- Email: [email protected]
Ultraharmonic myocardial contrast imaging: In vivo experimental and clinical data from a novel technique
Ultraharmonic Myocardial Contrast Imaging: In Vivo Experimental and Clinical Data from a Novel Technique Bettina Kuersten, MD, Thippeswamy H. Murthy, MD, Peng Li, MD, Zheng Liu, MD, Elizabeth Locricchio, BS, RDCS, Cheryl Baisch, BS, Patrick Rafter, MS, and Mani Vannan, MBBS, Ann Arbor, Michigan, and Andover, Massachusetts
Myocardial contrast echocardiography (MCE) with high–mechanical-index (MI), triggered harmonic imaging is the best-established technique to date for the assessment of myocardial perfusion. A high signal-to-noise ratio, which is significantly influenced by precontrast tissue signals, is an important prerequisite. Our goal was to evaluate the efficacy of ultraharmonic MCE, a technique that rejects tissue signals by receiving signals beyond the second but below the third harmonic. Imaging was performed in 6 closed-chest dogs and in 15 healthy volunteers (11 of whom also had dipyridamole stress). Analyses of
M
yocardial contrast echocardiography (MCE) has been shown to effectively assess myocardial microcirculation in experimental and initial clinical studies.1-4 The technique of MCE that is best established to date is high–mechanical-index (MI), triggered harmonic imaging. However, because of the considerable myocardial tissue signal before contrast, background subtraction is often required to enhance the signal-to-noise ratio. This ratio is significantly influenced by precontrast tissue signals because postcontrast enhancement is limited by the bandwidth of the receive elements in the transducer.A number of efforts have been made to reject tissue signals more efficiently while enhancing microbubble signals by using techniques such as pulse-inversion5 and sub-harmonic6 imaging.Another new technique
videointensity (VI) confirmed uniformly low precontrast tissue VI, a significant increase of postcontrast VI (before and after dipyridamole), and a significant decrease in VI after microbubble destruction. We conclude that ultraharmonic MCE produces low precontrast tissue signals, thus optimizing postcontrast myocardial opacification, and exhibits efficient microbubble destruction with use of multipleframe triggering. Thus this new technique opens up a new possibility of further optimizing coronary microcirculation imaging with microbubbles. (J Am Soc Echocardiogr 2001;14:910-6.)
is ultraharmonic imaging, which exploits the differential signals between tissue and microbubbles by receiving the returning signals at a frequency beyond the second harmonic but below the third harmonic frequency of the incident ultrasound waves. This frequency is part of the wideband frequencies that are backscattered by disrupting microbubbles, thus indicating the presence of microbubbles. In contrast, the myocardial tissue returns virtually no signals at this frequency. The aim of this study was to evaluate the efficacy of ultraharmonic imaging for myocardial opacification after intravenous administration of microbubbles.
METHODS Population
From the Department of Internal Medicine, Division of Cardiology, University of Michigan Health System, Ann Arbor, and Agilent Technologies, Andover, Mass (P.R.). Supported by an Agilent Technologies research grant. Reprint requests: Mani Vannan, MBBS, MCP Hahnemann University, School of Medicine, Department of Medicine, Mail Stop 470, 245 N 15th St, Philadelphia, PA (E-mail: [email protected]). Copyright © 2001 by the American Society of Echocardiography. 0894-7317/2001/$35.00 + 0 27/1/113257 doi:10.1067/mje.2001.113257
910
Dogs. Six healthy dogs (4 male and 2 female mongrels) were studied. Their weight ranged from 18.2 to 28.2 kg. The study was approved by the University Committee for the Use and Care of Animals (UCUCA) at the University of Michigan and conformed to the American Heart Association guidelines for the use of animals in research. Human beings. A total of 15 healthy volunteers with no history of coronary artery disease or any other cardiopulmonary symptoms and with good apical echocardiographic
Journal of the American Society of Echocardiography Volume 14 Number 9
Kuersten et al 911
windows were included in the study.The mean age was 27 years (range 21 to 39 years); 11 were men and 4 were women. The protocol was approved by the institutional review board of the University of Michigan Health System, and all subjects signed informed, written consent statements. Ultraharmonic Myocardial Contrast Echocardiography Dogs. The ultrasonic contrast agent MRX-225 (ImaRx Pharmaceutical Corp, Tucson, Ariz) that we used in the 6 dogs is composed of a lipid shell filled with perfluorobutane gas. It was administered by a single slow bolus injection through a peripheral intravenous line, followed by a saline flush. Echocardiographic imaging was performed by placing the closed-chest dogs in left lateral position and transthoracically scanning the 4-chamber view. Human beings. In the human subjects we used intravenous Definity (DuPont Pharmaceuticals Company Medical Imaging, Billerica, MA), a suspension of phospholipidcoated perfluorocarbon microbubbles. It was prepared by diluting 1300 µL of Definity in 25 mL of 0.9% saline solution.This solution was administered as a continuous infusion at a rate of 4.0 to 6.5 mL/min through a peripheral vein with the use of a syringe pump (Razel Scientific Instruments, Stamford, Conn).The maximal total dose was 3900 µL per subject.All human subjects underwent resting MCE, and 11 of them also underwent dipyridamole stress MCE. Images at stress were acquired 4 minutes after termination of the 4-minute infusion of dipyridamole at a rate of 0.56 mg/kg with isometric handgrip. We used a commercially available ultrasonographic platform (Sonos 5500, Agilent Technologies, Andover, Mass) equipped with an ultraharmonics modality and a wideband phased-array transducer (S3). In ultraharmonic imaging the transmitted ultrasound frequency is 1.3 MHz, and the receive frequency is 3.6 MHz. Thus the receive frequency lies between the second and third harmonic frequency of 2.6 and 3.9 MHz, at a point at which the insonated tissue returns only minimal signals while the signals emitted by the contrast agent are still much stronger (Figure 1).
