| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1
Department of Clinical Physiology, Turku University Hospital, Kiinamyllynkatu
4-8, 20520 Turku, Finland.
2
Turku PET Centre, University of Turku, Kiinamyllynkatu 4-8, 20520 Turku,
Finland.
3
Department of Radiology, Mie University Hospital, 2174 Edobashi, Tsu, Mie
514-8507, Japan.
4
Department of Radiology, Turku University Hospital, Kiinamyllynkatu 4-8, 20520
Turku, Finland.
Received December 4, 2000;
accepted after revision May 22, 2001.
Supported by the Turku University Central Hospital Research Foundation,
Instrumentarium Research Foundation, and the Finnish Medical Society
Duodecim.
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. We measured myocardial blood flow first at baseline and then after dipyridamole-induced hyperemia in 20 patients with coronary artery disease. Myocardial blood flow as revealed by MR imaging was calculated by dividing coronary sinus flow by the left ventricular mass. Coronary flow reserve was calculated by dividing the rate of hyperemic flow by the rate of baseline flow.
RESULTS. Using MR imaging, myocardial blood flow at baseline was 0.73 ± 0.23 mL·min-1·g-1, and at hyperemia the blood flow was 1.43 ± 0.37 mL·min-1·g-1, yielding an average coronary flow reserve of 1.99 ± 0.47. Using PET, myocardial blood flow was 0.89 ± 0.21 mL·min-1·g-1 at baseline and 1.56 ± 0.42 mL·min-1·g-1 at hyperemia, yielding an average coronary flow reserve of 1.77 ± 0.36. The correlation of myocardial blood flow and coronary flow reserve measurements for these two methods was an r of 0.80 (p < 0.01) and an r of 0.50 (p < 0.05), respectively.
CONCLUSION. This study shows that myocardial blood flow measurements obtained using MR imaging have a good correlation with corresponding PET measurements. Coronary flow reserve measurements obtained using MR imaging had only moderate correlation with PET-obtained measurements. Our results suggest that MR imaging flow quantification could potentially be used for measuring global myocardial blood flow in patients in whom interventional treatment for coronary artery disease is being evaluated.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Recent advances in imaging technology have made it possible to accurately measure blood flow in coronary vessels using a short imaging time [9,10]. Initially, MR imaging was used to visually evaluate blood flow using gray-scale images [11]. The first quantitative MR flow was based on the time-of-flight method [12], and since 1993, it has been used to image coronary artery flow [13]. Flow can also be measured with techniques that assess phase shift. In measuring flow velocity and flow volume rate, velocity-encoded cine imaging has been shown to be feasible in the aorta [14], pulmonary artery [15], mitral valve [16], left anterior descending coronary artery [8], and coronary sinus [10,17].
Coronary flow reserve is defined as the ratio of coronary flow under maximally induced hyperemia to baseline flow. Previous studies have found a decreased coronary flow reserve value to be associated with elevated cardiovascular risk status [18], hypercholesterolemia [19], hypertension [20], and diabetes [21]. Coronary flow reserve has been also regarded as a useful index of the functional impact of coronary artery stenosis [22], and it has been shown to be more sensitive in predicting major cardiac events than the measurement of the percentage of diameter reduction provided by coronary angiography [23]. Therefore, a reliable noninvasive imaging modality for measurement of myocardial blood flow is needed to improve identification of individuals who are at high risk of developing acute complications of coronary artery disease and who might benefit from aggressive therapy.
Approximately 96% of the blood flow from the left ventricular myocardium runs through the coronary sinus, making it a practical site to assess global myocardial blood flow [24]. Several methods have been established to measure regional coronary flow reserve in humans [25], but only invasive catheterization of the coronary sinus and positron emission tomography (PET) have been previously validated as methods of measuring global myocardial blood flow and global coronary flow reserve in patients with coronary artery disease. Currently, PET with 15O-labeled water is regarded as the gold standard for measuring coronary flow reserve [26]. However, both the invasive catheterization and PET procedures expose the patient to radiation.
