Article Text

Download PDFPDF

Cardiac imaging and non-invasive testing
Absolute myocardial blood flow determination using real-time myocardial contrast echocardiography during adenosine stress: comparison with single-photon emission computed tomography
  1. S S Abdelmoneim1,
  2. A Dhoble1,
  3. M Bernier1,
  4. S Moir1,
  5. M E Hagen1,
  6. S A C Ness1,
  7. S S Abdel-Kader2,
  8. P A Pellikka1,
  9. S L Mulvagh1
  1. 1
    Division of Cardiovascular Diseases, Mayo Clinic, Rochester, Minnesota, USA
  2. 2
    Department of Cardiovascular Medicine, Assiut University, Assiut, Egypt
  1. Correspondence to Dr Sharon L Mulvagh, Division of Cardiovascular Diseases, Mayo Clinic, 200 First Street SW, Rochester, MN 55905, USA; smulvagh{at}mayo.edu

Abstract

Objective: To assess the feasibility and diagnostic accuracy of real-time myocardial contrast echocardiography (MCE)-derived absolute myocardial blood flow for detection of myocardial perfusion abnormalities compared with simultaneous technetium 99 m sestamibi single-photon emission computed tomography (SPECT).

Design: Prospective study.

Setting: Tertiary-care medical institution.

Patients: 79 patients with known or suspected coronary artery disease.

Interventions: Simultaneous SPECT and real-time MCE during adenosine stress.

Main outcome measures: Absolute myocardial blood flow (MBF, ml/min/g), microbubble velocity (β, min−1), and reserve values. Endpoints included sensitivity, specificity, positive likelihood ratio (LR+) or negative likelihood ratio (LR−) and area under the curve (AUC) of the receiver operating characteristic curve.

Results: Reserve measurements were feasible in 975 of 1343 segments (73%); of these, 130 segments (13%) were abnormal by SPECT. MCE perfusion parameters clearly distinguished abnormal from normal segments for β reserve (1.13 (0.99) vs 2.22 (1.36), p<0.001) and MBF reserve (1.80 (2.29) vs 3.69 (2.79), p<0.001). The β reserve cut-off of 1.60 provided the following: AUC, 0.787; sensitivity, 82%; specificity, 66%; LR+, 2.40; and LR−, 0.28. The MBF reserve cut-off of 1.90 provided the following: AUC, 0.779; sensitivity, 73%; specificity, 72%; LR+, 2.69; and LR−, 0.37. MBF reserve had an AUC of 0.773 for the left anterior descending coronary artery, 0.885 for the left circumflex coronary artery and 0.739 for the right coronary artery.

Conclusions: Real-time MCE-derived absolute MBF, β, and reserve values are feasible and accurate for detecting myocardial perfusion abnormalities as defined by SPECT.

View Full Text

Statistics from Altmetric.com

Quantitative myocardial contrast echocardiography (MCE) has been developed to measure myocardial perfusion with kinetics data from replenishment curves (1). The video intensity y (t) is described by an exponential function:

y  =  A (1−eβt)

where A (dB) represents the plateau acoustic intensity (reflecting myocardial blood volume), β (s−1) represents the rate of increase in acoustic intensity (reflecting microbubble velocity) and the product of Aβ (dB/s) is a semiquantitative estimate of myocardial blood flow (MBF).1 MCE-derived semiquantitative MBF correlates well with intracoronary Doppler flow wire measurements in animals and invasive coronary flow reserve measurements in humans.2 3 In addition, previous studies reported on the diagnostic accuracy of quantitative real-time MCE imaging during adenosine stress in patients with known or suspected coronary artery disease (CAD) and used different reference tests for coronary stenosis detection, including anatomical localisation of stenosis by coronary angiography4 5 6 and functional localisation of perfusion defects with technetium 99m sestamibi single-photon emission computed tomography (SPECT).7

Recently, MCE-derived MBF measurements in absolute units of blood volume per time in relation to myocardial mass (ml/min/g) have been developed by Vogel and colleagues.8 Absolute MBF measurement has been validated against positron emission tomography and quantitative coronary angiography in various patient populations.8 9 10 11 12 Our prospective study was performed to evaluate the diagnostic accuracy of quantitative adenosine stress real-time MCE-derived absolute MBF and its reserve in patients with known or suspected CAD who were referred for clinically indicated stress SPECT testing.

