Relation Between Air-Filled Albumin Microbubble and Red Blood Cell Rheology in the Human Myocardium
Influence of Echocardiographic Systems and Chest Wall Attenuation
Background We have previously shown that the intravascular rheology of sonicated air-filled albumin microbubbles is similar to that of red blood cells (RBCs) and that their myocardial transit rate is also similar to that of RBCs in the beating canine heart. In the present study, we tested the hypothesis that the myocardial transit rates of these microbubbles reflect those of RBCs in humans at different coronary flow rates.
Methods and Results RBC and microbubble transit rates were measured in 17 patients undergoing coronary angiography: in 8, measurements were made only at rest, whereas in 9, they were performed both at rest and during a pacing-induced increase in coronary blood flow. A γ-variate function was used to derive mean RBC and microbubble transit rates from the time-activity and time-intensity plots after the left main injection of RBCs and microbubbles, respectively. There was linear correlation between the myocardial transit rates with both tracers with the slope of the correlation determined by the specific echocardiographic system that was used. Microbubble transit rate consistently overestimated RBC transit rate due to artificial narrowing of the time-intensity curves caused by chest wall attenuation of the echocardiographic signal, which was confirmed through in vitro experiments.
Conclusions There is close correlation between air-filled albumin microbubbles and RBC rheology in the human myocardium. The use of these microbubbles in the cardiac catheterization laboratory could, therefore, provide further insights into myocardial blood flow/myocardial blood volume relations in humans.
Myocardial contrast echocardiography (MCE) offers real-time assessment of myocardial perfusion.1 In humans, microbubbles can be injected directly into a coronary artery to define the spatial distribution of myocardial perfusion.2 3 In addition, the transit rate of air-filled albumin microbubbles through the myocardium can provide a quantitative estimation of the myocardial blood flow/blood volume relation.4
Microbubbles made by sonicating 5% human albumin solution have been shown to have no adverse effects on coronary blood flow and systemic hemodynamics when injected directly into the coronary circulation in both animals5 and humans.6 With high-powered microscopy, we have previously demonstrated that the intravascular rheology of these microbubbles is similar to that of RBCs in the microcirculation.7 More recently, we have shown that the mean transit rate of these microbubbles derived with the use of MCE is similar to that of RBCs in the beating canine myocardium.8 In the present study, we therefore hypothesized that the mean transit rates of these microbubbles would reflect RBC transit rates in the human myocardium during different coronary flow rates.
Patient Population and Clinical Protocol
Twenty-four patients undergoing diagnostic cardiac catheterization were enrolled in the present study. The study was approved by both the US Food and Drug Administration and the Human Investigation Committee at the University of Virginia. All patients provided written informed consent for participation in the study.
After diagnostic catheterization, the left main artery was engaged with a 7F Judkins catheter, the proximal end of which was connected to a power injector (model 3000, Liebel Flarsheim) via 2-mL saline-filled plastic tubing. The miniature γ-probe was placed on the anterior chest wall. The coronary catheter was then gently hand-filled with radiolabeled RBCs, and the entire contents of the tubing were immediately injected into the left main artery over 1 second (injection rate of 3 mL/s) with the power injector while the time-activity data were recorded. The left main artery was then disengaged.
Five minutes later, the left main artery was reengaged, 1 mL of microbubble solution was injected into it using the same technique as that used for RBCs, and the data were recorded. Images were evaluated on-line, and if contrast enhancement was judged to be poor, another injection of 2 mL of the microbubble solution was made. Conversely, if attenuation was judged to be present, the dose of microbubble solution was reduced to 0.5 mL. The duration of injection and, therefore, the input function in all instances were constant (1 second) and were identical to those of the radiolabeled RBCs.
