Author + information
- Received March 11, 2009
- Revision received June 16, 2009
- Accepted July 29, 2009
- Published online November 1, 2009.
- Kyle S. McCommis, BS⁎,
- Haosen Zhang, PhD⁎,
- Thomas A. Goldstein, MS⁎,
- Bernd Misselwitz, PhD‡,
- Dana R. Abendschein, PhD†,
- Robert J. Gropler, MD⁎ and
- Jie Zheng, PhD⁎,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Jie Zheng, Mallinckrodt Institute of Radiology, Campus Box 8225, 510 South Kingshighway Boulevard, St. Louis, Missouri 63110
Understanding the oxygen consumption of the left ventricular myocardium provides important insight into the relationship between myocardial oxygen supply and demand. In other territories, cardiac magnetic resonance has been utilized to measure myocardial oxygen consumption with a blood level oxygen dependent (BOLD) technique. The BOLD technology requires repetitive sampling of stationary tissues and is frequently implemented in areas such as the brain. A limitation to utilizing BOLD cardiac magnetic resonance techniques in the heart has been cardiac motion. In this study, we document a methodology for acquiring BOLD images in the heart and demonstrate the utility of the technique for identifying associations between myocardial oxygen consumption and blood flow.
- cardiac magnetic resonance
- myocardial blood flow
- myocardial blood volume
- myocardial oxygen consumption
- blood oxygen level dependent
Myocardial ischemia occurs when the supply of oxygen is inadequate for the metabolic demand of the myocardium. Measurements of myocardial perfusion (O2 supply) and myocardial oxygen consumption (MVO2) may provide accurate assessments of this balance in the heart. Two important parameters for oxygen delivery are myocardial blood flow (MBF) and myocardial blood volume (MBV). The addition of MBV measurements increases the accuracy of perfusion assessments because MBV has been shown to be altered in situations of increased MVO2 (1). The MBV is composed of vessels ≤200 μm, of which 90% are capillaries. Because only ∼50% of capillaries are functional at rest, altering the amount of functional capillaries can drastically change the tissue oxygenation levels.
Positron emission tomography (PET) is currently the only imaging modality capable of absolute quantification of regional myocardial perfusion and oxygen metabolism. The use of PET permits the accurate quantification of MBF with 15O-water and MVO2 with 11C-acetate. However, low spatial resolution, relatively long acquisition times, limited availability, relatively high cost, and ionizing radiation discourage the widespread use of PET for these purposes. Furthermore, PET cannot measure MBV. Myocardial contrast echocardiography (MCE) has the capability of measuring MBF and MBV but not MVO2.
Cardiac magnetic resonance (CMR) is a noninvasive imaging modality that provides excellent image spatial resolution and soft-tissue contrast, does not require ionizing radiation, and is widely available. CMR can measure absolute MBF and MBV via first-pass methods (2) and the oxygen extraction fraction (OEF) by the use of the blood oxygen level-dependent (BOLD) technique (3). Fick's law states that MVO2 ∝ MBF × OEF; therefore, CMR also can provide an estimation of MVO2. The objective of this study was to apply these comprehensive CMR techniques to evaluate MBF, MBV, OEF, and MVO2 in a canine model with control, moderate (75%), and severe (86% to 95%) coronary artery stenosis during dipyridamole or dobutamine hyperemia. We hypothesize that direct measurements of these parameters by CMR will facilitate a comprehensive ischemic assessment.
All animal procedures were approved by the Animal Studies Committee at Washington University. A total of 21 (weight 24.7 ± 3.0 kg) mongrel dogs were used, and they were divided into 6 groups (Table 1). A thoracotomy was performed in the fourth intercostal space and the pericardium incised. The left anterior descending coronary artery (LAD) was dissected free distal to the first diagonal branch. The artery was instrumented in a proximal-distal order with a Doppler flow probe, a pneumatic occluder, and a CMR-compatible stenosis clamp. The procedure for setting the stenosis severity has been described previously. Serial 20-s occlusions were performed to delineate the hyperemic flow responses. After tightening the stenosis clamp, another occlusion was performed to assess the decrease in hyperemic flow. After attaining the desired level of stenosis defined by reduction in hyperemic flow, the occluder was removed. The dogs remained open-chest and were moved to the magnetic resonance imaging suite. Control dogs were omitted from thoracotomy surgery.
