Author + information
- Received August 2, 2000
- Revision received October 24, 2000
- Accepted November 29, 2000
- Published online March 1, 2001.
- Thomas R Porter, MD, FACCa,* (, )
- Feng Xie, MDa,
- Mary Silver, RN, BSNa,
- David Kricsfeld, BSa and
- Edward O’Leary, MDa
- ↵*Reprint requests and correspondence: Dr. Thomas R. Porter, University of Nebraska Medical Center, 981165 Nebraska Medical Center, Omaha, Nebraska 68198–1165
We sought to determine how successful pulse inversion Doppler (PID) imaging would be in detecting myocardial perfusion defects during dobutamine stress echocardiography.
By transmitting multiple pulses of alternating polarity (PID) at a low mechanical index, myocardial contrast enhancement from intravenously injected microbubbles can be detected using real-time frame rates.
Pulse inversion Doppler imaging was performed in 117 patients during dobutamine stress echocardiography by using an intravenous bolus of a perfluorocarbon-filled, albumin- (Optison: n = 98) or liposome- (Definity: n = 19) encapsulated microbubble and a mechanical index of <0.3. The visual identification of myocardial contrast defects and wall motion abnormalities was determined by blinded review. Forty of the patients had quantitative angiography (QA) performed to correlate territorial contrast defects with stenosis diameter >50%.
There was a virtual absence of signal from the myocardium before contrast injections in all patients. Bright myocardial opacification at peak stress was observed in at least one coronary artery territory at frame rates up to 25 Hz in 114 of the 117 patients during dobutamine stress echocardiography. Regional myocardial contrast defects at peak stress were observed in all 30 patients with >50% stenosis in at least one vessel (13 with single-vessel and 17 with multivessel disease). Contrast defects were observed in 17 territories subtended by >50% diameter stenosis that had normal wall motion at peak stress. Overall agreement between QA and myocardial contrast enhancement on a territorial basis was 83%, as compared with 72% for wall motion.
Pulse inversion Doppler imaging allows the detection of myocardial perfusion abnormalities in real-time during stress echocardiography and will further add to the quality and sensitivity of this test.
Recent discoveries have shown that myocardial contrast enhancement after intravenous perfluorocarbon-containing microbubbles can be achieved in real-time with harmonic imaging at a low mechanical index. Unlike the myocardial contrast achieved with intermittent imaging at higher mechanical indexes (1–3), a lower index reduces microbubble destruction and permits the detection of harmonic myocardial contrast enhancement at frame rates of 10 to 20 Hz (4,5).
A lower mechanical index can also be used with pulse inversion Doppler (PID) imaging by examining nonlinear scatter in the Doppler frequency domain instead of the radiofrequency domain (6). With PID imaging, three to five ultrasound pulses are transmitted into the tissue, with each sequential pulse containing the opposite configuration of the preceding pulse. In vitro studies have demonstrated a significantly greater intensity of the contrast signal obtained with PID imaging as compared with harmonic imaging (6). In addition, there is no linear scatter overlap in the Doppler frequency domain; thus, enhanced real-time myocardial contrast enhancement can be achieved with reduced background tissue interference (7). We hypothesized that PID imaging would be clinically useful during dobutamine stress testing, where both wall motion and perfusion are important in detecting angiographically significant coronary artery disease (8).
The study involved 117 consecutive patients undergoing dobutamine stress echocardiography. The mean age of the patients was 62 ± 12 years (60 women). The studies were done in patients who had an intermediate pretest probability of significant coronary artery disease based on cardiac risk factors and clinical presentation (9). Forty-eight of the patients had atypical chest pain with risk factors, 38 had typical chest pain and 31 had multiple risk factors without chest discomfort. Thirty-three patients (28%) had diabetes, 78 (67%) had hypertension, 25 (21%) had a history of myocardial infarction and 34 (29%) had previous coronary artery bypass graft surgery or percutaneous revascularization. The contrast agents used for the study were either a ‘perfluorocarbon-containing albumin-coated microbubble (Optison, Mallinckrodt, St. Louis, Missouri) or a lipid-encapsulated perfluoropropane-filled microbubble (Definity by DuPont Pharmaceuticals, Wilmington, Delaware). Left ventricular ejection fraction was calculated using an adaptation of the cylinder hemi-ellipsoid formula, as described by Vuille and Weyman (10), and was >40% in all patients. The entire study was approved by the Institutional Review Board at the University of Nebraska Medical Center.