Imaging Protocol We acquired images in the 4-chamber view in all dogs and human beings. In the closed-chest dogs, the transducer was fixed in a stable position with use of a flexible mechanical arm.The dynamic range was set at 50 dB, and the transmit focus was positioned at the apex (upper third of the imaging field). Other instrument settings (gain, compensation, depth, etc) at baseline were optimized for each study but were kept constant throughout any given study to preclude changes in videointensity caused by changes in machine settings. We performed multiple-frame triggered
Figure 1 In ultraharmonic myocardial contrast imaging the contrast agent signals are separated from the tissue signals by receiving at a frequency between the second and third harmonic.
imaging at an MI of 1.4 to 1.7, with the time trigger set at end systole and the triggering interval at every fourth cardiac cycle. MCE Image Analysis For image analysis we digitized the loops from S-VHS videotape with use of Adobe Premiere 5.1 software (Adobe Systems Incorporated, San Jose, Calif). After alignment of the fill and destruction frames of these digitized clips, the videointensities of 6 manually drawn regions of interest (ROI) were quantitated with the software MCE par 2.1 (University of Virginia, Charlottesville, Va). The 6 ROIs were placed in the basal (BS) and mid septum (MS), the apex, the mid (ML) and basal lateral left ventricular wall (BL), and in the cavity. Statistical Analysis The statistical analyses were performed with Analyse-It for Microsoft Excel software (version 1.48, Analyse-It Software Ltd, UK). Videointensity values and values of the differences between videointensities were expressed as mean ± SEM.The differences in the means were compared with a 2-tailed, paired Student t test. Results were considered statistically significant at P < .05. To test for a difference in the means of the different ROIs, we applied a 1way analysis of variance (ANOVA) that computed contrasts with Tukey error protection and a confidence interval of 95%.
RESULTS In the 6 dogs a total of 30 ROIs were analyzed at baseline and a total of 29 ROIs after contrast administration. One BL segment had to be excluded after contrast because of an obvious BL artifact.
Journal of the American Society of Echocardiography September 2001
912 Kuersten et al
Table 1 Videointensities of the regions of interest in the 5 different myocardial segments, in the cavity, and in all 5 wall segments combined in human beings Baseline (n = 14)
Contrast fill frame (n = 15)
Destruction frame (n = 15)
Dipyridamole fill frame (n = 11)
34 (±3.1) 29 (±2.1) 27 (±1.5) 28 (±2.6) 29 (±1.4) 22 (±1.3) 29.7 (±1.10)
92 (±11.6) 116 (±11.9) 95 (±13.5) 75 (±10.7) 59 (±6.8) 201 (±7.4) 87.8 (±5.46)
52 (±8.2) 52 (±6.5) 40 (±4.2) 26 (±3.8) 34 (±4.5) 202 (±6.8) 41.3 (±2.80)
108 (±12.3) 131 (±12.9) 112 (±12.3) 91 (±12.6) 80 (±15.3) 210 (±5.9) 103.6 (±6.12)
Human beings
Basal septal Mid septal Apex Mid lateral Basal lateral Cavity All wall segments
Destruction frame (n = 11)
58 67 49 35 42 205 49.3
(±10.2) (±9.4) (±6.5) (±5.9) (±9.0) (±6.9) (±3.90)
Data are means (±SEM). No statistically significant difference was found between the different regions of interest at baseline, in the fill frames, and in the destruction frames after contrast injection before and after dipyridamole stress, with use of the 1-way-analysis of variance and Tukey error protection.
Table 2 Videointensities of the regions of interest in the 5 different myocardial segments, in the cavity, and in all 5 wall segments combined in dogs Baseline (n = 6)
Dogs
Basal septal Mid septal Apex Mid lateral Basal lateral Cavity All wall segments
27 24 40 26 35 21 30.0
(±3.7) (±1.8) (±7.5) (±2.9) (±7.3) (±1.2) (±2.30)
Contrast fill frame (n = 6)
102 96 74 80 68 207 84.6
(±13.3) (±9.8) (±8.9) (±9.8) (±10.8) (±5.3) (±5.01)
Destruction frame (n = 6)
61 36 39 33 37 211 41.1
(±5.8) (±4.8) (±6.2) (±6.6) (±9.2) (±9.8) (±4.24)
Data are means (±SEM). No statistically significant difference was found between the different regions of interest at baseline, in the fill frame, or in the destruction frame, with use of the 1-way analysis of variance and Tukey error protection.
In the 15 human subjects, a total of 70 ROIs could be analyzed at baseline (in 1 patient the baseline image was not available).After contrast, the presence of lateral wall artifact in 2 subjects meant that a total of 71 ROIs could be analyzed. All 55 ROIs of the 11 subjects who underwent dipyridamole stress MCE were included in the analyses. Baseline Images Before administration of the contrast agent, the ultraharmonic frequency produced by the myocardium resulted in a mean videointensity (VI) of 30.0 ± 2.3 in all myocardial segments in closed-chest dogs and a mean VI of 29.7 ± 1.10 in the human subjects.The absolute VIs in the separate ROIs in dogs ranged from 24 ± 1.8 in the MS to 40 ± 7.5 in the apex and in human subjects from 27 ± 1.5 in the apex to 34 ± 3.1 in the BS (Tables 1 and 2).The VIs among the various ROIs were not significantly different at a confidence interval of 95% in dogs or in human beings (Figures 2 and 3).
Myocardial Opacification with Microbubbles After administration of a myocardial contrast agent, the mean VI of all ROIs combined increased to values of 84.6 ± 5.01 in dogs and 87.8 ± 5.46 in human beings before dipyridamole, and to 103.6 ± 6.12 in human beings after dipyridamole. This increase was highly significant in dogs (P < .0001) as well as in human beings (P < .0001). The analyses of the separate ROIs in dogs after contrast confirmed a significant increase in absolute VI in all segments (BS: P < .005; MS: P < .0007; apex: P < .02; ML: P < .002; BL: P < .03). From baseline the VI increased to a minimum of 68 ± 10.8 in the BL and to a maximum of 102 ± 13.3 in the BS (Table 2, Figure 2). Expressed as percentages, the increase in VI from baseline was 309% ± 79.6% in the BS, 303% ± 47.5% in the MS, 132% ± 40.5% in the apex, 211% ± 36.9% in the ML, and 159% ± 66.9% in the BL. Analysis of variance showed no significant difference in absolute VIs among the various ROIs. In human beings, the opacification of each myocardial ROI before dipyridamole was significant (BS: P < .0001; MS: P < .0001; apex: P < .0003; ML: P < .002; BL: P < .004), with an absolute VI increase ranging from 59 ± 6.8 in the BL to 116 ± 11.9 in the MS. Comparable opacification was obtained with dipyridamole stress with absolute VIs ranging from 80 ± 15.3 in the BL to 131 ± 12.9 in the MS ROI (BS: P < .0001; MS: P < .0001; apex: P < .0001; ML: P < .0003; BL: P < .009), resulting in no significant difference of VI in the ROIs before and after dipyridamole administration (Table 1, Figure 3).The percentage increases of opacification in the ROIs from baseline were 182% ± 28.6% and 225% ± 41.4% (before and after dipyridamole stress, respectively) in the BS, 325% ± 43.2% and 381% ± 53.9% in the MS, 262% ± 54.2% and 334% ± 49.5% in the apex, 223% ± 53.8% and 276% ± 52.6% in the ML, and 104% ± 27.0% and 187% ± 52.7% in the BL.Analysis of variance showed
Journal of the American Society of Echocardiography Volume 14 Number 9
Kuersten et al 913
Figure 2 Myocardial opacification with triggered imaging in a 1 to 4 interval for the 6 dogs in each of the 5 regions of interest. The values of videointensity (VI) of each individual dog are given at baseline and in the fill and destruction frames after contrast application. The red curves represent the mean values in each region of interest. Error bars represent 1 SEM.