CMR was performed at rest and during dipyridamole-induced vasodilation or dobutamine-induced hyperemia. Dipyridamole (Bedford Laboratories, Bedford, Ohio) was injected intravenously at a dose of 0.14 mg/kg/min for 4 min. Dobutamine (Hospira Inc., Lake Forest, Illinois) was started at 5 μg/kg/min and titrated at 5-μg/kg/min increments every 5 min until heart rate reached >130 beats/min (maximum of 30 μg/kg/min).
Study timeline is shown in Figure 1. Imaging was performed on a 1.5-T Sonata scanner (Siemens Medical Solutions, Erlanger, Germany). A 4-element phased array coil placed around the chest was used for signal reception, and a body coil was used as a transmitter. Scout imaging was performed to obtain a short-axis image of the left ventricle (LV) at the middle level of the papillary muscles. During scans, respiratory motion was reduced by turning off ventilation to simulate breath holding.
First-Pass Perfusion CMR Protocol
Images during the bolus injection of Gadomer (Bayer Schering Pharma AG, Berlin, Germany), an intravascular contrast agent, were sequestered by a saturation-prepared turbo fast low-angle shot sequence. The short-axis slice of the LV was acquired during mid-diastole, triggered by the R-wave of the electrocardiogram. A total of 60 to 80 dynamic images were gathered, and images were collected at every RR interval. Other imaging parameters included the following: time of repetition = 2.5 ms; time of echo = 1.2 ms; time of inversion = 90 ms; flip angle = 18°; field of view = 220 × 138 mm2; matrix size = 128 × 80%; slice thickness = 8 mm; and image acquisition time window per cardiac cycle = 150 ms.
Protocol for BOLD CMR
The BOLD effect was detected by a multicontrast 2-dimensional segmented turbo spin-echo sequence. Double inversion-recovery preparation yielded black-blood images. The sequence was triggered by electrocardiogram with the turbo spin-echo train placed in mid-diastole to minimize cardiac motion and match the first-pass perfusion images. Parameters included the following: field of view = 220 × 131 mm2; matrix size 256 × 156; slice thickness = 8 mm; inversion time = 350 to 500 ms, depending on the RR interval; and data acquisition time = 24 × RR, or 14.4 s for a typical 600-ms RR interval. Three echo times, TE1 = 24, TE2 = 48, and TE3 = 72, were used. At rest, this sequence was run twice with 2 different echo spacings (τ = 8 and τ = 12) and at τ = 8 during hyperemia. The 2 T2 maps with 2 echo spacings at rest were used to determine model parameters for the calculation of OEF during hyperemia.
First-pass perfusion images were analyzed with a JAVA program (Java V5.0, Sun Microsystems, Santa Clara, California) created in our lab. Images were denoised and subjected to a validated perfusion quantification algorithm (2). This algorithm created both MBF and MBV maps, on which regions of interest (ROIs) could be drawn. We determined MBV by MBF divided by mean transit time. The mean transit time was determined by the area under the impulse curve, and images before the second contrast pass were removed. Details of this estimation were shown in a previous report.
We analyzed BOLD T2-weighted images with a MATLAB graphics program (The MathWorks, Natick, Massachusetts). Pixel-by-pixel maps of the myocardial T2 decay constants were calculated from the signal intensities, and then OEF maps during hyperemia were determined with our previously described model (3), on which ROIs similar to the first-pass perfusion map ROIs were drawn. A resting OEF of 0.6 was assumed, which is based on arterial and coronary sinus blood sampling measurements in control dogs at rest (R2 = 0.90), as well as PET measurements in dogs with moderate stenosis (R2 = 0.75). Sample perfusion and OEF maps are shown in Figure 2.
Calculation of MVO2
Once MBF and OEF were determined, MVO2 was calculated using Fick's law:(1) The constant [O2]a is defined as the total oxygen content of arterial blood, and a value of 7.99 μmol/ml was used.
The MBF, MBV, OEF, and MVO2 data are presented as mean ± SD. Percentage change (from rest to hyperemia) was determined. A paired and unpaired t test was used to compare between rest and pharmacologic stress and between stenosed and control dogs, respectively. Linear correlations between these parameters were expressed as the coefficients of determination. Comparison tests between 2 R2 values were performed by use of a Z test. A p value <0.05 signified significant differences.
Table 2 displays the hemodynamic changes. As expected, dipyridamole caused only slight changes in rate-pressure product (p = NS), whereas dobutamine produced significant increases in rate-pressure product (p < 0.05).
Absolute MBF and MBV values
The absolute MBF and MBV values, as well as percent change values for all groups, are displayed in Tables 3 and 4,⇓⇓ respectively. During dipyridamole, control dogs achieved 2- to 3-fold increases in MBF. Stenosis attenuated these MBF reserves in the LAD region and in the remote left circumflex coronary artery (LCX) region with severe 95% LAD stenosis. The dobutamine groups had similar trends in MBF. The trend for attenuation of increased MBF in the remote LCX region was not statistically significant but is in agreement with other reports.