Pulse inversion Doppler imaging set-up
Pulse inversion Doppler imaging was performed using an HDI 5000 scanner (Advanced Technology Laboratories, Bothell, Washington) and a 1.7-MHz transducer. The mechanical index was set between 0.12 and 0.18 for all Optison studies. A slightly higher mechanical index of 0.24 to 0.26 was chosen for Definity, based on previous observations in humans showing lesser degrees of attenuation and greater myocardial contrast at this mechanical index as compared with the range 0.12 to 0.18 (11). Time gain compensation, scan line density and frame rates for both agents are displayed in Table 1.
Dobutamine stress protocols
Baseline contrast-enhanced images were obtained in all three apical windows (four-chamber, two-chamber and long-axis views) after either a 0.3-ml bolus injection of Optison or a 0.15-ml injection of Definity. Dobutamine was then infused in 3- to 5-min stages of 5, 10, 20, 30 and, if necessary, 40 μ/kg per min. Atropine (up to 2 mg) was added to achieve >85% predicted maximal heart rate. Repeat contrast-enhanced images were obtained during the peak dobutamine infusion using the same apical views that were acquired under rest conditions, as well as the same doses of ultrasound contrast agent. The rest and peak dobutamine contrast-enhanced images were displayed on a quad screen format for blinded review of contrast enhancement. In 55 patients, contrast enhancement during an earlier low dose of dobutamine (10 to 20 μ/kg per min) was also assessed.
Analysis of contrast enhancement and wall motion
One independent reviewer (F.X.), who had no knowledge of the patients’ history, angiographic findings or other clinical data, examined the videotape and digitized images of all studies and graded both contrast enhancement and the location of any defects. Myocardial contrast enhancement and wall motion were evaluated according to segments and coronary artery distribution. Anteroseptal, apical and anterior segments were considered to be in the left anterior descending coronary artery (LAD) distribution, lateral segments in the left circumflex artery (LCx) and inferior segments in the right coronary artery (RCA) territory. The posterior wall was considered an overlap region, which was assigned to either the LCx or RCA distribution, as determined by quantitative angiography (QA). The period at which a contrast defect appeared after the bolus injection of contrast agent was also calculated from videotape review. The time (s) of when a defect first appeared in relation to the peak period of contrast enhancement until it disappeared from all segments was measured for all visually discernible defects.
The degree of myocardial contrast enhancement was graded in the basal, mid and distal segments of each wall, as well as the apex, by using the following scoring criteria: 0 = no enhancement; 1+ = mild enhancement; 2+ = bright contrast enhancement. If a score of 0 or 1+ was given to a segment, the reviewer determined whether the lack of enhancement was due to attenuation or a contrast defect. A contrast defect was considered present when two contiguous segments failed to exhibit contrast enhancement as compared with other segments at the same depth in the same view, and by comparing the contrast defect with side by side baseline images. Attenuation from contrast or lung interference was considered present if the segment could not be visualized or was not distinguishable from surrounding tissue.
A wall motion abnormality consistent with ischemia was considered present whenever hypokinesia, akinesia or dyskinesia was observed only at peak stress. A segment or territory was scored as an infarction whenever hypokinesia, akinesia or dyskinesia was present at rest and did not change during the low dose or peak dobutamine stress imaging.