Methods

Study population

We prospectively studied 79 patients with known or suspected CAD from July 2005 to August 2007, who were referred to the Nuclear Cardiology Laboratory at Mayo Clinic, Rochester, Minnesota, for clinically indicated adenosine SPECT. Patients were excluded for any of the following: age younger than 18 years, pregnant or lactating, substantial valvular heart disease, congenital heart defects, pulmonary embolism, end-stage diseases or contraindications to contrast agent or adenosine. The study was approved by the Mayo Clinic institutional review board.

Imaging protocol

Rest and stress MCE were performed with SPECT testing on the same day. For rest SPECT, technetium 99m sestamibi (mean dose (SD) 11.8 (1.9) mCi) was injected and imaging was initiated 30 minutes later. For stress SPECT, adenosine (Adenoscan; Astellas Pharma US, Inc, Deerfield, IL, USA) was infused over 6 minutes at 140 μg/kg/min and technetium 99m sestamibi (mean dose (SD)), 46.5 (4.2) mCi) was injected at a mean (SD) of 3.05 (0.4) minutes of adenosine infusion. Stress images were acquired 15–60 minutes after the injection.

An ultrasonography system (Sonos 7500 or iE33; Philips Medical Systems, Andover, MA, USA) equipped with a broadband transducer was used. Apical two-chamber, three-chamber and four-chamber views and short-axis views were acquired with use of a power modulation contrast setting with the adjusted mechanical index at 0.2 or less, the frame rate at about 20 Hz, and the focus at the mitral valve level. Destruction-replenishment imaging was used with a transient high mechanical index (1.2) imaging “flash” (about 15 frames) to clear myocardial microbubbles; replenishment was then observed over 15 cycles. If myocardial microbubble clearance was not complete, another image was acquired after adjustments of flash intensity and duration. Imaging was stored digitally for offline analysis. Continuous infusion of Definity injectable suspension (Lantheus Medical Imaging, Inc; North Billerica, MA, USA) was administered through a separate intravenous line. One vial of Definity diluted with 0.9% saline (30 ml) was infused at 250 ml/h with an infusion pump (Baxter, Deerfield, IL, USA). Contrast agent infusion was started 1 minute before MCE acquisition at rest and was kept constant. Stress MCE was begun 3 minutes after the adenosine infusion began; stress MCE was completed within 1 minute after the adenosine infusion was discontinued.

Image interpretation

SPECT images were interpreted independently by clinicians blinded to the MCE data. SPECT regional tracer uptake was visually graded as absent, markedly reduced, moderately reduced, mildly reduced, or normal. When rest and stress images were compared, perfusion was graded as normal or abnormal. Segmental abnormal perfusion was further classified as a reversible defect (seen at stress only) or as a fixed defect (seen at rest and after stress).

Digitised MCE images were analysed offline for quantitative MCE by one observer (SSA) who was blinded to the clinical and SPECT data. Images were evaluated from the acquired apical views focusing on end-systolic frames that were manually selected with QLAB version 5.0 (Philips Medical Systems). Segmental regions of interest of standard size and shape were placed on end-systolic frames starting in the frame immediately after the flash and were automatically copied to subsequent end-systolic frames but were manually adjusted to avoid overlapping cavity signals. The software automatically corrected for non-contrast signals arising from the tissue by subtracting the signal intensity of the first frame after bubble destruction. Time-intensity replenishment curves were generated and subsequently fitted to the following monoexponential function conventional equation: y  =  A (1−eβt).1 Absolute MBF (ml/min/g) was calculated with the model described by Vogel et al8 as follows. Additional regions of interest (of the same size as myocardial regions of interest) were manually tracked in the adjacent left ventricular cavity, and the left ventricular intensity (ALV) was obtained from the averaging of adjacent left ventricular signal intensities of all end-systolic frames except the ones during the flash and the first after the flash. The relative blood volume (rBV) (ml/ml) was calculated from the following formula:

rBV  =  A/ALV

The replenishment curve parameter, β (s−1), was then converted to β (min−1) to reflect the exchange flow velocity. Absolute MBF (ml/min/g) was calculated from the product of rBV×β divided by myocardial tissue density (1.05 g/ml).

