- Email: [email protected]
Contrast superharmonic imaging: A feasibility study
Ultrasound in Med. & Biol., Vol. 29, No. 4, pp. 547–553, 2003 Copyright © 2003 World Federation for Ultrasound in Medicine & Biology Printed in the USA. All rights reserved 0301-5629/03/$–see front matter
doi:10.1016/S0301-5629(03)00012-7
● Original Contribution CONTRAST SUPERHARMONIC IMAGING: A FEASIBILITY STUDY AYACHE BOUAKAZ,*‡ BOUDEWIJN J. KRENNING,*† WIM B. VLETTER,† FOLKERT J. TEN CATE*† and NICO DE JONG*‡ *Experimental Echocardiography and †Cardiology Department, Erasmus MC, Thoraxcenter, Rotterdam, The Netherlands; and ‡Interuniversity Cardiology Institute of the Netherlands (ICIN), Utrecht, The Netherlands (Received 31 July 2002; in final form 23 December 2002)
Abstract—Harmonic imaging provided significant improvement in image quality by taking advantage of the scattered second harmonic (2H) component from contrast bubbles. However, differentiation between contrast and tissue (usually termed contrast-to-tissue ratio, CTR) is sometimes cumbersome and this is mainly due to tissue contamination. We have previously demonstrated, using simulations and in vitro measurements, that CTR increases as a function of the order of the harmonic number. A new contrast imaging method based on the detection of the higher harmonics was developed and termed superharmonic (SH). This technique has been shown to be more sensitive to contrast by increasing the signal from contrast and suppressing that from tissue (high CTR). The purpose of this study was to determine the clinical feasibility and usefulness of SH in patients using a commercially available contrast agent (SonoVue威) for quantification of myocardial perfusion. A total of 10 patients with various cardiac diseases were assessed. Apical four-chamber views were acquired using SH in triggered mode before and after contrast injection. The superharmonic was performed with a newly developed probe transmitting at 0.8 MHz with a mechanical index of 0.2. Myocardial perfusion was determined visually and analyzed quantitatively using radiofrequency (RF) processing from different regions of interest. The results showed that, before contrast injection, SH was totally blinded to tissue and no superharmonic components were generated in the image view. After administration of SonoVue威, myocardial opacification was visualized by SH after contrast entered the myocardium. An increase of more than 15 dB in the myocardial bubbles echo compared to tissue echo was measured. In addition, the technique was used to visualize myocardial perfusion after myocardial septal ablation for hypertrophic cardiomyopathy. The clinical results showed the ability of contrast SH imaging in differentiating low and normal perfusion areas, demonstrating the high sensitivity and specificity of the technique. (E-mail: [email protected]) © 2003 World Federation for Ultrasound in Medicine & Biology. Key Words: Superharmonic, Harmonic, Contrast-to-tissue ratio, Ultrasound contrast, Patients, Perfusion.
by the presence of additional materials at the gas-liquid interface. In addition, by substituting different types of perfluorocarbon gases for air, the stability and plasma longevity of the agents have been markedly improved, usually lasting more than 5 min (Grayburn 2002; Kabalnov et al. 1998). With the emergence of new contrast bubble generations, new US technologies have been developed to improve microbubble detection. Improving both the intensity and longevity of the detection procedure was the consequence of a better understanding of the microbubble-US wave interaction and technological advances. The introduction of contrast harmonic imaging represents a major step in US contrast imaging (Burns et al. 1992; de Jong et al. 1994a, 1994b). This technology is now available in all medical scanners and has been introduced into clinical echocardiographic practice (Burns 2002). Harmonic imaging enhances the detection
INTRODUCTION Ultrasound (US) contrast agents (UCA) have been shown to improve the image quality of sonography by increasing the backscattered echoes from the desired regions (Burns et al. 1992; Goldberg et al. 1994, 2001). To extend the utility of US imaging, there has been active research into the use of gas microbubble contrast agents for diagnostic use. Over the past decade, research has been actively focused on developing efficacious US contrast agents and new contrast-specific imaging techniques. UCA consist of small free or encapsulated microbubbles of gas. A shell enveloping the gas is used to stabilize the bubble against dissolution and coalescence Address correspondence to: Ayache Bouakaz, Ph.D., Erasmus MC, Exp Echo, Ee2302, P.O. Box 1738. Rotterdam 3000 DR The Netherlands. E-mail: [email protected] 547
548
Ultrasound in Medicine and Biology
of UCA within the cardiac chambers and is readily used in patients undergoing echocardiography. It exploits the differences between the response of gas microbubbles and tissue to US irradiation (de Jong et al. 1994b). Soft tissues are known to be linear reflectors, whereas contrast bubbles exhibit a different behavior when interacting with US waves (Averkiou 2000). The response of a gas bubble to a US interrogation is asymmetrical, giving rise to nonlinear echoes that contain new acoustic information. This nonlinear bubble oscillation signifies a response containing multiples or harmonics of the scanning frequency. Differentiation between perfused and nonperfused tissue is now mainly based on these nonlinear characteristics. This differentiation is usually expressed by the ratio of the scattered power from the contrast agent to the scattered power from the tissue and is termed “contrast-to-tissue ratio” (CTR). The response of UCA at the second harmonic frequency is currently utilized in clinical echocardiographic imaging to enhance the contrast relative to the surrounding body tissue and, hence, the CTR. Second harmonic detection-based techniques showed significant improvements compared with conventional fundamental imaging modes. However, this technology has certain limitations. The major drawback is that it is not sensitive enough in discriminating between UCA and tissue because harmonic frequencies are also generated during the propagation of the wave through the tissue and are, therefore, reflected by the tissue itself (Frinking et al. 2000; Hamilton and Blackstock 1998). Initially, it was believed that harmonic imaging would allow entire elimination of tissue echoes, considering that tissue does not scatter nonlinearly and that the harmonic frequencies generated by wave distortion would be eliminated by attenuation. Nonetheless, it has been demonstrated that the process of propagation of US waves through tissue is nonlinear, so harmonics are generated and build up as the acoustic beam penetrates tissue. Thus, signals from the myocardium and surrounding tissue are not absent in the harmonic image. Therefore, a tissue image is present even before the injection of a contrast agent. All these findings have motivated thorough research into the interaction of UCA with US waves to better understand the behavior of gas microbubbles in a US field (Goldberg et al. 2001). Presently, it is believed that newly discovered acoustic signatures of UCA microbubbles would make the detection methods more sensitive and, subsequently, improve image quality. These new detection methods will certainly influence the design of phased-array transducers and instrumentation to attain the highest imaging performances with UCA. We have shown previously, using simulations and phantom measurements (Bouakaz et al. 2002), that, when gas