Quantitative measurements were made by an experienced interventional cardiologist (E.O.) using a hand-held electronic caliper (Tesa S.A., Renens, Switzerland) operated with custom-developed PC software (12). Measurements were expressed as the percent diameter narrowing using the diameter of the nearest normal-appearing region as a reference. Coronary artery narrowings of >50% and 70% diameter in the native or graft vessel were both used as cut-off values. The interventional cardiologist also estimated the coronary artery territory subtended by each stenosis with PID imaging or wall motion.
Kappa values were calculated to compare segmental and territorial contrast enhancement and wall thickening with angiographic measurements. Interobserver agreement was also calculated by having a second independent reviewer analyze the segmental and territorial wall motion and contrast enhancement in 45 patients. Contingency tables were used to compare differences in contrast enhancement or concordance with angiography when using Definity versus Optison.
Assessment of segmental contrast enhancement with pulse inversion Doppler imaging
Using the presets displayed in Table 1, bright 2+ myocardial contrast was observed in at least two contiguous segments of one coronary artery territory in 109 of the patients (93%) under rest conditions and in 114 patients (97%) at peak dobutamine stress. Overall, attenuation was considered present in 335 (16%) of the 2,073 segments, with 304 of these (91%) in basal segments. The incidence of segmental attenuation was higher with intravenous Definity (24% of the total segments) than with Optison (16% of total segments; p < 0.001).
Sixty of the patients were considered to have normal myocardial contrast enhancement in all coronary artery territories at rest and at peak dobutamine stress, whereas 57 patients (total of 83 coronary artery territories) had a contrast defect at peak dobutamine stress (75 inducible, 8 fixed). Wall motion and PID imaging were in agreement in 91% of the territories at peak stress (kappa 0.70). However, wall motion was considered normal in 31 (37%) of the 83 coronary artery territories with contrast defects.
Eleven patients (9%) had rest wall motion abnormalities. Seven of these patients did not exhibit an improvement in wall motion during low dose dobutamine infusion, and therefore were classified as having an infarction. Four patients improved during low dose dobutamine, but had deterioration in wall motion at peak stress, and thus were classified as having ischemia. Ten of these 11 patients exhibited rest contrast defects by PID imaging in the regions where wall motion was abnormal.
Angiographic results in patients with contrast defects on PID imaging
Ten of the 40 patients had <50% diameter stenoses in all vessels. Maximal percent diameter stenosis was between 20% and 45% in six of these patients, and in these patients both wall motion and myocardial contrast with PID imaging were normal in five. In four patients, the percent diameter stenosis was <10% in all vessels. Pulse inversion Doppler imaging was abnormal in all four of these patients and wall motion was abnormal in three at peak stress. These false positive contrast defects were in the anteroseptal and apical regions in two patients; in the inferior, lateral and posterior regions in one patient; and in the lateral wall in one patient.
Thirty patients undergoing angiography had significant coronary artery stenosis (13 with single-vessel and 17 with multivessel disease). Contrast defects with PID imaging were visually evident in at least one territory at peak stress in all patients. The time of first visualization of these contrast defects, however, was after the peak period of myocardial contrast. The mean time to the first visual appearance of an inducible contrast defect was 4 ± 3 s (range 0 to 10) after peak stress. An example of the delayed appearance of an inducible contrast defect is demonstrated in Figure 1, showing the PID image of a patient who had an inducible anteroseptal and apical wall motion abnormality during dobutamine stress. At the peak period of left ventricular cavity contrast (Fig. 1, top panel), myocardial contrast enhancement is transiently evident. At 5 s after peak myocardial contrast enhancement (Fig. 1, bottom panel), the contrast defect appears. In comparison, in the seven coronary territories that were considered as having infarction, on the basis of rest and low dose dobutamine wall motion responses, all but one exhibited contrast defects, even at the peak period of contrast enhancement. In coronary artery territories that were subtended by <50% diameter stenosis, contrast enhancement persisted for 17 ± 6 s (range 10 to 27 s) after peak enhancement.
