- Email: [email protected]
Assessment of resting perfusion with myocardial contrast echocardiography: Theoretical and practical considerations
Assessment of resting perfusion with myocardial contrast echocardiography: Theoretical and practical considerations Jonathan R. L i n d n e r , MD, F l o r d e l i z a S. Villanueva, MD,a J o h n M. Dent, MD, K e v t n Wel, MD, J i r i Sklenar, PhD, and Sanjiv Kaul, MD Charlottesville, Va, and Pittsburgh, Pa
Background
The aim of this study was to perform a quantitative comparison between myocardial contrast echocardiography (MCE) and single-photon emission computed tomography (SPECT) in patients with prior myocardial infarction (MI). We also wanted to determine the optimal method for the intravenous administration of an ultrasound contrast agent in the clinical setting.
Methods and Results Seventeen patients with resting perfusion defects in a single vascular territory on SPECT were studied. MCE was performed with intermittent harmonic imaging during continuous infusions of a second-generation ultrasound contrast agent (Sonovue, Bracco Diagnostics) in all 17 patients and after bolus injection in 8 of them. During continuous infusions, the video intensity (VI) ratio between the abnormal and normal myocardium at a pulsing interval (PI) of 8 cardiac cycles correlated well with the activity ratio between these segments on SPECT (r = 0.73, P < .01 ). When information regarding microbubble velocity (MV) denoted as change in VI with increasing PIs was added, the correlation with SPECT activity ratio improved (P < .05) significantly (r = 0.87, P < .0001 ). Higher microbubble doses resulted in higher VI during continuous infusions with good myocardial opacification and no far-field attenuation until the highest dose was reached. With bolus injections, the VI ratio between the abnormal and normal myocardium at PI of 1 and 5 cardiac cycles showed a modest correlation (r = 0.46 and r = 0.48, respectively, P < .05) with activity ratios between these regions on SPECT. When a dose of microbubbles administered as a bolus produced adequate myocardial opacification, it invariably resulted in far-field attenuation. (.onclusions
In patients with prior MI, quantitative assessment of resting perfusion defects on MCE correlates well with regional activity on SPECT. Continuous infusions offer an advantage over bolus injections because they can provide an assessment of both relative VI and MV. Adjustment of the microbubble infusion rate produces adequate myocardial opacification without attenuation. (Am Heart J 2000;139:231-40.)
W h e n the relation b e t w e e n myocardial m i c r o b u b b l e concentration and amplitude of ultrasound backscatter is within the linear range, myocardial video intensity (VI) on myocardial contrast e c h o c a r d i o g r a p h y (MCE) reflects the concentration of microbubbles in that region. 1 The 3 main determinants of myocardial microbubble concentration are (1) myocardial blood volume (MBV); (2) the fraction of the MBV within the
From the Cardiovascular Divisions, University of Virginia, and the °Universily of Pittsburgh. Supported in part by grants (ROI-HL48890 to Dr Kaul and R29-HL58865 to Dr Vinanueva) from the National Institutes of Health, Bethesda, Md, and Bracco Diagnostics Inc, Princeton, N J. Drs Lindner and Wei are recipients of the Mentored Clini. ca/ScientistDevelopment Award (KOS-HL03810 and KOS.HL03909) from the National Institutes of Health. Submitted December29, 1998; accepted May 12, 1999. Reprint requests: Sanjiv Kaul, MD, Cardiovascular Division, Box 158, Medical Center, Charlottesville, VA 22908. E-mail: [email protected] Copyright ¢) 2000 by Mosby, Inc. 0002-8703/2000/$12.00+0 4/I/100123
ultrasound beam that is filled with microbubbles2; and (3) the concentration of microbubbles in blood, t-3 At sufficiently high acoustic pressures, ultrasound destroys microbubbles. 2,4,5 W h e n the myocardial microbubble concentration is low, a single s w e e p of ultrasound can destroy enough microbubbles. O n continuous imaging at 30 Hz, therefore, by the next frame (33 ms later) myocardial VI is the same as that of the precontrast image. 6 However, if m o r e time is allowed before the next ultrasound frame, microbubbles present in the blood will partially or completely replenish the myocardium. During a continuous infusion, in w h i c h the blood concentration of microbubbles is constant, myocardial VI progressively increases with prolongation of the interval b e t w e e n the ultrasound frames (PI) until the MBV within the entire ultrasound beam is filled, at w h i c h time VI reflects MBV. 2 The rate of change of VI from l o w e r to higher PIs represents microbubble velocity (MV). 2 Myocardial VI is also independently influenced by the
AmericanHeartjournal February2000
232 Lindner et al
Table I. Clinical characteristics
Figure 1
Basilar Short-Axis
Mid Short-Axis
Vertical Long-Axis
Age (y, median)
M/F
Weight (kg) Height (cm) Heart rate (beats/rain)
Precontrast
anterior septal ~
lateral
inferior
antero-a~
Postcontrast Systolic blood pressure (mm Hg) Precontrast Postcontrast Diastolic blood pressure (mm Hg) Precontrast Postcontrast Pulse oximetr,/(% saturation) Precontrast Postcontrast
52
13/4 80 + 13 175 ± 6 64 4- 16 66 ± 16 132 ± 22 129 ± 21 77 ± 10 77 ± 10 97 ± 2 97 ± 2
infero-apical
Model used to compare MCE and SPECT data. Although measurements in any segment could be made from any view, they were always depicted in the 14 segments shown here. This example shows SPECT data represented in model.
of the infarct-related artery a n d / o r magnitude of collateral flow. xo,t t
Methods Patient population blood concentration of microbubbles. 2,3 During a continuous infusion, higher doses produce greater VI for any given MBV and PI. After a bolus injection, however, the blood concentration of microbubbles varies, not only with the dose but also the duration of injection and the cardiac output. 7 In this situation, VI truly reflects MBV only if microbubbles are not destroyed by ultrasound. 3 During a bolus injection, the blood concentration of microbubbles needs to be high to produce adequate myocardial opacification. Thus the myocardial concentration of microbubbles is usually higher during the brief period of opacification after a bolus injection compared with a continuous infusion. Because in this setting only a fraction of the microbubbles are likely to be destroyed by ultrasound, VI could still predominantly reflect MBV.3 However, because the concentration of microbubbles in the blood changes rapidly after a bolus injection, changes in VI at higher PI do not represent MV. We postulated that the assessment of regional myocardial perfusion with the use of MCE during continuous m i c r o b u b b l e infusion will correlate with myocardial isotope activity on single-photon emission c o m p u t e d t o m o g r a p h y (SPECT). We also w a n t e d to d e t e r m i n e the optimal m e t h o d (continuous infusion vs bolus injection) for the intravenous administration of an ultrasound contrast agent in the clinical setting. Accordingly, w e studied patients w i t h prior myocardial infarction (MD in w h o m MBV varies d e p e n d i n g on the degree of microvascular destruction.8,9 MV in these patients also varies, d e p e n d i n g on the patency
The study was approved by the Human Investigation Committees at the universities of Virginia and Pittsburgll. It was also approved as a phase II study by the United States Food and Drug Administration, with the primary objective of determining the effective dose range for myocardial opacification. The secondary objective was to compare MCE with SPECT in patients with prior MI who have a resting perfusion defect in a single vascular territory. All patients gave written informed consent. Ultrasound image quality needed to be adequate to evaluate all myocardial segments in the view in which the perfusion defect was best seen on SPECT. Exclusion criteria included unstable or post-Ml angina, hemodynamic instability, pregnancy, or lactation.
Single-photonemissioncomputedtomography Patients were injected at rest with either 2°lTl- (Mallinckrodt Medical, St Louis, Mo) or 99mTc-sestamibi (Dupont Pharmaceuticals, North Billerica, Mass). Data were acquired 10 minutes after injection of 2°IT1and 1 hour after administration of 99mTc-sestamibi. 20tT1 imaging was performed with a dual-headed ganmaa camera (Vertex, ADAC, Milpitas, CalL0, with 32 projections of 60 seconds, each acquired over a 180 degree orbit. 99mTc-sestamibi imaging was performed with a 3-headed gamma camera (Prism 3000S, Picker, Cleveland, Ohio), with each head acquiring 20 projections of 40 seconds, each over a 120 degree orbit. After filtered back-projection, the reconstructed 3-dimensional data were represented in a 14-segment model (Figure 1). t2 Activity in each segment was normalized to the segment with the highest activity in that view. t2
Myocardial contrast echocardiography The principle underlying the MCE assessment of MV and MBV during continuous infusion of microbubbles has been previously described. 2 After steady state is achieved, ultra-
American Hearl Journal Volume 139, Number 2, Part
Lindner et al 2 3 3
Figure 2
".
°
-
-
-
~
"2.
.
X
-2
2
X
• ,-~2. t _t ~ r , m - # J, , ? ~ t, =-t •
t
I~
b.. ~
"
i~t "~
~ .
. :.t,q) J
m
,,
*.. ,
r
m'
Ill
. .-_--..