In nine patients, coronary blood flow was increased through atrial pacing. A pacemaker lead (Hemaquet II 6468, CR Bard, Inc) was introduced through a 6F right femoral vein sheath, and its tip was positioned in the right atrium. This lead was attached to a demand pulse generator (model 5375, Medtronic Interventional Vascular Inc). The left main artery was engaged, and pacing was initiated at ≈150% of the patient's resting heart rate. One minute later, either radiolabeled RBCs or microbubbles were injected into the left main artery, and data were acquired. The left main artery was disengaged; 3 minutes later, the above procedure was repeated with either radiolabeled RBCs or microbubbles, whichever was not used the first time. Pacing was discontinued, and the pacemaker lead was removed. By protocol design, no more than three doses of microbubbles were injected into each patient.
In Vitro Experiment
In vitro experiments were performed to study the influence of the thickness of the interposing tissue between the transducer and the region from which the time-intensity curves were generated. The two ends of a segment of the femoral vein from a dog were connected to plastic tubing as previously described.9 The vein segment was immersed in a beaker of saline, and the echocardiographic transducer was fixed above it, also within the saline solution. Flow of saline through this system was set at a rate of 30 mL/min with a microprocessor-based peristaltic roller pump (system 77, Harvard Apparatus).
At each stage, 0.5 mL of the microbubble solution was introduced into the tubing via a side arm over 0.3 second with the use of the power injector, and echocardiographic data were recorded. Six such injections were performed, the first one without any interposing tissue between the transducer and the bath. At each subsequent injection, one or more pieces of left ventricular myocardium from a dog were juxtaposed between the transducer and the saline, without changing any instrument settings. This step was undertaken to simulate various thicknesses of tissue that normally interpose between the transducer and the heart in different patients. The width of the myocardial pieces exceeded the footprint of the transducer.
MCE was performed in patients with 1 of 3 echocardiographic systems: (1) RT5000 with a 5-MHz transducer (General Electric Medical Systems), (2) 128XP with a 3-MHz transducer (Acuson Corp), or (3) ND256 with a 5-MHz transducer (Biosound Corp). The first two systems are phased-array, whereas the third is mechanical. Different systems were used because of the known differences in their microbubble concentration–versus–video-intensity relations and the effect that these relations have on the estimation of mean microbubble transit rates.10 One of the aims of the present study was to determine the influence of various echocardiographic systems on the estimation of mean microbubble transit rates in humans. For the in vitro experiment, we used the GT5000 system.
The patients were positioned in the left lateral decubitus position, and transthoracic images of the left ventricle were obtained with either the parasternal long-axis or the midpapillary muscle short-axis view. A depth setting of 8 cm was used to focus on the anterior wall or the interventricular septum; with the Acuson system, the high-resolution mode was used for the same purpose. Log compression was optimized at the beginning of the examination to provide as linear a relation as possible between microbubble concentration and video intensity for the specific echocardiographic system that was used. The gain and power settings (maximizing gain and minimizing power) were adjusted individually for each patient for optimal image quality and then were held constant throughout. The same procedure was used for optimizing the system settings in the in vitro experiment.
Albunex (Molecular Biosystems, Inc) was used as the microbubble solution for contrast echocardiography.5 This agent has microbubbles with a mean diameter of 4.3±0.3 μm (<90% of microbubbles are <10 μm in diameter). The images were recorded on 1.25-cm VHS videotape.
Off-line analysis of the contrast echocardiographic images was performed with a dedicated computer (Mipron, Kontron Electronics) with methods that we have described previously.4 In brief, the steps involved are (1) transfer of data from videotape to video memory of the computer, (2) identification of frames from four or five frames before contrast appearance until its disappearance (end-diastolic images for patient studies and all consecutive images for the in vitro experiment), (3) alignment of the end-diastolic images from the patient studies with the use of automatic cross-correlation (because there was no motion of the vein in the in vitro experiment, alignment was not necessary), and (4) placement of a region-of-interest (ROI) of ≈4000 pixels for the derivation of a time-intensity plot from the images selected from the injection sequence. In the clinical studies, the ROI was placed over the anterior myocardium or the anterior interventricular septum, whereas in the in vitro experiment, it was placed over the image of the vein. The mean pixel intensity in the images before contrast appearance in the myocardium was considered to represent background.