Volume 29, Number 4, 2003
bubbles are scanned at proper acoustic pressure and frequency, significant energy is reflected not only at the second harmonic frequency but also at the higher harmonics (third, fourth, fifth and ultraharmonics). In addition, the higher harmonic components generated in tissue by nonlinear propagation effects are substantially reduced. The higher harmonic components from contrast contain individually less energy than the second harmonic component. However, appropriate combination of all the components together, named here “superharmonic,” increases extensively the amount of detected energy. To selectively image the superharmonic, we developed a new phased-array transducer with a frequency band of 145%. This wide frequency band is combined with a wide band filter to listen beyond the traditional second harmonic frequency band and, therefore, the echoes are received from contrast bubbles only, preventing tissue from being mistaken for a contrast agent (Bouakaz et al. 2002). This new imaging approach is shown to be more sensitive to UCA and “blinded” to soft tissue giving, hence, a higher CTR. In addition to superior contrast detection, the superharmonic improves the image quality by giving clean and sharply defined structures. The purpose of this study was to demonstrate the usefulness and clinical feasibility of superharmonic imaging to quantify myocardial perfusion in patients, using a commercially available contrast agent. METHODS Phased-array transducer: Dual frequency probe To perform superharmonic imaging, a new phasedarray transducer has been constructed (Bouakaz et al. 2002). The transducer consists of two different types of elements arranged in an interleaved pattern (odd and even elements). The total number of elements used was 96. The width of each of the elements was 0.2 mm. The elements could operate separately and at a distinct frequency, enabling separate transmission and reception modes. The odd elements (48) operated at around 2.8MHz center frequency and 80% band width. The even elements (48) had a center frequency of 900 kHz with a band width of 50%. With this specific array design, different operating modes were possible, depending on transmitting and receiving elements, because the lowfrequency and the high-frequency parts of the probe could be controlled individually and separately. The transducer was connected to a Vivid 5 GE-Vingmed scanner (Horten, Norway) and dedicated software was developed to drive the probe. For contrast superharmonic imaging, the low-frequency elements were set in transmission and the high-frequency elements were configured in receive mode. The transmitted signal had a center
Clinical contrast superharmonic imaging ● A. BOUAKAZ et al.
549
frequency of 0.8 MHz and contained two cycles. To assemble the higher harmonic components collectively, a wide band pass filter was applied to the received signals. The center frequency of the filter was 2.95 MHz and the width was 90%. Such parameters were selected because they showed an optimal superharmonic detection. Optimal superharmonic detection was quantified using the CTR curves obtained from radiofrequency (RF) data. In vitro study In vitro measurements were first carried out to find the scanning settings that were most optimal for contrast superharmonic imaging. A tissue-mimicking flow phantom was used, allowing a solution of contrast agent to circulate at a defined velocity. Various settings were assessed by varying the applied acoustic pressure and scanning frequency, as well as the receiving filter characteristics. RF data were recorded and analyzed to calculate the CTR. The settings that gave the highest CTR were used for patient studies. Patient study A total of 10 patients aged from 35 to 65 years with various cardiac diseases (mainly myocardial infarction, dilated cardiomyopathy and hypertrophic cardiomyopathy) were assessed to determine the applicability in a broad clinical spectrum. Apical four-chamber views were acquired end-diastolic in triggered mode before and after an IV bolus injection of 0.5 mL of SonoVue威. RF data were acquired and used to quantify myocardial perfusion. Imaging was performed in grey-scale B-mode, as well as in the angiographic color mode. Contrast harmonic imaging was determined visually and quantitatively using RF processing from regions-ofinterest (ROI). For RF processing, the ROI were acquired from the left ventricular cavity and from the myocardium before and after contrast injection. The visual evaluation of the left ventricular segments was done according to American Society of Echocardiography guidelines. The left ventricle in apical four-chamber view was divided into six segments, and each segment was scored using a three-point scale: 0 ⫽ nonvisible segment, 1 ⫽ moderately visible, and 2 ⫽ visible. Ultrasound contrast agent Sonovue威 is a second-generation contrast agent. It is composed of sulfur hexafluoride gas bubbles coated by a highly elastic phospholipid monolayer shell. The size distribution of Sonovue威 bubbles ranges from 1 to 12 m with a total number of 2 ⫻ 108/mL (Schneider et al. 1995). Sonovue威 has recently been approved in Europe and is now available in many countries in the European Union. It is approved for left ventricular border enhancement and opacification. In all patients, a 0.5-mL bolus IV injection was used.
Fig. 1. B-mode image of Sonovue威 contrast bubbles flowing in a phantom obtained using Vivid 5 scanner and the dual-frequency probe operating in superharmonic mode.
RESULTS AND DISCUSSION All the results shown here (in vitro and clinically) were obtained with the dual-frequency probe operating in the superharmonic mode, transmitting a burst of 0.8 MHz, 2 cycles (Bouakaz et al. 2002), and receiving a wide frequency band extending from the third harmonic up to the fifth harmonic. The acoustic pressure of the burst at the focus was measured separately in water using a polyvinylidene fluoride (PVDF) hydrophone (Force Institute, Brøndby, Denmark) and corresponded to a mechanical index approaching 0.2 (Bouakaz et al. 2002). In vitro measurements were carried out in a triggered mode corresponding to a repetition rate of 5 Hz. Patient recordings were carried out in triggered mode at 1 frame per heartbeat. In vitro Figure 1 shows a still frame 2-D B-mode image of Sonovue威 contrast agent flowing in the phantom. Because tissue superharmonics (third to fifth) were barely generated at this frequency as a result of nonlinear propagation, the phantom is displayed by a black and uniform background level. However, because the contrast bubbles vibrate nonlinearly and scatter at the superharmonic frequencies, the flow of the bubbles is clearly visualized on the image, which gives a high CTR. The UCA is displayed with a much higher contrast than the background tissue, giving a binary-like image. ROI were selected from the contrast bubbles and from the surrounding tissue at the same depths for further analysis. Figure 2A displays the scattered power from the contrast (—) and tissue (----) and the CTR is displayed in Fig. 2B as a function of the harmonic number. Figure 2A dem-
550
Ultrasound in Medicine and Biology
Volume 29, Number 4, 2003
Fig. 3. 2-D echo images of four-chamber views using superharmonic mode (A) before contrast injection; (B) contrast in the left ventricular cavity; (C) appearance of contrast bubbles into the myocardium.
