and MI 5 0.97 (the highest output
Creating Phase-Shift Droplets from Commercial Microbubbles d P. S. SHEERAN et al.
available). Between vaporization captures, an untested sample volume was brought into the imaging plane by restoring forward flow. Output pressures were determined by measuring the center frequency of the flash by membrane hydrophone (Model 804, Sonora Medical Systems, Longmont, CO, USA) and correcting the nominal (derated) value, which assumes an attenuation coefficient of 0.3 dB/(cm $ MHz) set by the US Food and Drug Administration. Here, a calculated free-field peak negative pressure was obtained by negating this attenuation over 2.2 cm at 4.7 MHz, the measured flash frequency. The peak negative pressure inside the tube was estimated by correcting the free-field value to account for attenuation over 0.24 mm of latex according to the frequency dependence characterized by Rickey et al. (1995). Losses due to reflection were neglected due to the similar speed of sound and density between water and latex. Droplets created from the pre-clinical microbubble MicroMarker (designed for small animal imaging) were tested using a VisualSonics Vevo2100 system with an MS400 (30 MHz) imaging probe (VisualSonics) in standard B-mode acquisition. The imaging focus was placed in the center of the tube at 9 mm depth. Due to lack of a manually triggered flash, sample flow in the circuit was stopped and videos of droplets in the imaging plane were captured as a function of increasing B-mode output power between the lowest and highest outputs available (Sheeran et al. 2016a). Here, as above, peak negative pressure inside the tube was estimated by correcting the measured free-field value to account for attenuation over the latex tube wall, with reflection losses neglected. At each output power, 50 frames were collected at 20 fps. Controls were performed by collecting additional videos under the same conditions with only water (i.e., no contrast agents). Dilutions were determined based on preliminary tests with similar in-house droplet formulations, and set to be within approximately one order of magnitude of the concentration used for microbubble characterization in vitro with similar vessel models (Tremblay-Darveau et al. 2016). In each case, a 120-mL bolus of the droplet suspension being tested was drawn from the cooled vial and mixed into 1 L of gas equilibrated deionized water in the sample container heated to the desired temperature. Forward flow was restored until the sample was present in the tube, after which flow was stopped and testing begun immediately. As an exception, a 60-mL bolus of the octafluoropropane droplets was added to the 1 L of deionized water at temperature below 32 C, where preliminary results suggested increased stability and very high contrast production compared to temperatures 32 C and above. Capturing the data over the entire test range per sample typically took 45 s to 1 min. This process was repeated
3
for the same sample an additional two times in the following minutes to characterize in vitro decay of that sample over time. This process was repeated for three to four vials of independently prepared samples of both droplet types. Video analysis Video captures obtained on the Philips iU22 were analyzed offline in QLab quantification software (Philips Ultrasound). Manually drawn regions of interest were placed inside the tube in the region of activation and the change in mean pixel dB after the flash frame was calculated. These results were averaged for each temperature for the three to four vials tested. Decay rates were calculated by fitting an exponential decay curve to the linear intensity values obtained over the three runs for each sample at the two highest flash pressures. These decay rates were averaged for the three to four vials tested at each temperature. Videos obtained from the Vevo2100 were analyzed offline in MATLAB (The MathWorks, Inc., Natick, MA, USA) by first converting the pixel values to linearized intensity (Williams et al. 2013). Manually drawn regions of interest were placed inside the test region and mean pixel values were calculated. These values were converted to dB and adjusted by subtracting the background dB level obtained from the water-only control samples at the same power level. The final values obtained for each sample were averaged for all vials at each temperature. During tests, it was noted that droplets did not decay during the test period, and so fitting with exponential decay was not performed in this case. In vivo vaporization demonstrations All in vivo experiments were approved by the animal care committee of the Sunnybrook Research