a significant difference in the increase of absolute VI between the MS and the BL before dipyridamole. For all other regions, ANOVA did not show any interregional differences in VI. After dipyridamole, no significant differences existed in absolute VI among regions. Destruction of the Microbubbles In the destruction frames, the mean VIs of all 5 ROIs combined decreased to a value of 41.1 ± 4.24 in dogs, to 41.3 ± 2.80 in human beings before dipyridamole stress, and to 49.3 ± 3.90 in human beings after dipyridamole stress. In both dogs (P < .0001) and human beings (P < .0001) this decrease was highly significant. The analyses in dogs confirmed a significant decrease of absolute VI in each myocardial ROI from 33 ± 6.6 in the ML to 61 ± 5.8 in the BS (BS: P < .005; MS: P < .003; apex: P < .0002; ML: P < .0003; BL: P < .04) (Table 2, Figure 2). The mean decrease in absolute VI ranged from a minimum of 31 in the BL to a maximum of 61 in the MS (Figure 2). The percentage decrease of VI in the ROIs from the fill to the destruction frame was 38% ± 4.5% in the BS, 61% ±
6.7% in the MS, 49% ± 3.7% in the apex, 59% ± 4.0% in the ML, and 44% ± 12.0% in the BL. In human beings the absolute VI in the ROIs after microbubble destruction before dipyridamole decreased to near baseline values from 26 ± 3.8 in the ML to 52 ± 8.2 and 52 ± 6.5 in the BS and MS, respectively (BS: P < .0001; MS: P < .0001; apex: P < .0001; ML: P < .0002; BL: P < .0001). Also, after dipyridamole, significantly decreased absolute VI could be measured in the destruction frames, ranging from 35 ± 5.9 in the ML to 67 ± 9.4 in the MS (BS: P < .0001; MS: P < .0001; apex: P < .0001; ML: P < .07; BL: P < .002) (Table 1, Figure 3). The mean decrease in absolute VI in the ROIs ranged from a minimum of 25 and 38 in the BL to a maximum of 64 and 64 (before and after dipyridamole) in the MS. The percentage decrease of VI in the ROIs was 45% ± 3.3% and 49% ± 4.5% (before and at dipyridamole stress, respectively) in the BS, 55% ± 2.8% and 49% ± 4.4% in the MS, 53% ± 4.0% and 55% ± 3.8% in the apex, 62% ± 3.6% and 62% ± 2.5% in the ML, and 40% ± 5.1% and 47% ± 4.6% in the BL. Before dipyridamole, a significant difference in the decrease of VI was seen between the ML and the BS and MS. For all other
914 Kuersten et al
Journal of the American Society of Echocardiography September 2001
Figure 3 Myocardial opacification with triggered imaging in a 1 to 4 interval for the 15 healthy human subjects in each of the 5 regions of interest. The values of videointensity (VI) of each individual subject are given at baseline, in the fill and destruction frames after contrast application, and in the fill and destruction frames after dipyridamole stress. The red curves represent the mean values in each region of interest. Error bars represent 1 SEM.
regions before dipyridamole, ANOVA did not show any interregional differences in VI. After dipyridamole stress, no significant differences in VI existed between regions.
DISCUSSION In this study we have demonstrated that ultraharmonic MCE produces significant and homogeneous myocardial opacification after intravenous injection of contrast agents in closed-chest dogs and in healthy human subjects. Furthermore, we have shown that the precontrast tissue signals with ultraharmonic imaging are negligible and that multiple-frame high– MI triggering with this technique results in efficient bubble destruction (Figure 4). The technique of ultraharmonic imaging combines the dual advantage of an efficient rejection of tissue signals while enhancing the microbubble signals, as described previously in the Methods section. This is an important prerequisite for a technique that relies on high–MI, triggered imaging. At a high MI,
myocardial tissue produces harmonic signals, socalled tissue harmonics, which results in enhanced visualization of the myocardium.5,7 This phenomenon is counterproductive if the aim is to visualize microbubbles. In a B-mode representation, visual discrimination of myocardial enhancement by microbubbles is not reliable and may not be possible in some circumstances. To this end, when using the conventional second harmonic high–MI, triggered technology, it is imperative that precontrast myocardial signals are subtracted from the postcontrast images to reliably assess myocardial opacification.8 Unlike the problem of enhanced tissue harmonics at a high MI, when imaging with ultraharmonic technology, the myocardium in fact appears darker. This is because the returning frequencies are examined beyond the second harmonic of the transmitted frequency but before the third harmonic, at which point myocardial tissue has very few signals. In our study it is noteworthy that the mean VI in all myocardial regions in closed-chest dogs and in healthy human subjects was comparably low. For example, at the MS it was 24 ± 1.8 in dogs and 29 ± 2.1 in human
Journal of the American Society of Echocardiography Volume 14 Number 9
A
Kuersten et al 915
B
Figure 4 Ultraharmonic myocardial contrast imaging in a healthy human volunteer (A) and in a healthy dog (B). The fill frame showing the myocardial enhancement by intravenous myocardial contrast agent (left side) is displayed side by side with the destruction frame (right side).