MR imaging of the coronary sinus has been validated against PET in assessing global myocardial blood flow in healthy volunteers [10,27], but, to our knowledge, has not been studied in patients with coronary artery disease. With the MR protocol we present, global myocardial blood flow, coronary flow reserve, and left ventricular function can be measured almost simultaneously. The purpose of this study was to investigate whether velocity-encoded cine MR imaging of coronary sinus blood flow could accurately measure global myocardial blood flow and global coronary flow reserve in patients with coronary artery disease as compared with PET.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Study Design
The patients first underwent MR imaging and PET at rest and then after IV
administration of dipyridamole (Persantin; Boehinger Ingelheim, Ingelheim,
Germany) at 0.56 mg/kg per 4 min so that we could obtain blood flow data at
baseline and after maximal vasodilation. Coronary sinus flow was considered to
represent global myocardial blood flow in the left ventricular myocardium
[24]. Hyperemic myocardial
blood flow was recorded 6 min after the onset of IV infusion of dipyridamole
both in MR imaging and PET studies. An investigator who analyzed the MR
imaging studies was unaware of the corresponding PET results. The average
delay between MR imaging and PET studies was 8 ± 10 days.
The patients were instructed to avoid large meals for 4 hr before the studies and to avoid caffeine (tea, coffee, colas), alcohol, and tobacco for 12 hr before the studies. One catheter was inserted in the antecubital vein for the injections of dipyridamole and also for 15O-labeled water in PET. Another venous catheter was inserted in the opposite arm for blood sampling in the PET studies. During the studies, each patient's ECG measurements were monitored continuously, and blood pressure readings were obtained at 2-min intervals.
MR Imaging
MR imaging was performed with a 1.5-T clinical imager (Signa Horizon LX;
General Electric Medical Systems, Milwaukee, WI). Patients were examined in a
supine position with a 5-inch (12.7-cm) receiver coil attached to the chest
and with a general purpose flexible surface coil on the backside. ECG leads
were attached to the chest for cardiac gating. Axial scout MR images were
obtained for localizing the coronary sinus (Fig.
1A,1B,1C,1D).
|
|
|
|
Breath-hold velocity-encoded fast gradient-echo cine sequences with k-space segmentation were used to measure coronary sinus flow. The following imaging parameters were used: field of view, 28 x 21 cm; matrix 256 x 96, interpolated to 256 x 192; slice thickness, 7 mm (no gaps); pixel size, 1.1 x 2.2 mm; trigger window, 10%; TR/TE, 12/4.9; k-space segmentation with 4 views per segment; and flip angle, 15°. Flow sensitivity was set to ±70 cm/sec; maximum gradient strength of 23 mT/(m·sec), maximum gradient slew rate of 120 T/(m·sec), and receiver bandwidth of ±15.6 kHz were used. Temporal resolution of 48 msec allowed between nine and 19 flow measurements per cardiac cycle, depending on the volunteer's heart rate (breath-hold lasted 24 heartbeats). Three adjacent cross-sectional image sets were obtained perpendicular to the coronary sinus at 7-mm intervals without gaps. The optimal imaging plane was chosen from these three slices to ensure that the coronary sinus was visible throughout the cardiac cycle. After the baseline imaging was completed, the dipyridamole infusion was given. Two minutes after the end of the 4-min infusion, flow measurements were again taken in the same three adjacent slices.
The coronary sinus was traced on the magnitude images. Then, the regions of interest were corrected to avoid misregistration of flow areas by windowing both the positive and negative flow areas visible on the phase images. To compensate for the through-plane motion, a second region of interest was determined for each phase image on the myocardial tissue close to the vessel. Cardiac flow analysis software (Cardiac Flow Analysis Package; General Electric Medical Systems) was used to calculate the mean flow volume rates and velocities. Two independent observers using the software analyzed the images of all patients. The interobserver intraclass correlation for measurements of myocardial blood flow using MR imaging was 0.98. The flow analysis program was validated using a flow phantom and the same imaging sequence used to image the coronary sinus. An excellent correlation (r = 0.998, p < 0.0001) was found between the weighed flow volumes and the measured MR imaging flow volumes.
