Multidetector Computed Tomography Myocardial Perfusion Imaging During Adenosine Stress

Multidetector Computed Tomography Myocardial Perfusion Imaging During Adenosine Stress

Journal of the American College of Cardiology © 2006 by the American College of Cardiology Foundation Published by Elsevier Inc. Vol. 48, No. 1, 2006...

388KB Sizes 0 Downloads 129 Views

Journal of the American College of Cardiology © 2006 by the American College of Cardiology Foundation Published by Elsevier Inc.

Vol. 48, No. 1, 2006 ISSN 0735-1097/06/$32.00 doi:10.1016/j.jacc.2006.04.014

Multidetector Computed Tomography Myocardial Perfusion Imaging During Adenosine Stress Richard T. George, MD,* Caterina Silva, MD,* Marco A. S. Cordeiro, MD, PHD,* Anthony DiPaula, BS,* Douglas R. Thompson, PHD,储 William F. McCarthy, PHD,储 Takashi Ichihara, PHD,¶ Joao A. C. Lima, MD, FACC,*§ Albert C. Lardo, PHD, FAHA*†‡§ Baltimore, Maryland; and Otawara, Japan The purpose of this study is to validate the accuracy of multidetector computed tomography (MDCT) to measure differences in regional myocardial perfusion during adenosine stress in a canine model of left anterior descending (LAD) artery stenosis, during first-pass, contrast-enhanced helical MDCT. BACKGROUND Myocardial perfusion imaging by MDCT may have significant implications in the diagnosis and treatment of coronary artery disease. METHODS Eight dogs were prepared with a LAD stenosis, and contrast-enhanced MDCT imaging was performed 5 min into adenosine infusion (0.14 to 0.21 mg/kg/min). Images were analyzed using a semiautomated approach to define the regional signal density (SD) ratio (myocardial SD/left ventricular blood pool SD) in stenosed and remote territories, and then compared with microsphere myocardial blood flow (MBF) measurements. RESULTS Mean MBF in stenosed versus remote territories was 1.37 ⫾ 0.46 ml/g/min and 1.29 ⫾ 0.48 ml/g/min at baseline (p ⫽ NS) and 2.54 ⫾ 0.93 ml/g/min and 8.94 ⫾ 5.74 ml/g/min during adenosine infusion, respectively (p ⬍ 0.05). Myocardial SD was 92.3 ⫾ 39.5 HU in stenosed versus 180.4 ⫾ 41.9 HU in remote territories (p ⬍ 0.001). There was a significant linear association of the SD ratio with MBF in the stenosed territory (R ⫽ 0.98, p ⫽ 0.001) and between regional myocardial SD ratio and MBF ⬍8 ml/g/min, slope ⫽ 0.035, SE ⫽ 0.007, p ⬍ 0.0001. Overall, there was a significant non-linear relationship over the range of flows studied (LR chi-square [2 degrees of freedom] ⫽ 31.8, p ⬍ 0.0001). CONCLUSIONS Adenosine-augmented MDCT myocardial perfusion imaging provides semiquantitative measurements of myocardial perfusion during first-pass MDCT imaging in a canine model of LAD stenosis. (J Am Coll Cardiol 2006;48:153– 60) © 2006 by the American College of Cardiology Foundation OBJECTIVES

Myocardial perfusion imaging (MPI) has proven to be a useful and reliable tool in the diagnosis and prognosis of patients with coronary artery disease (CAD) (1). Pharmacologic-induced coronary vasodilation with dipyridamole or adenosine during the infusion of radionuclide tracers (Tl-201 or Tc-99m-sestamibi) has been shown to be as accurate as exercise stress testing with single-photon emission tomography in diagnosing coronary disease (2– 6). In addition, MPI can assess the physiologic significance of a stenosis and appropriately risk stratify patients with intermediate stenoses (7). Multidetector computed tomography (MDCT) is a rapidly evolving technology with growing application in the noninvasive diagnosis of CAD. The rapid advancement of MDCT technology over the past 5 years has greatly improved its

spatial/temporal resolution and has expanded its application beyond coronary angiography towards a more comprehensive evaluation of cardiovascular disease, including function, viability, and perfusion (8 –10). In its current form, noninvasive coronary angiography by MDCT lacks nuclear stress testing’s ability to assess the functional significance of coronary lesions as supported by a recent study (11). In light of these observations, a novel method to assess myocardial perfusion during first-pass, contrast-enhanced MDCT would be of great value in the diagnosis and treatment of CAD. The purpose of this study is two-fold: 1) to validate the ability of MDCT to detect a reduction in perfusion in a myocardial territory supplied by a stenosed vessel during adenosine stress; and 2) to correlate MDCT-derived myocardial signal density (SD) with myocardial blood flow (MBF) measurements determined by microspheres.