Reserve values were calculated for each perfusion parameter as stress values divided by baseline values. Failure of curve fitting either at rest or at stress resulted in excluding the segment from the analysis, and the segment was considered not analysable. Curve fitting usually fails if the input image quality is poor, if the segment is not visualised entirely from the endocardial border to the epicardial border or if the intensity values do not model well with the monoexponential function equation. Overall feasibility of quantitative MCE was based on the available reserve values from all segments. On the territorial level, quantitative MCE reserve parameters were defined as feasible if one or more segments were analysable in each territory; on the patient level, quantitative MCE was defined as feasible if at least five of the 17 segments per patient were analysable. Both MCE and SPECT analyses used the American Society of Echocardiography 17-segment model. Segments were ascribed to coronary territories: the four apical segments (apex, base-mid anterior wall, base-mid anteroseptum and mid inferior septum) were assigned to the left anterior descending coronary artery (LAD), the base-mid anterolateral and base-mid inferolateral walls were assigned to the left circumflex coronary artery (Cx) and the base-mid inferior wall and base inferior septum were assigned to the right coronary artery (RCA).13 Biplane Simpson quantitation of ejection fraction was performed from digitised images at rest and at peak stress.

Coronary angiography

Coronary angiography was evaluated in a patient subset within 1 month of MCE and SPECT. Significant CAD was defined as a stenosis with a visually assessed diameter of 50% or more in a major coronary artery.

Statistical analysis

Continuous data were reported as mean (SD). The paired t test or the Wilcoxon signed rank test was used to compare continuous data, including haemodynamic and quantitative MCE values at rest and at stress. The independent t test was used for comparison of independent data (between patients). Frequencies were used to report categorical variables and were compared with the χ2 test or the Fisher exact test. Receiver operating characteristic curve analysis was used to determine the best cut-off values for perfusion for predicting abnormal perfusion by SPECT, and the area under the curve (AUC) was calculated for each parameter. Evaluation of the diagnostic accuracy of MCE compared with SPECT included sensitivity, specificity and likelihood of positive or negative results. The positive likelihood ratio (LR+) demonstrates how much the odds of the disease increase when a result is positive; it is calculated as sensitivity/(1 − specificity). The negative likelihood ratio (LR−) demonstrates how much the odds of the disease decrease when a result is negative; it is calculated as (1 − sensitivity)/specificity. All diagnostic accuracy measures were calculated on the segmental level with SPECT as the reference test and were reported with the 95% confidence interval (CI). In the subset of patients who underwent coronary angiography within 1 month of MCE and SPECT, the diagnostic accuracy of quantitative MCE and SPECT, compared with coronary angiography, was also calculated.

Interobserver variability for quantitative MCE analysis was assessed among 30 patients by two independent observers (SSA and AD) who were blinded to clinical and SPECT data. Intraobserver variability for quantitative MCE was assessed in the same patients 4 months apart by one observer (SSA). The method of analysis was standardised for accurate comparison. Variability for quantitative MCE was assessed by linear regression and by reporting the correlation coefficient (r) and mean difference (SE) for each measurement. In all analyses, significant levels were set at a two-tailed p<0.05. All analysis was performed with JMP version 7.0 (SAS Institute Inc, Cary, NC, USA).

Results

Clinical and haemodynamic data

Clinical characteristics of the study group are summarised in table 1. Comparison of resting and stress values shows that adenosine administration resulted in a significant increase in heart rate (67.3 (11.1) vs 79.7 (14.7) beats per minute; p<0.001); decrease in systolic blood pressure (144.1 (21.6) vs 136.7 (18.4) mm Hg; p<0.001); decrease in diastolic blood pressure (77.7 (12.5) vs 72.6 (14.2) mm Hg; p<0.001); and increase in rate-pressure product (9672.4 (2044.6) vs 10 983.9 (2302.9) beats per minute × mm Hg; p<0.001). Mild adverse effects from adenosine infusion were reported in 73 patients (92%); all resolved within 2 minutes of discontinuation of the infusion. Adverse effects included flushing in 46 (58%), dyspnoea in 36 (46%), chest pressure in 28 (35%) and headache in 14 (18%). Second degree atrioventricular block occurred in two patients (2.5%). There was a statistically significant increase in mean ejection fraction between rest and peak stress (52.2% (7.3%) vs 61.6% (12.6%); p = 0.003).