beings. In dogs, apical precontrast VI was greater than in other regions because of transducer artifact at the apex. This was a constant factor before and after contrast administration. After contrast administration, videointensities increased from 104% to 325% in both closed-chest dogs and in human subjects.The magnitude of this increase in VI over baseline values makes it possible for visual appreciation of contrast enhancement akin to black becoming white after contrast. Ultraharmonic technology is able to attain this degree of sensitivity because it also combines a multiple-pulse transmission technique that induces maximum nonlinearity and causes maximum bubble disruption. It is well recognized that the use of a multiple-pulse technology technique causes pronounced distortion of microbubble morphology and ultimately its disruption at high MIs, which results in high-intensity signals.9,10 In our study, the use of an MI of 1.4 to 1.7 and end-systolic triggering enabled us to attain marked opacification of the myocardium with intravenous contrast agents. Multiple-frame triggering is used for 2 purposes. The first purpose is to confirm that the signals seen in the first triggered frame are from microbubbles and not confounded by either excessive receive gain or tissue signals. The second purpose is to attain complete microbubble destruction so that the microbubble refilling kinetics can be used to quantitate myocardial blood flow.11 In our study, in both dogs and human subjects, ultraharmonic dual-frame trigger produced up to a 62% decrease in VI from the peak VI in the first triggered frame. This degree of efficiency in bubble destruction may be the result of the multiple-pulse technology incorporated in ultraharmonic imaging. Multiple pulses cause maximum bubble destruction and hence a very pronounced contrast effect in the first triggered frame, and in the next triggered frame, multiple pulses cause near
complete destruction of any residual microbubbles. It is possible that we may have attained further microbubble destruction if we had used 2 destruction frames after the first triggered frame rather than 1, as we had done in our studies. However, the second destruction frame may have been as long as 90 to 100 ms after the first triggered frame and therefore may have crossed over into early diastole. Hence, the use of 2 frames, the first representing the triggered fill frame and the second representing the triggered destruction frame as in our study, represents a practical approach to the multiple-frame trigger technique. Another notable feature of ultraharmonic imaging is its ability to combine all the favorable physics with a display with the highest resolution, namely B-mode display. Similar to experimental data, vasodilation with dipyridamole in our study did not produce an additional significant increase in VI. This is consistent with the notion that myocardial blood volume, indicated by peak VI, does not alter significantly after dipyridamole.1,12 The increase in myocardial blood flow after vasodilatation is predominantly the result of an increase in blood velocity rather than in blood volume. For example, in our study, VI in peak VI in ultraharmonic MCE in human beings at the MS was 116 ± 11.9 and 131 ± 12.9 before and after dipyridamole. The difference was statistically not significant. We did not do incremental triggering and thus could not assess flow kinetics. Study Limitations One of the major limitations of this study is that we acquired only the apical 4-chamber view and so we did not study the effect of incremental triggering intervals on regional videointensities. We chose the apical 4-chamber view because we were able to get comparable views in dogs and human beings. Short-
916 Kuersten et al
axis views could not be analyzed because of attenuation caused by contrast in the right ventricle. With respect to other apical views in human beings, we expect similar efficacy data with ultraharmonic imaging. Regional heterogeneity in peak VI is still noticeable with ultraharmonic technology. This reflects the nonuniform power in the transmitted ultrasound field, and the same reason applies to the relative lack of signals in the BL in the apical 4-chamber view.Therefore, in other apical views we expect artifacts in the distal field similar to those seen in the apical 4-chamber view in ultraharmonic imaging. However, it must be noted that the amount of decrease in VI with dual-frame ultraharmonics was relatively comparable across all regions in the apical 4chamber view. Although we have demonstrated the efficacy of ultraharmonic MCE, we did not directly compare its performance against other available techniques in MCE in this study. But it is well recognized that tissue harmonics confound the value of high–MI, triggered conventional B-mode harmonic imaging insofar as visual assessment of myocardial opacification is concerned. Conclusions These preliminary data demonstrate that ultraharmonic MCE is a unique and novel high–MI, triggered technique that combines high-resolution, microbubble-sensitive B-mode images with efficient rejection of tissue and optimal disruption of microbubbles. The resultant increase in the signal-to-noise ratio makes visual assessment of myocardial opacification possible, and the multiple-frame triggered technique may allow more reliable quantification of myocardial blood flow. Direct comparative data of this novel technique with other available MCE techniques and the ability of ultraharmonic MCE to detect abnormal myocardial perfusion patterns remain to be examined in future studies.
Journal of the American Society of Echocardiography September 2001
REFERENCES 1. Wei K, Jayaweera AR , Firoozan S, Linka A, Skyba DM, Kaul S. Quantification of myocardial blood flow with ultrasoundinduced destruction of microbubbles administered as a constant venous infusion. Circulation 1998;97:473-83. 2. Rovai D, L’Abbate A , Lombardi M, Nissen SE, Marzilli M, Distante A, et al. Nonuniformity of the transmural distribution of coronary blood flow during the cardiac cycle. In vivo documentation by contrast echocardiography. Circulation 1989;79:179-87. 3. Lindner JR, Villanueva FS, Dent JM, Wei K, Sklenar J, Kaul S. Assessment of resting perfusion with myocardial contrast echocardiography: theoretical and practical considerations. Am Heart J 2000;139:231-40. 4. Meza MF, Ramee S, Collins T, Stapleton D, Milani RV, Murgo JP, et al. Knowledge of perfusion and contractile reserve improves the predictive value of recovery of regional myocardial function postrevascularization: a study using the combination of myocardial contrast echocardiography and dobutamine echocardiography. Circulation 1997;96:3459-65. 5. Burns PN, Hope SD, Averkiou MA. Nonlinear imaging. Ultrasound Med Biol 2000;26(Suppl 1):S19-22. 6. Shankar PM, Dala KP , Newhouse VL. Advantages of subharmonic over second harmonic backscatter for contrast-to-tissue echo enhancement. Ultrasound Med Biol 1998;24:395-9. 7. Thomas JD, Rubin DN. Tissue harmonic imaging: why does it work? J Am Soc Echocardiogr 1998;11:803-8. 8. Pelberg RA, Wei K, Kamiyama N, Sklenar J, Bin J, Kaul S. Potential advantage of flash echocardiography for digital subtraction of B-mode images acquired during myocardial contrast echocardiography. J Am Soc Echocardiogr 1999;12:85-93. 9. Heinle SK, Noblin J, Goree-Best P, Mello A, Ravad G, Mull S, et al. Assessment of myocardial perfusion by harmonic power Doppler imaging at rest and during adenosine stress: comparison with (99m)Tc-sestamibi SPECT imaging. Circulation 2000;102:55-60. 10. Burns PN. Harmonic imaging with ultrasound contrast agents. Clin Radiol 1996;51(Suppl 1):50-5. 11. Kaul S. Myocardial contrast echocardiography: 15 years of research and development. Circulation 1997;96:3745-60. 12. Wei K, Jayaweera AR, Firoozan S, Linka A, Skyba DM, Kaul S. Basis for detection of stenosis using venous administration of microbubbles during myocardial contrast echocardiography: bolus or continuous infusion? J Am Coll Cardiol 1998; 32:252-60.