Measurement of RBC Transit Rate in Patients
A miniature γ-probe consisting of a CsI2 scintillation crystal optically coupled to a photodiode and a 1-cm-long converging collimator (Oakfield Instruments) was used for the measurement of RBC transit rates.8 11 The photodiode is connected to a preamplifier/amplifier unit that was designed to detect the 99mTc photopeak and connected to a personal computer via an RS-232 interface. This system samples every 0.5 second, and its relation between activity and response rate is linear below ≈20 000 counts per 0.5 second, which is the range in which data were collected. The probe was placed on the anterior chest wall, and its position in relation to the anterior surface of the heart was confirmed with fluoroscopy.
Five milliliters of venous blood was obtained from each patient 1 hour before cardiac catheterization, and it was centrifuged to separate the RBCs, which were labeled ex vivo with 99mTc.12 These RBCs were washed several times, so that free unbound 99mTc constituted <0.5% of the entire activity within the RBC suspension. RBCs, constituting ≈100 μCi of radioactivity, were injected into the left main artery to obtain time-activity plots from the myocardium. The volume of injectate and the input function were the same as those of Albunex during that stage.
Data were analyzed with RS/1 software (Bolt, Beranek, and Newman) with a minicomputer (VAX4000-system 90, Digital Equipment Corp). A γ-variate function (y=Ate−αt) with a least-squares fit was applied to the background-subtracted RBC time-activity and MCE time-intensity plots, where A is a scaling factor, t is time, and α/2 represents the mean transit rate.13 The correlations between RBC and microbubble transit rates were performed with linear regression analysis (least-squares fit). The null hypothesis that the intercept of the correlation is equal to zero was tested with a t test.
Of the 24 patients entered in the study, 9 were women and 15 were men, and their ages ranged from 40 to 78 years. Eight patients had normal coronary angiograms; one each had only right coronary and left circumflex artery disease, respectively; and the remaining 14 had left anterior descending coronary artery stenosis (mean diameter narrowing, 75±16%). Of patients with left anterior descending coronary artery disease, 4 had three-vessel, 8 had two-vessel, and 2 had one-vessel disease. The mean left ventricular ejection fraction of these 24 patients was 0.62±0.09.
A total of 36 radiolabeled RBC and 60 microbubble injections were made in the 24 patients. None of the patients experienced adverse effects during or after the injections. The ECGs in all patients remained unchanged at 30 minutes and 1 day later.
Time-activity plots were obtained from 23 of the 24 patients; in 1 patient, data were not acquired due to technical difficulties with the miniature probe. MCE images of sufficient quality to allow derivation of time-intensity curves were obtained from 17 patients. In 4 patients, the baseline and all subsequent images were of poor quality; in 1 patient, suboptimal enhancement was achieved at all of the microbubble doses; and in 2 patients, attenuation occurred even at the smallest dose. Adequate myocardial opacification in the other 17 patients was achieved with a mean microbubble dose of 0.9±0.2 mL.
Fig 1⇓ illustrates the anterior portion of the heart in a modified long-axis view with the high-resolution feature of the Acuson system. Fig 1A⇓ depicts the interventricular septum before injection of contrast, whereas Fig 1B⇓ and 1C⇓ illustrates the septum at maximal contrast effect during baseline and pacing, respectively. Fig 1D⇓ illustrates the method of placement of the ROI to obtain time-intensity plots. Increased contrast effect and increased transit rate of microbubbles were consistently observed visually during pacing compared with baseline.
Fig 2⇓ depicts representative RBC time-activity and microbubble time-intensity plots obtained from the anterior myocardium of one patient. It is evident that the microbubble time-intensity plot is narrower than the RBC time-activity plot. Fig 3⇓ shows the correlation between the mean RBC (x axis) and microbubble (y axis) transit rates for all injections with the three echocardiographic systems. There is a linear correlation between the RBC and microbubble transit rates with each system, but the slope of the correlation is different for each system. It is closest to the line of identity for the Biosound system, almost 1.5-fold greater for the Acuson system, and approximately twofold greater for the General Electric system. The correlations for all three systems have intercepts that tend to be higher than zero, although because of small numbers of patients, this does not reach statistical significance. Because of this effect, as shown in Table 1,⇓ the microbubble transit rate overestimated the RBC transit rate for all three echocardiographic systems, including the Biosound system.
