Transient Myocardial Contrast After Initial Exposure to Diagnostic Ultrasound Pressures With Minute Doses of Intravenously Injected Microbubbles
Demonstration and Potential Mechanisms
Background We have observed a transient but significant increase in myocardial contrast intensity with intravenously injected perfluorocarbon-exposed sonicated dextrose albumin (PESDA) microbubbles that occurs on initial exposure to pulsed ultrasound (transient-response imaging). The characteristics and magnitude of this response were examined in the present study.
Methods and Results In 14 dogs, the myocardial contrast intensity produced by transient-response imaging (TRI) was compared with conventional 30-Hz imaging (CI) after a 0.005 to 0.030 mL/kg intravenous injection of PESDA. TRI was produced either by measuring myocardial contrast during triggered (1 pulse per cardiac cycle) ultrasound or by withholding real time ultrasound transmission until after microbubbles had entered the myocardium after intravenous injection. Both first-harmonic imaging (2.0 to 3.5 MHz) and second-harmonic imaging (2.0 to 2.5 MHz fundamental, 4.0 to 5.0 MHz received) were used. TRI produced over three times the anterior myocardial contrast intensity of CI (36±12 U TRI versus 11±11 U CI; P<.01), with visually better anterior and posterior myocardial contrast. The spatial extent of myocardial ischemia was easily visualized after intravenous PESDA by use of TRI and correlated closely with risk area as measured with Monastral blue (r=.99, P=.002).
Conclusions TRI produces significantly greater myocardial contrast than CI and may dramatically enhance the ability of intravenous ultrasound contrast agents to identify myocardial perfusion abnormalities.
Myocardial contrast can be produced from intravenous ultrasound contrast agents, but the effect of transducer output and frequency in determining the degree of contrast is unknown. The acoustic output of a diagnostic ultrasound transducer may cause transient changes in microbubble size1 2 3 as well as significant increases in microbubble destruction rate.4 5 6 Either one of these bubble responses would have a significant effect on the scattering properties of an ultrasound contrast agent. Different exposure times to diagnostic ultrasound, therefore, may alter the amount of contrast seen from a given intravenous injection of microbubbles. In the present study, we demonstrate the markedly improved myocardial contrast effect produced by intravenously injected perfluorocarbon-exposed sonicated dextrose albumin (PESDA) microbubbles when exposed to brief, interrupted pulses of ultrasound as opposed to standard ultrasound imaging.
Fourteen mongrel dogs were used for the study. The entire study was approved by the Institutional Animal Care and Use Committee and was in compliance with the position of the American Heart Association on research animal use. Each dog was placed under general anesthesia with intravenous sodium pentobarbital, intubated, and placed on a respirator. A left lateral thoracotomy was performed, and the heart was suspended in a pericardial cradle. Two-millimeter transit-time ultrasound coronary flow cuffs (S-series; Transonics, Inc) were placed around the left anterior descending and left circumflex coronary arteries to monitor coronary flow. A 7F thermodilution pulmonary artery catheter and a 7F pigtail catheter were used to monitor pulmonary artery and left ventricular pressures, respectively.
Ultrasound images in seven dogs were obtained using a 2.0-, 2.5-, or 3.5-MHz ultrasound transducer connected to a commercially available scanner (Hewlett-Packard Sonos 1500). The peak negative pressure generated by these transducers at the acoustic output used in this study at a focal depth of 4 to 6 cm is 0.8 Megapascals, and the spatial peak pulse average intensity is 66 W/cm2. Whereas tissue and blood reflect ultrasound at the transmitted or fundamental frequency, microbubbles undergo both linear and nonlinear oscillations in the presence of ultrasound. These have been referred to as harmonics,7 and in seven dogs in this study, a prototype transducer connected to a Hewlett-Packard Sonos 2500 scanner was used to study this phenomenon. This transducer transmitted at 2.0-MHz (4.0-MHz received frequency) in six dogs and 2.5-MHz (5.0-MHz received frequency) in two dogs.
The transducer was placed in a warm-water bath overlying the anterior surface of the heart and positioned to produce a short-axis view of the left ventricle at the mid–papillary muscle level. A 2-0 silk ligature was placed around either the left anterior descending or left circumflex artery in eight of the dogs to create a critical coronary narrowing (>50% diameter by quantitative angiography). In five of these dogs, this same ligature was also used to create total vessel occlusion.