Institute. Octafluoropropane droplets (generated from Definity) were investigated in the kidney of a New Zealand white rabbit using the Philips iU22 scanner and L9-3 probe attached to a fixed mechanical arm. The rabbit was inducted with dexmedetomidine and maintained under anesthesia with isoflurane according to protocols approved at Sunnybrook Research Institute. The kidney was imaged at 2.5 cm of depth using a side-by-side contrast and B-mode scan, with the focus placed at 2.25 cm. The MI was set to 0.15 for contrast mode and 0.04 for B-mode at a frame rate of 13 Hz. A 200-mL bolus of octafluoropropane droplets was added to 1 mL saline and manually injected over 20 s via ear-vein catheterization, followed by a 2.5-mL saline flush, which falls within previously published doses for microbubbles in this animal model (McDannold et al. 2006; Tremblay-Darveau et al. 2016). Wash-in of the bolus (bubbles produced by spontaneous vaporization) was monitored and a manually
4
Ultrasound in Medicine and Biology
triggered flash frame at an MI of 1.31 was triggered at the peak bolus enhancement. The resulting video was analyzed offline in QLab. Decafluorobutane droplets (generated from MicroMarker) were investigated in the liver of a C3H mouse using the Vevo2100 system and MS400 probe. The mouse was anesthetized by isoflurane according to protocols approved at Sunnybrook Research Institute. The center of the liver was positioned at the transducer’s elevational focus, 9 mm from the transducer face. Baseline scans were taken before agent injection in B-mode (20 fps) at 1% and 100% power using a two-focus imaging setting with foci at 8 mm and 9 mm for a larger region of vaporization. Following a previously developed protocol, a 100-mL bolus of undiluted decafluorobutane droplets was administered by syringe pump over 1 min, followed by a saline flush at the same volume and rate
Volume -, Number -, 2016
(Sheeran et al. 2016a). Five min after injection, scans were taken again at 1% and 100% power. RESULTS AND DISCUSSION Droplet generation and sizing Microbubbles were prepared by either mechanical agitation or reconstitution (Fig. 1). After condensation, the appearance of the emulsion changed from a milky solution to a more translucent appearance (Fig. 1). In the case of droplets generated from Definity, a clear bubble layer persisted at the top of the emulsion after condensation and centrifugation while the droplets generated from MicroMarker appeared uniform throughout—highlighting the reduced stability of octafluoropropane droplets compared to decafluorobutane droplets. In both cases, condensation produced primarily sub-micron
Fig. 1. Commercial microbubble agents Definity and MicroMarker shown as packaged in vial. After mechanical agitation (Definity) or reconstitution (MicroMarker), a microbubble condensation technique can be used to convert microbubbles to sub-micron phase-shift droplets.
Creating Phase-Shift Droplets from Commercial Microbubbles d P. S. SHEERAN et al.
droplet sizes. The example sizing of one vial of octafluoropropane droplets on the NanoSight produced a peak diameter, 10th percentile diameter (90% of the distribution above this size) and 90th percentile diameter (90% of the distribution below this size) of 216 nm 6 6 nm, 191 nm 6 8 nm and 476 nm 6 65 nm, respectively, and a concentration of 2.90 3 1011 particles/mL 6 0.93 3 1011 particles/mL. The decafluorobutane droplet sample produced a peak diameter, 10th percentile and 90th percentile diameters of 264 nm 6 12 nm, 230 nm 6 6 nm and 626 nm 6 43 nm, respectively, and a concentration of 1.08 3 1011 particles/mL 6 0.12 3 1011 particles/mL. In vitro results: octafluoropropane droplets Octafluoropropane droplet samples diluted to between 0.006% and 0.012% were exposed to manually triggered flash frames on the Philips iU22 to test droplet vaporization as a function of both peak negative pressure and ambient temperature. At 25 C, the droplet samples provided no significant ultrasound contrast within the tube before the flash frame (Fig. 2a). Once exposed to single flash frames with nominal mechanical indices above 0.5, the sample volume interrogated produced a high level of B-mode contrast, which decayed steadily over the course of 20–30 s. As temperature increased, the ‘‘preflash’’ signal magnitude increased in a generally linear fashion from a mean value of 7.6 dB at 25 C to 15.6 dB at 37 C, indicating a greater level of spontaneous vaporization with increasing temperature. At 32 C and 35 C, though bubbles were clearly present in the pre-flash segment, vaporization could still be observed as a marked increase in brightness after the flash frame. However, at
5