The left ventricle was imaged with a breath-hold fast gradient-echo cine sequence. The whole left ventricle was covered from apex to base by short-axis images with a slice thickness of 10 mm. The left ventricle endo-and epicardial borders were traced in systole and in diastole for the calculation of left ventricle volume. The left ventricle mass was calculated using a specific left ventricle mass coefficient (1.05), which was multiplied by the left ventricle volume. The global myocardial blood flow was calculated as flow volume divided by 96% of myocardial mass and then expressed in milliliters per minute per gram.
PET
For production of 15O-labeled isotopes, a low-energy deuteron
accelerator was used (Cyclonen 3; Ion Beam Application, Louvaine-la-Neuve,
Belgium). 15O-labeled carbon monoxide was produced in the
conventional way [28].
15O-labeled water was produced using dialysis techniques in a
continuously working water module
[29]. Sterility and pyrogenity
tests were performed to verify the purity of the products.
A 15-slice tomograph (ECAT 931-08; Siemens-CTI, Knoxville, TN) was used with a measured axial resolution of 6.7 mm and plane resolution of 6.5 mm. Each patient was positioned in the equipment so that the heart was within the gantry. After a transmission scan, the patient's nostrils were closed, and he inhaled 15O-labeled carbon monoxide for 2 min through a three-way inhalation flap-valve mechanism (0.14% carbon monoxide mixed with room air, mean dose 3010 ± 140 MBq [81 ± 4 mCi]). Carbon monoxide was allowed to combine with hemoglobin in RBCs for 2 min before a 4-min static scan was started. During the scan period, three blood samples were taken at 2-min intervals, and blood radioactivity concentration was measured immediately with a well-type sodium iodide (Tl) detector (3MW3/3; Bicron, Newbury, OH). Before taking the flow measurements, we waited 10 min to allow 15O-labeled carbon monoxide radioactivity to decay.
Blood flow was measured at baseline and 2 min after administration of dipyridamole (0.56 mg/kg per 4 min). At baseline, 1.670 ± 70 MBq (45 ± 2 mCi) of 15O-labeled water was injected in 2 min, and 1.660 ± 70 MBq (45 ± 2 mCi) of 15O-labeled water was injected after dipyridamole administration. Then dynamic scanning was performed for 6 min. All data were corrected for dead time and for decay and photon attenuation, and were reconstructed in a 128 x 128 matrix. The final in-plane resolution in reconstructed and Hann-filtered (0.3 cycles per second) images was 9.5 mm full width at half-maximum.
Regions of interest were placed on representative axial ventricular slices in each study covering the left ventricle. The regions of interest were drawn on the four mid ventricular images obtained with the patients at rest and copied to the images obtained after dipyridamole administration. The correct placement of the regions was confirmed for the each image set. Values of global myocardial blood flow (expressed in milliliters per gram of tissue per minute) were calculated according to previously published methods using the single-compartment model [30,31]. Mean blood flow values at baseline and after dipyridamole administration were calculated and used in the subsequent analysis. The arterial input function was obtained from the left ventricular time—activity curve using a previously validated method [30], in which corrections were made for the limited recovery of the left ventricular region of interest and the spillover from the myocardial signals.
Data Analysis
For correlation analysis, we calculated Spearman's correlation
coefficients. The intraclass correlation
[32] between the measurements
of myocardial blood flow using MR imaging and using PET was calculated as well
as intraclass correlation for repeated measurements of myocardial blood flow
using MR imaging. Coefficients of variation were calculated by dividing the
standard deviation by the mean and were expressed as percentage values.
Agreement between the methods was also demonstrated by Bland-Altman plots
[33]. All statistical tests
were done using SAS software (version 6.03; SAS Institute, Cary, NC).