Table 1

Clinical characteristics of study population (n = 79)

Feasibility of quantitative MCE

Quantitative MCE and SPECT analyses were feasible for all 79 patients, in whom a total of 237 coronary territories and 1343 segments were evaluated. Quantitative MCE was feasible in 224 territories at rest (95%; 79 LAD, 69 Cx, and 76 RCA) and in 225 at stress (95%; 79 LAD, 67 Cx, and 79 RCA); reserve measurements were feasible in 219 coronary territories (92%; 79 LAD, 64 Cx and 76 RCA). Segmental quantitative MCE was feasible for analysis in 1060 segments (79%) at rest and 1075 (80%) at stress and in 975 (73%) for reserve measurements. The 975 segments feasible for reserve were distributed as follows: 605 of 790 LAD segments (77%), 194 of 256 Cx segments (76%) and 176 of 228 RCA segments (77%). Failure to obtain quantitative MCE data was related to failed curve fitting or non-visualisation of the myocardial segments from endocardium to epicardium.

All 975 segments available for reserve quantitative MCE analysis were analysed with SPECT. On the segmental level, SPECT findings were abnormal in 130 segments (13%; 81 of 605 LAD segments, 20 of 194 Cx segments and 29 of 176 RCA segments). Perfusion characterisation by SPECT on the segmental level was as follows: normal perfusion in 845 segments, reversible perfusion defect in 93 segments and fixed or mixed perfusion defect in 37 segments. On the patient level, SPECT findings were abnormal in 38 patients (48%), including 21 abnormal LAD territories, 14 abnormal Cx territories and 20 abnormal RCA territories. Perfusion characterisation by SPECT on the patient level was as follows: normal perfusion in 41 patients, reversible perfusion in 21 patients and fixed or mixed perfusion in 17 patients.

Quantitative MCE perfusion parameters compared with SPECT imaging

Baseline perfusion parameters significantly increased during peak stress: rBV increased from 0.27 (0.13) to 0.39 (0.17) ml/ml; β increased from 6.81 (3.50) to 13.13 (8.13) min−1; and absolute MBF increased from 1.72 (1.16) to 5.03 (3.68) ml/min/g. In segments with normal perfusion by SPECT imaging, all quantitative MCE parameters were significantly increased during peak stress compared with rest; however, in segments with abnormal perfusion by SPECT, the increase was not statistically significant (table 2). Consequently, segments with abnormal perfusion by SPECT had significantly lower values for reserve quantitative parameters compared with segments with normal perfusion (rBV reserve: 1.44 (0.89) vs 1.73 (0.98), p<0.001; β reserve: 1.13 (0.99) vs 2.22 (1.36), p<0.001; and MBF reserve: 1.80 (2.29) vs 3.69 (2.79), p<0.001).

Table 2

Quantitative myocardial contrast echocardiographic perfusion parameters at the segmental level†

Segments with reversible and fixed or mixed (or both) perfusion defects by SPECT had lower quantitative perfusion parameters compared with normally perfused segments. In segments with normal perfusion, rBV increased significantly at peak stress (0.26 (0.11) vs 0.40 (0.18) ml/ml, p<0.001), but rBV did not increase significantly in segments with reversible perfusion defects (0.28 (0.16) vs 0.30 (0.10) ml/ml, p = 0.409) or in segments with fixed perfusion defects (0.31 (0.36) vs 0.35 (0.12) ml/ml, p = 0.501). Similarly, β increased at peak stress in segments with normal perfusion (7.24 (3.51 vs 14.55 (7.67) min−1, p<0.001) but did not increase significantly in segments with reversible perfusion defects (3.87 (1.69) vs 3.76 (3.39) min−1, p = 0.753) or in segments with fixed or mixed (or both) perfusion defects (4.4 (1.87) vs 4.09 (4.26) min−1, p = 0.656). Absolute MBF increased at peak stress in segments with normal perfusion (1.81 (1.14) vs 5.61 (3.56) ml/min/g, p<0.001) but did not increase significantly in segments with reversible perfusion defects (1.06 (0.76) vs 1.12 (1.23) ml/min/g, p = 0.713) or in segments with fixed or mixed (or both) perfusion defects (1.35 (1.75) vs 1.44 (1.82) ml/min/g, p = 0.829).