Both PESDA and room air–containing sonicated dextrose albumin were prepared by methods similar to those previously described.8 Briefly, 8 mL of perfluorobutane or room air were hand-agitated with a 3:1 mixture of 5% dextrose with 5% human albumin. This mixture then underwent electromechanical sonication for 80 seconds.
To test how much better the transient contrast produced by brief or gated exposure to pulsed ultrasound (transient-response imaging, or TRI) was than conventional imaging, the protocol was performed in one of two ways. One method involved measuring the peak myocardial videointensity observed in real time immediately after freezing transducer output for 10 to 60 seconds after intravenous injection of a low dose (0.005 to 0.03 mL/kg) of PESDA (method 1). This interval was computed from a previous test injection that counted the time required for contrast to reach the myocardium after venous injection and for acoustic shadowing in the left ventricular cavity to resolve. A second method of producing transient contrast enhancement was employed by delivering ultrasound pulses gated (triggered) to only one part of the cardiac cycle (method 2).
In six of the dogs in the study, the peak myocardial videointensity produced by TRI of intravenous room air–containing sonicated dextrose albumin was also determined. Four of these determinations were made with second-harmonic imaging, and two were made with first-harmonic imaging.
In 4 dogs, comparisons of anterior (left anterior descending artery distribution) and posterior (left circumflex distribution) peak myocardial videointensity between conventional imaging and TRI were performed by use of the second harmonic, whereas comparisons of TRI with conventional imaging in the first harmonic were performed in 10 dogs. Comparisons of TRI with conventional imaging were made under baseline conditions and during alterations in coronary flow produced by a coronary stenosis (>50% diameter by quantitative angiography) during incremental dobutamine stress in 7-minute stages of 5, 10, 20, and 30 μg · kg−1 · min−1. In 5 dogs, a final injection of PESDA was given after total occlusion of the coronary vessel. The contrast defect observed with TRI in these dogs was compared with the planimetered risk area as determined by use of Monastral blue staining as previously described.9
Background-subtracted peak myocardial videointensity (PMVI) from the midanterior and posterior portions of the myocardium (corresponding to left anterior descending and left circumflex perfusion beds) was measured off-line at end-systole from high-fidelity videotape images. Gray-scale software (Tom-Tec Review Station), which quantifies videointensity (0 to 255 scale) versus time, was used to measure the contrast intensity. Instrument settings, as well as intravenous PESDA dose, were kept constant in each dog for each comparison. Visual grading of myocardial contrast enhancement produced by TRI versus conventional imaging in both the anterior and posterior myocardium was assessed in eight dogs using a scale of 0 (no contrast enhancement), 1+ (mild contrast enhancement), and 2+ (bright contrast enhancement). Although these measurements were made in an unblinded manner, interobserver variability on these visual measurements was determined by comparing myocardial contrast intensity as measured by two independent, experienced observers following 12 injections in three dogs. The comparison of PMVI produced with TRI versus conventional imaging was evaluated by use of a Mann-Whitney rank sum test. When second-harmonic imaging was available, comparisons of PMVI produced by conventional imaging and TRI with the first versus the second harmonic were made by use of ANOVA. Comparisons of the myocardial contrast produced by room air–containing sonicated dextrose albumin and PESDA with TRI were made by use of unpaired t testing. Contingency tables were created to compare differences in visual grading of myocardial contrast. Linear regression was used to compare area at risk with TRI and risk area as determined by use of Monastral blue and to compare coronary flow with PMVI using TRI.
A total of 31 comparisons (62 injections) of TRI with conventional imaging were made with intravenous PESDA under baseline conditions. None of these injections resulted in any significant change in heart rate, left ventricular systolic and end-diastolic pressure, or mean pulmonary artery pressure. Cardiac output also did not change after any injection. For these 31 comparisons, the PMVI produced with TRI was over three times greater than that produced by conventional imaging (36±12 versus 11±11, respectively; P<.0001).
In the dogs in which second-harmonic imaging was available, TRI made by use of the first versus the second harmonic was compared with conventional first-harmonic and second-harmonic imaging (Table⇓). Fig 1⇓ shows sonograms taken after an intravenous injection of PESDA at peak intensity made by use of conventional imaging compared with TRI under baseline conditions. In both first- and second-harmonic imaging, the myocardial contrast seen with TRI was three to seven times greater than conventional 30-Hz frame rate imaging (Table⇓). The low dose of PESDA required to produce this contrast with TRI (0.005 to 0.01 mL/kg) also prevented acoustic shadowing from left ventricular cavity contrast. When the peak anterior myocardial videointensity produced by TRI in the first and second harmonic were compared, there was significantly higher contrast with the second harmonic (23±8 for first-harmonic TRI versus 60±25 for second-harmonic TRI; P<.001).