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Global Myocardial Blood Flow, Left Ventricular Myocardial Mass, and
Global Coronary Flow Reserve as Revealed by MR Imaging
Based on blood flow volume rates in coronary sinus, global myocardial blood
flow was 139 ± 36 mL·min-1 at baseline and 272
± 81 mL·min-1 at hyperemia. The mean left ventricular
mass was 196 ± 45 g. The calculated global myocardial blood flow was
0.73 ± 0.23 mL·min-1·g-1 at
baseline and 1.43 ± 0.37
mL·min-1·g-1 at hyperemia. These data
yielded an average coronary flow reserve of 1.99 ± 0.47.
Myocardial Blood Flow and Coronary Flow Reserve as Revealed by PET
and the Comparison of PET and MR Imaging
Using PET, we found the mean coronary flow in the left ventricle to be 0.89
± 0.21 mL·min-1·g-1 at baseline and
1.56 ± 0.42 mL·min-1·g-1 at
hyperemia. These data yielded an average coronary flow reserve of 1.77
± 0.36.
The Spearman's correlation coefficients of myocardial blood flow and coronary flow reserve between these two methods were 0.80 and 0.50, respectively (Figs. 2 and 3). The intraclass correlation of myocardial blood flow between PET and MR imaging was 0.68. Coefficients of variation for baseline and hyperemic myocardial blood flow and coronary flow reserve between the methods were 21.2% ± 12.1%, 18.3% ± 13.1%, and 16.0% ± 9.7%. Bland-Altman plots illustrate the validity of MR measurement (Figs. 4 and 5). In univariate analysis, heart rate correlated statistically significantly with myocardial blood flow in MR imaging (r2 = 0.50) and in PET (r2 = 0.50).
|
|
|
|
Hemodynamic Parameters
Dipyridamole infusion increased baseline heart rates of the patients at MR
imaging from 64 to 77 and at PET from 54 to 75 beats per minute. Baseline
rate-pressure products of the patients at MR imaging and at PET were 8940
± 2560 and 7930 ± 1800, respectively, and rate-pressure products
during hyperemia were 10,690 ± 2440 and 10,110 ± 2230,
respectively. The differences in baseline hemodynamic states between the
methods were statistically significant in the two-way analysis of variance
test. However, the calculated intraclass correlation between the studies was
only 0.73 for rate-pressure products at baseline and during hyperemia, which
indicates the personal variability in response between the two study sessions.
Myocardial blood flow corrected for rate-pressure product was calculated for
MR imaging and PET measurements. Coefficients of variation for corrected
baseline and hyperemic myocardial blood flow and coronary flow reserve between
the methods were 23.8% ± 12.6%, 17.2% ± 12.5% and 18.3% ±
11.9%. Variability between the methods can not be explained by dissimilar
hemodynamic responses, although rate-pressure—product correcting does
not make responses directly comparable. Dipyridamole infusion caused mainly
minor symptoms, such as headache and mild discomfort; only three patients
reported chest pain.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Several critical differences exist between our imaging protocol and the ones that had the same study design but used healthy volunteers and orthotopic heart transplant recipients as subjects [10, 27]. We used prospective gating whereas they used retrospective gating, leading to a decrease from 320 to 24 heartbeats at image acquisition. Their protocol produced data that were more complicated to analyze but had better temporal resolution. Correspondingly, different fields of view (280 x 210 mm and 200 x 200 mm) and matrices (256 x 96 and 256 x 160) were used in the previous studies. Therefore, spatial resolution differed from 1.1 x 2.2 mm in our study compared with 0.78 x 1.25 mm in their study. Respiration during retrospective scans on their protocol produced artifacts and blurred images, which seem to be disadvantages that outweigh improved resolution. We were not able to use smaller fields of view because of aliasing. In the previous studies, the myocardium flow drained by the coronary sinus was estimated using different techniques, but neither technique was shown to be absolutely superior to the other because the effect of abnormal coronary anatomy was not taken into account. The earlier researchers reported a slightly better correlation between MR imaging and PET. However, a relatively low r value of coronary flow reserve measurements in this study (when compared with the result in healthy subjects in the previous studies) might be a consequence of coronary flow reserve in patients with coronary artery disease. The flow in these patients is generally low, and typically only a small range of coronary flow reserve in such patients is observed.