Diagnostic accuracy of quantitative MCE perfusion parameters by receiver operator characteristic curves

On the segmental and patient levels, receiver operator characteristic curves were constructed for rBV, β, and MBF reserves for comparison with SPECT. On the segmental level, β and MBF reserves were accurate in the detection of abnormal myocardial perfusion by SPECT. The optimal β reserve cut-off of 1.60 provided the following: AUC, 0.787; sensitivity, 82% (95% CI 75% to 87%; 106 of 130 segments); specificity, 66% (95% CI, 65% to 69%; 558 of 845 segments); LR+, 2.40 (95% CI 2.13 to 2.63); and LR−, 0.28 (95% CI 0.19 to 0.39). The MBF reserve cut-off of 1.90 provided the following: AUC, 0.779; sensitivity, 73% (95% CI 66% to 80%; 95 of 130 segments); specificity, 72% (95% CI 71% to 74%; 615 of 845 segments); LR+, 2.69 (95% CI 2.31 to 3.04); and LR−, 0.37 (95% CI 0.28 to 0.48). Receiver operator characteristic curves for reserve parameters compared with SPECT are shown in figure 1. The diagnostic accuracy of reserve parameters on segments grouped into coronary territories is shown in table 3.

Figure 1

Receiver operating characteristic curve. The area under the curve (AUC) is shown for reserve perfusion parameters for the detection of abnormal perfusion in myocardial segments by single-photon emission computed tomography. The three parameters are relative myocardial blood volume (rBV), myocardial blood flow velocity (β) and myocardial blood flow (MBF). AUC values for the three parameters are the following: rBV, 0.609; β, 0.787; and MBF, 0.779.

Table 3

Diagnostic accuracy of quantitative myocardial contrast echocardiographic perfusion parameters at the territorial level

On the patient level, use of the same cut-off value for β and MBF reserves provided accurate detection of abnormal myocardial perfusion by SPECT. The β reserve provided the following: AUC, 0.651; sensitivity, 66% (95% CI 55% to 75%; 25 of 38 patients); specificity, 71% (95% CI, 60% to 80%; 29 of 41 patients); LR+, 2.25 (95% CI 1.37 to 3.69); and LR−, 0.48 (95% CI 0.31 to 0.76). The MBF reserve provided the following: AUC, 0.806; sensitivity, 71% (95% CI, 60% to 79%; 27 of 38 patients); specificity, 83% (95% CI 73% to 90%; 34 of 41 patients); LR+, 4.16 (95% CI 2.24 to 8.04); and LR−, 0.35 (95% CI 0.23 to 0.54).

Quantitative MCE perfusion parameters and SPECT imaging compared with coronary angiography

In a subset of 11 patients who underwent clinically indicated coronary angiography within 1 month of MCE and SPECT, significant CAD (⩾50% stenosis) was detected in eight patients. SPECT findings were concordant in the presence or absence of significant CAD in nine patients, with sensitivity of 88% (95% CI 72% to 96%; seven of eight patients), specificity of 67% (95% CI 25% to 91%; two of three patients), LR+ of 2.63 (95% CI 0.95 to 10.2) and LR– of 0.19 (95% CI 0.04 to 1.15). Optimal values of β and MBF reserves for detecting CAD by coronary angiography revealed an AUC of 0.667 and 0.625, respectively. The cut-off for β reserve provided a sensitivity of 63% (95% CI 44% to 63%; five of eight patients), specificity of 100% (95% CI 52% to 100%; three of three patients), LR+ of 4.89 (95% CI 0.81 to 49.0), and LR– of 0.44 (95% CI 0.43 to 1.21). The cut-off for MBF reserve provided a sensitivity of 63% (95% CI 47% to 73%; five of eight patients), specificity of 67% (95% CI 24% to 93%; two of three patients), LR+ of 1.88 (95% CI, 0.61 to 10.9) and LR– of 0.56 (95% CI 0.29 to 2.24).