Fig. 2. Data from Fig. 1. (A) Scattered power spectra from (—) contrast bubbles and from (----) tissue; (B) CTR.
onstrates that the major difference in scattering capabilities occurs in the frequency band of the higher harmonic components. This is sharply translated in the CTR curve where the highest ratio is obtained beyond the second harmonic. The CTR decreases around the fundamental frequency and at higher frequencies where the upper frequency band limit of the probe is reached. As previously described (Bouakaz et al. 2002), the CTR obtained with a regular second harmonic probe showed values at least 30 dB lower than the values of the CTR obtained with the dual-frequency probe in combination with superharmonic imaging. Patients Figure 3 shows four-chamber views of the heart of a patient recorded at three different time instants. Figure
3A was recorded before contrast administration and should demonstrate only echoes scattered from tissue at superharmonic frequencies. Clearly, the image shows totally black video intensity. The absence of tissue visualization is due to the fact that the superharmonic component is entirely blinded to tissues and, therefore, hardly any tissue scattering is detected in this frequency band. After contrast has been injected (Fig. 3B), we can appreciate a clear and a well-pronounced homogeneous opacification of the left ventricular cavity. The sensitivity of the superharmonic to only contrast bubbles echoes renders the borders accurately defined with a clear differentiation between the cavity and the myocardium. Because the contrast bubbles have not yet reached the contracting myocardium in this recording, this area is still depicted with black video level, which makes its distinction with the cavity easier. A few seconds later, a new recording was performed (Fig. 3C). We see that the contrast has arrived in the myocardium, giving an appreciable perfusion. The perfusion is discernible and manifest in all the segments of the heart. These findings are confirmed by the RF data recorded from different regions of the view, as shown in Fig. 4. The quantification of the left ventricle opacification and myocardial perfusion is demonstrated in Fig. 4A, where we display the scattered power from three different regions corresponding to myocardium before contrast injection ( · · · · ), myocardium after contrast injection (----) and in the left ventricular cavity after contrast injection (—). Using the scattered power signals, the CTR is calculated and plotted for the myocardium (----) and for the left ventricular cavity (—) as a function of the harmonic number in Fig. 4B. The scattered power from the myocardium before con-
Clinical contrast superharmonic imaging ● A. BOUAKAZ et al.
551
Table 1. Contrast to tissue ratio at the fundamental and harmonic frequencies obtained from RF data in a patient study.
Fundamental Second harmonic Third harmonic Fourth harmonic Fifth harmonic Superharmonic
Fig. 4. Data from Fig. 3. (A) Scattered power spectra from contrast bubbles in the (—) LV; (-----) myocardium; and from (····) tissue. (B) CTR in (—) the LV and in (-----) the myocardium.
trast injection shows second harmonic generation through the tissue due to nonlinear propagation effects, even though the transmit acoustic pressure in this case was low. The higher harmonics are however, absent, and entirely immersed in the noise with a level more than 40 dB below the fundamental. After contrast injection, the scattered power echoes recorded in both the left ventricle and the myocardium demonstrate a significant enhancement occurring in the frequency band corresponding to the higher harmonics. The increase in scattered power in the cavity is higher than that observed in the myocardium because the concentration or number of scatterers (contrast bubbles) present in the cavity can be up to 10 times higher than that contained in the myocardium. The CTR curves for both the left ventricle and the myocardium are
Left ventricle
Myocardium
10.47 23.31 27.63 31.63 23.69 28.97
7.4 12.61 19.85 18.58 13.95 18.01
given in Fig. 4B as a function of the harmonic number. For both the cavity and the myocardium, the CTR curves show a tendency to increase at the higher harmonic frequencies. Maximal ratios are reached close to the fourth harmonic with values roughly up to 30 dB and 20 dB in the left ventricle and the myocardium, respectively. A precise quantification of the CTR is given in Table 1. The CTR is given at the fundamental frequency and at the harmonic frequencies up to the fifth harmonic. In addition, an estimation of the CTR at the superharmonics is given based on combination of the higher harmonics from the third to the fifth component. We see that the CTR at the second harmonic is higher than at the fundamental and is 23.3 dB in the left ventricle cavity and 12.6 dB in the myocardium. The CTR increases further at the third and the fourth harmonic components, but then declines at the fifth harmonic frequency. At this higher harmonic frequency, the upper frequency limit of the probe is reached and, therefore, the sensitivity becomes lower. At the superharmonic, the CTR shows a significant augmentation compared to the second harmonic by almost 6 dB for both the left ventricle cavity and the myocardium. Such an increase in CTR can be compared virtually to an increase in the injected concentration of the bubbles or the transmitted acoustic power by a factor of 4 (a factor of 2 in transmitted acoustic pressure) if the superharmonic is selected instead of the second harmonic. Note that the overall difference between the scattered power curves in both the left ventricle cavity and the myocardium can go as high as 10 dB. This corresponds to a difference in the bubble concentration between the two regions of up to a factor of 10.
Table 2. Scoring results using superharmonic imaging mode before and after contrast agent injection obtained from 10 patients. Before
After
0
1
2
0
1
2
60
0
0
9
4
47
552
Ultrasound in Medicine and Biology
Fig. 5. 2-D echo images of four-chamber views using superharmonic imaging. (A) Entrance of contrast into the right ventricle; (B) appearance of contrast in the left ventricular cavity; (C) appearance of contrast into the distal septal myocardium; observe no contrast in the proximal septal myocardium; (D) consecutive frame from (C) when the contrast begins disappearing from the myocardium, but is still in the LV cavity.