37 C, only a decrease in brightness in the region of interest was observed due to microbubble destruction from the flash frame (Fig. 2a). The change in mean pixel dB as a function of flash pressure and temperature produced a clear picture of regimes in which the flash frame can increase contrast (‘‘flash activation’’ of droplets) or destroy contrast (‘‘flash destruction’’ of microbubbles present by spontaneous vaporization) (Fig. 3a). At 25 C, the flash produced as much as 12 dB of enhancement at peak negative pressures above 1.5 MPa. At 32 C and 35 C, bubbles are clearly present in-plane before the flash frame, and low flash pressures tended to destroy rather than create contrast. Positive contrast was observed at pressures above 0.9 MPa and 0.5 MPa for 32 C and 35 C, respectively. The magnitude of the contrast enhancement decreased with increasing temperature as a result of a greater background signal before the flash. At 37 C, the pre-flash signal appeared to be dominated by microbubbles (Fig. 2a), and the flash frame only produced microbubble destruction, resulting in a decrease in measured signal strength. In vitro results: decafluorobutane droplets Decafluorobutane droplet samples diluted to 0.012% were exposed to increasing output power on the Vevo 2100 to test droplet vaporization as a function of both peak negative pressure and ambient temperature. At 25 C, droplet samples provided no significant contrast within the tube at any output power (Fig. 2b). At temperatures higher than 35 C, droplet vaporization produced individually distinguishable microbubbles that translated axially due to acoustic streaming and radiation force over
Fig. 2. Ultrasound imaging and activation of droplet suspensions in a latex tube. (a) Imaged with a Philips iU22 and L9-3 probe, octafluoropropane droplets generated from Definity show high stability at 25 C and can be vaporized with the manual flash to form contrast-providing microbubbles. At 37 C, however, droplet show a high level of spontaneous vaporization before the flash, and as a result the flash frame diminishes enhancement. (b) Imaged with a VisualSonics Vevo2100 and MS-400 probe, decafluorobutane droplets generated from MicroMarker show high stability at all temperatures, and could not be vaporized at 25 C at the maximum output of the probe. At higher temperatures, vaporization produced echogenic microbubbles within the tube.
6
Ultrasound in Medicine and Biology
Volume -, Number -, 2016
a
b
Fig. 3. Contrast enhancement of droplets as a function of temperature and estimated peak negative pressure inside the latex tube. Markers represent average value across samples, while bars represent the range of the individual measurements at each pressure. (a) Octafluoropropane droplets were vaporized at 4.7 MHz using a manually triggered flash frame on the iU22, with contrast enhancement calculated relative to the signal just before the flash. For clarity, the 25 C and 35 C series have been shifted left and right, respectively, by 20 kPa. (b) Decafluorobutane droplets were vaporized at 30 MHz by increasing the imaging output power on the Vevo2100 and the contrast enhancement calculated relative to agent-free controls. For clarity, the 42 C and 37 C series have been shifted left and right, respectively, by 25 kPa.
the 50 frames. No significant change was observed in the signal for the lowest output powers at each temperature, indicating the decafluorobutane droplets were stable against spontaneous vaporization over the entire temperature range—consistent with other reports at these temperatures (Mountford et al. 2015b; Sheeran et al. 2012b, 2016a).
Calculating the increase in mean pixel dB over the no-agent controls illustrates the relationship of temperature and pressure to vaporization of decafluorobutane droplets (Fig. 3b). At temperatures of 35 C and higher, vaporization produced maximum B-mode contrast on the order of 10–12 dB within the tube. The peak negative pressure required to vaporize droplets at 37 C occurred
Creating Phase-Shift Droplets from Commercial Microbubbles d P. S. SHEERAN et al.
Fig. 4. In vivo demonstrations of droplet performance. (a) Octafluoropropane droplets imaged with the iU22 showed substantial spontaneous vaporization upon intravenous injection and imaging in the rabbit kidney (contrast mode shown with kidney region manually outlined), but the manually triggered flash produced a modest increase in contrast as a result of droplet vaporization, calculated in the plot below. (b) Vaporization of decafluorobutane droplets in the mouse liver with the Vevo2100 produced a bright band of enhancement above the focal depth compared to the same region before injection.