In control dogs, dipyridamole and dobutamine induced moderate 30% and >50% changes in MBV, respectively. Stenosis >86% dramatically attenuated the MBV increases in the LAD region during either dipyridamole vasodilation or dobutamine hyperemia. Such attenuation also was observed in the remote LCX region, especially with dobutamine injection. However, for 75% area stenosis during dobutamine, the attenuation of MBV was much less, although significant changes were still observed relative to the resting values.
Values for OEF and MVO2
The hyperemic OEF and MVO2 values, as well as the percent change values, are presented in Tables 5 and 6,⇓⇓ respectively. As expected, in control dogs dipyridamole caused large decreases in OEF. These decreases also were observed in the remote normal LCX regions of the dogs with coronary stenosis. Dobutamine increased both MVO2 and MBF, which in control dogs produced no significant change in OEF. Only small OEF changes were observed in the stenotic and remote regions.
In control dogs, dipyridamole caused moderate 30% to 70% increases in MVO2 (4). The LAD stenosis attenuated the MVO2 increases in the LAD- and LCX-perfused regions. Dobutamine resulted in significantly larger increases in MVO2 in control dogs. However, even 75% stenoses significantly reduced the change in MVO2 in both the LAD- and LCX-perfused regions.
Relationship between myocardial perfusion and oxygen utilization
Correlations between MVO2 reserve versus MBF or MBV reserve for dogs with and without coronary artery stenosis in the LAD and/or LCX (in control dogs) subtended regions are plotted in Figure 3. In control dogs, MBV reserve shows mild-to-moderate correlation with MVO2 reserve with dobutamine stress but not with dipyridamole vasodilation (Fig. 3A). In dogs with stenosis, MBF reserve appears to correlate well with MVO2 reserve, and similar correlations were observed between dipyridamole and dobutamine hyperemia. With stenosis, MBV reserve again appears much less correlated with MVO2 during dipyridamole vasodilation. However, with the larger increases in MVO2 with dobutamine, MBV reserve appears to be strongly correlated with MVO2 reserve, although the p values for the correlation coefficients were not statistically different as the result of limited data points. It is noted that during dobutamine, the slope of MBV reserve versus MVO2 reserve is also greater than MBF reserve versus MVO2 reserve, indicating a better association with MBV reserve. From this point of view, MBV may be a significant source of O2 supply when MBF supply becomes exhausted with increased O2 demand in settings of coronary artery stenosis.
The purpose of this study was to apply our noninvasive CMR techniques to directly assess regional microcirculatory changes that occur during dipyridamole and dobutamine stress. To our knowledge, this is the first study to show that regional MBF, MBV, OEF, and MVO2 can be assessed noninvasively in a single imaging session. The role of MBV during increased O2 demand is also confirmed in this study.
Dogs during dipyridamole
By using blood sampling techniques in control dogs, Hoffman et al. (4) observed MVO2 increases of 70% with adenosine vasodilation, which is comparable with our 30% to 70% increases in MVO2 using dipyridamole. In comparison with the control groups, resting MBF of LAD regions in the 86% and 95% stenosis groups significantly reduced 38% (p < 0.01) and 42% (p = 0.02), respectively. However, after normalizing the resting MBF by rate-pressure product, there is no statistical difference between control and either stenosis groups for the resting MBF (MBF/rate-pressure product × 104: 0.46 ± 0.43 ml/g/min in control vs. 0.78 ± 0.09 ml/g/min in the 86% stenosis group or vs. 0.61 ± 0.24 ml/g/min in the 95% stenosis group). Similar decreases in MBF reserve have been observed in dogs and clinically with MCE and PET. Although they showed ∼130% increases in MBF in normal regions of animals with stenosis (we showed 140% to 200%), they also observed only ∼24% increases in MBF in the zone distal to the stenosis (we showed 12% to 29%). The changes in MVO2 in the LAD stenotic region during dipyridamole vasodilation were observed at a similar level for both 86% and 95% severe stenosis.
It has been suspected that a primary method of reducing MBV is a reduction in perfusion bed size. These data support the notion that MBV plays a mediating role in the match/mismatch of MBF and MVO2 (1). Our MBV findings during dipyridamole conform to other reports (1) in that no significant relationship exists between MBV and small changes in MVO2 caused by chronotropic stimulation alone.