From the *Department of Medicine, Division of Cardiology, †Department of Biomedical Engineering, ‡Department of Surgery, and the §Department of Radiology, Johns Hopkins University School of Medicine, Baltimore, Maryland; 储Maryland Medical Research Institute, Baltimore, Maryland; and the ¶Toshiba Medical Systems Corporation, Otawara, Japan. Dr. Ichihara is an employee of Toshiba American Medical Systems, Inc. Drs. Lardo and Lima are, in part, funded by a research grant from Toshiba American Medical Systems, Inc. The terms of this arrangement are being managed by Johns Hopkins University in accordance with its conflict of interest policies. Research support: Donald W. Reynolds Foundation, Las Vegas, Nevada; Toshiba American Medical Systems, Tufton, California. Markus Schwaiger, MD, acted as Guest Editor for this paper. Manuscript received November 9, 2005; revised manuscript received February 28, 2006, accepted March 8, 2006.

METHODS Animal preparation. The Animal Care and Use Committee of the Johns Hopkins University School of Medicine approved all procedures. A total of 8 mongrel dogs (24.5 to 29.5 kg) were anesthetized with intravenous thiopental, intubated, and mechanically ventilated with isoflurane anesthesia during preparation and MDCT scanning. After femoral cut downs, 8-F sheaths were placed in both femoral veins, right femoral artery, and the right internal jugular

154

George et al. MDCT Perfusion Imaging

Abbreviations and Acronyms CAD ⫽ coronary artery disease EBCT ⫽ electron beam computed tomography LAD ⫽ left anterior descending artery LV ⫽ left ventricle/ventricular MBF ⫽ myocardial blood flow MBV ⫽ myocardial blood volume MDCT ⫽ multidetector computed tomography MPI ⫽ myocardial perfusion imaging SD ⫽ signal density

vein. A left thoracotomy was performed in the fifth intercostal space, and the pericardium was excised. A catheter was then placed into the proximal descending aorta for microsphere sampling, and neutron-activated microspheres (STERIspheres, BioPAL, Inc., Worcester, Massachusetts) were injected using a left atrial catheter to document baseline MBF. The proximal to mid-left anterior descending artery (LAD) was isolated and instrumented with an electromagnetic flow meter, and reactive hyperemia was tested in the vessel with an inflatable external occluder. Suture was then secured around the LAD and tightened to produce a graded stenosis aimed to maintain baseline flow, but attain a 50% or more reduction in hyperemic flow. The thoracotomy was closed, and the animal was transported to the MDCT suite using a portable ventilator. MDCT imaging protocol. Each animal was placed on an electrocardiographic monitor in a 64-detector MDCT scanner (Aquilion 64, Toshiba Medical Systems Corporation, Otawara, Japan). Animals received intravenous propranolol (5 to 20 mg) to achieve a heart rate ⬍100 beats/min. In order to study a wide range of flows, adenosine was infused for 5 min at 0.14 mg/kg/min in three experiments, 0.21 mg/kg/min in three experiments, and one animal did not receive adenosine. After scout film acquisition, intravenous contrast iodixanol (Visipaque 320 mg iodine/ml, Amersham Health, Amersham, United Kingdom) was infused at a rate of 2.5 ml/s for a total of 100 ml. Using bolus tracking, MDCT scanning was initiated when a threshold of 180 HU was detected in the ascending aorta. Respiration was then suspended with the airway open to air and imaging performed using a retrospectively gated MDCT protocol with the following parameters: gantry rotation time ⫽ 400 ms,