7
8
Ultrasound in Medicine and Biology
near 3 MPa, which is in agreement with a recent study testing similar in-house preparations (Sheeran et al. 2016a). In vivo results Upon intravenous injection of octafluoropropane droplets (generated from Definity), a bolus wash-in of microbubbles was observed in the kidney as a result of spontaneous vaporization. Unlike the in vitro results at physiologic temperatures, the manually triggered flash at peak wash-in produced an increase in contrast on the order of 5 dB within the kidney region, which dissipated over the course of several seconds (Fig. 4a). The difference between in vitro performance (Fig. 2a, high initial microbubble signal diminished by flash) and in vivo performance (enhancement over microbubble signal upon flash) is likely a result of clearance of microbubbles from circulation and filtration by the lungs en route to the kidney, allowing the un-vaporized droplets to be activated and visualized. This is consistent with the results for inhouse preparations of octafluoropropane droplets, which show that some portion of the original population remains unvaporized at body temperature within the test period (Mountford et al. 2015b; Sheeran et al. 2012b, 2014, 2015). In this study, attempts to produce additional vaporization in the minutes following administration were less successful, indicating a relatively short circulation profile consistent with previous reports (Porter et al. 2016; Sheeran et al. 2015). For decafluorobutane droplets, imaging was performed 5 min after injection of the bolus, similar to the protocol in Sheeran et al. (2016a). The liver was first imaged in B-mode at 1% power, and no droplet vaporization was observed. Immediately following this, output power was increased to 100%, producing a clear, bright band within the liver superficial to the depth of focus as a result of droplet vaporization (Fig. 4b). Enhancement at later time points was not tested. In vitro results: temperature dependence and instability We expect droplets to decay due to a combination of droplet dissolution and spontaneous vaporization, which both depend on ambient temperature. The observed spontaneous vaporization as a function of temperature in this study matches closely with that recently reported by Mountford et al. (2015b) for octafluoropropane droplets, with general stability against spontaneous vaporization at temperatures below 30 C and a high degree of spontaneous vaporization at temperatures near 37 C. The similarity in the stability properties between this study (observed acoustically) and that of Mountford et al. (measured optically) despite differences in size and encapsulation supports the notion that the spontaneous vaporization of encapsulated perfluorocarbon droplets
Volume -, Number -, 2016
primarily depends on the nearness of the compound’s critical temperature to the ambient temperature (Eberhart 1976). At 25 C, the octafluoropropane droplets generated from Definity produced an exponential decay rate of 0.05 min21 6 0.06 min21 (half-life of approximately 14.7 min). Upon exposure to higher temperatures, the decay rate increased more than an order of magnitude to 1.1 min21 6 0.13 min21 (half-life of approximately 0.6 min) at 35 C. In contrast, the decafluorobutane droplets generated from MicroMarker did not produce a discernible decay over the test period. Future testing of this sort may be useful in order to experimentally derive the dissolution rate of each droplet core compound as a function of temperature, encapsulation and concentration, which has not been previously characterized over a temperature range relevant to medical ultrasound. Limitations and future directions As a proof-of-principle, this study gave only preliminary evidence of droplet thresholds and size, which may be refined in future investigations. In order to demonstrate activation of the two droplet types, which had substantially different thermal stability and activation thresholds, two different commercial imaging systems were required. The system-dependent imaging properties and the wide separation in center frequency used to assess activation made it impossible to directly compare some details, such as relative contrast enhancement between the two droplet types and the pressures required for conversion—for which frequency dependence is a complex phenomenon (Kripfgans et al. 2000; Sheeran et al. 2016a; Shpak et al. 2014; Vlaisavljevich et al. 2015; Williams et al. 2013). Regardless, some general conclusions can be drawn. Sub-micron droplets with decafluorobutane cores were not able to be activated at the peak output of the L9-3 probe at 37 C (nominal peak MI of 0.97 in B-mode), although they could be activated at the frequencies available on small-animal imaging systems (Sheeran et al. 2016a). Conversely, droplets with octafluoropropane cores could be vaporized using the clinical system, but exhibited poor thermal stability at the ideal temperatures. The low vaporization threshold of these droplets prevented characterization on the small-animal imaging system, as the lowest imaging pressures available exceeded the vaporization threshold at temperatures of 32 C or higher, leaving no option for imaging without activation. The fact that the more stable decafluorobutane droplets could not be activated on the iU22 with the L9-3 probe should not be interpreted to mean that they require activation pressures beyond the diagnostic limit. In fact, most experimental characterizations of these droplets indicate activation at 5–8 MHz is possible with
Creating Phase-Shift Droplets from Commercial Microbubbles d P. S. SHEERAN et al.