Dogs during dobutamine
In control dogs, we observed 157% to 194% increases in MVO2 with dobutamine. This large O2 demand was accounted for by 192% to 204% increases in MBF and 51% to 60% increases in MBV. These findings are similar to Le et al. (1), who found ∼200% increases in MBF and 90% to 150% increases in MVO2 in normal dogs with varying dobutamine doses. With 75% area stenosis, the region distal to the stenosis had dramatic attenuation of MVO2 increase. This finding was associated with a similar rate of reduction of MBF reserve in the stenotic LAD region, and OEF remained similar to rest. The MBV increase was slightly attenuated in the 75% area stenosis regions than in control dogs. With further increase in stenosis severity, both MBF and MBV reserve were attenuated, and OEF again remained the same as at rest, resulting in less MVO2 increase. Similar trends were observed in the remote LCX region.
A recent study (5) in which the authors used radiolabeled microspheres in dogs showed the remote regions of dogs with coronary stenosis had a 216% MBF increase, whereas the region distal to the stenosis showed only a 20% increase in MBF. By using blood sampling techniques, they also showed large increases in MVO2 with dobutamine before stenosis (120%) and significantly attenuated MVO2 increases with dobutamine after stenosis (20%).
The MBV has a close relationship with MVO2 when inotropic stimulation (as with dobutamine) induces relatively large changes in MVO2. This can be observed from Figure 3; when MVO2 was only moderately increased with dipyridamole, MBF reserve had a closer relationship to MVO2 reserve than MBV reserve. When MVO2 was more significantly altered with dobutamine, MBV reserve was more closely associated with MVO2 reserve. These results are in agreement with MCE findings showing that minor increases in MVO2 can be met by increases in MBF alone but major MVO2 increases require increases in MBV as well.
There are several factors that affect these quantification methods for OEF and MVO2. First, there is no CMR quantification method for absolute OEF at rest. Thus, we must assume a rest OEF of 0.6 and determine the hyperemic OEF based on change in T2 during hyperemia. Although this may induce systematic errors, the hyperemic OEF was closely correlated with other gold standards, either by blood sampling (3) or by PET. Second, spatial resolution and signal-to-noise ratio are limited to differentiate endomyocardium and epimyocardium. These transmural gradients in myocardial perfusion and oxygenation are important for the diagnosis of myocardial ischemia. Although MBF and MBV maps could be analyzed for this purpose, the major limitation lies in the OEF map that is created by a T2 map using a 3-echo fitting procedure. Such a low number of echoes created “mapping noise” that caused a large spatial variation in T2. Other T2-weighted methods or more echo numbers may be needed to reduce the noise. Finally, LV wall motion may hamper the accuracy of myocardial T2 measurement. Although images are acquired at mid-diastole, i.e., the relatively motionless period within one cardiac cycle, we have observed LV motion among images with different echo times. This cardiac motion often occurs with increased heart rates during dobutamine hyperemia. Because we used double inversion recovery technique to minimize blood flow signal in the LV, this cardiac motion also reduced the efficiency of the blood flow signal suppression. Therefore, combined adverse effects of this motion may attenuate or lengthen the echo train of the T2 decay curve. Given the complexity of these motion effects, no simulation was yet performed to estimate the error from these effects. However, on the basis of our experience, the possible error could be up to 5% of the myocardial T2, which is close to the precision we observed in normal dogs (3). This would lead to an error of 14% in the estimation of OEF. Efforts to reduce this cardiac motion are one of our laboratory's ongoing projects.
The combination of the CMR methods used in this study facilitates the comprehensive evaluation of microcirculatory pathophysiology caused by coronary stenosis. Notably, MBV appears to correlate with significantly increased MVO2 in both normal and ischemic myocardial regions. Because even a moderate stenosis during stress effects not only O2 delivery (MBF and MBV) but also oxygen utilization (MVO2 and OEF), it is important to be able to assess all these parameters to evaluate myocardial ischemia.
This work was supported by NIH grant R01 HL74019-01. Dr. Misselwitz is an employee of Bayer Schering Pharma, which produces the Gadomer contrast agent used in this study.
- Abbreviations and Acronyms
- blood oxygen level-dependent
- cardiac magnetic resonance
- left anterior descending coronary artery
- left circumflex coronary artery
- left ventricle
- myocardial blood flow
- myocardial blood volume
- myocardial contrast echocardiography
- myocardial oxygen consumption
- oxygen extraction fraction
- positron emission tomography
- region of interest
- Received March 11, 2009.
- Revision received June 16, 2009.
- Accepted July 29, 2009.
- American College of Cardiology Foundation