I,y
Method of manual image alignment. Top panels, Images to be aligned. Bottom panels, "Ghost images" in green and magenta before (left) and after (right) successful alignment.
s o u n d is u s e d to destroy m i c r o b u b b l e s p r e s e n t widlin the m y o c a r d i n m . T h e i r rate of u l t r a s o u n d b e a m r e p l e n i s h m e n t is t h e n m e a s u r e d by gradu~dly increasing t h e interval b e t w e e n u l t r a s o u n d pulses. At a s h o r t PI, only a p o r t i o n o f t h e ultras o t m d b e a m is r e p l e n i s h e d , resulting in a sm,'dl increase in VI. As t h e PI progressively increases, g r e a t e r r e p l e n i s h m e n t o f die u l t r a s o u n d b e a m o c c u r s w i t h g r e a t e r increases in VI. W h e n t h e u l t r a s o u n d b e a m has b e e n c o m p l e t e l y r e p l e n i s h e d , f u r t h e r i n c r e a s e s in PI do n o t result in any c h a n g e in VI. T h e rate of increase in VI w i t h i n c r e m e n t a l PI d e n o t e s MV, w h e r e a s t h e plateau VI r e p r e s e n t s MBV. 2 Bec.'mse o f t h e limited n u m b e r o f PIs u s e d in this study, w e w e r e n o t able to fit ~m e x p o n e n t i a l
f u n c t i o n to t h e data as previously described.-' W e t h e r e f o r e fitted a line to t h e data, a n d based o n t h e slope o f t h e line determ i n e d w h e t h e r V1 progressively increased or did not increase w i t h i n c r e m e n t s in PI. T h e VI at t h e h i g h e s t PI (8 cardiac cycles) w a s a s s u m e d to r e p r e s e n t MBV. For this study, intermittent h a r m o n i c imaging w a s performed widl a phased-array system (Sonos 5500, Agilent Technologies, Andover, Mass) by u s e of transmit a n d receive freq u e n c i e s o f 1.8 a n d 3.6 MHz, respectively. Acoustic p o w e r a n d c o m p r e s s i o n w e r e m a x i m i z e d a n d gain settings w e r e o p t i m i z e d at t h e o n s e t o f e a c h s t u d y a n d h e l d c o n s t a n t t h r o u g h o u t . As p e r t h e p h a s e II p r o t o c o l , i m a g e s w e r e
234
[indner et al
American Heart Journal Februor"12000
Figure 3
Examples of perfusion defects on SPECT and MCE (latter acquired at PIs of 1, 3, and 8 cardiac cycles) in a patient with an anteroapical MI (top panel) and in a patient with an inferior MI (bottom panel). Top panel shows images in 4-chamber views; bottom panel shows images in 2-chamber views (with reorientation of corresponding SPECT image).
acquired in a single view (parasternal short-axis or apical) that corresponded to the view that best s h o w e d the p r e s e n c e of a defect on SPECT. Baseline images w e r e acquired before administration of bubbles for background subtraction. The ultrasound contrast agent used for this study was Sonovue (Bracco Diagnostics), w h i c h is c o m p o s e d of phospholipid microbubbles containing sulfur hexafluoride, with a m e a n diameter of 2.5 }am and a m e a n concentration of 2 • 108 mL-l. 13 To determine the dose range for optimal myocardial opacification, the infusion rates varied from 0.03 to 0.08 mL • kg -1 • rain -I as per the phase I1 protocol. After 2 minutes allowing for steady state, 2 at least 8 end-systolic images were acquired at each PI of 1, 3, and 8 cardiac cycles with the ultrasound transmission gated to the T wave of the electrocardiogram. t If the image orientation changed, continuous imaging was performed to regain the original orientation, after w h i c h intermittent imaging was again initiated. As per the phase II protocol, bolus injections of 0.3, 0.6, 1.2, or 1.8 mL • kg q of Sonovue were also administered in 8 patients w h o received the lowest dose for continuous infusion. These injections were delivered over a period of 20 seconds, followed by a 5-mL saline flush. On resolution of farfield attenuation from left ventricular (LV) cavity contrast, at least 8 end-systolic images were acquired at PI of 1 followed by 5 cardiac cycles. Each patient received 1 or more doses of continuous infusions and/or bolus injections, with the total dose not exceeding 0.52 mL • kg -].
Data w e r e analyzed off-line by a single observer blinded to the electrocardiographic and SPECT results, with customdesigned software. 14 Images were transferred from videotape to a computer, w h e r e precontrast images and several contrastenhanced images at each PI were selected. Each of these image sets w a s separately aligned by use of o n e of t w o m e t h o d s . Alignment was first attempted with the use of a previously described automated method.14 The gray-scale amplitude in each pixel in the image to be aligned is correlated with the gray-scale amplitude from the same pixel (same x and y coordinates) in the index image. The image is t h e n shifted 1 pixel at a time in the x and then the y direction, and the correlation is repeated at each step. The image to be aligned is then automatically shifted in the x and y directions to the coordinates w h e r e the best correlation is achieved with the index image. Because gray-scale intensity in the contrast-enhanced images obtained at different intervals is so different, automated alignm e n t may not always work. In that case, w e used a manual m e t h o d depicted in Figure 2. "Ghost images" of the frames to be aligned (panels A and B) are produced, which can be distinguished by their 2 different colors: magenta and green (panel C). One image can t h e n be moved over the other with a handheld m o u s e or by using the arrow keys until alignment is optimal (panel D). Each set of images (background and those at different P1) were t h e n separately averaged. The averaged background image was digitally subtracted from the averaged contrast-
American Heart Journal Volume 139, Number 2, Parl 1
enhanced image for each PI. The VI scale of the digitally subtracted image was expanded to 256 gray levels, and each pixel was assigned a color on the basis of a heated object algorithm, in which shades of red to orange to yellow to white represent incremental contrast enhancement. 14 The LV cavity was masked out. VI was measured in 6 regions of interest defined over the myocardium of color-coded images at each PI. These regions of interest were approximately the same in size and corresponded to the regions in the 14-segment model (Figure l). Each region of interest was made as large as possible, with care taken to exclude the specular endocardial and epicardial borders, the LV cavity, or any areas with attenuaUon or other artifacts.
Lindner et al
Figure 4
50. 40. 30. to
20.
"o
10.
Comparison of MCE and SPECTimages
0.
The activity in each myocardial segment on SPECT was compared with a normal database. The values in this database are specific for each segment and to patient sex (to account for breast attenuation in women). A segment was considered abnormal if its activity was <2 SD of the mean value for that segment derived from a large n u m b e r of normal subjects.12 The activities within segments considered abnormal or normal on SPECT and MCE were separately averaged. Segments with attenuation on MCE or SPECT were not included in the process of averaging. The change in VI with increasing PI was derived from these averaged values from which a slope was calculated by linear regression.
:rval
40.
30. -~
Statistical methods Comparisons of VI at different PIs during continuous infusions were made using either 1-way or 2-way analysis of variance, whereas VI at the 2 PIs during bolus injections were compared by use of the paired t test. CorrelaUons between the VI ratios on MCE between abnormal and normal segments and the radionuclide activity ratios between these segments on SPECT were performed with the use of least-squares linear regression. A value of P < .05 (2-sided) was considered statistically significant.
Results MCE a n d SPECT w e r e p e r f o r m e d w i t h i n 4 8 h o u r s o f e a c h o t h e r in all p a t i e n t s , a n d n o clinical e v e n t s o c c u r r e d b e t w e e n t h e t w o e x a m i n a t i o n s . Analysis o f MCE data d u r i n g c o n t i n u o u s i n f u s i o n s c o u l d n o t b e perf o r m e d in 7 o f t h e 24 p a t i e n t s b e c a u s e o f s h a d o w i n g o v e r t h e p e r f u s i o n d e f e c t s at t h e h i g h e s t d o s e s r e q u i r e d b y t h e p r o t o c o l . In 6 o f t h e s e p a t i e n t s , t h e p a r a s t e r n a l short-axis v i e w w a s u s e d b e c a u s e t h e d e f e c t w a s b e s t visualized in t h i s v i e w o n SPECT. In all t h e s e p a t i e n t s , MI w a s l o c a t e d inferiorly. T h e clinical c h a r a c t e r i s t i c s o f t h e r e m a i n i n g 17 p a t i e n t s , w h o also i n c l u d e d t h e 8 r e c e i v i n g b o l u s i n j e c t i o n s , are listed in T a b l e I. T h e b a s e l i n e VI b e f o r e m i c r o b u b b l e a d m i n i s t r a t i o n w a s h i g h e r in t h e a b n o r m a l c o m p a r e d w i t h t h e n o r m a l r e g i o n s (42 + 12% vs 36 + 12%, P < .05)
Continuous infusions MCE a n d SPECT i m a g e s f r o m 2 p a t i e n t s ( o n e w i t h a n apical a n d a n o t h e r w i t h a n i n f e r i o r d e f e c t ) are s h o w n
20.
O o
0.
Pulsing Interval
Effect of microbubble dose (x-axis) and PI (z-axis) on VI (y-axis) measured from normal myocardium during (top panel) continuous infusions and (bottom panel) bolus injections.
in Figure 3. B e c a u s e o f g r e a t e r r e p l e n i s h m e n t o f t h e b e a m e l e v a t i o n , VI in t h e n o r m a l m y o c a r d i u m w a s h i g h e r at a PI o f 8 c o m p a r e d w i t h 1 a n d 3 c a r d i a c cycles. T h e r e l a t i o n b e t w e e n VI a n d PI in t h e n o r m a l m y o c a r d i u m in all 17 p a t i e n t s at d i f f e r e n t m i c r o b u b b l e d o s e s is d e p i c t e d in Figure 4, A. As e x p e c t e d , for a n y g i v e n dose, VI w a s h i g h e r at l o n g e r PIs. For t h e s a m e PI, a h i g h e r d o s e also r e s u l t e d in a h i g h e r m y o c a r d i a l VI, b u t far-field a t t e n u a t i o n w a s s e e n o v e r at least 1 m y o c a r d i a l s e g m e n t at t h e h i g h e s t d o s e in all cases. T h e t i m e t a k e n to c o m p l e t e l y fill t h e u l t r a s o u n d b e a m a f t e r b u b b l e d e s t r u c t i o n also d e p e n d s o n t h e ultrasound beam elevation. For the ultrasound probe used in this study, t h e n o r m a l m y o c a r d i u m fills in 5 t o 6 seco n d s at rest. 2 T h e r e f o r e , b e c a u s e t h e m e a n h e a r t rate o f p a t i e n t s in this s t u d y w a s a p p r o x i m a t e l y 6 0 b e a t s / m i n
235
AmericanHeartJournal February2000
'~36 Lindneret al
Figure 6
Figure 5 0
o
•
oo
o PI1 ~=-6.17x+13.2 EE =8.14; r=O.08
0.8-
• PI3
~--0.33x+0.77 EE --0.35, t=-0.11
O 0.6-
o o >
•• to
•o
0.4•
0.2 0.2
0
•
Examples of background-subtracted color-coded images in apical 3-chamber view in patient with prior inferior MI during (left) continuous infusion at PI of 8 cardiac cycles and (right) during bolus injection at PI of 5 cardiac cycles. Infusion rate was 0.06 mL. kg -1 • min-1; bolus dose was 2.4 mL. kg -].