Quantification of myocardial contrast perfusion was also performed visually from all the available segments using the ASE guideline. Table 2 displays the scores obtained before and after contrast injection. Before the bubbles have been injected, there was no tissue image seen, indicating clearly the blindness of superharmonic to tissue. Because higher harmonics are barely generated from nonlinear propagation effects at these selected settings, a black image is obtained when contrast bubbles are not present. This characteristic can be compared with a virtual background subtraction technique. After the bubbles have reached the myocardium, even very small concentrations of bubbles will be detected using the superharmonic. The scoring shows that more than 85% of the myocardial perfusion was distinguished. More than 78% were clearly visible in the superharmonic image and 6.6% were moderately seen. The rest of the segments (15%) were not visible, and this was mainly due to their unavailability on the image views and not to sensitivity limitations. The results obtained here demonstrate that the superharmonic is selectively sensitive to contrast bubbles only and, thus, even low concentrations of bubbles would give a scattered superharmonic echo that is readily detectable because the background tissue image has been virtually subtracted. This property is of significant importance and makes this technique capable of detecting and differentiating even different levels of myocardial perfusion, which is a difficult task with conventional second harmonic imaging mode due to tissue contamination. To demonstrate this capability of the contrast superharmonic approach, a particular case was
Volume 29, Number 4, 2003
studied. Figure 5 shows an example obtained in a 55year-old woman referred to the hospital for a hypertrophic obstructive cardiomyopathy. For treatment, the patient had undergone a percutaneous transmural septal myocardial alcohol ablation (PTSMA). This procedure, known as alcohol ablation, consists of injecting a small volume of ethanol into the artery that supplies the enlarged part of the myocardium. Therefore, two septal side branches of the left anterior descending coronary artery (LAD) were injected with 96% ethanol. This causes a precisely localized infarction and regression of the myocardial tissue. For this particular clinical case, we focused our investigation to obtain a clear visualization of the hypertrophic septal region. Figure 5 shows four views of the patient’s heart in a four-chamber view. The recordings for this study were carried out in color angiographic superharmonic mode. The first recording (Fig. 5A) was performed immediately after Sonovue威 injection. We can appreciate the arrival of the bubbles into the right side of the heart. The left side of the heart, not yet filled with contrast bubbles, appears in black due to lack of superharmonic echo generation in tissue. Figure 5B was recorded 5 s later and shows complete and homogeneous filling of the left ventricular cavity. In addition, the borders are sharply defined, demonstrating a distinct delineation of the cavity walls. Furthermore, we can clearly distinguish the hypertrophic area, which is still displayed in black because the bubbles have not reached it yet. Figure 5C was recorded 10 s later and demonstrates two distinct parts of the septal wall with two perfusion levels. The distal septal myocardium is fully and considerably perfused, whereas the proximal septal myocardium treated by PTSMA remains almost with the same level as before contrast arrival, indicating a very low level of perfusion. This clearly confirms the presence of normal and low perfusion levels in the same image. In Fig. 5D, the bubbles start disappearing from the perfused segment, but are still present vigorously in the left ventricle. Using coronary arteriography, it has been demonstrated that the absent septal side branches of the LAD correlates with the area of low perfusion and corresponds to the therapeutically ablated segment of the myocardium. The contrast bubbles can, therefore, not be present. CONCLUSIONS The availability of new generations of stable contrast bubbles has induced a significant step in the clinical applications of US. It has also encouraged the production of various new techniques of imaging methods, such as harmonic imaging, which is now evident in commercial systems. One of the main requirements of new imaging methods is the ability to identify and to display echoes from contrast agent bubbles and to reject echoes from
Clinical contrast superharmonic imaging ● A. BOUAKAZ et al.
surrounding tissue, offering, therefore, a new way of detecting microvessels. Using echo contrast agents and harmonic imaging, noninvasive assessment of myocardial perfusion can now be performed. Unfortunately, the presence of tissue harmonics is still a major limitation for the evaluation of the presence or absence of microbubbles and, especially, the ability to distinguish different levels of perfusion, which requires higher mechanical index scanning. Reliable identification of different perfusion states is a primary goal for the evaluation of myocardial perfusion by contrast echocardiography. Based on scattering properties of bubbles at the higher harmonics, we have demonstrated previously that the detection capabilities increase with the harmonic order. Superharmonic imaging is a new contrast detection approach based on selective listening to the higher harmonics from the third harmonic up to the fifth harmonic. With appropriate scanning settings, the scattered power from tissue before contrast injection is highly damped and, therefore, no tissue image is available. After contrast injection, only gas bubbles scatter energy in the superharmonic frequency band. The blindness of superharmonic to tissue renders this technique more sensitive to contrast bubbles. The advantages of superharmonic imaging, in addition to increased CTR, is that it profits from optimized imaging conditions. A better penetration is obtained by using lower than normal transmitted frequencies and a better resolution is achieved by using higher and wider than normal received frequencies. The tissue structures are, therefore, nicely defined and the noise and clutter are reduced. To use superharmonic imaging techniques, we developed a new wide band transducer that allows the reception of higher harmonic frequencies. This prelimi-
553
nary study demonstrates the advantages of the technique and its clinical feasibility. Furthermore, improvements are underway to bring such a method to clinical practice. Acknowledgments—The authors thank Delft Instruments for making the dual-frequency array transducer and GE-Vingmed, Norway for providing the driving software.
REFERENCES Averkiou M. Tissue harmonic imaging. Proc. IEEE Ultrason Sympos 2000;2:1530–1541. Bouakaz A, Frigstad S, Ten Cate F, de Jong N. Super harmonic imaging: A new technique for improved contrast detection. Ultrasound Med Biol 2002;28:59–68. Burns P. Instrumentation for contrast echocardiography. Echocardiography 2002;19:241–258. Burns PN, Powers JE, Fritzsch T. Harmonic imaging: A new imaging and Doppler method for contrast enhanced ultrasound. Radiology 1992;185(P):142. de Jong N, Cornet R, Lancee CT. Higher harmonics of vibrating gas filled microspheres. Part one: Simulations. Ultrasonics 1994a;32: 447–453. de Jong N, Cornet R, Lancee CT. Higher harmonics of vibrating gas filled microspheres. Part two: Measurements. Ultrasonics 1994b; 32:455–459. Frinking P, Bouakaz A, Kirkhorn J, Ten Cate F, de Jong N. Ultrasound contrast imaging: Current and new potential methods. Ultrasound Med Biol 2000;26:965–975. Goldberg BB, Liu JB, Forsberg F. Ultrasound contrast agents: A review. Ultrasound Med Biol 1994;20:319–333. Goldberg B, Raichlen J, Forsberg F. Ultrasound contrast agents. Basic principles and clinical applications. In: Dunitz M, ed. Ultrasound contract agents: Basic principles and clinical applications. 2nd ed. London: Martin Dunitz Ltd., 2001. Grayburn P. Current and future contrast agents. Echocardiography 2002;19:259–265. Hamilton M, Blackstock D. Nonlinear acoustics. San Diego, New York, Boston, et al.: Academic Press, 1998. Kabalnov A, Klein D, Pelura T, Schutt E, Weers J. Dissolution of multicomponent microbubbles in the bloodstream: 1. Theory. Ultrasound Med Biol 1998;24:739–749. Schneider M, Arditi M, Barrau MB, et al. BR1: A new ultrasonographic contrast agent based on sulfur hexafluoride-filled microbubbles. Invest Radiol 1995;30:451–457.