The duration of transient myocardial contrast in real time using method 1 averaged 3.3±2.4 seconds (Fig 2A⇓). When method 2 was used (triggered pulses once per cardiac cycle), the peak myocardial contrast was just as high as for method 1 (47±19 U for method 2 versus 46±23 U for method 1; n=16 comparisons) but was demonstrable for the entire injection period. An example of this is shown in Fig 2⇓, in which the same dose of PESDA is given twice in the same dog, once using method 1 and once using method 2. Because method 2 delivered a frame each cardiac cycle beginning immediately after injection, the increased contrast with TRI could be demonstrated earlier and with longer duration.
The visual grade of anterior myocardial contrast after intravenous PESDA was 2+ after 9 of 18 injections made by use of first-harmonic TRI and after all 8 of 8 injections made by use of second-harmonic TRI. When conventional first-harmonic imaging was used, 2+ myocardial contrast was not observed after any injection using the first harmonic and was seen after only 3 of 9 injections with second-harmonic conventional imaging (P<.001 by Fisher’s exact test). The two independent reviewers agreed on the degree of visual myocardial contrast in 10 of 12 random comparisons.
Intravenous room air–containing sonicated dextrose albumin (up to 8 mL) did not produce myocardial contrast when first-harmonic TRI was used or with conventional first- or second-harmonic imaging. In three of the four dogs when room air–containing microbubbles were given using second-harmonic TRI, however, visually evident anterior myocardial contrast was produced. The anterior myocardial PMVI produced with 8-mL intravenous injections of room air–containing sonicated dextrose albumin (34±24 U) was still significantly smaller than that produced with 0.005 to 0.01 mL/kg intravenous PESDA in these same dogs (83±16 U; P<.0001).
TRI During Myocardial Ischemia
A total of 21 additional intravenous injections were given to test the ability of TRI with intravenous PESDA to quantify coronary flow changes. The range of coronary flows in this group of dogs ranged from 0 (during total occlusion) to 172 mL/min (peak dobutamine stress). The range of peak myocardial videointensities after intravenous PESDA with TRI under these conditions ranged from 0 to 100 U. There was a good correlation within dogs between peak videointensity as measured by TRI and coronary flow (mean correlation coefficient, .87; range, 0.79 to 0.94). Despite the higher peak videointensity seen with second-harmonic TRI, the correlation with coronary flow was good for both harmonics (r=.87, first harmonic; r=.87, second harmonic; P<.01 for both).
In the five dogs in which Monastral blue was given during acute ischemia, the stain could be visualized in four. In one dog, the Monastral blue did not stain any region of myocardium, and this dog was excluded from analysis. In the remaining four dogs, there was a close correlation between risk area as measured with Monastral blue and that measured with TRI (r=.99; P<.001; range of risk areas, 4.0 to 7.4 cm2). Fig 3⇓ demonstrates area at risk using first- and second-harmonic TRI in two dogs during left circumflex ischemia.
In the present study, we demonstrated that very low intravenous doses of PESDA (0.005 to 0.01 mL/kg) could produce dramatically greater myocardial contrast than conventional imaging when ultrasound pulses were not delivered until microbubbles had entered the myocardium and left ventricular cavity acoustic shadowing had resolved after intravenous injection. The duration of this contrast with real-time imaging was only 3.3±2.4 seconds but could be observed throughout the injection period when ultrasound was delivered at only one point in the cardiac cycle (Fig 2⇑). Furthermore, we demonstrated that second-harmonic imaging could improve this transient contrast response even further. Second-harmonic imaging improves the ratio of blood echo intensity to tissue echo intensity7 and thus significantly improves the signal-to–background noise ratio. These advantages were evident in the present study, in which we observed that background-subtracted PMVI produced by TRI in the second harmonic was significantly higher than in first-harmonic imaging (Table⇑).