Myocardial contrast echocardiography (MCE) with high–mechanical-index (MI), triggered harmonic imaging is the best-established technique to date for the assessment of myocardial perfusion. A high signal-to-noise ratio, which is significantly influenced by precontrast tissue signals, is an important prerequisite. Our goal was to evaluate the efficacy of ultraharmonic MCE, a technique that rejects tissue signals by receiving signals beyond the second but below the third harmonic. Imaging was performed in 6 closed-chest dogs and in 15 healthy volunteers (11 of whom also had dipyridamole stress). Analyses of
M
yocardial contrast echocardiography (MCE) has been shown to effectively assess myocardial microcirculation in experimental and initial clinical studies.1-4 The technique of MCE that is best established to date is high–mechanical-index (MI), triggered harmonic imaging. However, because of the considerable myocardial tissue signal before contrast, background subtraction is often required to enhance the signal-to-noise ratio. This ratio is significantly influenced by precontrast tissue signals because postcontrast enhancement is limited by the bandwidth of the receive elements in the transducer.A number of efforts have been made to reject tissue signals more efficiently while enhancing microbubble signals by using techniques such as pulse-inversion5 and sub-harmonic6 imaging.Another new technique
videointensity (VI) confirmed uniformly low precontrast tissue VI, a significant increase of postcontrast VI (before and after dipyridamole), and a significant decrease in VI after microbubble destruction. We conclude that ultraharmonic MCE produces low precontrast tissue signals, thus optimizing postcontrast myocardial opacification, and exhibits efficient microbubble destruction with use of multipleframe triggering. Thus this new technique opens up a new possibility of further optimizing coronary microcirculation imaging with microbubbles. (J Am Soc Echocardiogr 2001;14:910-6.)
is ultraharmonic imaging, which exploits the differential signals between tissue and microbubbles by receiving the returning signals at a frequency beyond the second harmonic but below the third harmonic frequency of the incident ultrasound waves. This frequency is part of the wideband frequencies that are backscattered by disrupting microbubbles, thus indicating the presence of microbubbles. In contrast, the myocardial tissue returns virtually no signals at this frequency. The aim of this study was to evaluate the efficacy of ultraharmonic imaging for myocardial opacification after intravenous administration of microbubbles.
METHODS Population
From the Department of Internal Medicine, Division of Cardiology, University of Michigan Health System, Ann Arbor, and Agilent Technologies, Andover, Mass (P.R.). Supported by an Agilent Technologies research grant. Reprint requests: Mani Vannan, MBBS, MCP Hahnemann University, School of Medicine, Department of Medicine, Mail Stop 470, 245 N 15th St, Philadelphia, PA (E-mail: [email protected]). Copyright © 2001 by the American Society of Echocardiography. 0894-7317/2001/$35.00 + 0 27/1/113257 doi:10.1067/mje.2001.113257
910
Dogs. Six healthy dogs (4 male and 2 female mongrels) were studied. Their weight ranged from 18.2 to 28.2 kg. The study was approved by the University Committee for the Use and Care of Animals (UCUCA) at the University of Michigan and conformed to the American Heart Association guidelines for the use of animals in research. Human beings. A total of 15 healthy volunteers with no history of coronary artery disease or any other cardiopulmonary symptoms and with good apical echocardiographic
Journal of the American Society of Echocardiography Volume 14 Number 9
Kuersten et al 911
windows were included in the study.The mean age was 27 years (range 21 to 39 years); 11 were men and 4 were women. The protocol was approved by the institutional review board of the University of Michigan Health System, and all subjects signed informed, written consent statements. Ultraharmonic Myocardial Contrast Echocardiography Dogs. The ultrasonic contrast agent MRX-225 (ImaRx Pharmaceutical Corp, Tucson, Ariz) that we used in the 6 dogs is composed of a lipid shell filled with perfluorobutane gas. It was administered by a single slow bolus injection through a peripheral intravenous line, followed by a saline flush. Echocardiographic imaging was performed by placing the closed-chest dogs in left lateral position and transthoracically scanning the 4-chamber view. Human beings. In the human subjects we used intravenous Definity (DuPont Pharmaceuticals Company Medical Imaging, Billerica, MA), a suspension of phospholipidcoated perfluorocarbon microbubbles. It was prepared by diluting 1300 µL of Definity in 25 mL of 0.9% saline solution.This solution was administered as a continuous infusion at a rate of 4.0 to 6.5 mL/min through a peripheral vein with the use of a syringe pump (Razel Scientific Instruments, Stamford, Conn).The maximal total dose was 3900 µL per subject.All human subjects underwent resting MCE, and 11 of them also underwent dipyridamole stress MCE. Images at stress were acquired 4 minutes after termination of the 4-minute infusion of dipyridamole at a rate of 0.56 mg/kg with isometric handgrip. We used a commercially available ultrasonographic platform (Sonos 5500, Agilent Technologies, Andover, Mass) equipped with an ultraharmonics modality and a wideband phased-array transducer (S3). In ultraharmonic imaging the transmitted ultrasound frequency is 1.3 MHz, and the receive frequency is 3.6 MHz. Thus the receive frequency lies between the second and third harmonic frequency of 2.6 and 3.9 MHz, at a point at which the insonated tissue returns only minimal signals while the signals emitted by the contrast agent are still much stronger (Figure 1).