In the 30 patients with angiographic disease, there were a total of 55 coronary artery territories supplied by >50% diameter stenosis and 40 territories supplied by >70% diameter stenosis. Contrast defects at peak stress were observed in 46 (84%) of the 55 territories supplied by >50% diameter stenosis and in 34 (85%) of the 40 territories supplied by >70% diameter stenosis. Figures 1 through 3demonstrate inducible perfusion defects in the LAD (Fig. 1), LCx (Fig. 2)and RCA (Fig. 3)territories of patients with >50% diameter stenosis supplying each territory. In each of these cases, there were no wall motion abnormalities at peak stress.
Using the 50% diameter cut-off value, overall concordance between myocardial contrast and QA on a territorial basis was 83% (kappa 0.65). The overall concordance between wall motion and angiography was 72% (p = 0.07 vs. PID agreement with angiography). The PID imaging results were in agreement with the QA results in 83% of the LAD territories, 80% of the RCA territories and 85% of the LCx territories. In comparison, wall motion was in agreement with QA in 68% of the LAD territories. In the 22 territories that exhibited abnormal contrast enhancement at peak stress, but were still considered to have normal wall motion, 17 (77%) had >50% coronary stenosis subtending the territory. There was no differences in the rate–pressure product achieved between patients who had inducible contrast defects without wall motion abnormalities (21,500 ± 4,200) and those with both contrast defects and wall motion abnormalities (19,900 ± 3,500), as compared with patients who had both inducible wall motion abnormalities and contrast defects (n = 22 territories in 16 patients). There was also no difference in the percent diameter stenosis subtending the coronary artery territory between the two groups (73 ± 14% vs. 79 ± 17%; p = 0.34). However, contrast defects were observed before the development of wall motion abnormalities. In the 22 patients who underwent QA and who had both inducible wall motion abnormalities and contrast defects at peak stress, 12 (55%) had contrast defects during low dose dobutamine infusion (Fig. 4).
There was 84% agreement between the reviewers who scored contrast enhancement (kappa 0.63), and 91% agreement (kappa 0.64) between reviewers who scored wall motion.
In this study, we demonstrated that the simultaneous assessment of myocardial perfusion and wall thickening could be performed with PID imaging. The advantages of PID imaging over B-mode harmonic imaging include the virtual elimination in background signal noise before injection (demonstrated in Fig. 4) and the enhanced nonlinear backscatter produced by the pulse inversion waveform at this very low mechanical index. In >90% of the patients in this study, we observed bright myocardial opacification from small intravenous bolus injections of either Optison or Definity.
Rationale for a low mechanical index with PID imaging
Regardless of the imaging technique or location of microbubble injection, myocardial contrast enhancement from microbubbles has been shown to reflect myocardial blood volume (13,14). Myocardial blood volume distal to >50% diameter coronary stenosis is reduced during pharmacologic stress, resulting in the appearance of contrast defects (3). The regional location of contrast defects observed during pharmacologic stress after intravenous bolus injections of microbubbles in humans is closely correlated with radionuclide uptake (3,4).
Until recently, detection of these defects has required intermittent imaging to reduce microbubble destruction. More recent investigations in humans have shown that nonlinear scatter from microbubbles can be observed at faster frame rates if a mechanical index of <0.4 is used (4,5,15). These initial studies were performed with B-mode harmonic imaging, during which linear echoes from tissue at baseline often mask or interfere with the nonlinear scatter from microbubbles. This interference from linear echoes was not present with PID imaging, because in this setting, wall filters can be used to completely eliminate any linear scatter from tissue (6). Furthermore, in vitro studies have demonstrated that the nonlinear response from microbubbles, when using PID imaging, is >10 dB higher than the response with harmonic imaging (7). Therefore, we observed both bright myocardial contrast from intravenous Optison or Definity and significantly less background tissue interference.