Intraobserver and interobserver variability

For intraobserver variability, the mean differences (SE) and the correlation coefficients (r) for rBV, β reserve and MBF reserve were, respectively, 5.85 (0.81) (r = 0.371, p<0.001); 1.34 (0.04) (r = 0.623, p<0.001); and 5.96 (0.17) (r = 0.544, p<0.001). For interobserver variability, the mean differences (SE) and the correlation coefficients (r) for rBV, β reserve, and MBF reserve were, respectively, 4.26 (0.61) (r = 0.308, p<0.001); 0.57 (0.05) (r = 0.528, p<0.001); and 1.72 (0.06) (r = 0.528, p<0.001). Previous studies from our institution reported high agreement between interobserver SPECT readings.14

Figure 2 shows MCE images with corresponding raw data from the exponentially fitted time-intensity curves for a 67-year-old man with hypertension, dyslipidaemia and typical chest pain.

Figure 2

Myocardial contrast echocardiographic images with corresponding raw data from the exponentially fitted time-intensity curves for a 67-year-old man with hypertension, dyslipidaemia and typical chest pain. Corresponding exponentially fitted time-acoustic intensity curves were obtained at peak stress, showing the region of interest (ROI) in the apical segment (yellow ROI) and in the mid segment of the anterior wall (blue ROI). During adenosine stress, parameter A in the apical segment was 11.01 dB and parameter β was 0.75 s−1, whereas in the mid segment of the anterior wall they were higher (11.14 dB and 1.19 s−1, respectively). Single-photon emission computed tomographic (SPECT) images revealed an apical defect (arrows) during stress only in the apex and apical anterior wall. Lat, lateral; RST, rest; Sep, septum; STR, stress.

Discussion

The present study demonstrates that quantitative determination of real-time MCE-derived absolute MBF (ml/min/g) and β (min−1) was feasible, permitting accurate assessment of perfusion abnormalities as validated by technetium 99m sestamibi SPECT during adenosine stress. In addition, quantitative perfusion parameters accurately differentiated reversible and fixed defects as characterised by SPECT.

To our knowledge, this article is the first to compare MCE-derived absolute MBF (ml/min/g) with adenosine technetium 99m SPECT in a routine clinical scenario. In addition, we included a larger population and performed MCE and SPECT simultaneously during the same day. In this study, we used MCE-derived MBF in absolute units of blood volume per time in relation to myocardial mass (ml/min/g), an approach developed recently by Vogel and colleagues.8 MCE-determined absolute MBF has been validated against positron emission tomography using a volumetric model at rest and during intravenous adenosine infusion.8 Further validation by quantitative coronary angiography in various patient populations was subsequently demonstrated.9 10 11 12 Absolute MBF measurements have also been shown to be reproducible, with good interobserver variability.9 10

Quantification of MBF and β reserve was feasible in 92% of coronary territories. This was in agreement with previous studies.4 10 Nevertheless, lower feasibility (73%) has been reported at the segmental level as a result of either failure of replenishment curve fitting or imaging artefacts and dropouts (frequently observed in the basal-mid anterior wall segments).10

During adenosine administration, we demonstrated an increase in MBF that resulted mainly from a large increase in β and a small increase in rBV, as observed by Vogel and colleagues.10 The rBV increase can be explained by an associated increase in myocardial oxygen consumption as a result of an increased rate-pressure product after intravenous administration of adenosine1 15 and associated capillary recruitment.16 17

In the present study, significantly lower MBF reserve was observed in segments with abnormal perfusion compared with normally perfused segments by SPECT. The lower MBF reserve in abnormally perfused segments is due to a suboptimal increase of β or to a failure of the rBV to increase during peak adenosine stress (that is, a failure of capillary recruitment). In this phenomenon, known as “capillary derecruitment”, the decrease in coronary driving pressure (after intravenous administration of adenosine) causes a decrease in capillary perfusion pressure. To maintain a constant capillary hydrostatic pressure, capillary resistance increases from derecruitment of capillary units.16 17 The decrease in reserve measurements in our study was in agreement with the results of Vogel and colleagues,10 who documented a reduction in β and MBF reserve in territories supplied by diseased coronary arteries compared with territories of disease-free vessels as measured by quantitative coronary angiography.