Imaging Protocol We acquired images in the 4-chamber view in all dogs and human beings. In the closed-chest dogs, the transducer was fixed in a stable position with use of a flexible mechanical arm.The dynamic range was set at 50 dB, and the transmit focus was positioned at the apex (upper third of the imaging field). Other instrument settings (gain, compensation, depth, etc) at baseline were optimized for each study but were kept constant throughout any given study to preclude changes in videointensity caused by changes in machine settings. We performed multiple-frame triggered
Figure 1 In ultraharmonic myocardial contrast imaging the contrast agent signals are separated from the tissue signals by receiving at a frequency between the second and third harmonic.
imaging at an MI of 1.4 to 1.7, with the time trigger set at end systole and the triggering interval at every fourth cardiac cycle. MCE Image Analysis For image analysis we digitized the loops from S-VHS videotape with use of Adobe Premiere 5.1 software (Adobe Systems Incorporated, San Jose, Calif). After alignment of the fill and destruction frames of these digitized clips, the videointensities of 6 manually drawn regions of interest (ROI) were quantitated with the software MCE par 2.1 (University of Virginia, Charlottesville, Va). The 6 ROIs were placed in the basal (BS) and mid septum (MS), the apex, the mid (ML) and basal lateral left ventricular wall (BL), and in the cavity. Statistical Analysis The statistical analyses were performed with Analyse-It for Microsoft Excel software (version 1.48, Analyse-It Software Ltd, UK). Videointensity values and values of the differences between videointensities were expressed as mean ± SEM.The differences in the means were compared with a 2-tailed, paired Student t test. Results were considered statistically significant at P < .05. To test for a difference in the means of the different ROIs, we applied a 1way analysis of variance (ANOVA) that computed contrasts with Tukey error protection and a confidence interval of 95%.
RESULTS In the 6 dogs a total of 30 ROIs were analyzed at baseline and a total of 29 ROIs after contrast administration. One BL segment had to be excluded after contrast because of an obvious BL artifact.
Journal of the American Society of Echocardiography September 2001
912 Kuersten et al
Table 1 Videointensities of the regions of interest in the 5 different myocardial segments, in the cavity, and in all 5 wall segments combined in human beings Baseline (n = 14)
Contrast fill frame (n = 15)
Destruction frame (n = 15)
Dipyridamole fill frame (n = 11)
34 (±3.1) 29 (±2.1) 27 (±1.5) 28 (±2.6) 29 (±1.4) 22 (±1.3) 29.7 (±1.10)
92 (±11.6) 116 (±11.9) 95 (±13.5) 75 (±10.7) 59 (±6.8) 201 (±7.4) 87.8 (±5.46)
52 (±8.2) 52 (±6.5) 40 (±4.2) 26 (±3.8) 34 (±4.5) 202 (±6.8) 41.3 (±2.80)
108 (±12.3) 131 (±12.9) 112 (±12.3) 91 (±12.6) 80 (±15.3) 210 (±5.9) 103.6 (±6.12)
Human beings
Basal septal Mid septal Apex Mid lateral Basal lateral Cavity All wall segments
Destruction frame (n = 11)
58 67 49 35 42 205 49.3
(±10.2) (±9.4) (±6.5) (±5.9) (±9.0) (±6.9) (±3.90)
Data are means (±SEM). No statistically significant difference was found between the different regions of interest at baseline, in the fill frames, and in the destruction frames after contrast injection before and after dipyridamole stress, with use of the 1-way-analysis of variance and Tukey error protection.
Table 2 Videointensities of the regions of interest in the 5 different myocardial segments, in the cavity, and in all 5 wall segments combined in dogs Baseline (n = 6)
Dogs
Basal septal Mid septal Apex Mid lateral Basal lateral Cavity All wall segments
27 24 40 26 35 21 30.0
(±3.7) (±1.8) (±7.5) (±2.9) (±7.3) (±1.2) (±2.30)
Contrast fill frame (n = 6)
102 96 74 80 68 207 84.6
(±13.3) (±9.8) (±8.9) (±9.8) (±10.8) (±5.3) (±5.01)
Destruction frame (n = 6)
61 36 39 33 37 211 41.1
(±5.8) (±4.8) (±6.2) (±6.6) (±9.2) (±9.8) (±4.24)
Data are means (±SEM). No statistically significant difference was found between the different regions of interest at baseline, in the fill frame, or in the destruction frame, with use of the 1-way analysis of variance and Tukey error protection.
In the 15 human subjects, a total of 70 ROIs could be analyzed at baseline (in 1 patient the baseline image was not available).After contrast, the presence of lateral wall artifact in 2 subjects meant that a total of 71 ROIs could be analyzed. All 55 ROIs of the 11 subjects who underwent dipyridamole stress MCE were included in the analyses. Baseline Images Before administration of the contrast agent, the ultraharmonic frequency produced by the myocardium resulted in a mean videointensity (VI) of 30.0 ± 2.3 in all myocardial segments in closed-chest dogs and a mean VI of 29.7 ± 1.10 in the human subjects.The absolute VIs in the separate ROIs in dogs ranged from 24 ± 1.8 in the MS to 40 ± 7.5 in the apex and in human subjects from 27 ± 1.5 in the apex to 34 ± 3.1 in the BS (Tables 1 and 2).The VIs among the various ROIs were not significantly different at a confidence interval of 95% in dogs or in human beings (Figures 2 and 3).
Myocardial Opacification with Microbubbles After administration of a myocardial contrast agent, the mean VI of all ROIs combined increased to values of 84.6 ± 5.01 in dogs and 87.8 ± 5.46 in human beings before dipyridamole, and to 103.6 ± 6.12 in human beings after dipyridamole. This increase was highly significant in dogs (P < .0001) as well as in human beings (P < .0001). The analyses of the separate ROIs in dogs after contrast confirmed a significant increase in absolute VI in all segments (BS: P < .005; MS: P < .0007; apex: P < .02; ML: P < .002; BL: P < .03). From baseline the VI increased to a minimum of 68 ± 10.8 in the BL and to a maximum of 102 ± 13.3 in the BS (Table 2, Figure 2). Expressed as percentages, the increase in VI from baseline was 309% ± 79.6% in the BS, 303% ± 47.5% in the MS, 132% ± 40.5% in the apex, 211% ± 36.9% in the ML, and 159% ± 66.9% in the BL. Analysis of variance showed no significant difference in absolute VIs among the various ROIs. In human beings, the opacification of each myocardial ROI before dipyridamole was significant (BS: P < .0001; MS: P < .0001; apex: P < .0003; ML: P < .002; BL: P < .004), with an absolute VI increase ranging from 59 ± 6.8 in the BL to 116 ± 11.9 in the MS. Comparable opacification was obtained with dipyridamole stress with absolute VIs ranging from 80 ± 15.3 in the BL to 131 ± 12.9 in the MS ROI (BS: P < .0001; MS: P < .0001; apex: P < .0001; ML: P < .0003; BL: P < .009), resulting in no significant difference of VI in the ROIs before and after dipyridamole administration (Table 1, Figure 3).The percentage increases of opacification in the ROIs from baseline were 182% ± 28.6% and 225% ± 41.4% (before and after dipyridamole stress, respectively) in the BS, 325% ± 43.2% and 381% ± 53.9% in the MS, 262% ± 54.2% and 334% ± 49.5% in the apex, 223% ± 53.8% and 276% ± 52.6% in the ML, and 104% ± 27.0% and 187% ± 52.7% in the BL.Analysis of variance showed
Journal of the American Society of Echocardiography Volume 14 Number 9
Kuersten et al 913
Figure 2 Myocardial opacification with triggered imaging in a 1 to 4 interval for the 6 dogs in each of the 5 regions of interest. The values of videointensity (VI) of each individual dog are given at baseline and in the fill and destruction frames after contrast application. The red curves represent the mean values in each region of interest. Error bars represent 1 SEM.