JACC Vol. 48, No. 1, 2006 July 4, 2006:153–60

detector collimation ⫽ 0.5 mm ⫻ 32 (isotropic voxels ⫽ 0.5 ⫻ 0.5 ⫻ 0.5 mm3 ⫺ 13 linepairs/cm), helical pitch ⫽ variable depending on heart rate (range: 6.4 to 8.8), beam pitch ⫽ variable depending on heart rate (range: 0.224 to 0.275), tube voltage ⫽ 120 kV, tube current ⫽ 400 mA, and a display field of view ⫽ 13.2 cm. Imaging started at the aortic root and stopped caudal to all cardiac structures. Before discontinuation of the adenosine infusion, neutronactivated microspheres were again injected to document MBF during adenosine infusion. A timeline summary of experimental and imaging protocols is shown in Figure 1. Sample processing. After imaging, the thoracotomy was reopened and the LAD suture was tightened to occlude the vessel. Monastral blue dye was injected into the left atrium to stain myocardium remote to the LAD territory. The animal was then euthanized with a saturated solution of potassium chloride, and the heart was excised and divided into five equal slices perpendicular to the short axis. Myocardial samples (0.87 to 3.24 g) were excised and processed according to the technique described by Reinhardt et al. (12) from the anterior, anteroseptal, lateral, inferolateral, inferior, and inferoseptal walls from base to apex according to the 17-segment system recommended by the American Heart Association Writing Group on Myocardial Segmentation and Registration for Cardiac Imaging (13). Semi-automated, volumetric regional image analysis. Segmented reconstruction was performed, and images were examined at 10% intervals throughout the R-R interval to determine the optimal phase for reconstruction and analysis. Using multiplanar reconstruction, images were reconstructed in the cardiac short axis at a 4-mm slice thickness. Using hand planimetry, the remote myocardial SD was determined throughout the myocardial volume containing the inferoseptal, inferior, posterior, and inferolateral walls. These results were used to define a perfusion deficit SD threshold as myocardium having an SD one standard deviation below the mean SD of the remote region. Then, using custom cardiac function/perfusion software, with images reconstructed in the cardiac short axis, endocardial and epicardial borders were defined using an automated border detection algorithm (Toshiba Medical, Inc.). The perfusion deficit SD threshold was entered into the custom software that used a step based algorithm to: 1) detect voxels

Figure 1. Experimental protocol for multidetector computed tomography (MDCT) myocardial perfusion imaging.

George et al. MDCT Perfusion Imaging

JACC Vol. 48, No. 1, 2006 July 4, 2006:153–60

within the defined SD range; 2) check for voxel continuity in the X, Y, and Z directions; 3) generate clusters of voxels that meet the perfusion deficit definition; and 4) define the perfusion deficit as the largest cluster detected. Using this method, the software determined the mean SD of the perfusion defect in HU and calculated its volume and mass. Left ventricular (LV) blood pool SD was determined by calculating the mean SD of the entire LV cavity. In order to compare myocardial SD measurements with microsphere MBF measurements, all myocardial signal intensities were normalized as follows: Myocardial SD Ratio ⫽

MeanMYOSD MeanLVSD

where MYOSD ⫽ myocardial SD and LVSD ⫽ LV blood pool SD. Myocardial SD ratio determined in the perfusion deficit was compared with microsphere-derived MBF. Additionally, myocardial SD ratio of the perfusion deficit and remote myocardium were compared with microsphere-derived MBF over the range of flows studied. Slice-by-slice image analysis. Using segmented reconstruction described above, images were reconstructed in the axial plane with a 4-mm slice thickness and a cardiac phase of 80%. Using hand planimetry, regions of interest were drawn in the stenosed and remote territories for each slice, and the SD was determined. In order to determine if there was SD variation from base to apex secondary to the time the axial slice was acquired, SD for each slice from the base to LV apex was examined to verify that SD did not vary beyond the signal noise (standard deviation) for each slice. Sliceby-slice SD was then normalized to the LV blood pool SD (myocardial SD ratio) as described above. Slice-by-slice myocardial SD ratio was then compared with slice-by-slice microsphere-derived MBF.

155

Statistical analysis. Myocardial blood flow and myocardial SD were expressed as mean ⫾ standard deviation. Volumetric data were expressed as mean and range. Mean MBF and MDCT signal densities were compared using the paired t test. When considering a single SD measurement in the stenosed territory per experiment and its relationship to MBF, linear regression analysis was performed to estimate the slope and statistical significance of the relationship, and a Pearson correlation coefficient was calculated. After the method of Bowman and Waller (14), analyses involving multiple measures at different locations within animals used mixed models to account for the dependence among measures taken from the same animal. In all models, polynomial terms were included when they contributed significantly to model fit based on likelihood ratio tests. Likelihood ratio tests were also used to test whether there were statistically significant associations between SD ratios and MBF. All tests were performed at an alpha level of p ⬍ 0.05. Statistical analyses were performed using SAS version 9.1 (SAS Institute, Cary, North Carolina).