To determine the reason for the offset in the intercept, we performed in vitro studies with interposition of tissue of different dimensions between the transducer and the ROI. Fig 4⇓ illustrates the influence of the thickness of the interposing tissue (x axis) on the derivation of mean microbubble transit rates (y axis). Thus, at the same flow rate and with the same input function and dose of microbubbles, the mean microbubble transit rate increased as the thickness of the interposing tissue increased.
Table 2⇓ depicts baseline and pacing data in eight of the nine patients in whom complete data were obtained both at baseline and during pacing. One patient developed angina during pacing, and it was decided not to repeat the microbubble injection. The heart rate during pacing increased by 38±27% compared with baseline. There was no significant change in the diastolic or systolic blood pressure, and the double product increased by 36±28%. Although during both baseline and pacing the mean microbubble transit rate overestimated the mean RBC transit rate, the percent increases in transit rates with the use of both methods were equivalent and of the same magnitude as the increase in the heart rate and double product (37±15% for radiolabeled RBCs and 40±55% for sonicated albumin microbubbles).
This study demonstrates, for the first time, that the myocardial rheology of air-filled albumin microbubbles correlates with that of RBCs in humans at different coronary flow rates. The slope of the correlation between microbubble and RBC transit rates is influenced by the echocardiographic system used. Microbubbles consistently overestimate RBC transit rates due to artifi-cial narrowing of the MCE time-intensity curves caused by chest wall attenuation of the echocardiographic signal.
Comparison With Experimental Animal Data
Although the results of this study are similar to those of an experimental study performed in dogs in which we also found a close correlation between microbubble and RBC transit rates,8 there are certain notable differences. In our experimental study, we used the Biosound system and found that the slope of the correlation between the microbubble and RBC transit rates was close to unity.8 In the present study, we found the same results when the Biosound system was used; however, for the other two systems, the slopes of the correlations were greater than unity (Fig 3⇑). These results are similar to those of our in vitro experiments in which we found that the transit rates obtained with the Biosound system were more accurate than those obtained with the General Electric system.10 In those experiments, we explained these disparities based on the differences in the shapes of the microbubble concentration–versus–video-intensity relations between the various echocardiographic systems.10 The more linear the relation, the better was the prediction of the actual transit rate. In those experiments, we did not use the Acuson system.
In our experimental canine study, the intercept of the regression line of the RBC-versus-microbubble transit rates was close to zero, which did not occur in the present study, even when the echocardiographic system that was used in the experimental study was used (Biosound). On the basis of this phenomenon alone, for any given RBC transit rate, the microbubble transit rate appeared to be faster in the present study for each of the three echocardiographic systems regardless of the slope of the specific relations between RBC and microbubble transit rates, as discussed (Fig 3⇑).
There are three possible explanations for the offset in the intercept seen in Fig 3⇑. First, the mean size of the microbubbles (4.3 μm) is significantly smaller than that of RBCs in humans (8.3 μm). Because smaller intravascular particles have a more axial distribution within arterioles and have higher velocities than those on the margins of the vessels,14 15 the mean microbubble transit rate could be higher than that of RBCs. The magnitude of the difference between RBC and microbubble transit rates noted in this study cannot, however, be explained on the basis of this phenomenon alone.
The second reason could be the stability of microbubbles. Because the sonicated albumin microbubbles used in this study are composed of air, which is highly diffusible, they can shrink, particularly at high ambient pressures present within the myocardium, particularly in systole.16 Because the backscatter from microbubbles is related to the sixth power of their radius,17 even a small change in size could reduce backscatter and potentially make the time-intensity curve narrower.8 Microbubbles can also be destroyed rapidly when imaged at their resonant frequency.18 In this study, microbubbles were introduced proximal to the site of imaging (left main artery), while imaging was performed over a single cross section of the left ventricular myocardium. They were, therefore, exposed very briefly to ultrasound and were unlikely to be affected by it.