0 014
016
0[8
SPECT
1-
0.8-
/ 00 /
y=l.17x-0.16 SEE--0.12;r=0.73 p
O/
/
o 0.6-
o
ID
>
o
o
0.4-
0.2 0.2
o
I
I
I
0.4
0.6
0.8
I
1
SPECT
Effect of PI ( 1 and 3 in upper panel and 8 in lower panel) on relation between VI ratio on MGE (y-axis) and activity ratio on SPECT (x-axis) between abnormal and normal segments in same view.
(Table I), VI at 8 cardiac cycles probably represents relative MBV in the normal myocardium. It is for this reason that the VI ratios b e t w e e n the abnormal and normal segments obtained at this PI had a better correlation with activity ratios b e t w e e n these segments on SPECT (P < .01) than those obtained at PIs of 1 and 3 cardiac cycles (Figure 5). Figure 3 demonstrates that relative MV can also be assessed on sequential MCE images acquired at different PI during continuous infusions. VI in the apical region did not increase despite increases in PI, indicating virtually no flow to that region. In comparison, VI in the inferior wail increased at longer PIs, indicating some flow to that region. The rate of change of VI with prolongation of PI, however, was lower in the inferior compared with the anterior region, indicating slower MV and hence
lower myocardial blood flow in the inferior compared with the anterior bed. Of the 17 patients, 10 demonstrated an increase in VI (>5 VI units) at longer PI within the abnormal region, whereas 7 s h o w e d no change (VI change of <5 units). Patients with no change in VI within the infarct zone on MCE had worse (P < .01) perfusion defects on SPECT (48% + 6% activity) c o m p a r e d with those showing increase in VI at higher PI (57% + 9% activity). The correlation b e t w e e n the relative rate of change of VI in the infarct bed on MCE and radionuclide activity ratio on SPECT was r = 0.47 (P < .01). VI ratio b e t w e e n normal and abnormal beds at PI of 8 cycles (MBV ratio) and rate of change in VI in the infarct bed (MV) w e r e evaluated with other variables (including VI ratio at PIs of 1 and 3 cardiac cycles) to determine the best MCE correlates of radionuclide activity ratio. The only MCE variables that correlated with radionuclide activity ratio b e t w e e n the infarct and normal beds w e r e VI ratio at 8 cardiac cycles (a) and w h e t h e r VI increased within the infarct bed at longer PI (b). W h e n the two w e r e combined in a multiple linear regression model, the correlation with radionuclide activity improved significantly (P < .05) to r -- 0.87 (P < .0001). The final regression equation was y = 0.44a + O.11b + 0.31.
Bolus injections Figure 6 illustrates images obtained during a continuous infusion and a bolus injection in the same patient. The image during continuous infusion was acquired at a PI of 8 cycles, and that after bolus injection was acquired at a PI o f 5 cardiac cycles w h e n s h a d o w i n g o v e r the m y o c a r d i u m had just r e c e d e d . The relative VI in the abnormal and normal beds appears to be very similar with the use of both methods of microbubble administration and in both cases represents MBV. The VI ratio b e t w e e n the abnormal and normal segments on MCE
American Heorl Journal Volume 139. Number 2. Part I
Lindner el al
Figure 7
O ° ~,,,t
/••us
A
- - /-- / . .I.b. o l u s B . . . . . . ~ -
~D
............
atte_n_uatio_nth_reshold
O
O ,L3 ,.Q O z) o~,,~
t
t2
t3
Time Differences between continuous infusions and bolus injections in terms of optimal myocardial opacification. Region between dotted lines represents VI range in which relation between microbubble concentration and VI is linear. Bolus A and bolus B represent 2 bolus doses. Intervals t2-t ) and t4-ta are periods during which optimal myocardial opacification can be achieved with bolus injections during upslope or downslope of curve.
correlated modestly (P < .05) with the activity ratio between these segments on SPECT (r = 0.46 and r = 0.48 for PI of 1 and 5 cardiac cycles, respectively), probably because of the small ntunber of observations (n -- 8). Figure 4 illustrates the effect of PI and microbubble dose on VI measured from the normal myocardium during bolus injections. Because the longer PI was used w h e n the blood pool concentration of microbnbbles was declining, VI at 5 c o m p a r e d with 1 cardiac cycle was not significantly different, although it tended to be higher at 5 cardiac cycles. Regardless of the PI, VI was significantly lower, at a dose of 0.6 mL- kg -1 c o m p a r e d with higher doses. This dose also did not result in farfield shadowing. VI in the normal myocardium was identical at the 3 higher doses, all of w h i c h p r o d u c e d far-field shadowing.
Discussion With all other variables being constant, the relative myocardial uptake of a radionuclide tracer is determined by the cross-sectional area through w h i c h it can diffuse ( w h i c h is proportional to the capillary density or MBV 15) and the rate at w h i c h it is delivered (velocity of the tracer, w h i c h if m i x e d adequately in blood should equal MV). 16 Our results s h o w that w h e n both MBV and MV are measured with MCE, the correlation with relative activity on SPECT is good in patients with
prior MI. Correlation of MBV alone (VI at high PI during continuous infusion) with SPECT activity is also reasonable. Thus w h e n c o u p l e d with intermittent harmonic imaging with varying PI, continuous infusions of microbubbles provide an accurate quantification of resting perfusion defects on MCE.
Continuous infusion versus bolus injection Continuous infusions of m i c r o b u b b l e s are m o r e desirable than bolus injections because by changing the PI, information regarding both MBV and MV can be obtained.3A7 After a bolus injection, change in VI at a higher PI cannot denote MV because the concentration of microbubbles in the blood changes. Therefore at best, a bolus injection can only provide an approximation of MBV. 3 At steady state achieved during a continuous infusion of microbubbles, myocardial VI at any given PI is influenced by both MV and MBV. Thus p o o r opacification within a myocardial region could result from either reduced MV (and h e n c e myocardial blood flow) or decreased MBV (as occurs with MI). If the latter case, myocardial viability is unlikely, s-1 ] If MV is decreased and MBV is normal, myocardial viability is more likely. Thus by the use of continuous infusion, greater information can be obtained than by the use of bolus injection. There are o t h e r reasons for using continuous infu-
237
American HeartJournal February 2000
2 3 8 Lindner et al
Figure 8 % 4¢
z
4q,' ,
, 21
50"
' ; Iq
'
MCE images in a patient with prior inferior MI before (left panel) and after (right panel) infusion of microbubbles. Top panels show images alone; bottom panels depict VI in 6 myocardial segments in this view.
sions. A narrow range exists in which the relation between microbubble concentration and VI is linear (approximately 50 to 60 background-subtracted VI units).l For better signal-to-noise ratio, it is desirable to use the higher end of this range (the region b e t w e e n the dotted lines in Figure 7), which can be easily achieved by titrating the rate of continuous infusion. Higher doses produce far-field attenuation, whereas lower doses preclude adequate myocardial opacification (Figure 7). Once the infusion dose has b e e n optimized, there is adequate time to interrogate each myocardial segment in detail. A single operator can start the infusion, adjust the rate of administration, and perform the entire examination. If the orientation of the heart changes during the examination, imaging can be changed to real-time, the view can be reoriented, and imaging can then be resumed. Finally, poor regional myocardial opacification can result from either a true perfusion defect or modest attenuation (in which the myocardium is still visualized). In this situation, change in VI with change in PI is more likely to reflect an artifact, especially if the rate of change is the same as in the normal myocardium. If a single PI is used, it is more difficult to differentiate an artifact from a real perfusion defect.
After a bolus injection, the blood concentration of microbubbles varies over time. Microbubble concentration often exceeds the level in which its relation with VI is no longer linear (bolus A in Figure 7). The high concentration of microbubbles in the blood within the LV cavity also produces far-field shadowing. The period of optimal myocardial opacification, in which shadowing has not yet started (t2-t l) or has receded (t4-t 3) and w h e n the signal-to-noise ratio within the myocardium is favorable, is very brief. Ideally, one could design a bolus dose that will result in myocardial microbubble concentrations only within the effective range (bolus B in Figure 5). However, because myocardial VI depends on several factors (such as cardiac output, 7 acoustic impedance of the chest wall, m homogeneity of the ultrasotmd field, etc), an ideal bolus dose cannot be predicted. The rate of a constant infusion, however, can be adjusted during the examination to accotmt for these factors. Bolus injections pose other problems as well. Because the imaging plane must be correctly aligned and because motion artifacts caused by respiration are common, more than 1 injection may be required per view. During continuous infusion, images with motion artifacts can be recognized easily, and more images can be acquired at the s;mle PI. Unsuitable frames can later be rejected. Bolus injections require one operator to acquire the images and another to inject the contrast agent. In our study, better myocardial opacification was achieved at higher m i c r o b u b b l e doses, until a dose was reached w h e r e despite better opacification of the near-field myocardium, far-field shadowing occurred. This dose was required by the protocol for dose ranging. In the apical view, higher blood pool concentrations could be used before shadowing affected myocardial regions. Shadowing was limited over the left atrium, whereas good myocardial opacification was achieved. Thus the optimal blood c o n c e n t r a t i o n s during c o n t i n u o u s infusions of m i c r o b u b b l e s can vary with different views. A fine balance must be achieved b e t w e e n good myocardial opacification and the least a m o u n t of shadowing, w h i c h can be accomplished by adjusting the rate of infusion. Because this was a dose-ranging study, w e could not adjust the rate of infusion. During bolus injections, VI was measured only w h e n shadowing over the myocardium was receding (proximity of t 3 in Figure 7). VI in the normal myocardium was therefore similar for all doses that produced shadowing (bolus A in Figure 7). The lowest dose that did not cause attenuation produced a lower VI (diagrammatically represented as bolus B in Figure 7). Even if VI were measured over the myocardium in the near field not affected by shadowing, VI probably would have reached a plateau at higher doses because of signal saturation at those doses. This could result
American Heart Journal Volume 139, Number 2, Part 1
in apparent myocardial opacifications in regions with low perfusion.