mechanical indices on the order of 1.3 (Sheeran et al. 2013, 2015). However, this does not mean a clinical machine and transducer will necessarily be able to produce these absolute pressures/MIs within tissue. For example, most soft tissue imaging settings assume a lower than normal tissue attenuation of 0.3 dB/(cm $ MHz), as opposed to values near 0.5 dB/(cm $ MHz) or higher (Goss et al. 1979) in order to ensure the actual pressures are not higher than expected. So a nominal MI of 1.3 at 5 MHz in 2 cm of tissue may produce an actual MI on the order of 1.0 or less, below the activation threshold. Here, the estimated peak negative pressure inside the tube reached a maximum of 2.67 MPa, or MI 5 1.23—just below where activation likely begins for the decafluorobutane droplets at this frequency. Future studies are necessary to identify clinical systems with ideal vaporization frequencies and output pressures to efficiently convert sub-micron droplets in vivo. Thus, for use with low clinical frequencies (1–5 MHz) within the regulatory output limits, this study highlights that there is motivation for development of phase-shift droplets with properties in between the octafluoropropane and decafluorobutane droplets tested here—ideally with reduced vaporization thresholds while maintaining high thermal stability. This might be accomplished by choosing an alternative compound that has similar properties (hydrophobic, inert, non-toxic, low solubility in aqueous media) but with a more ideal boiling point/critical temperature, by incorporating nucleation seeds to reduce activation thresholds (Lee et al. 2015; Matsuura et al. 2009), or through mixtures (Kawabata et al. 2005; Sheeran et al. 2012b). However, recent studies suggest that mixtures of such highly superheated compounds will not behave ideally in vivo (Mountford et al. 2015a). It may also be possible to reduce the vaporization pressure simply by choosing a decafluorobutane microbubble with a larger mean diameter than MicroMarker, as the vaporization threshold is inversely related to droplet diameter (Kripfgans et al. 2004; Sheeran et al. 2012b). This study provides proof-of-principle that one can develop phase-shift droplets from existing commercial agents through simple adjustment of ambient conditions, converting them into a new type of agent with advantages compared to microbubbles for specific ultrasoundmediated applications. The ability to produce phaseshift droplets without the necessity of in-house bubble or droplet synthesis expertise may facilitate preclinical development and exploration of new applications. That these droplets can be generated from precursors already approved for the clinic and with existing good manufacturing protocols may also provide new opportunities for commercialization and translation, although assessment of pharmacokinetics and safety would be
9
necessary. The two different commercial microbubbles were selected in order to demonstrate the applicability of the condensation technique for different perfluorocarbon cores as well as differences in microbubble preparation (mechanical agitation vs. reconstitution). Although MicroMarker is intended for pre-clinical use only, it is similar in encapsulation, size distribution, concentration and method of reconstitution to currently available clinical microbubbles. Thus, it can be reasonably expected that the condensation technique can be applied to those formulations as well. Other perfluorocarbon-based commercial microbubbles that may potentially be converted into droplets include Optison (GE Healthcare, Milwaukee, WI, USA), Sonazoid (GE Healthcare/Daiichi Sankyo, Tokyo, Japan), BR38 and the VEGFR2-targeted agent BR55 (Bracco). In particular, we propose that those with decafluorobutane cores, such as Sonazoid, may provide the most promise moving forward. CONCLUSIONS In this study, we have proposed that ultrasonically vaporized phase-shift droplets can be generated directly from commercially available microbubble contrast agents by utilizing a microbubble condensation technique. This presents a unique strategy to create new imaging and therapy agents from existing ones. The results demonstrate the role of perfluorocarbon choice in the trade-off between thermal stability and vaporization threshold, and suggest that commercial microbubbles with decafluorobutane cores may be ideal for this approach. Acknowledgments—This work has been supported by a grant from the Canadian Institutes of Health Research (MOP119346 and MOP125994) and the Ontario Institute for Cancer Research. P .S. appreciates the generous support of the Banting Postdoctoral Fellowship, administered by the Canadian Institutes of Health Research. We thank Megan Thompson and Linda Nghiem for assistance with animal studies, Charles Tremblay-Darveau and Zachary Zajac for assistance with animal experiments and data processing and Chelsea Munding for technical discussions.