RESULTS One of eight animals prepared for testing became hemodynamically unstable and could not be imaged. Mean preadenosine MBF by microspheres was 1.37 ⫾ 0.46 ml/g/min in the stenosed territory and 1.29 ⫾ 0.48 ml/g/min in the remote territory (p ⫽ NS). During adenosine infusion, MBF increased to 2.54 ⫾ 0.93 ml/g/min in stenosed and 8.94 ⫾ 5.74 ml/g/min in remote territories (p ⱕ 0.05). Perfusion deficits showed marked differences in myocardial SD compared with remote regions on visual inspection (Fig. 2). Myocardial SD in stenosed and remote territories measured 92.3 ⫾ 39.5 HU and 180.4 ⫾ 41.9 HU (p ⫽ 0.001), respectively, and there were statistically significant differences in myocardial SD in ischemic versus remote regions in

Figure 2. Mid-ventricular slice in the axial plane showing a perfusion deficit (arrows) in the anteroseptal, anterior, and anterolateral myocardial territory supplied by the stenosed left anterior descending artery (left). Multiplanar reconstruction showing the extent of the perfusion deficit (arrows) extending from the anteroseptal and anterior walls to the apex (right).

156

George et al. MDCT Perfusion Imaging

Figure 3. Differences in myocardial signal density measured in myocardial territories supplied by a stenosed left anterior descending coronary artery (open bars) versus remote territories (solid bars) in experiments 1 through 6 receiving adenosine (p ⬍ 0.05). No significant difference noted in the control.

each experiment (Fig. 3). The defect relative mass was on average 14.7% (range 4.1% to 30.0%) of the LV myocardium. Flow reserve, calculated by dividing flow during adenosine infusion by that territory’s baseline flow, was 1.9 ⫾ 0.6 in stenosed versus 7.3 ⫾ 4.7 in remote territories (p ⬍ 0.05). The ratio of the myocardial SD and the LV cavity blood pool SD (MYOSD/LVSD) was 0.17 ⫾ 0.07 for stenosed versus 0.33 ⫾ 0.08 for remote territories (p ⬍ 0.01). The semiautomated volumetric analysis algorithm accurately identified the location of the perfusion defect in the myocardial territory supplied by the stenosed LAD in all animals during adenosine infusion (Fig. 4). For the semiautomatic volumetric analysis method, there was a significant linear association of MYOSD/LVSD and microspherederived MBF in the stenosed territory, slope ⫽ 0.070, SE ⫽

JACC Vol. 48, No. 1, 2006 July 4, 2006:153–60

Figure 5. Perfusion deficit myocardial signal density (SD) divided by the left ventricular (LV) blood pool signal density (y-axis) versus microspherederived myocardial blood flow (x-axis) in the stenosed left anterior descending artery territory determined by a semiautomated approach that defines the perfusion deficit as one standard deviation below the mean SD of the remote myocardial territory. (y ⫽ 0.07037x ⫺ 0.00681, Pearson R ⫽ 0.98, p ⬍ 0.001; LR chi-square [1 degree of freedom] ⫽ 19.21, SE ⫽ 0.007, p ⬍ 0.0001) (n ⫽ 6).

0.007, p ⬍ 0.0001 (Pearson R ⫽ 0.98, p ⫽ 0.001) as shown in Figure 5. There was a significant non-linear relationship between regional SD ratio (MYOSD/LVSD) and absolute MBF in both the ischemic and remote regions over the entire range of flows studied and is shown in Figure 6A (y ⫽ ⫺0.00168x2 ⫹ 0.04701x ⫹ 0.07604, LR chi-square [2 degrees of freedom] ⫽ 31.8, p ⬍ 0.0001). However, when considering absolute MBF ⬍8 ml/g/min, as shown in Figure 6B, there was a significant linear association between SD ratio and MBF (slope ⫽ 0.035, SE ⫽ 0.007, p ⬍ 0.0001). A slice-by-slice analysis showed there was no significant variation in SD from slice to slice from the base to the LV apex when restricting the analysis within the remote portion of the myocardial wall. The relationship of SD ratio versus flow in the slice-by-slice analysis was non-linear over the

Figure 4. (A) Shows a mid-ventricular slice of myocardium clearly showing remote territories stained with Monastral blue dye and the unstained left anterior descending coronary artery territory. (B) Shows the corresponding multiplanar reconstructed image in short axis. Using semiautomated function/perfusion software (Toshiba, Inc.), myocardium meeting the perfusion deficit signal density threshold of one standard deviation below the remote myocardial signal density is designated in blue.