Lindner el al
Figure 9
Advantages of quantification An), measurement should be more accurate than a subjective assessment. Because acoustic pressures within the ultrasound field are inhomogeneous, microbubble destruction and hence myocardial VI may vary b e t w e e n different beds in the same view and within the same bed in different views. Quantification provides a comparison with values obtained in normal individuals, thus increasing both the sensitivity and specificity for detecting perfusion defects.19 The interobserver error is also minimized and reproducibility is maximized, z0 Another reason for quantification is that myocardial regions with prior MI frequently have a higher VI during harmonic imaging at baseline, which is not seen on fundamental imaging. The reason for this increase in VI on harmonic imaging is unclear but may be related to either better signal-to-noise3 or nonlinear properties of infarcted tissue, el This increase in myocardial VI in infarcted tissue makes the change in VI after microbubble administration more difficult to appreciate. The left panel in Figure 8 shows a higher VI measured in the posterior compared with the anterior wall before microbubble injection in a patient with prior inferior MI, which is not as apparent visually. Both visually and quantitaUvely VI is equal during microbubble infusion in both walls at a PI of 8 cardiac cycles. However, w h e n both the precontrast and contrast-enhanced images are evaluated side by side, the change in VI is less in the posterior than in the anterior wall, indicating a lower MBV (and hence a prior MI) in the posterior bed. The same result can be achieved by background subtraction. The top panels in Figure 9 depict color-coded MCE images obtained after background subtraction of the left from the right image in Figure 8. As can be noted from the numeric values in the top right panel, the color differences also denote the relative changes in VI in different myocardial regions b e t w e e n the two sets of images. The corresponding color-coded SPECT images are shown in the bottom panels in Figure 9. The advantage of color over gray scale is that although the human eye can discern only a few shades of gray, it can discriminate between thousands of hues of color. Thus perception of VI differences is easier. In panels C and D in Figure 7 and in Figure 3, perfusion defects visually appear similar w h e n MCE and SPECT data are color-coded.
Comparison with previous studies Although an experimental study from our laboratory has compared continuous infusions with bolus injections for stenosis detection, 17 to our knowledge this is the first study that has compared continuous infusions with bolus injections of microbubbles in the setting of
~,
-~
~l
l
~li
Ski
Left panels show color-coded MCE images obtained after digital subtraction of images in left panels from that in right in Figure 8. Top panel shows image without VI measurements; bottom panel shows same image with VI measurements. These measurements reflect difference in VI measured in same segments on gray-scale images (bottom panels in Figure 8). Right panels show corresponding color-coded SPECTimages without (top) and with (bottom) quantitative measurements performed in different myocardial segments corresponding to similar segments on MCE. Values on sides represent average for normal and abnormal segments, respectively.
prior MI. Perfusion defects on MCE have been shown to correlate well with those on SPECT with bolus injections, z2 but a quantitative assessment has not been performed. Continuous infusions and variable PI have also been used for detecting reversible perfusion defects with MCE performed in conjunction with dobutamine echocardiography.23
Limitations of the study The n u m b e r of patients is small. However, it is adequate to address the principal aims of the study. Because we were limited by the total deliverable dose, we were unable to acquire data at PI >8 cardiac cycles during c o n t i n u o u s infusions. On the basis of our previous observations, we believe that VI reached a plateau within all normal myocardial beds at or before 8 cardiac cycles. However, complete replenishment of the ultrasound beam may not have occurred in abnormal segments, particularly if resting flow was markedly
239
American Hearl Journal February 2000
2 4 0 Lindner el al
reduced in these segments. Thus w e cannot make an assessment regarding lack of myocardial viability in segments showing decreased VI at 8 cardiac cycles. Ideally, images should have been acquired at several PIs so that ,an exponential function could be fitted to the PI versus VI data. 2,17 We w e r e tmable to do so with the limited n u m b e r of PIs used in this study. During bolus injections, w e have a limited time in w h i c h to acquire data and w e r e thus unable to go beyond a PI of 5 cardiac cycles. Some patients received 2°tTl instead of 99mTc. Because the latter does not redistribute, it is ideal for comparison with MCE. Although imaging was initiated very early after 201Tl injection, redistribution could still have occurred. Substantial redistribution in patients w i t h p r i o r MI and no e v i d e n c e of resting ischemia is, however, unlikely. 24 Although the correlation b e t w e e n activity ratios on SPECT and VI ratios on MCE b e t w e e n the abnormal and normal regions w e r e not different for 2°IT1 or 99mTc, the numbers were too small to afford a meaningful comparison. Because of the limit imposed on the dose of microbubbles that could be administered in individual patients, w e only obtained a single view on MCE. Finally, because w e selected MCE views based on SPECT results, our study design was biased in favor of finding a perfusion defect on MCE. For this reason, w e limited ourselves to a quantitative analysis.
We thank Anthony Marano, MD, Bruce Helmly, MD, and Raju Modg MD, f o r patient recruitment as well as A m a n d a Doss, MEd, Nancy DLxon, and Virginia Schneider, RN, f o r technical assistance.
References 1. Skyba DM, Jayaweera AR, Goodman NC, et aJ. Quantification of myocardial perfusion with myocardial contrast echocardiography from left atrial injection of contrast: implications for venous injections. Circulation 1994;90:1513-21. 2. Wei K, Jayoweera AR, Firoozan S, et al. Quantification of myocardial blood flow with ultrasound-induced destruction of microbubbles administered as a constant venous infusion. Circulation 1998;97:473-83. 3. Firschke C, Lindner JR, Wei K, et al. Myocardial perfusion imaging in the setting of coronary artery stenosis and acute myocardial infarction using venous injection of a second-generation echocardiographic contrast agent. Circulation 1997;96:959-67. 4. Wei K, Skyba DM, Firschke C, et al. Interaction between microbubbles and ultrasound: in vitro and in viva observations. J Am Coil Cordial 1997;29:1081-8. 5. Skyba DM, Price RJ, Linka AZ, et al. Direct in viva visualization of microvessel rupture and tissue injury caused by intravascular destruction of microbubbles by ultrasound. Circulation 1998;98:290-3. 6. Pelberg RA, Wei K, Kamiyama N, et al. Potential advantage of 'flash echocardiography' for digital subtraction of B-mode images acquired during myocardial contrast echocardiography. J Am Sac Echocardiogr 1999; 12:86-93. 7. Villanueva FS, Glasheen WP, Sklenar J, et al. Successful and reproducible myocardial opacificalion during 2-dimensional echocardio-
8.
9.
10.
11.
12.
! 3.
14.
15. 16.
17.
18.
19.
20.
21. 22.
23.
24.
graphy from right heart injection of contrast. Circulation 1992;85: 1557-64. Villanueva FS, Glasheen WP, Sklenar J, et al. Characterization of spatial patterns of flow within the reperfused myocardium using myocardial contrast echocardiography: implications in determining the extent of myocardial salvage. Circulation 1993;88:2596-606. Villanueva FS, Camarano G, Ismail S, et al. Coronary reserve abnormalities during post-infarct reperfusion: implications for the timing of myocardial contrast echocardiography to assess myocardial viability. Circulation 1996;94:748-54. Sabia PJ, Powers ER, Jayaweera AR, et al. Functional significance of collateral blood flow in patients with recent acute myocardial infarction: a study using myocardial contrast echocardiography. Circulation 1992;85:2080-9. Sabia PJ, Powers ER, Ragosta M, et al. An association between collateral blood flow and myocardial viability in patients with recent myocardial infarction. N Engl J Med 1992;372:1825-31. Smith WH, Kastner P,J,Calnon DA, et aL Quantitative gated single photon emission computed tomography imaging: a counts-based method for display and measurement of regional and global ventricular systolic function. J Nucl Cardiol 1997;4:451-63. Schneider M, Arditi M, Barrau MB, et al. BR1: a new ultrasonographic contrast agent based on sulfur hexafluoride-filled microbubbles. Invest Radial 1995;30:451-7. Jayaweera AR, Sklenar J, Kaul S. Quantification of images obtained during myocardial contrast echocardiography. Echocardiography 1994; 1 1:385-96. Kassab GS, Lin DH, Fung YB. Morphometry of pig coronary venous system. Am J Physiol 1994;267:H2100-13. Leppo JA. Cardiac transport of single photon myocardial perfusion agents. In: Zaret BA, Belier GA, editors. Nuclear cardiology: state of the art and future directions. St Louis, Mo: Mosby; 1993. p. 35-44. Wei K, Linka A, Jayaweera AR, et aL Basis of stenosis detection by myocardial contrast echocardiography during venous administration of microbubbles: bolus or continuous infusion? J Am Coil Cardiol 1998;32:252-60. Ismail S, Jayaweera AR, Camarano G, et al. Relation between airfilled albumin microbubble and red blood cell rheology in the human myocardium: influence of echocardiographic systems and chest wall attenuation. Circulation 1996;94:445-51. Kaul S, Boucher CA, Newell JB, et al. Determination of quantitative thallium parameters that optimize detection of coronary artery disease. J Am Coil Cardiol 1986;7:527-3. Kaul S, Chesler DA, Okada RD, et al. Computer versus visual analysis of exercise thallium-2Ol images: a critical appraisal in 325 patients with chest pain. Am Heart J 1987; 114:1129-37. Kaul S. Assessing the myocardium after attempted reperfusion: should we bother? Circulation 1998;98:625-7. Kaul S, Senior R, Dittrich H, et al. Detection of coronary artery disease with myocardial contrast echocardiography: comparison with 9¢'mTc-sestamibisingle-photon emission computed tomography. Circulation 1997;96:785-92. Porter TS, Li S, Kriscfield D, et al. Detection of myocardial perfusion in multiple echocardiographic windows with one intravenous injection of microbubbles using transient response imaging. J Am Coil Cardiol 1997;29:791-9. Sabia PJ, Powers ER, Ragosta M, et al. Role of quantitative planar thallium-201 imaging for predicting viability in patients with acute myocardial infarction and a totally occluded infarct-related artery. J Nucl Med 1993;34:728-36.