The significantly increased myocardial contrast with TRI may occur for one of two reasons. The first possibility is that transient cavitation is occurring, similar to that which has been seen in vitro with sonicated albumin. Transient increases in sonicated albumin microbubble size occur in the presence of conventional ultrasound transmission.10 These increases in microbubble size have been brief (milliseconds in duration), followed by a return to steady state. Since microbubble reflectance is exponentially related to its size, this would result in transient contrast enhancement. Although the duration of transient cavitation observed in vitro with sonicated albumin was much shorter than the duration of increased contrast we observed, there are two potential reasons why transient cavitation may have lasted longer with the microbubbles used in the present study. First, we could not detect transient contrast in the first harmonic with our room air–containing microbubbles but were able to identify it with fluorocarbon-containing microbubbles. It is possible that these less-soluble gases may prolong the duration of transient cavitation. Next, harmonic resonances have been observed with microbubbles that transiently cavitate in response to diagnostic ultrasound.2 This may explain why we also observed transient myocardial contrast when a second-harmonic frequency was used (Table⇑).
Another explanation for the transient contrast may be that standard-imaging pulse rates destroy the microbubbles. By delaying ultrasound transmission or triggering pulses to one part of the cardiac cycle, more bubbles may reach the myocardium and hence cause greater contrast. Microbubble destruction in response to diagnostic ultrasound pressures has been observed with sonicated albumin4 5 6 but has not been detected with fluorocarbon-containing microbubbles. To distinguish between these two potential mechanisms for TRI, passive or active acoustic detection systems will need to be used that can detect whether transient cavitation of microbubbles is occurring.
TRI with intravenous ultrasound contrast agents could potentially overcome three limitations that presently exist with intravenous ultrasound contrast agents. First, it produces dramatically better myocardial contrast than what is achieved with conventional imaging. Second, the doses required to achieve myocardial contrast are lower and thus safer than what is required for conventional imaging. Finally, because much lower doses of intravenous contrast are required, it reduces the amount of posterior myocardial attenuation caused by left ventricular cavity contrast.
The peak myocardial contrast produced by TRI was able to delineate resting and stress-induced coronary flow changes (Fig 3⇑). Peak myocardial videointensity with TRI correlated with flow in these situations because changes in flow were induced by changes in myocardial blood volume, and it has been demonstrated that peak videointensity correlates more closely with myocardial blood volume than with flow.11 Additionally, the improved contrast signal obtained with TRI could be used to visualize more easily the risk area or regions of relative hypoperfusion and thus improve the detection of coronary artery disease and quantification of the spatial extent of ischemic burden.12 It may also delineate regions of no reflow after successful reperfusion and thus improve previous methods of detecting myocardial viability with intravenously injected ultrasound contrast.13
In conclusion, dramatically improved myocardial ultrasound contrast can be produced with very low doses of intravenous PESDA by use of TRI. Further in vivo study is needed to determine the mechanism of this phenomenon, as well as to apply this imaging modality to transthoracic imaging in humans.
The authors would like to acknowledge David Kricsfeld for his assistance with data analysis, Karen Reinert for secretarial assistance, and Hewlett-Packard Co for engineering assistance with this manuscript.
- Received June 29, 1995.
- Revision received September 6, 1995.
- Accepted September 8, 1995.
- Copyright © 1995 by American Heart Association
Holt RG, Crum LA. Acoustically forced oscillations of air bubbles in water: experimental results. J Acoust Soc Am. 1992;91(part 1):1924-1932.
Mor-Avi V, Robinson K, Shroff S, Lang RM. Stability of albunex microspheres under ultrasonic irradiation: an in vitro study. J Am Soc Echocardiogr. 1994;7:S29.
Wray RA, Zoghbi WA, Quinones MA, Cheirif J. Contrast echocardiography: relation of acoustic power and time gain compensation to contrast intensity duration. J Am Soc Echocardiogr. 1992;4:286.
Roy RA, Church CC, Calabrese A. Cavitation produced by short pulses of ultrasound. In: Hamilton AF, Blackstock DT, eds. Frontiers of Nonlinear Acoustics: Proceedings of 12th International Symposium of Nonlinear Acoustics. London: Elsevier Science Publishing Co; 1990:476-481.
Skyba DM, Jayaweera AR, Goodman NC. Quantification of myocardial contrast echocardiography during left atrial injection of contrast: implications for venous injection. Circulation. 1994;90:1513.
Ismail S, Jayaweera AR, Goodman NC, Camarano GP, Skyba DM, Kaul S. Detection of coronary stenoses and quantification of the degree and spatial extent of blood flow mismatch during coronary hyperemia with myocardial contrast echocardiography. Circulation. 1995;91:821-830.
Villanueva FS, Glasheen WP, Sklenar J, Kaul S. Characterization of spatial patterns of flow within the reperfused myocardium by myocardial contrast echocardiography: implications in determining extent of myocardial salvage. Circulation. 1993;88:2596-2606.