doi:10.1016/S0301-5629(03)00012-7
● Original Contribution CONTRAST SUPERHARMONIC IMAGING: A FEASIBILITY STUDY AYACHE BOUAKAZ,*‡ BOUDEWIJN J. KRENNING,*† WIM B. VLETTER,† FOLKERT J. TEN CATE*† and NICO DE JONG*‡ *Experimental Echocardiography and †Cardiology Department, Erasmus MC, Thoraxcenter, Rotterdam, The Netherlands; and ‡Interuniversity Cardiology Institute of the Netherlands (ICIN), Utrecht, The Netherlands (Received 31 July 2002; in final form 23 December 2002)
Abstract—Harmonic imaging provided significant improvement in image quality by taking advantage of the scattered second harmonic (2H) component from contrast bubbles. However, differentiation between contrast and tissue (usually termed contrast-to-tissue ratio, CTR) is sometimes cumbersome and this is mainly due to tissue contamination. We have previously demonstrated, using simulations and in vitro measurements, that CTR increases as a function of the order of the harmonic number. A new contrast imaging method based on the detection of the higher harmonics was developed and termed superharmonic (SH). This technique has been shown to be more sensitive to contrast by increasing the signal from contrast and suppressing that from tissue (high CTR). The purpose of this study was to determine the clinical feasibility and usefulness of SH in patients using a commercially available contrast agent (SonoVue威) for quantification of myocardial perfusion. A total of 10 patients with various cardiac diseases were assessed. Apical four-chamber views were acquired using SH in triggered mode before and after contrast injection. The superharmonic was performed with a newly developed probe transmitting at 0.8 MHz with a mechanical index of 0.2. Myocardial perfusion was determined visually and analyzed quantitatively using radiofrequency (RF) processing from different regions of interest. The results showed that, before contrast injection, SH was totally blinded to tissue and no superharmonic components were generated in the image view. After administration of SonoVue威, myocardial opacification was visualized by SH after contrast entered the myocardium. An increase of more than 15 dB in the myocardial bubbles echo compared to tissue echo was measured. In addition, the technique was used to visualize myocardial perfusion after myocardial septal ablation for hypertrophic cardiomyopathy. The clinical results showed the ability of contrast SH imaging in differentiating low and normal perfusion areas, demonstrating the high sensitivity and specificity of the technique. (E-mail: [email protected]) © 2003 World Federation for Ultrasound in Medicine & Biology. Key Words: Superharmonic, Harmonic, Contrast-to-tissue ratio, Ultrasound contrast, Patients, Perfusion.
by the presence of additional materials at the gas-liquid interface. In addition, by substituting different types of perfluorocarbon gases for air, the stability and plasma longevity of the agents have been markedly improved, usually lasting more than 5 min (Grayburn 2002; Kabalnov et al. 1998). With the emergence of new contrast bubble generations, new US technologies have been developed to improve microbubble detection. Improving both the intensity and longevity of the detection procedure was the consequence of a better understanding of the microbubble-US wave interaction and technological advances. The introduction of contrast harmonic imaging represents a major step in US contrast imaging (Burns et al. 1992; de Jong et al. 1994a, 1994b). This technology is now available in all medical scanners and has been introduced into clinical echocardiographic practice (Burns 2002). Harmonic imaging enhances the detection
INTRODUCTION Ultrasound (US) contrast agents (UCA) have been shown to improve the image quality of sonography by increasing the backscattered echoes from the desired regions (Burns et al. 1992; Goldberg et al. 1994, 2001). To extend the utility of US imaging, there has been active research into the use of gas microbubble contrast agents for diagnostic use. Over the past decade, research has been actively focused on developing efficacious US contrast agents and new contrast-specific imaging techniques. UCA consist of small free or encapsulated microbubbles of gas. A shell enveloping the gas is used to stabilize the bubble against dissolution and coalescence Address correspondence to: Ayache Bouakaz, Ph.D., Erasmus MC, Exp Echo, Ee2302, P.O. Box 1738. Rotterdam 3000 DR The Netherlands. E-mail: [email protected] 547
548
Ultrasound in Medicine and Biology
of UCA within the cardiac chambers and is readily used in patients undergoing echocardiography. It exploits the differences between the response of gas microbubbles and tissue to US irradiation (de Jong et al. 1994b). Soft tissues are known to be linear reflectors, whereas contrast bubbles exhibit a different behavior when interacting with US waves (Averkiou 2000). The response of a gas bubble to a US interrogation is asymmetrical, giving rise to nonlinear echoes that contain new acoustic information. This nonlinear bubble oscillation signifies a response containing multiples or harmonics of the scanning frequency. Differentiation between perfused and nonperfused tissue is now mainly based on these nonlinear characteristics. This differentiation is usually expressed by the ratio of the scattered power from the contrast agent to the scattered power from the tissue and is termed “contrast-to-tissue ratio” (CTR). The response of UCA at the second harmonic frequency is currently utilized in clinical echocardiographic imaging to enhance the contrast relative to the surrounding body tissue and, hence, the CTR. Second harmonic detection-based techniques showed significant improvements compared with conventional fundamental imaging modes. However, this technology has certain limitations. The major drawback is that it is not sensitive enough in discriminating between UCA and tissue because harmonic frequencies are also generated during the propagation of the wave through the tissue and are, therefore, reflected by the tissue itself (Frinking et al. 2000; Hamilton and Blackstock 1998). Initially, it was believed that harmonic imaging would allow entire elimination of tissue echoes, considering that tissue does not scatter nonlinearly and that the harmonic frequencies generated by wave distortion would be eliminated by attenuation. Nonetheless, it has been demonstrated that the process of propagation of US waves through tissue is nonlinear, so harmonics are generated and build up as the acoustic beam penetrates tissue. Thus, signals from the myocardium and surrounding tissue are not absent in the harmonic image. Therefore, a tissue image is present even before the injection of a contrast agent. All these findings have motivated thorough research into the interaction of UCA with US waves to better understand the behavior of gas microbubbles in a US field (Goldberg et al. 2001). Presently, it is believed that newly discovered acoustic signatures of UCA microbubbles would make the detection methods more sensitive and, subsequently, improve image quality. These new detection methods will certainly influence the design of phased-array transducers and instrumentation to attain the highest imaging performances with UCA. We have shown previously, using simulations and phantom measurements (Bouakaz et al. 2002), that, when gas
Volume 29, Number 4, 2003