Background
The aim of this study was to perform a quantitative comparison between myocardial contrast echocardiography (MCE) and single-photon emission computed tomography (SPECT) in patients with prior myocardial infarction (MI). We also wanted to determine the optimal method for the intravenous administration of an ultrasound contrast agent in the clinical setting.
Methods and Results Seventeen patients with resting perfusion defects in a single vascular territory on SPECT were studied. MCE was performed with intermittent harmonic imaging during continuous infusions of a second-generation ultrasound contrast agent (Sonovue, Bracco Diagnostics) in all 17 patients and after bolus injection in 8 of them. During continuous infusions, the video intensity (VI) ratio between the abnormal and normal myocardium at a pulsing interval (PI) of 8 cardiac cycles correlated well with the activity ratio between these segments on SPECT (r = 0.73, P < .01 ). When information regarding microbubble velocity (MV) denoted as change in VI with increasing PIs was added, the correlation with SPECT activity ratio improved (P < .05) significantly (r = 0.87, P < .0001 ). Higher microbubble doses resulted in higher VI during continuous infusions with good myocardial opacification and no far-field attenuation until the highest dose was reached. With bolus injections, the VI ratio between the abnormal and normal myocardium at PI of 1 and 5 cardiac cycles showed a modest correlation (r = 0.46 and r = 0.48, respectively, P < .05) with activity ratios between these regions on SPECT. When a dose of microbubbles administered as a bolus produced adequate myocardial opacification, it invariably resulted in far-field attenuation. (.onclusions
In patients with prior MI, quantitative assessment of resting perfusion defects on MCE correlates well with regional activity on SPECT. Continuous infusions offer an advantage over bolus injections because they can provide an assessment of both relative VI and MV. Adjustment of the microbubble infusion rate produces adequate myocardial opacification without attenuation. (Am Heart J 2000;139:231-40.)
W h e n the relation b e t w e e n myocardial m i c r o b u b b l e concentration and amplitude of ultrasound backscatter is within the linear range, myocardial video intensity (VI) on myocardial contrast e c h o c a r d i o g r a p h y (MCE) reflects the concentration of microbubbles in that region. 1 The 3 main determinants of myocardial microbubble concentration are (1) myocardial blood volume (MBV); (2) the fraction of the MBV within the
From the Cardiovascular Divisions, University of Virginia, and the °Universily of Pittsburgh. Supported in part by grants (ROI-HL48890 to Dr Kaul and R29-HL58865 to Dr Vinanueva) from the National Institutes of Health, Bethesda, Md, and Bracco Diagnostics Inc, Princeton, N J. Drs Lindner and Wei are recipients of the Mentored Clini. ca/ScientistDevelopment Award (KOS-HL03810 and KOS.HL03909) from the National Institutes of Health. Submitted December29, 1998; accepted May 12, 1999. Reprint requests: Sanjiv Kaul, MD, Cardiovascular Division, Box 158, Medical Center, Charlottesville, VA 22908. E-mail: [email protected] Copyright ¢) 2000 by Mosby, Inc. 0002-8703/2000/$12.00+0 4/I/100123
ultrasound beam that is filled with microbubbles2; and (3) the concentration of microbubbles in blood, t-3 At sufficiently high acoustic pressures, ultrasound destroys microbubbles. 2,4,5 W h e n the myocardial microbubble concentration is low, a single s w e e p of ultrasound can destroy enough microbubbles. O n continuous imaging at 30 Hz, therefore, by the next frame (33 ms later) myocardial VI is the same as that of the precontrast image. 6 However, if m o r e time is allowed before the next ultrasound frame, microbubbles present in the blood will partially or completely replenish the myocardium. During a continuous infusion, in w h i c h the blood concentration of microbubbles is constant, myocardial VI progressively increases with prolongation of the interval b e t w e e n the ultrasound frames (PI) until the MBV within the entire ultrasound beam is filled, at w h i c h time VI reflects MBV. 2 The rate of change of VI from l o w e r to higher PIs represents microbubble velocity (MV). 2 Myocardial VI is also independently influenced by the
AmericanHeartjournal February2000
232 Lindner et al
Table I. Clinical characteristics
Figure 1
Basilar Short-Axis
Mid Short-Axis
Vertical Long-Axis
Age (y, median)
M/F
Weight (kg) Height (cm) Heart rate (beats/rain)
Precontrast
anterior septal ~
lateral
inferior
antero-a~
Postcontrast Systolic blood pressure (mm Hg) Precontrast Postcontrast Diastolic blood pressure (mm Hg) Precontrast Postcontrast Pulse oximetr,/(% saturation) Precontrast Postcontrast
52
13/4 80 + 13 175 ± 6 64 4- 16 66 ± 16 132 ± 22 129 ± 21 77 ± 10 77 ± 10 97 ± 2 97 ± 2
infero-apical
Model used to compare MCE and SPECT data. Although measurements in any segment could be made from any view, they were always depicted in the 14 segments shown here. This example shows SPECT data represented in model.
of the infarct-related artery a n d / o r magnitude of collateral flow. xo,t t
Methods Patient population blood concentration of microbubbles. 2,3 During a continuous infusion, higher doses produce greater VI for any given MBV and PI. After a bolus injection, however, the blood concentration of microbubbles varies, not only with the dose but also the duration of injection and the cardiac output. 7 In this situation, VI truly reflects MBV only if microbubbles are not destroyed by ultrasound. 3 During a bolus injection, the blood concentration of microbubbles needs to be high to produce adequate myocardial opacification. Thus the myocardial concentration of microbubbles is usually higher during the brief period of opacification after a bolus injection compared with a continuous infusion. Because in this setting only a fraction of the microbubbles are likely to be destroyed by ultrasound, VI could still predominantly reflect MBV.3 However, because the concentration of microbubbles in the blood changes rapidly after a bolus injection, changes in VI at higher PI do not represent MV. We postulated that the assessment of regional myocardial perfusion with the use of MCE during continuous m i c r o b u b b l e infusion will correlate with myocardial isotope activity on single-photon emission c o m p u t e d t o m o g r a p h y (SPECT). We also w a n t e d to d e t e r m i n e the optimal m e t h o d (continuous infusion vs bolus injection) for the intravenous administration of an ultrasound contrast agent in the clinical setting. Accordingly, w e studied patients w i t h prior myocardial infarction (MD in w h o m MBV varies d e p e n d i n g on the degree of microvascular destruction.8,9 MV in these patients also varies, d e p e n d i n g on the patency
The study was approved by the Human Investigation Committees at the universities of Virginia and Pittsburgll. It was also approved as a phase II study by the United States Food and Drug Administration, with the primary objective of determining the effective dose range for myocardial opacification. The secondary objective was to compare MCE with SPECT in patients with prior MI who have a resting perfusion defect in a single vascular territory. All patients gave written informed consent. Ultrasound image quality needed to be adequate to evaluate all myocardial segments in the view in which the perfusion defect was best seen on SPECT. Exclusion criteria included unstable or post-Ml angina, hemodynamic instability, pregnancy, or lactation.
Single-photonemissioncomputedtomography Patients were injected at rest with either 2°lTl- (Mallinckrodt Medical, St Louis, Mo) or 99mTc-sestamibi (Dupont Pharmaceuticals, North Billerica, Mass). Data were acquired 10 minutes after injection of 2°IT1and 1 hour after administration of 99mTc-sestamibi. 20tT1 imaging was performed with a dual-headed ganmaa camera (Vertex, ADAC, Milpitas, CalL0, with 32 projections of 60 seconds, each acquired over a 180 degree orbit. 99mTc-sestamibi imaging was performed with a 3-headed gamma camera (Prism 3000S, Picker, Cleveland, Ohio), with each head acquiring 20 projections of 40 seconds, each over a 120 degree orbit. After filtered back-projection, the reconstructed 3-dimensional data were represented in a 14-segment model (Figure 1). t2 Activity in each segment was normalized to the segment with the highest activity in that view. t2
Myocardial contrast echocardiography The principle underlying the MCE assessment of MV and MBV during continuous infusion of microbubbles has been previously described. 2 After steady state is achieved, ultra-
American Hearl Journal Volume 139, Number 2, Part
Lindner et al 2 3 3
Figure 2
".
°
-
-
-
~
"2.
.
X
-2
2
X
• ,-~2. t _t ~ r , m - # J, , ? ~ t, =-t •
t
I~
b.. ~
"
i~t "~
~ .
. :.t,q) J
m
,,
*.. ,
r
m'
Ill
. .-_--..