a significant difference in the increase of absolute VI between the MS and the BL before dipyridamole. For all other regions, ANOVA did not show any interregional differences in VI. After dipyridamole, no significant differences existed in absolute VI among regions. Destruction of the Microbubbles In the destruction frames, the mean VIs of all 5 ROIs combined decreased to a value of 41.1 ± 4.24 in dogs, to 41.3 ± 2.80 in human beings before dipyridamole stress, and to 49.3 ± 3.90 in human beings after dipyridamole stress. In both dogs (P < .0001) and human beings (P < .0001) this decrease was highly significant. The analyses in dogs confirmed a significant decrease of absolute VI in each myocardial ROI from 33 ± 6.6 in the ML to 61 ± 5.8 in the BS (BS: P < .005; MS: P < .003; apex: P < .0002; ML: P < .0003; BL: P < .04) (Table 2, Figure 2). The mean decrease in absolute VI ranged from a minimum of 31 in the BL to a maximum of 61 in the MS (Figure 2). The percentage decrease of VI in the ROIs from the fill to the destruction frame was 38% ± 4.5% in the BS, 61% ±
6.7% in the MS, 49% ± 3.7% in the apex, 59% ± 4.0% in the ML, and 44% ± 12.0% in the BL. In human beings the absolute VI in the ROIs after microbubble destruction before dipyridamole decreased to near baseline values from 26 ± 3.8 in the ML to 52 ± 8.2 and 52 ± 6.5 in the BS and MS, respectively (BS: P < .0001; MS: P < .0001; apex: P < .0001; ML: P < .0002; BL: P < .0001). Also, after dipyridamole, significantly decreased absolute VI could be measured in the destruction frames, ranging from 35 ± 5.9 in the ML to 67 ± 9.4 in the MS (BS: P < .0001; MS: P < .0001; apex: P < .0001; ML: P < .07; BL: P < .002) (Table 1, Figure 3). The mean decrease in absolute VI in the ROIs ranged from a minimum of 25 and 38 in the BL to a maximum of 64 and 64 (before and after dipyridamole) in the MS. The percentage decrease of VI in the ROIs was 45% ± 3.3% and 49% ± 4.5% (before and at dipyridamole stress, respectively) in the BS, 55% ± 2.8% and 49% ± 4.4% in the MS, 53% ± 4.0% and 55% ± 3.8% in the apex, 62% ± 3.6% and 62% ± 2.5% in the ML, and 40% ± 5.1% and 47% ± 4.6% in the BL. Before dipyridamole, a significant difference in the decrease of VI was seen between the ML and the BS and MS. For all other
914 Kuersten et al
Journal of the American Society of Echocardiography September 2001
Figure 3 Myocardial opacification with triggered imaging in a 1 to 4 interval for the 15 healthy human subjects in each of the 5 regions of interest. The values of videointensity (VI) of each individual subject are given at baseline, in the fill and destruction frames after contrast application, and in the fill and destruction frames after dipyridamole stress. The red curves represent the mean values in each region of interest. Error bars represent 1 SEM.
regions before dipyridamole, ANOVA did not show any interregional differences in VI. After dipyridamole stress, no significant differences in VI existed between regions.
DISCUSSION In this study we have demonstrated that ultraharmonic MCE produces significant and homogeneous myocardial opacification after intravenous injection of contrast agents in closed-chest dogs and in healthy human subjects. Furthermore, we have shown that the precontrast tissue signals with ultraharmonic imaging are negligible and that multiple-frame high– MI triggering with this technique results in efficient bubble destruction (Figure 4). The technique of ultraharmonic imaging combines the dual advantage of an efficient rejection of tissue signals while enhancing the microbubble signals, as described previously in the Methods section. This is an important prerequisite for a technique that relies on high–MI, triggered imaging. At a high MI,
myocardial tissue produces harmonic signals, socalled tissue harmonics, which results in enhanced visualization of the myocardium.5,7 This phenomenon is counterproductive if the aim is to visualize microbubbles. In a B-mode representation, visual discrimination of myocardial enhancement by microbubbles is not reliable and may not be possible in some circumstances. To this end, when using the conventional second harmonic high–MI, triggered technology, it is imperative that precontrast myocardial signals are subtracted from the postcontrast images to reliably assess myocardial opacification.8 Unlike the problem of enhanced tissue harmonics at a high MI, when imaging with ultraharmonic technology, the myocardium in fact appears darker. This is because the returning frequencies are examined beyond the second harmonic of the transmitted frequency but before the third harmonic, at which point myocardial tissue has very few signals. In our study it is noteworthy that the mean VI in all myocardial regions in closed-chest dogs and in healthy human subjects was comparably low. For example, at the MS it was 24 ± 1.8 in dogs and 29 ± 2.1 in human
Journal of the American Society of Echocardiography Volume 14 Number 9
A
Kuersten et al 915
B
Figure 4 Ultraharmonic myocardial contrast imaging in a healthy human volunteer (A) and in a healthy dog (B). The fill frame showing the myocardial enhancement by intravenous myocardial contrast agent (left side) is displayed side by side with the destruction frame (right side).