REFERENCES Bardin D, Martz TD, Sheeran PS, Shih R, Dayton PA, Lee AP. Highspeed, clinical-scale microfluidic generation of stable phase-change droplets for gas embolotherapy. Lab Chip 2011;11: 3990–3998. Doinikov AA, Sheeran PS, Bouakaz A, Dayton PA. Vaporization dynamics of volatile perfluorocarbon droplets: A theoretical model and in vitro validation. Med Phys 2014;41:102901. Eberhart JG. The thermodynamic and the kinetic limits of superheat of a liquid. J Colloid Interface Sci 1976;56:262–269. Fabiilli M, Lee J, Kripfgans O, Carson P, Fowlkes J. Delivery of watersoluble drugs using acoustically triggered perfluorocarbon double emulsions. Pharm Res 2010;27:2753–2765. Giesecke T, Hynynen K. Ultrasound-mediated cavitation thresholds of liquid perfluorocarbon droplets in vitro. Ultrasound Med Biol 2003;29:1359–1365.
10
Ultrasound in Medicine and Biology
Goss SA, Frizzell LA, Dunn F. Ultrasonic absorption and attenuation in mammalian tissues. Ultrasound Med Biol 1979;5:181–186. Helfield BL, Huo X, Williams R, Goertz DE. The effect of preactivation vial temperature on the acoustic properties of Definity. Ultrasound Med Biol 2012;38:1298–1305. Hudson JM, Williams R, Tremblay-Darveau C, Sheeran PS, Milot L, Bjarnason GA, Burns PN. Dynamic contrast enhanced ultrasound for therapy monitoring. Eur J Radiol 2015;84:1650–1657. Kawabata K, Sugita N, Yoshikawa H, Azuma T, Umemura S. Nanoparticles with multiple perfluorocarbons for controllable ultrasonically induced phase shifting. Jpn J Appl Phys 2005;44:5. Kripfgans OD, Fabiilli ML, Carson PL, Fowlkes JB. On the acoustic vaporization of micrometer-sized droplets. J Acoust Soc Am 2004; 116:272–281. Kripfgans OD, Fowlkes JB, Miller DL, Eldevik OP, Carson PL. Acoustic droplet vaporization for therapeutic and diagnostic applications. Ultrasound Med Biol 2000;26:1177–1189. Lee JY, Carugo D, Crake C, Owen J, De Saint Victor M, Seth A, Coussios C, Stride E. Nanoparticle-loaded protein–polymer nanodroplets for improved stability and conversion efficiency in ultrasound imaging and drug delivery. Adv Mater 2015;27: 5484–5492. Matsuura N, Williams R, Gorelikov I, Chaudhuri J, Rowlands J, Hynynen K, Foster S, Burns P, Resnik N. Nanoparticle-loaded perfluorocarbon droplets for imaging and therapy. In: IEEE International Ultrasonics Symposium (IUS) 2009:5–8. McDannold N, Vykhodtseva N, Hynynen K. Targeted disruption of the blood–brain barrier with focused ultrasound: Association with cavitation activity. Phys Med Biol 2006;51:793. Mountford PA, Smith WS, Borden MA. Fluorocarbon nanodrops as acoustic temperature probes. Langmuir 2015a;31:10656–10663. Mountford PA, Thomas AN, Borden MA. Thermal activation of superheated lipid-coated perfluorocarbon drops. Langmuir 2015b;31: 4627–4634. Porter TR, Arena C, Sayyed S, Lof J, High RR, Xie F, Dayton PA. Targeted transthoracic acoustic activation of systemically administered nanodroplets to detect myocardial perfusion abnormalities. Circ Cardiovasc Imaging 2016;9:e003770. Rapoport N. Phase-shift, stimuli-responsive perfluorocarbon nanodroplets for drug delivery to cancer. Wiley Interdiscip Rev Nanomed Nanobiotechnol 2012;4:492–510. Rapoport NY, Kennedy AM, Shea JE, Scaife CL, Nam KH. Controlled and targeted tumor chemotherapy by ultrasound-activated nanoemulsions/microbubbles. J Control Release 2009;138:268–276. Raymond JL, Haworth KJ, Bader KB, Radhakrishnan K, Griffin JK, Huang SL, McPherson DD, Holland CK. Broadband attenuation measurements of phospholipid-shelled ultrasound contrast agents. Ultrasound Med Biol 2014;40:410–421. Rickey DW, Picot PA, Christopher DA, Fenster A. A wall-less vessel phantom for Doppler ultrasound studies. Ultrasound Med Biol 1995;21:1163–1176. Seo M, Gorelikov I, Williams R, Matsuura N. Microfluidic assembly of monodisperse, nanoparticle-incorporated perfluorocarbon microbubbles for medical imaging and therapy. Langmuir 2010;26: 13855–13860.