I,y
Method of manual image alignment. Top panels, Images to be aligned. Bottom panels, "Ghost images" in green and magenta before (left) and after (right) successful alignment.
s o u n d is u s e d to destroy m i c r o b u b b l e s p r e s e n t widlin the m y o c a r d i n m . T h e i r rate of u l t r a s o u n d b e a m r e p l e n i s h m e n t is t h e n m e a s u r e d by gradu~dly increasing t h e interval b e t w e e n u l t r a s o u n d pulses. At a s h o r t PI, only a p o r t i o n o f t h e ultras o t m d b e a m is r e p l e n i s h e d , resulting in a sm,'dl increase in VI. As t h e PI progressively increases, g r e a t e r r e p l e n i s h m e n t o f die u l t r a s o u n d b e a m o c c u r s w i t h g r e a t e r increases in VI. W h e n t h e u l t r a s o u n d b e a m has b e e n c o m p l e t e l y r e p l e n i s h e d , f u r t h e r i n c r e a s e s in PI do n o t result in any c h a n g e in VI. T h e rate of increase in VI w i t h i n c r e m e n t a l PI d e n o t e s MV, w h e r e a s t h e plateau VI r e p r e s e n t s MBV. 2 Bec.'mse o f t h e limited n u m b e r o f PIs u s e d in this study, w e w e r e n o t able to fit ~m e x p o n e n t i a l
f u n c t i o n to t h e data as previously described.-' W e t h e r e f o r e fitted a line to t h e data, a n d based o n t h e slope o f t h e line determ i n e d w h e t h e r V1 progressively increased or did not increase w i t h i n c r e m e n t s in PI. T h e VI at t h e h i g h e s t PI (8 cardiac cycles) w a s a s s u m e d to r e p r e s e n t MBV. For this study, intermittent h a r m o n i c imaging w a s performed widl a phased-array system (Sonos 5500, Agilent Technologies, Andover, Mass) by u s e of transmit a n d receive freq u e n c i e s o f 1.8 a n d 3.6 MHz, respectively. Acoustic p o w e r a n d c o m p r e s s i o n w e r e m a x i m i z e d a n d gain settings w e r e o p t i m i z e d at t h e o n s e t o f e a c h s t u d y a n d h e l d c o n s t a n t t h r o u g h o u t . As p e r t h e p h a s e II p r o t o c o l , i m a g e s w e r e
234
[indner et al
American Heart Journal Februor"12000
Figure 3
Examples of perfusion defects on SPECT and MCE (latter acquired at PIs of 1, 3, and 8 cardiac cycles) in a patient with an anteroapical MI (top panel) and in a patient with an inferior MI (bottom panel). Top panel shows images in 4-chamber views; bottom panel shows images in 2-chamber views (with reorientation of corresponding SPECT image).
acquired in a single view (parasternal short-axis or apical) that corresponded to the view that best s h o w e d the p r e s e n c e of a defect on SPECT. Baseline images w e r e acquired before administration of bubbles for background subtraction. The ultrasound contrast agent used for this study was Sonovue (Bracco Diagnostics), w h i c h is c o m p o s e d of phospholipid microbubbles containing sulfur hexafluoride, with a m e a n diameter of 2.5 }am and a m e a n concentration of 2 • 108 mL-l. 13 To determine the dose range for optimal myocardial opacification, the infusion rates varied from 0.03 to 0.08 mL • kg -1 • rain -I as per the phase I1 protocol. After 2 minutes allowing for steady state, 2 at least 8 end-systolic images were acquired at each PI of 1, 3, and 8 cardiac cycles with the ultrasound transmission gated to the T wave of the electrocardiogram. t If the image orientation changed, continuous imaging was performed to regain the original orientation, after w h i c h intermittent imaging was again initiated. As per the phase II protocol, bolus injections of 0.3, 0.6, 1.2, or 1.8 mL • kg q of Sonovue were also administered in 8 patients w h o received the lowest dose for continuous infusion. These injections were delivered over a period of 20 seconds, followed by a 5-mL saline flush. On resolution of farfield attenuation from left ventricular (LV) cavity contrast, at least 8 end-systolic images were acquired at PI of 1 followed by 5 cardiac cycles. Each patient received 1 or more doses of continuous infusions and/or bolus injections, with the total dose not exceeding 0.52 mL • kg -].
Data w e r e analyzed off-line by a single observer blinded to the electrocardiographic and SPECT results, with customdesigned software. 14 Images were transferred from videotape to a computer, w h e r e precontrast images and several contrastenhanced images at each PI were selected. Each of these image sets w a s separately aligned by use of o n e of t w o m e t h o d s . Alignment was first attempted with the use of a previously described automated method.14 The gray-scale amplitude in each pixel in the image to be aligned is correlated with the gray-scale amplitude from the same pixel (same x and y coordinates) in the index image. The image is t h e n shifted 1 pixel at a time in the x and then the y direction, and the correlation is repeated at each step. The image to be aligned is then automatically shifted in the x and y directions to the coordinates w h e r e the best correlation is achieved with the index image. Because gray-scale intensity in the contrast-enhanced images obtained at different intervals is so different, automated alignm e n t may not always work. In that case, w e used a manual m e t h o d depicted in Figure 2. "Ghost images" of the frames to be aligned (panels A and B) are produced, which can be distinguished by their 2 different colors: magenta and green (panel C). One image can t h e n be moved over the other with a handheld m o u s e or by using the arrow keys until alignment is optimal (panel D). Each set of images (background and those at different P1) were t h e n separately averaged. The averaged background image was digitally subtracted from the averaged contrast-
American Heart Journal Volume 139, Number 2, Parl 1
enhanced image for each PI. The VI scale of the digitally subtracted image was expanded to 256 gray levels, and each pixel was assigned a color on the basis of a heated object algorithm, in which shades of red to orange to yellow to white represent incremental contrast enhancement. 14 The LV cavity was masked out. VI was measured in 6 regions of interest defined over the myocardium of color-coded images at each PI. These regions of interest were approximately the same in size and corresponded to the regions in the 14-segment model (Figure l). Each region of interest was made as large as possible, with care taken to exclude the specular endocardial and epicardial borders, the LV cavity, or any areas with attenuaUon or other artifacts.
Lindner et al
Figure 4
50. 40. 30. to
20.
"o
10.
Comparison of MCE and SPECTimages
0.
The activity in each myocardial segment on SPECT was compared with a normal database. The values in this database are specific for each segment and to patient sex (to account for breast attenuation in women). A segment was considered abnormal if its activity was <2 SD of the mean value for that segment derived from a large n u m b e r of normal subjects.12 The activities within segments considered abnormal or normal on SPECT and MCE were separately averaged. Segments with attenuation on MCE or SPECT were not included in the process of averaging. The change in VI with increasing PI was derived from these averaged values from which a slope was calculated by linear regression.
:rval
40.
30. -~
Statistical methods Comparisons of VI at different PIs during continuous infusions were made using either 1-way or 2-way analysis of variance, whereas VI at the 2 PIs during bolus injections were compared by use of the paired t test. CorrelaUons between the VI ratios on MCE between abnormal and normal segments and the radionuclide activity ratios between these segments on SPECT were performed with the use of least-squares linear regression. A value of P < .05 (2-sided) was considered statistically significant.
Results MCE a n d SPECT w e r e p e r f o r m e d w i t h i n 4 8 h o u r s o f e a c h o t h e r in all p a t i e n t s , a n d n o clinical e v e n t s o c c u r r e d b e t w e e n t h e t w o e x a m i n a t i o n s . Analysis o f MCE data d u r i n g c o n t i n u o u s i n f u s i o n s c o u l d n o t b e perf o r m e d in 7 o f t h e 24 p a t i e n t s b e c a u s e o f s h a d o w i n g o v e r t h e p e r f u s i o n d e f e c t s at t h e h i g h e s t d o s e s r e q u i r e d b y t h e p r o t o c o l . In 6 o f t h e s e p a t i e n t s , t h e p a r a s t e r n a l short-axis v i e w w a s u s e d b e c a u s e t h e d e f e c t w a s b e s t visualized in t h i s v i e w o n SPECT. In all t h e s e p a t i e n t s , MI w a s l o c a t e d inferiorly. T h e clinical c h a r a c t e r i s t i c s o f t h e r e m a i n i n g 17 p a t i e n t s , w h o also i n c l u d e d t h e 8 r e c e i v i n g b o l u s i n j e c t i o n s , are listed in T a b l e I. T h e b a s e l i n e VI b e f o r e m i c r o b u b b l e a d m i n i s t r a t i o n w a s h i g h e r in t h e a b n o r m a l c o m p a r e d w i t h t h e n o r m a l r e g i o n s (42 + 12% vs 36 + 12%, P < .05)
Continuous infusions MCE a n d SPECT i m a g e s f r o m 2 p a t i e n t s ( o n e w i t h a n apical a n d a n o t h e r w i t h a n i n f e r i o r d e f e c t ) are s h o w n
20.
O o
0.
Pulsing Interval
Effect of microbubble dose (x-axis) and PI (z-axis) on VI (y-axis) measured from normal myocardium during (top panel) continuous infusions and (bottom panel) bolus injections.
in Figure 3. B e c a u s e o f g r e a t e r r e p l e n i s h m e n t o f t h e b e a m e l e v a t i o n , VI in t h e n o r m a l m y o c a r d i u m w a s h i g h e r at a PI o f 8 c o m p a r e d w i t h 1 a n d 3 c a r d i a c cycles. T h e r e l a t i o n b e t w e e n VI a n d PI in t h e n o r m a l m y o c a r d i u m in all 17 p a t i e n t s at d i f f e r e n t m i c r o b u b b l e d o s e s is d e p i c t e d in Figure 4, A. As e x p e c t e d , for a n y g i v e n dose, VI w a s h i g h e r at l o n g e r PIs. For t h e s a m e PI, a h i g h e r d o s e also r e s u l t e d in a h i g h e r m y o c a r d i a l VI, b u t far-field a t t e n u a t i o n w a s s e e n o v e r at least 1 m y o c a r d i a l s e g m e n t at t h e h i g h e s t d o s e in all cases. T h e t i m e t a k e n to c o m p l e t e l y fill t h e u l t r a s o u n d b e a m a f t e r b u b b l e d e s t r u c t i o n also d e p e n d s o n t h e ultrasound beam elevation. For the ultrasound probe used in this study, t h e n o r m a l m y o c a r d i u m fills in 5 t o 6 seco n d s at rest. 2 T h e r e f o r e , b e c a u s e t h e m e a n h e a r t rate o f p a t i e n t s in this s t u d y w a s a p p r o x i m a t e l y 6 0 b e a t s / m i n
235
AmericanHeartJournal February2000
'~36 Lindneret al
Figure 6
Figure 5 0
o
•
oo
o PI1 ~=-6.17x+13.2 EE =8.14; r=O.08
0.8-
• PI3
~--0.33x+0.77 EE --0.35, t=-0.11
O 0.6-
o o >
•• to
•o
0.4•
0.2 0.2
0
•
Examples of background-subtracted color-coded images in apical 3-chamber view in patient with prior inferior MI during (left) continuous infusion at PI of 8 cardiac cycles and (right) during bolus injection at PI of 5 cardiac cycles. Infusion rate was 0.06 mL. kg -1 • min-1; bolus dose was 2.4 mL. kg -].