Our results of myocardial blood flow and coronary flow reserve measured using MR imaging are in accordance with the results of a previous report of using PET in patients with different stages of the coronary artery stenosis [34]. In the previous study using the coronary sinus method in healthy volunteers [10], the patients' global myocardial blood flow at rest and during hyperemia and global coronary flow reserve were 0.53 ± 0.14 mL·min-1·g-1, 2.27 ± 0.78 mL·min-1·g-1, and 4.3 ± 1.2, measurements that seem reasonable when compared with our results in patients with coronary artery disease.
PET with 15O-labeled water is currently regarded as the most
accurate noninvasive approach to measuring regional and global myocardial
blood flow [26]. Unlike other
methods, PET enables simultaneous measurements of regional and global
myocardial blood flow. 15O-labeled water is a freely diffusible
radionuclide that has simple kinetics. Its short half-life (
2 min) allows
sequential measurements of flow. On the other hand, PET is an expensive,
complex, and time-consuming method to determine coronary flow reserve. The
coefficient of variation of 14% ± 11% has been reported in myocardial
blood flow measurements using PET
[35]. Because of limited
availability, PET is not suitable for large intervention trials. Therefore, MR
imaging, as a noninvasive and radiation-free method, has great potential in
this field.
The blood flow velocity is low in the outer parts of the vessel lumen, which complicates the determination of the lumen-to-vessel interface. In addition, the partial volume effect may lead to an increase in calculated blood flow volume if the vessel area defined is larger than it should be. As reported earlier [10], extensive shifts of the coronary sinus were observed from diastole to systole. Contractility and heart rate increase during hyperemia, which may affect the location of the imaging plane in the coronary sinus; and thus flow data from middle cardiac vein might be registered unintentionally. Measurement of the coronary flow reserve also can be affected.
Coronary sinus cross-sectional area is approximately 80 mm2 and contains 40 pixels. Therefore, the spatial resolution in MR images is sufficient for quantifying the flow volume in coronary sinus. This method is still sensitive to beat-to-beat variations and blood flow—related artifacts, and such sensitivity might reduce to some extent the absolute blood-flow volume measurements. Images were obtained through the entire cardiac cycle with the exception of a small window at the end of the diastole. This short period (called the trigger window) could lead to an overestimation of the flow-volume rate because atrial contraction during late diastole may cause retrograde flow in the coronary sinus. Such a technical limitation in the imaging sequence might produce a slight difference between prospectively and retrospectively gated myocardial blood flow measurements, but both methods have advantages and disadvantages.
Breath-holding changes intrathoracic pressure, thus affecting cardiac output and venous blood flow. However, shallow inspiration of 20-30 sec without Valsalva's maneuver does not significantly affect the aortic blood flow, and therefore, the effect on venous return is marginal (Kawada N et al., presented at International Society of Magnetic Resonance in Medicine, May 1999). There were personal variations in basal hemodynamic states and hemodynamic responses among the patient studies as measured in rate-pressure products. The variation could be explained by normal individual variations or by dissimilar imaging environments. MR imaging studies were performed before PET studies, which may have had an effect on sympathetic tone and thus on myocardial blood flow. These physiologic differences may have had negative effects on measured correlations. In addition, coronary flow reserve value as a ratio is especially sensitive to variations in the measured baseline flow values, which are highly dependent on heart rate and rate-pressure product.
In conclusion, MR imaging allows noninvasive measurement of global myocardial blood flow in patients with coronary artery disease. Also, MR imaging may prove of value in measuring changes to myocardial blood flow in patients who are undergoing interventional treatment.
Acknowledgments
We thank the personnel of the PET Centre and Department of Radiology at
Turku University Hospital for their excellent technical assistance.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