bubbles are scanned at proper acoustic pressure and frequency, significant energy is reflected not only at the second harmonic frequency but also at the higher harmonics (third, fourth, fifth and ultraharmonics). In addition, the higher harmonic components generated in tissue by nonlinear propagation effects are substantially reduced. The higher harmonic components from contrast contain individually less energy than the second harmonic component. However, appropriate combination of all the components together, named here “superharmonic,” increases extensively the amount of detected energy. To selectively image the superharmonic, we developed a new phased-array transducer with a frequency band of 145%. This wide frequency band is combined with a wide band filter to listen beyond the traditional second harmonic frequency band and, therefore, the echoes are received from contrast bubbles only, preventing tissue from being mistaken for a contrast agent (Bouakaz et al. 2002). This new imaging approach is shown to be more sensitive to UCA and “blinded” to soft tissue giving, hence, a higher CTR. In addition to superior contrast detection, the superharmonic improves the image quality by giving clean and sharply defined structures. The purpose of this study was to demonstrate the usefulness and clinical feasibility of superharmonic imaging to quantify myocardial perfusion in patients, using a commercially available contrast agent. METHODS Phased-array transducer: Dual frequency probe To perform superharmonic imaging, a new phasedarray transducer has been constructed (Bouakaz et al. 2002). The transducer consists of two different types of elements arranged in an interleaved pattern (odd and even elements). The total number of elements used was 96. The width of each of the elements was 0.2 mm. The elements could operate separately and at a distinct frequency, enabling separate transmission and reception modes. The odd elements (48) operated at around 2.8MHz center frequency and 80% band width. The even elements (48) had a center frequency of 900 kHz with a band width of 50%. With this specific array design, different operating modes were possible, depending on transmitting and receiving elements, because the lowfrequency and the high-frequency parts of the probe could be controlled individually and separately. The transducer was connected to a Vivid 5 GE-Vingmed scanner (Horten, Norway) and dedicated software was developed to drive the probe. For contrast superharmonic imaging, the low-frequency elements were set in transmission and the high-frequency elements were configured in receive mode. The transmitted signal had a center
Clinical contrast superharmonic imaging ● A. BOUAKAZ et al.
549
frequency of 0.8 MHz and contained two cycles. To assemble the higher harmonic components collectively, a wide band pass filter was applied to the received signals. The center frequency of the filter was 2.95 MHz and the width was 90%. Such parameters were selected because they showed an optimal superharmonic detection. Optimal superharmonic detection was quantified using the CTR curves obtained from radiofrequency (RF) data. In vitro study In vitro measurements were first carried out to find the scanning settings that were most optimal for contrast superharmonic imaging. A tissue-mimicking flow phantom was used, allowing a solution of contrast agent to circulate at a defined velocity. Various settings were assessed by varying the applied acoustic pressure and scanning frequency, as well as the receiving filter characteristics. RF data were recorded and analyzed to calculate the CTR. The settings that gave the highest CTR were used for patient studies. Patient study A total of 10 patients aged from 35 to 65 years with various cardiac diseases (mainly myocardial infarction, dilated cardiomyopathy and hypertrophic cardiomyopathy) were assessed to determine the applicability in a broad clinical spectrum. Apical four-chamber views were acquired end-diastolic in triggered mode before and after an IV bolus injection of 0.5 mL of SonoVue威. RF data were acquired and used to quantify myocardial perfusion. Imaging was performed in grey-scale B-mode, as well as in the angiographic color mode. Contrast harmonic imaging was determined visually and quantitatively using RF processing from regions-ofinterest (ROI). For RF processing, the ROI were acquired from the left ventricular cavity and from the myocardium before and after contrast injection. The visual evaluation of the left ventricular segments was done according to American Society of Echocardiography guidelines. The left ventricle in apical four-chamber view was divided into six segments, and each segment was scored using a three-point scale: 0 ⫽ nonvisible segment, 1 ⫽ moderately visible, and 2 ⫽ visible. Ultrasound contrast agent Sonovue威 is a second-generation contrast agent. It is composed of sulfur hexafluoride gas bubbles coated by a highly elastic phospholipid monolayer shell. The size distribution of Sonovue威 bubbles ranges from 1 to 12 m with a total number of 2 ⫻ 108/mL (Schneider et al. 1995). Sonovue威 has recently been approved in Europe and is now available in many countries in the European Union. It is approved for left ventricular border enhancement and opacification. In all patients, a 0.5-mL bolus IV injection was used.
Fig. 1. B-mode image of Sonovue威 contrast bubbles flowing in a phantom obtained using Vivid 5 scanner and the dual-frequency probe operating in superharmonic mode.
RESULTS AND DISCUSSION All the results shown here (in vitro and clinically) were obtained with the dual-frequency probe operating in the superharmonic mode, transmitting a burst of 0.8 MHz, 2 cycles (Bouakaz et al. 2002), and receiving a wide frequency band extending from the third harmonic up to the fifth harmonic. The acoustic pressure of the burst at the focus was measured separately in water using a polyvinylidene fluoride (PVDF) hydrophone (Force Institute, Brøndby, Denmark) and corresponded to a mechanical index approaching 0.2 (Bouakaz et al. 2002). In vitro measurements were carried out in a triggered mode corresponding to a repetition rate of 5 Hz. Patient recordings were carried out in triggered mode at 1 frame per heartbeat. In vitro Figure 1 shows a still frame 2-D B-mode image of Sonovue威 contrast agent flowing in the phantom. Because tissue superharmonics (third to fifth) were barely generated at this frequency as a result of nonlinear propagation, the phantom is displayed by a black and uniform background level. However, because the contrast bubbles vibrate nonlinearly and scatter at the superharmonic frequencies, the flow of the bubbles is clearly visualized on the image, which gives a high CTR. The UCA is displayed with a much higher contrast than the background tissue, giving a binary-like image. ROI were selected from the contrast bubbles and from the surrounding tissue at the same depths for further analysis. Figure 2A displays the scattered power from the contrast (—) and tissue (----) and the CTR is displayed in Fig. 2B as a function of the harmonic number. Figure 2A dem-
550
Ultrasound in Medicine and Biology
Volume 29, Number 4, 2003
Fig. 3. 2-D echo images of four-chamber views using superharmonic mode (A) before contrast injection; (B) contrast in the left ventricular cavity; (C) appearance of contrast bubbles into the myocardium.