0 014
016
0[8
SPECT
1-
0.8-
/ 00 /
y=l.17x-0.16 SEE--0.12;r=0.73 p
O/
/
o 0.6-
o
ID
>
o
o
0.4-
0.2 0.2
o
I
I
I
0.4
0.6
0.8
I
1
SPECT
Effect of PI ( 1 and 3 in upper panel and 8 in lower panel) on relation between VI ratio on MGE (y-axis) and activity ratio on SPECT (x-axis) between abnormal and normal segments in same view.
(Table I), VI at 8 cardiac cycles probably represents relative MBV in the normal myocardium. It is for this reason that the VI ratios b e t w e e n the abnormal and normal segments obtained at this PI had a better correlation with activity ratios b e t w e e n these segments on SPECT (P < .01) than those obtained at PIs of 1 and 3 cardiac cycles (Figure 5). Figure 3 demonstrates that relative MV can also be assessed on sequential MCE images acquired at different PI during continuous infusions. VI in the apical region did not increase despite increases in PI, indicating virtually no flow to that region. In comparison, VI in the inferior wail increased at longer PIs, indicating some flow to that region. The rate of change of VI with prolongation of PI, however, was lower in the inferior compared with the anterior region, indicating slower MV and hence
lower myocardial blood flow in the inferior compared with the anterior bed. Of the 17 patients, 10 demonstrated an increase in VI (>5 VI units) at longer PI within the abnormal region, whereas 7 s h o w e d no change (VI change of <5 units). Patients with no change in VI within the infarct zone on MCE had worse (P < .01) perfusion defects on SPECT (48% + 6% activity) c o m p a r e d with those showing increase in VI at higher PI (57% + 9% activity). The correlation b e t w e e n the relative rate of change of VI in the infarct bed on MCE and radionuclide activity ratio on SPECT was r = 0.47 (P < .01). VI ratio b e t w e e n normal and abnormal beds at PI of 8 cycles (MBV ratio) and rate of change in VI in the infarct bed (MV) w e r e evaluated with other variables (including VI ratio at PIs of 1 and 3 cardiac cycles) to determine the best MCE correlates of radionuclide activity ratio. The only MCE variables that correlated with radionuclide activity ratio b e t w e e n the infarct and normal beds w e r e VI ratio at 8 cardiac cycles (a) and w h e t h e r VI increased within the infarct bed at longer PI (b). W h e n the two w e r e combined in a multiple linear regression model, the correlation with radionuclide activity improved significantly (P < .05) to r -- 0.87 (P < .0001). The final regression equation was y = 0.44a + O.11b + 0.31.
Bolus injections Figure 6 illustrates images obtained during a continuous infusion and a bolus injection in the same patient. The image during continuous infusion was acquired at a PI of 8 cycles, and that after bolus injection was acquired at a PI o f 5 cardiac cycles w h e n s h a d o w i n g o v e r the m y o c a r d i u m had just r e c e d e d . The relative VI in the abnormal and normal beds appears to be very similar with the use of both methods of microbubble administration and in both cases represents MBV. The VI ratio b e t w e e n the abnormal and normal segments on MCE
American Heorl Journal Volume 139. Number 2. Part I
Lindner el al
Figure 7
O ° ~,,,t
/••us
A
- - /-- / . .I.b. o l u s B . . . . . . ~ -
~D
............
atte_n_uatio_nth_reshold
O
O ,L3 ,.Q O z) o~,,~
t
t2
t3
Time Differences between continuous infusions and bolus injections in terms of optimal myocardial opacification. Region between dotted lines represents VI range in which relation between microbubble concentration and VI is linear. Bolus A and bolus B represent 2 bolus doses. Intervals t2-t ) and t4-ta are periods during which optimal myocardial opacification can be achieved with bolus injections during upslope or downslope of curve.
correlated modestly (P < .05) with the activity ratio between these segments on SPECT (r = 0.46 and r = 0.48 for PI of 1 and 5 cardiac cycles, respectively), probably because of the small ntunber of observations (n -- 8). Figure 4 illustrates the effect of PI and microbubble dose on VI measured from the normal myocardium during bolus injections. Because the longer PI was used w h e n the blood pool concentration of microbnbbles was declining, VI at 5 c o m p a r e d with 1 cardiac cycle was not significantly different, although it tended to be higher at 5 cardiac cycles. Regardless of the PI, VI was significantly lower, at a dose of 0.6 mL- kg -1 c o m p a r e d with higher doses. This dose also did not result in farfield shadowing. VI in the normal myocardium was identical at the 3 higher doses, all of w h i c h p r o d u c e d far-field shadowing.
Discussion With all other variables being constant, the relative myocardial uptake of a radionuclide tracer is determined by the cross-sectional area through w h i c h it can diffuse ( w h i c h is proportional to the capillary density or MBV 15) and the rate at w h i c h it is delivered (velocity of the tracer, w h i c h if m i x e d adequately in blood should equal MV). 16 Our results s h o w that w h e n both MBV and MV are measured with MCE, the correlation with relative activity on SPECT is good in patients with
prior MI. Correlation of MBV alone (VI at high PI during continuous infusion) with SPECT activity is also reasonable. Thus w h e n c o u p l e d with intermittent harmonic imaging with varying PI, continuous infusions of microbubbles provide an accurate quantification of resting perfusion defects on MCE.
Continuous infusion versus bolus injection Continuous infusions of m i c r o b u b b l e s are m o r e desirable than bolus injections because by changing the PI, information regarding both MBV and MV can be obtained.3A7 After a bolus injection, change in VI at a higher PI cannot denote MV because the concentration of microbubbles in the blood changes. Therefore at best, a bolus injection can only provide an approximation of MBV. 3 At steady state achieved during a continuous infusion of microbubbles, myocardial VI at any given PI is influenced by both MV and MBV. Thus p o o r opacification within a myocardial region could result from either reduced MV (and h e n c e myocardial blood flow) or decreased MBV (as occurs with MI). If the latter case, myocardial viability is unlikely, s-1 ] If MV is decreased and MBV is normal, myocardial viability is more likely. Thus by the use of continuous infusion, greater information can be obtained than by the use of bolus injection. There are o t h e r reasons for using continuous infu-
237
American HeartJournal February 2000
2 3 8 Lindner et al
Figure 8 % 4¢
z
4q,' ,
, 21
50"
' ; Iq
'
MCE images in a patient with prior inferior MI before (left panel) and after (right panel) infusion of microbubbles. Top panels show images alone; bottom panels depict VI in 6 myocardial segments in this view.
sions. A narrow range exists in which the relation between microbubble concentration and VI is linear (approximately 50 to 60 background-subtracted VI units).l For better signal-to-noise ratio, it is desirable to use the higher end of this range (the region b e t w e e n the dotted lines in Figure 7), which can be easily achieved by titrating the rate of continuous infusion. Higher doses produce far-field attenuation, whereas lower doses preclude adequate myocardial opacification (Figure 7). Once the infusion dose has b e e n optimized, there is adequate time to interrogate each myocardial segment in detail. A single operator can start the infusion, adjust the rate of administration, and perform the entire examination. If the orientation of the heart changes during the examination, imaging can be changed to real-time, the view can be reoriented, and imaging can then be resumed. Finally, poor regional myocardial opacification can result from either a true perfusion defect or modest attenuation (in which the myocardium is still visualized). In this situation, change in VI with change in PI is more likely to reflect an artifact, especially if the rate of change is the same as in the normal myocardium. If a single PI is used, it is more difficult to differentiate an artifact from a real perfusion defect.
After a bolus injection, the blood concentration of microbubbles varies over time. Microbubble concentration often exceeds the level in which its relation with VI is no longer linear (bolus A in Figure 7). The high concentration of microbubbles in the blood within the LV cavity also produces far-field shadowing. The period of optimal myocardial opacification, in which shadowing has not yet started (t2-t l) or has receded (t4-t 3) and w h e n the signal-to-noise ratio within the myocardium is favorable, is very brief. Ideally, one could design a bolus dose that will result in myocardial microbubble concentrations only within the effective range (bolus B in Figure 5). However, because myocardial VI depends on several factors (such as cardiac output, 7 acoustic impedance of the chest wall, m homogeneity of the ultrasotmd field, etc), an ideal bolus dose cannot be predicted. The rate of a constant infusion, however, can be adjusted during the examination to accotmt for these factors. Bolus injections pose other problems as well. Because the imaging plane must be correctly aligned and because motion artifacts caused by respiration are common, more than 1 injection may be required per view. During continuous infusion, images with motion artifacts can be recognized easily, and more images can be acquired at the s;mle PI. Unsuitable frames can later be rejected. Bolus injections require one operator to acquire the images and another to inject the contrast agent. In our study, better myocardial opacification was achieved at higher m i c r o b u b b l e doses, until a dose was reached w h e r e despite better opacification of the near-field myocardium, far-field shadowing occurred. This dose was required by the protocol for dose ranging. In the apical view, higher blood pool concentrations could be used before shadowing affected myocardial regions. Shadowing was limited over the left atrium, whereas good myocardial opacification was achieved. Thus the optimal blood c o n c e n t r a t i o n s during c o n t i n u o u s infusions of m i c r o b u b b l e s can vary with different views. A fine balance must be achieved b e t w e e n good myocardial opacification and the least a m o u n t of shadowing, w h i c h can be accomplished by adjusting the rate of infusion. Because this was a dose-ranging study, w e could not adjust the rate of infusion. During bolus injections, VI was measured only w h e n shadowing over the myocardium was receding (proximity of t 3 in Figure 7). VI in the normal myocardium was therefore similar for all doses that produced shadowing (bolus A in Figure 7). The lowest dose that did not cause attenuation produced a lower VI (diagrammatically represented as bolus B in Figure 7). Even if VI were measured over the myocardium in the near field not affected by shadowing, VI probably would have reached a plateau at higher doses because of signal saturation at those doses. This could result
American Heart Journal Volume 139, Number 2, Part 1
in apparent myocardial opacifications in regions with low perfusion.