Reductions of β and MBF reserves below the cut-offs of 1.60 and 1.90, respectively, identified abnormally perfused segments by SPECT with an acceptable diagnostic accuracy, but rBV reserves did not have an equivalent accuracy. This finding is in agreement with previous studies reporting that rBV correlates poorly with flow measurements with radiolabelled microspheres or intracoronary Doppler flow wire.18 19 In addition, this was also a concordant finding in previous quantitative MCE studies reporting that β and MBF reserves were more accurate than rBV reserve for detection of perfusion defects.7 20

The diagnostic performance of quantitative real-time MCE during adenosine stress has been previously evaluated in a series of smaller studies using either SPECT or coronary angiography as the reference tests.4 5 6 7 10 Reported diagnostic accuracy measures for β and MBF reserve are shown in table 4. The AUC for β reserve ranged from 0.787 to 0.899; the AUC for MBF reserve ranged from 0.779 to 0.928. For β reserve, the reported sensitivity ranged from 77% to 88% and the specificity from 66% to 97%; for MBF reserve, the sensitivity ranged from 73% to 89% and the specificity from 67% to 92%. This wide range of diagnostic accuracy can be explained by the different methodologies applied in each study, including use of various contrast agents, different cut-offs, and different reference tests. Nevertheless, all studies (including ours) had an acceptable diagnostic accuracy. Similarly, in agreement with the previous studies, our determination of MCE-derived quantitative perfusion parameters accurately identified segments with abnormal perfusion by SPECT in various coronary territories and differentiated reversible and fixed perfusion defects.

Table 4

Studies of quantitative real-time myocardial contrast echocardiography during adenosine stress in patients with known or suspected coronary artery disease

One limitation of the present study is that we chose to evaluate the diagnostic accuracy of quantitative MCE for predicting abnormal perfusion by SPECT as the reference test and did not correlate our findings with anatomical coronary artery stenosis identified with coronary angiography in all patients. The comparison was reported for only a subset of patients (14%) who underwent clinically indicated coronary angiography. Nevertheless, SPECT is considered the reference standard for the evaluation of myocardial perfusion. Another limitation is the comparison between two imaging modalities with different resolutions (1 mm for MCE vs 1 cm for SPECT); however, the use of quantitative MCE provided objective measurements, which decreased the likelihood of error. In addition, we used a unified method for analysis, including a systematic approach evaluating all left ventricular 17 segments by SPECT and quantitative MCE. Another limitation is the inclusion of patients with an intermediate to high pretest probability of CAD who were already referred for stress SPECT testing.

We also acknowledge that the visual assessment of SPECT images is inherently subjective. However, we used the standard clinical semiquantitative approach for grading tracer uptake in each segment with a 5-point scale. This scoring was shown in a large cohort to be an effective gatekeeper for coronary angiography, with the annual overall mortality rate of 1.9% for patients with normal images during 10 years of follow-up.21

Our study had many strengths, including masking of the data during evaluation of perfusion by SPECT and quantitative MCE. We systematically used the American Society of Echocardiography 17-segment model in analysing both SPECT and quantitative MCE. In addition, we included all patients irrespective of baseline echocardiographic image quality.

In conclusion, real-time MCE-derived absolute MBF (ml/min/g), β (min−1) and reserve values are feasible and accurate for detecting myocardial perfusion abnormalities compared with SPECT imaging during adenosine stress.

Acknowledgments

Contrast agent (Definity) was provided by Lantheus Medical Imaging Inc (North Billerica, MA, USA). Adenosine (Adenoscan) was provided by Astellas Pharma US, Inc (Deerfield, IL, USA).

REFERENCES

View Abstract

Footnotes

  • Competing interests SLM received a research grant from Lantheus Medical Imaging and Astellas Pharma. There are no conflicts of interest to disclose for the remaining authors.

  • Provenance and Peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.