beings. In dogs, apical precontrast VI was greater than in other regions because of transducer artifact at the apex. This was a constant factor before and after contrast administration. After contrast administration, videointensities increased from 104% to 325% in both closed-chest dogs and in human subjects.The magnitude of this increase in VI over baseline values makes it possible for visual appreciation of contrast enhancement akin to black becoming white after contrast. Ultraharmonic technology is able to attain this degree of sensitivity because it also combines a multiple-pulse transmission technique that induces maximum nonlinearity and causes maximum bubble disruption. It is well recognized that the use of a multiple-pulse technology technique causes pronounced distortion of microbubble morphology and ultimately its disruption at high MIs, which results in high-intensity signals.9,10 In our study, the use of an MI of 1.4 to 1.7 and end-systolic triggering enabled us to attain marked opacification of the myocardium with intravenous contrast agents. Multiple-frame triggering is used for 2 purposes. The first purpose is to confirm that the signals seen in the first triggered frame are from microbubbles and not confounded by either excessive receive gain or tissue signals. The second purpose is to attain complete microbubble destruction so that the microbubble refilling kinetics can be used to quantitate myocardial blood flow.11 In our study, in both dogs and human subjects, ultraharmonic dual-frame trigger produced up to a 62% decrease in VI from the peak VI in the first triggered frame. This degree of efficiency in bubble destruction may be the result of the multiple-pulse technology incorporated in ultraharmonic imaging. Multiple pulses cause maximum bubble destruction and hence a very pronounced contrast effect in the first triggered frame, and in the next triggered frame, multiple pulses cause near
complete destruction of any residual microbubbles. It is possible that we may have attained further microbubble destruction if we had used 2 destruction frames after the first triggered frame rather than 1, as we had done in our studies. However, the second destruction frame may have been as long as 90 to 100 ms after the first triggered frame and therefore may have crossed over into early diastole. Hence, the use of 2 frames, the first representing the triggered fill frame and the second representing the triggered destruction frame as in our study, represents a practical approach to the multiple-frame trigger technique. Another notable feature of ultraharmonic imaging is its ability to combine all the favorable physics with a display with the highest resolution, namely B-mode display. Similar to experimental data, vasodilation with dipyridamole in our study did not produce an additional significant increase in VI. This is consistent with the notion that myocardial blood volume, indicated by peak VI, does not alter significantly after dipyridamole.1,12 The increase in myocardial blood flow after vasodilatation is predominantly the result of an increase in blood velocity rather than in blood volume. For example, in our study, VI in peak VI in ultraharmonic MCE in human beings at the MS was 116 ± 11.9 and 131 ± 12.9 before and after dipyridamole. The difference was statistically not significant. We did not do incremental triggering and thus could not assess flow kinetics. Study Limitations One of the major limitations of this study is that we acquired only the apical 4-chamber view and so we did not study the effect of incremental triggering intervals on regional videointensities. We chose the apical 4-chamber view because we were able to get comparable views in dogs and human beings. Short-
916 Kuersten et al
axis views could not be analyzed because of attenuation caused by contrast in the right ventricle. With respect to other apical views in human beings, we expect similar efficacy data with ultraharmonic imaging. Regional heterogeneity in peak VI is still noticeable with ultraharmonic technology. This reflects the nonuniform power in the transmitted ultrasound field, and the same reason applies to the relative lack of signals in the BL in the apical 4-chamber view.Therefore, in other apical views we expect artifacts in the distal field similar to those seen in the apical 4-chamber view in ultraharmonic imaging. However, it must be noted that the amount of decrease in VI with dual-frame ultraharmonics was relatively comparable across all regions in the apical 4chamber view. Although we have demonstrated the efficacy of ultraharmonic MCE, we did not directly compare its performance against other available techniques in MCE in this study. But it is well recognized that tissue harmonics confound the value of high–MI, triggered conventional B-mode harmonic imaging insofar as visual assessment of myocardial opacification is concerned. Conclusions These preliminary data demonstrate that ultraharmonic MCE is a unique and novel high–MI, triggered technique that combines high-resolution, microbubble-sensitive B-mode images with efficient rejection of tissue and optimal disruption of microbubbles. The resultant increase in the signal-to-noise ratio makes visual assessment of myocardial opacification possible, and the multiple-frame triggered technique may allow more reliable quantification of myocardial blood flow. Direct comparative data of this novel technique with other available MCE techniques and the ability of ultraharmonic MCE to detect abnormal myocardial perfusion patterns remain to be examined in future studies.
Journal of the American Society of Echocardiography September 2001
REFERENCES 1. Wei K, Jayaweera AR , Firoozan S, Linka A, Skyba DM, Kaul S. Quantification of myocardial blood flow with ultrasoundinduced destruction of microbubbles administered as a constant venous infusion. Circulation 1998;97:473-83. 2. Rovai D, L’Abbate A , Lombardi M, Nissen SE, Marzilli M, Distante A, et al. Nonuniformity of the transmural distribution of coronary blood flow during the cardiac cycle. In vivo documentation by contrast echocardiography. Circulation 1989;79:179-87. 3. Lindner JR, Villanueva FS, Dent JM, Wei K, Sklenar J, Kaul S. Assessment of resting perfusion with myocardial contrast echocardiography: theoretical and practical considerations. Am Heart J 2000;139:231-40. 4. Meza MF, Ramee S, Collins T, Stapleton D, Milani RV, Murgo JP, et al. Knowledge of perfusion and contractile reserve improves the predictive value of recovery of regional myocardial function postrevascularization: a study using the combination of myocardial contrast echocardiography and dobutamine echocardiography. Circulation 1997;96:3459-65. 5. Burns PN, Hope SD, Averkiou MA. Nonlinear imaging. Ultrasound Med Biol 2000;26(Suppl 1):S19-22. 6. Shankar PM, Dala KP , Newhouse VL. Advantages of subharmonic over second harmonic backscatter for contrast-to-tissue echo enhancement. Ultrasound Med Biol 1998;24:395-9. 7. Thomas JD, Rubin DN. Tissue harmonic imaging: why does it work? J Am Soc Echocardiogr 1998;11:803-8. 8. Pelberg RA, Wei K, Kamiyama N, Sklenar J, Bin J, Kaul S. Potential advantage of flash echocardiography for digital subtraction of B-mode images acquired during myocardial contrast echocardiography. J Am Soc Echocardiogr 1999;12:85-93. 9. Heinle SK, Noblin J, Goree-Best P, Mello A, Ravad G, Mull S, et al. Assessment of myocardial perfusion by harmonic power Doppler imaging at rest and during adenosine stress: comparison with (99m)Tc-sestamibi SPECT imaging. Circulation 2000;102:55-60. 10. Burns PN. Harmonic imaging with ultrasound contrast agents. Clin Radiol 1996;51(Suppl 1):50-5. 11. Kaul S. Myocardial contrast echocardiography: 15 years of research and development. Circulation 1997;96:3745-60. 12. Wei K, Jayaweera AR, Firoozan S, Linka A, Skyba DM, Kaul S. Basis for detection of stenosis using venous administration of microbubbles during myocardial contrast echocardiography: bolus or continuous infusion? J Am Coll Cardiol 1998; 32:252-60.