Lindner el al
Figure 9
Advantages of quantification An), measurement should be more accurate than a subjective assessment. Because acoustic pressures within the ultrasound field are inhomogeneous, microbubble destruction and hence myocardial VI may vary b e t w e e n different beds in the same view and within the same bed in different views. Quantification provides a comparison with values obtained in normal individuals, thus increasing both the sensitivity and specificity for detecting perfusion defects.19 The interobserver error is also minimized and reproducibility is maximized, z0 Another reason for quantification is that myocardial regions with prior MI frequently have a higher VI during harmonic imaging at baseline, which is not seen on fundamental imaging. The reason for this increase in VI on harmonic imaging is unclear but may be related to either better signal-to-noise3 or nonlinear properties of infarcted tissue, el This increase in myocardial VI in infarcted tissue makes the change in VI after microbubble administration more difficult to appreciate. The left panel in Figure 8 shows a higher VI measured in the posterior compared with the anterior wall before microbubble injection in a patient with prior inferior MI, which is not as apparent visually. Both visually and quantitaUvely VI is equal during microbubble infusion in both walls at a PI of 8 cardiac cycles. However, w h e n both the precontrast and contrast-enhanced images are evaluated side by side, the change in VI is less in the posterior than in the anterior wall, indicating a lower MBV (and hence a prior MI) in the posterior bed. The same result can be achieved by background subtraction. The top panels in Figure 9 depict color-coded MCE images obtained after background subtraction of the left from the right image in Figure 8. As can be noted from the numeric values in the top right panel, the color differences also denote the relative changes in VI in different myocardial regions b e t w e e n the two sets of images. The corresponding color-coded SPECT images are shown in the bottom panels in Figure 9. The advantage of color over gray scale is that although the human eye can discern only a few shades of gray, it can discriminate between thousands of hues of color. Thus perception of VI differences is easier. In panels C and D in Figure 7 and in Figure 3, perfusion defects visually appear similar w h e n MCE and SPECT data are color-coded.
Comparison with previous studies Although an experimental study from our laboratory has compared continuous infusions with bolus injections for stenosis detection, 17 to our knowledge this is the first study that has compared continuous infusions with bolus injections of microbubbles in the setting of
~,
-~
~l
l
~li
Ski
Left panels show color-coded MCE images obtained after digital subtraction of images in left panels from that in right in Figure 8. Top panel shows image without VI measurements; bottom panel shows same image with VI measurements. These measurements reflect difference in VI measured in same segments on gray-scale images (bottom panels in Figure 8). Right panels show corresponding color-coded SPECTimages without (top) and with (bottom) quantitative measurements performed in different myocardial segments corresponding to similar segments on MCE. Values on sides represent average for normal and abnormal segments, respectively.
prior MI. Perfusion defects on MCE have been shown to correlate well with those on SPECT with bolus injections, z2 but a quantitative assessment has not been performed. Continuous infusions and variable PI have also been used for detecting reversible perfusion defects with MCE performed in conjunction with dobutamine echocardiography.23
Limitations of the study The n u m b e r of patients is small. However, it is adequate to address the principal aims of the study. Because we were limited by the total deliverable dose, we were unable to acquire data at PI >8 cardiac cycles during c o n t i n u o u s infusions. On the basis of our previous observations, we believe that VI reached a plateau within all normal myocardial beds at or before 8 cardiac cycles. However, complete replenishment of the ultrasound beam may not have occurred in abnormal segments, particularly if resting flow was markedly
239
American Hearl Journal February 2000
2 4 0 Lindner el al
reduced in these segments. Thus w e cannot make an assessment regarding lack of myocardial viability in segments showing decreased VI at 8 cardiac cycles. Ideally, images should have been acquired at several PIs so that ,an exponential function could be fitted to the PI versus VI data. 2,17 We w e r e tmable to do so with the limited n u m b e r of PIs used in this study. During bolus injections, w e have a limited time in w h i c h to acquire data and w e r e thus unable to go beyond a PI of 5 cardiac cycles. Some patients received 2°tTl instead of 99mTc. Because the latter does not redistribute, it is ideal for comparison with MCE. Although imaging was initiated very early after 201Tl injection, redistribution could still have occurred. Substantial redistribution in patients w i t h p r i o r MI and no e v i d e n c e of resting ischemia is, however, unlikely. 24 Although the correlation b e t w e e n activity ratios on SPECT and VI ratios on MCE b e t w e e n the abnormal and normal regions w e r e not different for 2°IT1 or 99mTc, the numbers were too small to afford a meaningful comparison. Because of the limit imposed on the dose of microbubbles that could be administered in individual patients, w e only obtained a single view on MCE. Finally, because w e selected MCE views based on SPECT results, our study design was biased in favor of finding a perfusion defect on MCE. For this reason, w e limited ourselves to a quantitative analysis.
We thank Anthony Marano, MD, Bruce Helmly, MD, and Raju Modg MD, f o r patient recruitment as well as A m a n d a Doss, MEd, Nancy DLxon, and Virginia Schneider, RN, f o r technical assistance.
References 1. Skyba DM, Jayaweera AR, Goodman NC, et aJ. Quantification of myocardial perfusion with myocardial contrast echocardiography from left atrial injection of contrast: implications for venous injections. Circulation 1994;90:1513-21. 2. Wei K, Jayoweera AR, Firoozan S, et al. Quantification of myocardial blood flow with ultrasound-induced destruction of microbubbles administered as a constant venous infusion. Circulation 1998;97:473-83. 3. Firschke C, Lindner JR, Wei K, et al. Myocardial perfusion imaging in the setting of coronary artery stenosis and acute myocardial infarction using venous injection of a second-generation echocardiographic contrast agent. Circulation 1997;96:959-67. 4. Wei K, Skyba DM, Firschke C, et al. Interaction between microbubbles and ultrasound: in vitro and in viva observations. J Am Coil Cordial 1997;29:1081-8. 5. Skyba DM, Price RJ, Linka AZ, et al. Direct in viva visualization of microvessel rupture and tissue injury caused by intravascular destruction of microbubbles by ultrasound. Circulation 1998;98:290-3. 6. Pelberg RA, Wei K, Kamiyama N, et al. Potential advantage of 'flash echocardiography' for digital subtraction of B-mode images acquired during myocardial contrast echocardiography. J Am Sac Echocardiogr 1999; 12:86-93. 7. Villanueva FS, Glasheen WP, Sklenar J, et al. Successful and reproducible myocardial opacificalion during 2-dimensional echocardio-
8.
9.
10.
11.
12.
! 3.
14.
15. 16.
17.
18.
19.
20.
21. 22.
23.
24.
graphy from right heart injection of contrast. Circulation 1992;85: 1557-64. Villanueva FS, Glasheen WP, Sklenar J, et al. Characterization of spatial patterns of flow within the reperfused myocardium using myocardial contrast echocardiography: implications in determining the extent of myocardial salvage. Circulation 1993;88:2596-606. Villanueva FS, Camarano G, Ismail S, et al. Coronary reserve abnormalities during post-infarct reperfusion: implications for the timing of myocardial contrast echocardiography to assess myocardial viability. Circulation 1996;94:748-54. Sabia PJ, Powers ER, Jayaweera AR, et al. Functional significance of collateral blood flow in patients with recent acute myocardial infarction: a study using myocardial contrast echocardiography. Circulation 1992;85:2080-9. Sabia PJ, Powers ER, Ragosta M, et al. An association between collateral blood flow and myocardial viability in patients with recent myocardial infarction. N Engl J Med 1992;372:1825-31. Smith WH, Kastner P,J,Calnon DA, et aL Quantitative gated single photon emission computed tomography imaging: a counts-based method for display and measurement of regional and global ventricular systolic function. J Nucl Cardiol 1997;4:451-63. Schneider M, Arditi M, Barrau MB, et al. BR1: a new ultrasonographic contrast agent based on sulfur hexafluoride-filled microbubbles. Invest Radial 1995;30:451-7. Jayaweera AR, Sklenar J, Kaul S. Quantification of images obtained during myocardial contrast echocardiography. Echocardiography 1994; 1 1:385-96. Kassab GS, Lin DH, Fung YB. Morphometry of pig coronary venous system. Am J Physiol 1994;267:H2100-13. Leppo JA. Cardiac transport of single photon myocardial perfusion agents. In: Zaret BA, Belier GA, editors. Nuclear cardiology: state of the art and future directions. St Louis, Mo: Mosby; 1993. p. 35-44. Wei K, Linka A, Jayaweera AR, et aL Basis of stenosis detection by myocardial contrast echocardiography during venous administration of microbubbles: bolus or continuous infusion? J Am Coil Cardiol 1998;32:252-60. Ismail S, Jayaweera AR, Camarano G, et al. Relation between airfilled albumin microbubble and red blood cell rheology in the human myocardium: influence of echocardiographic systems and chest wall attenuation. Circulation 1996;94:445-51. Kaul S, Boucher CA, Newell JB, et al. Determination of quantitative thallium parameters that optimize detection of coronary artery disease. J Am Coil Cardiol 1986;7:527-3. Kaul S, Chesler DA, Okada RD, et al. Computer versus visual analysis of exercise thallium-2Ol images: a critical appraisal in 325 patients with chest pain. Am Heart J 1987; 114:1129-37. Kaul S. Assessing the myocardium after attempted reperfusion: should we bother? Circulation 1998;98:625-7. Kaul S, Senior R, Dittrich H, et al. Detection of coronary artery disease with myocardial contrast echocardiography: comparison with 9¢'mTc-sestamibisingle-photon emission computed tomography. Circulation 1997;96:785-92. Porter TS, Li S, Kriscfield D, et al. Detection of myocardial perfusion in multiple echocardiographic windows with one intravenous injection of microbubbles using transient response imaging. J Am Coil Cardiol 1997;29:791-9. Sabia PJ, Powers ER, Ragosta M, et al. Role of quantitative planar thallium-201 imaging for predicting viability in patients with acute myocardial infarction and a totally occluded infarct-related artery. J Nucl Med 1993;34:728-36.