Fig. 2. Data from Fig. 1. (A) Scattered power spectra from (—) contrast bubbles and from (----) tissue; (B) CTR.
onstrates that the major difference in scattering capabilities occurs in the frequency band of the higher harmonic components. This is sharply translated in the CTR curve where the highest ratio is obtained beyond the second harmonic. The CTR decreases around the fundamental frequency and at higher frequencies where the upper frequency band limit of the probe is reached. As previously described (Bouakaz et al. 2002), the CTR obtained with a regular second harmonic probe showed values at least 30 dB lower than the values of the CTR obtained with the dual-frequency probe in combination with superharmonic imaging. Patients Figure 3 shows four-chamber views of the heart of a patient recorded at three different time instants. Figure
3A was recorded before contrast administration and should demonstrate only echoes scattered from tissue at superharmonic frequencies. Clearly, the image shows totally black video intensity. The absence of tissue visualization is due to the fact that the superharmonic component is entirely blinded to tissues and, therefore, hardly any tissue scattering is detected in this frequency band. After contrast has been injected (Fig. 3B), we can appreciate a clear and a well-pronounced homogeneous opacification of the left ventricular cavity. The sensitivity of the superharmonic to only contrast bubbles echoes renders the borders accurately defined with a clear differentiation between the cavity and the myocardium. Because the contrast bubbles have not yet reached the contracting myocardium in this recording, this area is still depicted with black video level, which makes its distinction with the cavity easier. A few seconds later, a new recording was performed (Fig. 3C). We see that the contrast has arrived in the myocardium, giving an appreciable perfusion. The perfusion is discernible and manifest in all the segments of the heart. These findings are confirmed by the RF data recorded from different regions of the view, as shown in Fig. 4. The quantification of the left ventricle opacification and myocardial perfusion is demonstrated in Fig. 4A, where we display the scattered power from three different regions corresponding to myocardium before contrast injection ( · · · · ), myocardium after contrast injection (----) and in the left ventricular cavity after contrast injection (—). Using the scattered power signals, the CTR is calculated and plotted for the myocardium (----) and for the left ventricular cavity (—) as a function of the harmonic number in Fig. 4B. The scattered power from the myocardium before con-
Clinical contrast superharmonic imaging ● A. BOUAKAZ et al.
551
Table 1. Contrast to tissue ratio at the fundamental and harmonic frequencies obtained from RF data in a patient study.
Fundamental Second harmonic Third harmonic Fourth harmonic Fifth harmonic Superharmonic
Fig. 4. Data from Fig. 3. (A) Scattered power spectra from contrast bubbles in the (—) LV; (-----) myocardium; and from (····) tissue. (B) CTR in (—) the LV and in (-----) the myocardium.
trast injection shows second harmonic generation through the tissue due to nonlinear propagation effects, even though the transmit acoustic pressure in this case was low. The higher harmonics are however, absent, and entirely immersed in the noise with a level more than 40 dB below the fundamental. After contrast injection, the scattered power echoes recorded in both the left ventricle and the myocardium demonstrate a significant enhancement occurring in the frequency band corresponding to the higher harmonics. The increase in scattered power in the cavity is higher than that observed in the myocardium because the concentration or number of scatterers (contrast bubbles) present in the cavity can be up to 10 times higher than that contained in the myocardium. The CTR curves for both the left ventricle and the myocardium are
Left ventricle
Myocardium
10.47 23.31 27.63 31.63 23.69 28.97
7.4 12.61 19.85 18.58 13.95 18.01
given in Fig. 4B as a function of the harmonic number. For both the cavity and the myocardium, the CTR curves show a tendency to increase at the higher harmonic frequencies. Maximal ratios are reached close to the fourth harmonic with values roughly up to 30 dB and 20 dB in the left ventricle and the myocardium, respectively. A precise quantification of the CTR is given in Table 1. The CTR is given at the fundamental frequency and at the harmonic frequencies up to the fifth harmonic. In addition, an estimation of the CTR at the superharmonics is given based on combination of the higher harmonics from the third to the fifth component. We see that the CTR at the second harmonic is higher than at the fundamental and is 23.3 dB in the left ventricle cavity and 12.6 dB in the myocardium. The CTR increases further at the third and the fourth harmonic components, but then declines at the fifth harmonic frequency. At this higher harmonic frequency, the upper frequency limit of the probe is reached and, therefore, the sensitivity becomes lower. At the superharmonic, the CTR shows a significant augmentation compared to the second harmonic by almost 6 dB for both the left ventricle cavity and the myocardium. Such an increase in CTR can be compared virtually to an increase in the injected concentration of the bubbles or the transmitted acoustic power by a factor of 4 (a factor of 2 in transmitted acoustic pressure) if the superharmonic is selected instead of the second harmonic. Note that the overall difference between the scattered power curves in both the left ventricle cavity and the myocardium can go as high as 10 dB. This corresponds to a difference in the bubble concentration between the two regions of up to a factor of 10.
Table 2. Scoring results using superharmonic imaging mode before and after contrast agent injection obtained from 10 patients. Before
After
0
1
2
0
1
2
60
0
0
9
4
47
552
Ultrasound in Medicine and Biology
Fig. 5. 2-D echo images of four-chamber views using superharmonic imaging. (A) Entrance of contrast into the right ventricle; (B) appearance of contrast in the left ventricular cavity; (C) appearance of contrast into the distal septal myocardium; observe no contrast in the proximal septal myocardium; (D) consecutive frame from (C) when the contrast begins disappearing from the myocardium, but is still in the LV cavity.