The third and most probable reason for the overestimation of RBC transit rates by microbubbles may be the “thresholding effect” of the echocardiographic systems that were used.19 20 Ideally, for any given flow, the transit rate of a tracer should be constant regardless of the amount injected. The low sensitivity of echocardiographic systems, however, results in the lower portion of the time-intensity curves not being detected. Consequently, as we and others demonstrated, the curves appear narrower at lower doses of microbubbles, which results in an artificial increase in the measured microbubble transit rate (Fig 5A⇓).19 20 Interposition of tissue between the transducer and the ROI results in an “apparent” decrease in the number of microbubbles due to attenuation that caused an artificial narrowing of the time-intensity curve and an increase in microbubble transit rate.
Identical results have been reported previously in a similar in vitro model.21 In that study, when system gain was increased in an attempt to increase microbubble signal, the microbubble transit rate was not affected,21 because by increasing system gain, both the background and the microbubble signals were increased with no effect on the signal-to-noise ratio (Fig 5B⇑ and 5C⇑). Increasing the amount of injected microbubbles also does not overcome the effect of tissue attenuation on mean microbubble transit rate because at higher doses the relation between microbubble concentration and video intensity becomes nonlinear (Fig 6A⇓), which can cause error in the calculation of mean transit rates (Fig 6B⇓).9 It is conceivable that instruments with better sensitivity, such as those with harmonic imaging,18 may improve the signal-to-noise ratio, resulting in decreased overestimation of mean transit rates.
Thus, unlike the experimental setting, in which open-chest dogs are used, machine factors and chest wall configuration are likely to influence the calculation of microbubble transit rates in humans with the use of MCE. Nevertheless, as indicated by the results of our pacing studies, changes in RBC transit rates are paralleled by changes in microbubble transit rates as long as the echocardiographic system remains unchanged. Our results with air-filled sonicated albumin microbubbles cannot, however, be extrapolated to either similar bubbles filled with other gases22 23 or bubbles made through other methods24 since the intravascular rheology of these bubbles may not mimic that of RBCs during intracoronary injection.
We were unable to obtain data of adequate quality in approximately one fourth of our patients, which is a major limitation of transthoracic echocardiography, particularly in the cardiac catheterization laboratory. It is possible that evolving ultrasound techniques, such as intracardiac echocardiography,25 will overcome this problem in the future.
When we injected microbubbles and RBCs into the left main coronary artery, they traveled to both the left anterior descending and left circumflex beds. Although we derived the microbubble time-intensity data from the anterior myocardium, we have no method of ensuring that our time-activity data for the RBCs were not contaminated from counts from the left circumflex bed. We believe this to be unlikely, however, because the activity from the posteriorly placed left circumflex bed would be negligible (it falls rapidly based on the inverse square law) compared with that from the left anterior descending bed. Because the collimator design prevents photons from being detected anywhere except the myocardium just below it, photons from the lateral and medial aspects of the left ventricle also would not contribute to the activity detected by the probe. The excellent correlation between the microbubble and RBC transit rates also argues against contamination of counts from other beds.
Because of the limit set on the number of injections that could be performed in each patient, we could not measure interinjection variability for the derivation of tracer transit rates. We have, however, shown both methods to be reproducible in the experimental laboratory.8 As expected, the reproducibility depends on the size of the ROI, sampling rate, and mathematical method used to derive the mean transit rate.9 In addition, image quality can significantly affect reproducibility, which is of particular concern with transthoracic echocardiography.