Volume -, Number -, 2016 Seo M, Matsuura N. Monodisperse, Submicrometer droplets via condensation of microfluidic-generated gas bubbles. Small 2012;8: 2704–2714. Seo M, Williams R, Matsuura N. Size reduction of cosolvent-infused microbubbles to form acoustically responsive monodisperse perfluorocarbon nanodroplets. Lab Chip 2015;15:3581–3590. Sheeran PS, Daghighi Y, Yoo K, Williams R, Cherin E, Foster FS, Burns PN. Image-guided ultrasound characterization of volatile sub-micron phase-shift droplets in the 20-40 MHz frequency range. Ultrasound Med Biol 2016a;42:795–807. Sheeran PS, Dayton PA. Phase-change contrast agents for imaging and therapy. Curr Pharm Des 2012a;18:2152–2165. Sheeran PS, Luois S, Dayton PA, Matsunaga TO. Formulation and acoustic studies of a new phase-shift agent for diagnostic and therapeutic ultrasound. Langmuir 2011a;27:10412–10420. Sheeran PS, Luois SH, Mullin LB, Matsunaga TO, Dayton PA. Design of ultrasonically-activatable nanoparticles using low boiling point perfluorocarbons. Biomaterials 2012b;33:3262–3269. Sheeran PS, Matsunaga TO, Dayton PA. Phase-transition thresholds and vaporization phenomena for ultrasound phase-change nanoemulsions assessed via high-speed optical microscopy. Phys Med Biol 2013;58:4513. Sheeran PS, Matsunaga TO, Dayton PA. Phase change events of volatile liquid perfluorocarbon contrast agents produce unique acoustic signatures. Phys Med Biol 2014;59:379. Sheeran PS, Matsuura N, Borden MA, Williams R, Matsunaga TO, Burns PN, Dayton PA. Methods of generating sub-micron phaseshift perfluorocarbon droplets for applications in medical ultrasonography. IEEE Trans Ultrason Ferroelectr Freq Control 2016b. in review. Sheeran PS, Rojas JD, Puett C, Hjelmquist J, Arena CB, Dayton PA. Contrast-enhanced ultrasound imaging and in vivo circulation kinetics with low boiling point nanoscale phase-change perfluorocarbon agents. Ultrasound Med Biol 2015;41:814–831. Sheeran PS, Wong VP, Luois S, McFarland RJ, Ross WD, Feingold S, Matsunaga TO, Dayton PA. Decafluorobutane as a phase-change contrast agent for low-energy extravascular ultrasonic imaging. Ultrasound Med Biol 2011b;37:1518–1530. Shpak O, Verweij M, Vos HJ, de Jong N, Lohse D, Versluis M. Acoustic droplet vaporization is initiated by superharmonic focusing. Proc Natl Acad Sci 2014;111:1697–1702. Tremblay-Darveau C, Williams R, Milot L, Bruce M, Burns PN. Visualizing the tumor microvasculature with a nonlinear plane-wave doppler imaging scheme based on amplitude modulation. IEEE Trans Med Imaging 2016;35:699–709. Vlaisavljevich E, Aydin O, Durmaz YY, Lin KW, Fowlkes B, ElSayed M, Xu Z. Effects of ultrasound frequency on nanodroplet-mediated histotripsy. Ultrasound Med Biol 2015;41: 2135–2147. Williams R, Wright C, Cherin E, Reznik N, Lee M, Gorelikov I, Foster FS, Matsuura N, Burns PN. Characterization of submicron phase-change perfluorocarbon droplets for extravascular ultrasound imaging of cancer. Ultrasound Med Biol 2013;39:475–489. Wilson SR, Burns PN. Microbubble-enhanced US in body imaging: What role? Radiology 2010;257:24–39.