Quantification of myocardial contrast perfusion was also performed visually from all the available segments using the ASE guideline. Table 2 displays the scores obtained before and after contrast injection. Before the bubbles have been injected, there was no tissue image seen, indicating clearly the blindness of superharmonic to tissue. Because higher harmonics are barely generated from nonlinear propagation effects at these selected settings, a black image is obtained when contrast bubbles are not present. This characteristic can be compared with a virtual background subtraction technique. After the bubbles have reached the myocardium, even very small concentrations of bubbles will be detected using the superharmonic. The scoring shows that more than 85% of the myocardial perfusion was distinguished. More than 78% were clearly visible in the superharmonic image and 6.6% were moderately seen. The rest of the segments (15%) were not visible, and this was mainly due to their unavailability on the image views and not to sensitivity limitations. The results obtained here demonstrate that the superharmonic is selectively sensitive to contrast bubbles only and, thus, even low concentrations of bubbles would give a scattered superharmonic echo that is readily detectable because the background tissue image has been virtually subtracted. This property is of significant importance and makes this technique capable of detecting and differentiating even different levels of myocardial perfusion, which is a difficult task with conventional second harmonic imaging mode due to tissue contamination. To demonstrate this capability of the contrast superharmonic approach, a particular case was
Volume 29, Number 4, 2003
studied. Figure 5 shows an example obtained in a 55year-old woman referred to the hospital for a hypertrophic obstructive cardiomyopathy. For treatment, the patient had undergone a percutaneous transmural septal myocardial alcohol ablation (PTSMA). This procedure, known as alcohol ablation, consists of injecting a small volume of ethanol into the artery that supplies the enlarged part of the myocardium. Therefore, two septal side branches of the left anterior descending coronary artery (LAD) were injected with 96% ethanol. This causes a precisely localized infarction and regression of the myocardial tissue. For this particular clinical case, we focused our investigation to obtain a clear visualization of the hypertrophic septal region. Figure 5 shows four views of the patient’s heart in a four-chamber view. The recordings for this study were carried out in color angiographic superharmonic mode. The first recording (Fig. 5A) was performed immediately after Sonovue威 injection. We can appreciate the arrival of the bubbles into the right side of the heart. The left side of the heart, not yet filled with contrast bubbles, appears in black due to lack of superharmonic echo generation in tissue. Figure 5B was recorded 5 s later and shows complete and homogeneous filling of the left ventricular cavity. In addition, the borders are sharply defined, demonstrating a distinct delineation of the cavity walls. Furthermore, we can clearly distinguish the hypertrophic area, which is still displayed in black because the bubbles have not reached it yet. Figure 5C was recorded 10 s later and demonstrates two distinct parts of the septal wall with two perfusion levels. The distal septal myocardium is fully and considerably perfused, whereas the proximal septal myocardium treated by PTSMA remains almost with the same level as before contrast arrival, indicating a very low level of perfusion. This clearly confirms the presence of normal and low perfusion levels in the same image. In Fig. 5D, the bubbles start disappearing from the perfused segment, but are still present vigorously in the left ventricle. Using coronary arteriography, it has been demonstrated that the absent septal side branches of the LAD correlates with the area of low perfusion and corresponds to the therapeutically ablated segment of the myocardium. The contrast bubbles can, therefore, not be present. CONCLUSIONS The availability of new generations of stable contrast bubbles has induced a significant step in the clinical applications of US. It has also encouraged the production of various new techniques of imaging methods, such as harmonic imaging, which is now evident in commercial systems. One of the main requirements of new imaging methods is the ability to identify and to display echoes from contrast agent bubbles and to reject echoes from
Clinical contrast superharmonic imaging ● A. BOUAKAZ et al.
surrounding tissue, offering, therefore, a new way of detecting microvessels. Using echo contrast agents and harmonic imaging, noninvasive assessment of myocardial perfusion can now be performed. Unfortunately, the presence of tissue harmonics is still a major limitation for the evaluation of the presence or absence of microbubbles and, especially, the ability to distinguish different levels of perfusion, which requires higher mechanical index scanning. Reliable identification of different perfusion states is a primary goal for the evaluation of myocardial perfusion by contrast echocardiography. Based on scattering properties of bubbles at the higher harmonics, we have demonstrated previously that the detection capabilities increase with the harmonic order. Superharmonic imaging is a new contrast detection approach based on selective listening to the higher harmonics from the third harmonic up to the fifth harmonic. With appropriate scanning settings, the scattered power from tissue before contrast injection is highly damped and, therefore, no tissue image is available. After contrast injection, only gas bubbles scatter energy in the superharmonic frequency band. The blindness of superharmonic to tissue renders this technique more sensitive to contrast bubbles. The advantages of superharmonic imaging, in addition to increased CTR, is that it profits from optimized imaging conditions. A better penetration is obtained by using lower than normal transmitted frequencies and a better resolution is achieved by using higher and wider than normal received frequencies. The tissue structures are, therefore, nicely defined and the noise and clutter are reduced. To use superharmonic imaging techniques, we developed a new wide band transducer that allows the reception of higher harmonic frequencies. This prelimi-
553
nary study demonstrates the advantages of the technique and its clinical feasibility. Furthermore, improvements are underway to bring such a method to clinical practice. Acknowledgments—The authors thank Delft Instruments for making the dual-frequency array transducer and GE-Vingmed, Norway for providing the driving software.
REFERENCES Averkiou M. Tissue harmonic imaging. Proc. IEEE Ultrason Sympos 2000;2:1530–1541. Bouakaz A, Frigstad S, Ten Cate F, de Jong N. Super harmonic imaging: A new technique for improved contrast detection. Ultrasound Med Biol 2002;28:59–68. Burns P. Instrumentation for contrast echocardiography. Echocardiography 2002;19:241–258. Burns PN, Powers JE, Fritzsch T. Harmonic imaging: A new imaging and Doppler method for contrast enhanced ultrasound. Radiology 1992;185(P):142. de Jong N, Cornet R, Lancee CT. Higher harmonics of vibrating gas filled microspheres. Part one: Simulations. Ultrasonics 1994a;32: 447–453. de Jong N, Cornet R, Lancee CT. Higher harmonics of vibrating gas filled microspheres. Part two: Measurements. Ultrasonics 1994b; 32:455–459. Frinking P, Bouakaz A, Kirkhorn J, Ten Cate F, de Jong N. Ultrasound contrast imaging: Current and new potential methods. Ultrasound Med Biol 2000;26:965–975. Goldberg BB, Liu JB, Forsberg F. Ultrasound contrast agents: A review. Ultrasound Med Biol 1994;20:319–333. Goldberg B, Raichlen J, Forsberg F. Ultrasound contrast agents. Basic principles and clinical applications. In: Dunitz M, ed. Ultrasound contract agents: Basic principles and clinical applications. 2nd ed. London: Martin Dunitz Ltd., 2001. Grayburn P. Current and future contrast agents. Echocardiography 2002;19:259–265. Hamilton M, Blackstock D. Nonlinear acoustics. San Diego, New York, Boston, et al.: Academic Press, 1998. Kabalnov A, Klein D, Pelura T, Schutt E, Weers J. Dissolution of multicomponent microbubbles in the bloodstream: 1. Theory. Ultrasound Med Biol 1998;24:739–749. Schneider M, Arditi M, Barrau MB, et al. BR1: A new ultrasonographic contrast agent based on sulfur hexafluoride-filled microbubbles. Invest Radiol 1995;30:451–457.