Although the close correlation between microbubble transit rate and coronary or myocardial blood flow has been previously described in animal models,26 27 there is controversy relating to this issue, particularly when air-filled microbubbles are used.28 The exact reasons for inconsistent results between different studies are not clear but may be related to the handling of these fragile air-filled microbubbles. For example, in our study, when the microbubbles were introduced into the tubing, they were not exposed to the coronary artery pressure until just before injection. The stopcock connecting the tubing to the catheter placed in the left main coronary artery was opened just before injection; not doing so can result in albumin microbubble destruction caused by high pressures.16
In the present study, we did not achieve very high flow rates since we increased the double product by only ≈40%. We therefore do not know the relation between RBC and microbubble transit rates at higher flow rates in humans. Our experimental results, however, indicate that the relation should remain linear at higher flows. The estimation of microbubble transit rates becomes more difficult at very high flows because the rapid myocardial transit of microbubbles results in only a few data points being available for curve fitting.8
Implications for Assessment of Myocardial Perfusion in the Cardiac Catheterization Laboratory
There is no routinely used myocardial perfusion imaging tracer that remains entirely within the intravascular space. In nuclear cardiology, the imaging tracer is extracted by myocytes (201Tl or sestamibi29 30 ), whereas for radiographic imaging (digital subtraction angiography and cine-computed tomography31 32 ), the radiopaque contrast agents enter the extravascular space. In comparison, air-filled albumin microbubbles act as true intravascular tracers that remain within the intravascular space during their transit through the myocardium.5 Their transit rate through the myocardium, therefore, reflects the myocardial blood flow/volume relation.4 Furthermore, this relation pertains to nutrient perfusion regardless of the source of blood flow (anterograde or collateral).
Despite the central role of myocardial blood volume in regulating myocardial blood flow,31 32 33 there is no technique that can measure myocardial blood volume in the catheterization laboratory. In a situation in which blood flow is measured with techniques such as Doppler flow wire, microbubble transit rates can provide an estimation of concomitant changes in myocardial blood volume and lead to a better understanding of coronary bed responsiveness under different physiological conditions and in different disease states. In the present study, for example, the increase in coronary blood flow resulted solely from an increase in myocardial oxygen consumption caused by an increase in heart rate. The increase in blood flow probably occurred due to an increase in blood volume mediated by metabolic and biochemical factors.33 Measurement of both myocardial blood flow and myocardial blood volume would therefore have provided more complete insight into the effect of pacing on myocardial perfusion.
In addition to blood flow through and blood volume of the myocardium in question, the transit rate of a tracer is influenced by the input function, which is the width of the bolus that the myocardium “sees.”4 We have previously shown that as long as the site and duration of injection remain constant, changes in transit rates correlate with changes in the myocardial blood flow/volume relations.8 Standardized injection protocols will therefore have to be implemented in the cardiac catheterization laboratory to allow meaningful information to be obtained from myocardial transit rates of air-filled albumin microbubbles. Attention to instrument settings and image quality as well as care in data acquisition will also determine the usefulness of the technique. Further studies are required to determine the clinical usefulness of measurement of microbubble transit rates in the cardiac catheterization laboratory in humans.
This work was supported in part by grants from the National Institutes of Health, Bethesda, Md (R01-HL-48890); the American Heart Association, Virginia Affiliate, Glen Allen, Va; and Molecular Biosystems, Inc, San Diego, Calif; and by an equipment grant from General Electric Medical Systems, Milwaukee, Wis. Dr Ismail was the recipient of a Fellowship Training Grant from the American Heart Association, Virginia Affiliate, and Dr. Camarano was the recipient of a fellowship training grant from the Mallinckrodt Corp, St Louis, Mo. Dr Kaul is an Established Investigator of the American Heart Association, Dallas, Tex.
- Received November 29, 1995.
- Revision received February 8, 1996.
- Accepted February 16, 1996.
- Copyright © 1996 by American Heart Association
Kaul S, Force T. Assessment of myocardial perfusion with contrast two-dimensional echocardiography. In: Weyman AE, ed. Principles and Practice of Echocardiography, 2nd ed. Lea & Febiger; 1993:687-720.
Sabia PJ, Powers ER, Ragosta M, Sarembock IJ, Burwell L, Kaul S. An association between collateral blood flow and myocardial viability in patients with recent myocardial infarction. N Engl J Med.. 1992;372:1825-1831.
Ragosta M, Camarano GP, Kaul S, Powers E, Sarembock IJ, Gimple LW. Microvascular integrity indicates myocellular viability in patients with recent myocardial infarction: new insights using myocardial contrast echocardiography. Circulation.. 1994;89:2562-2569.