Improved concordance with QA
Our study demonstrated that the addition of PID imaging to dobutamine stress testing may improve the sensitivity of detecting coronary artery disease. Although wall motion and contrast enhancement were in agreement in 90% of the coronary artery territories, there were a significant number of territories subtended by >50% diameter stenosis, which exhibited contrast defects despite normal wall motion. Similar findings have been observed with radionuclide single-photon emission computed tomographic studies comparing wall motion and perfusion with angiography during dobutamine stress (8). The patients who exhibited abnormal contrast enhancement but normal wall motion did not have lower rate–pressure products or lesser degrees of stenosis severity. The most likely explanation for the improved concordance observed with perfusion imaging is that the threshold for inducing a perfusion defect is lower. We observed perfusion defects during the lower dose dobutamine stages in 55% of the patients who had both wall motion abnormalities and perfusion defects at peak stress, which indicates that the target heart rate required to detect significant coronary artery disease with dobutamine perfusion imaging is lower.
Potential limitations of interpretation of PID imaging
Although we could easily identify perfusion defects in real-time with PID imaging, the dynamic range of the technique is a potential limitation. This was most evident with bolus injections during dobutamine stress, where visual myocardial contrast was observed even in ischemic regions that were transiently equivalent to normally perfused territories (Fig. 1). Because of the reduced dynamic range, the transient increase in microbubble concentration reaching abnormally perfused segments after a bolus injection was able to create a brief period of bright myocardial contrast enhancement. Therefore, it is essential to have a period of at least 5 s after the peak period of contrast enhancement before any image interpretation and digital acquisition.
Furthermore, we consistently observed basal segment attenuation with PID imaging at the low mechanical index—a problem that may reduce the ability of this technique to detect myocardial perfusion defects confined to these segments. This phenomenon was observed with both agents used in this study, but slightly more frequently with the liposome-encapsulated agent. Continuous infusions, as opposed to bolus injections, may reduce the far-field attenuation, as already observed with harmonic imaging (16). The size and shell of microbubbles may also play a role in the degree of attenuation observed, because more uniform polymer-coated microbubbles have demonstrated less attenuation, as compared with Optison, which is albumin-coated (17).
Conclusions and future directions
Pulse inversion Doppler imaging with a low mechanical index can be used to analyze both myocardial perfusion and wall motion simultaneously during dobutamine stress echocardiography. The advantages of PID imaging over harmonic imaging include an absence of background signal and a higher signal to noise ratio. Angiographically significant coronary artery stenoses were detected with PID imaging; these would have been considered normal on the basis of wall motion analysis alone. However, because not all patients had angiography in our study, we cannot assess the actual sensitivity and specificity of PID imaging. A multicenter study is needed to assess the feasibility of detecting myocardial perfusion defects with PID imaging during dobutamine stress echocardiography, and whether it will significantly improve the test’s sensitivity (without lowering specificity), as compared with wall motion analysis.
☆ Grant support was provided by DuPont Pharmaceuticals, Inc., Molecular Biosystems, Inc. and Advanced Technology Laboratories.
- left anterior descending coronary artery
- left circumflex coronary artery
- pulse inversion Doppler
- quantitative angiography
- right coronary artery
- Received August 2, 2000.
- Revision received October 24, 2000.
- Accepted November 29, 2000.
- American College of Cardiology
- Porter T.R,
- Xie F,
- Kricsfeld D,
- Armbruster R.W
- Porter T.R,
- Li S,
- Kricsfeld D,
- Armbruster R.W
- Kaul S,
- Senior R,
- Dittrich H,
- Raval U,
- Khattar R,
- Lahiri A
- Elhendy A,
- Geleijnse M.L,
- Roelandt J.R.T.C,
- et al.
- ↵Vuille C, Weyman AE. Left ventricle I: general considerations, assessment of chamber size and function. In: Weyman AE, editor. Principles and Practice of Echocardiography. 2nd ed. Lippincott Williams & Wilkins, 1994:575–624.
- Porter T.R,
- Xie F,
- Kricsfeld D,
- Silver M
- Uehata A,
- Matsuguchi T,
- Bittl J.A,
- et al.
- Ismail S,
- Jayaweera A,
- Goodman N.C,
- Camarano G.P,
- Skyba D.M,
- Kaul S