Ten Cate FJ, Widimsky P, Cornel JH, Waldstein DJ, Serruys PW, Waaler A. Intracoronary Albunex: its effect on left ventricular hemodynamics, function, and coronary sinus flow in humans. Circulation.. 1993;88:2123-2127.
Keller MW, Segal SS, Kaul S, Duling BR. The behavior of sonicated albumin microbubbles within the microcirculation: a basis for their use during myocardial contrast echocardiography. Circ Res.. 1989;65:458-467.
Jayaweera AR, Edwards N, Glasheen WP, Villanueva FS, Abbott RD, Kaul S. In-vivo myocardial kinetics of air-filled albumin microbubbles during myocardial contrast echocardiography: comparison with radiolabeled red blood cells. Circ Res.. 1994;74:1157-1165.
Ismail S, Jayaweera AR, Skyba DM, Sklenar J, Goodman NC, Kaul S. Integrated backscatter and digital acquisition for myocardial contrast echocardiography: is there an advantage over conventional echocardiography for determining the mean microbubble transit rate from intracoronary injections? J Am Soc Echocardiogr.. 1995;6:453-464.
Broadhurst P, Cashman P, Crawley J, Raftery E, Lahiri A. Clinical validation of a miniature nuclear probe system for continuous on-line monitoring of cardiac function and ST segment. J Nucl Med.. 1991;32:37-43.
Callahan RJ, Froelich JW, McKusick KA, Leppo J, Strauss HW. A modified method in the in-vivo labeling of red cells with Tc-99m. J Nucl Med.. 1982;23:315-317.
Thompson MK, Starmer CF, Whorten RE, McIntosh MD. Indicator transit time considered as a gamma variate. Circ Res.. 1964;14:502-514.
Albers VM. Underwater Acoustic Handbook. State College, Pa: The Pennsylvania State University Press; 1960.
Porter TR, Xie F. Transient myocardial contrast after initial exposure to diagnostic ultrasound pressures with minute doses of intravenously injected microbubbles: demonstration of potential mechanisms. Circulation.. 1995;92:2391-2395.
Skyba DM, Jayaweera AR, Goodman NC, Ismail S, Camarano GP, Kaul S. Quantification of myocardial perfusion with myocardial contrast echocardiography from left atrial injection of contrast: implications for venous injection. Circulation.. 1994;90:1513-1521.
Skyba DM, Goodman C, Jayaweera AR, Ismail S, Camarano G, Bales G, Kaul S. Hemodynamic and safety characteristics of FS-069, a new contrast agent capable of producing myocardial opacification from a venous injection. J Am Soc Echocardiogr.. 1994;7:S30. Abstract.
Arai A, Kenny A, Shiota T, Unger E, Yellowhair D, Passafini A, Matsunaga T, Grauer SE, Pantley G, Sahn DJ. Transpulmonary passage of AerosomesTM, a pressure stable, lipid-based echocardiographic contrast agent: studies in pigs. J Am Coll Cardiol.. 1994;23:25A. Abstract.
Maddahi J, Kiat H, Friedman JD, Berman DS, Van Train KF, Garcia EV. Technetium-99m-sestamibi myocardial perfusion imaging for evaluation of coronary artery disease. In: Zaret BL, Beller GA, eds. Nuclear Cardiology: State-of-the-Art and Future Directions. St Louis, MO: CV Mosby; 1993:191-200.
Wu X, Ewert DL, Liu Y, Ritman EL. In vivo relation of intramyocardial blood volume to myocardial perfusion: evidence supporting microvascular site for autoregulation. Circulation.. 1992;85:730-737.
Eigler NL, Schuhlen H, Whiting JS, Plaff JM, Zeiher A, Gu S. Digital angiographic impulse response analysis of regional myocardial perfusion: estimation or coronary flow, flow reserve, and distribution volume by compartmental transit time measurement in a canine model. Circ Res.. 1991;68:870-880.
Marcus ML. Metabolic regulation of coronary blood flow. In: Marcus ML. The Coronary Circulation in Health and Disease. New York, NJ: McGraw-Hill; 1983:65-92.