JACC Vol. 48, No. 1, 2006 July 4, 2006:153–60

Figure 6. (A) Regional myocardial signal density divided by the left ventricular blood pool signal density (y-axis) versus microsphere-derived myocardial blood flow (x-axis) in both the stenosed left anterior descending artery territory and remote myocardial territory using a semiautomated volumetric analysis over the range of flows studied (y ⫽ ⫺0.00168x2 ⫹ 0.04701x ⫹ 0.07604, LR chi-square (2 df) ⫽ 31.8, p ⬍ 0.0001). (B) Data representing flows less than 8 ml/g/min from (A) are shown in (B) (y ⫽ 0.03453x ⫹ 0.09437, LR chi-square [1 degree of freedom] ⫽ 17.0, SE ⫽ 0.007, p ⬍ 0.0001) (n ⫽ 7). SD ratio ⫽ myocardial signal density/left ventricular blood pool signal density.

George et al. MDCT Perfusion Imaging

157

we used a novel image analysis method along with objective criteria to define perfusion deficits that eliminated observer bias commonly associated with qualitative and quantitative hand planimetry-based analysis methods. To our knowledge, this is the first study to evaluate the relationship between myocardial SD and MBF using first-pass, helical 64-detector computed tomography. Dynamic versus helical MDCT perfusion imaging. Our study used MDCT “helical mode” scanning for MPI, which is fundamentally different from “dynamic” or “flow mode” scanning as is performed during electron beam computed tomography (EBCT). Unlike dynamic scanning where images are acquired in four to six fixed slice locations over a predetermined period of time to characterize the wash-in and wash-out of contrast in the myocardium, helical MDCT imaging implements a moving table that transports the subject through the plane of a rotating gantry that contains a grid of detectors opposite a radiation source. Imaging is initiated during first-pass contrast injection when a pre-defined concentration of contrast is detected in the aortic root and continues over a predetermined volume of the chest, starting cranial to the base of the heart and ending caudal to the apex. In this mode, imaging is performed during retrospective gating and generates a stack of 250 to 300 images, each with near isotropic voxel resolution (0.5-mm thickness). Our study is unique as it aims to measure differences in myocardial perfusion during helical MDCT imaging by capturing myocardial enhancement during a specific phase of the first-pass contrast enhancement curve (Fig. 8). Unlike myocardial perfusion studies using dynamic X-ray imaging, helical MDCT scanning cannot fully measure all phases of myocardial contrast kinetics. Instead, our study is designed to demonstrate that a first-pass helical MDCT coronary angiographic imaging protocol is capable of detecting myocardial perfusion patterns that take place in the early phase of first-pass contrast-enhancement and that relative SD

range of flows studied (y ⫽ ⫺0.00041x2 ⫹ 0.01575x ⫹ 0.1937, LR chi-square [2 degrees of freedom] ⫽ 14.1, p ⫽ 0.001) (Fig. 7).

DISCUSSION Main findings. The main findings of this study are: 1) first-pass, contrast-enhanced helical MDCT imaging can detect a LAD territory flow deficit during adenosine stress; and 2) MDCT SD ratios of stenosed and remote myocardial beds correlate well with microsphere-derived absolute MBF. These data are the first to demonstrate that myocardial perfusion information can be extracted from imaging protocols designed for first-pass coronary angiography. We used a well-established canine model of LAD stenosis to achieve lesions that were non-flow limiting at rest but flow limiting during pharmacologic stress. This allowed us to compare the sensitivity of relative myocardial MDCT measurements with a proven gold standard for MBF. Further,

Figure 7. Slice-by-slice myocardial signal density divided by the left ventricular blood pool signal density (y-axis) versus slice-by-slice microsphere-derived myocardial blood flow (x-axis) in both the stenosed left anterior descending artery territory and remote myocardial territory over the range of flows studied (y ⫽ ⫺0.00041x2 ⫹ 0.01575x ⫹ 0.1937 LR chi-square [2 degrees of freedom] ⫽ 14.1, p ⫽ 0.001) (n ⫽ 163). SD ratio ⫽ myocardial signal density/left ventricular blood pool signal density.

158

George et al. MDCT Perfusion Imaging

JACC Vol. 48, No. 1, 2006 July 4, 2006:153–60

Figure 8. Signal density-time curves in a canine model of left anterior descending artery stenosis during infusion of adenosine, intravenous iodinated contrast (2.5 ml/s), and dynamic multidetector computed tomography scanning. Myocardial curves were measured from the anterior myocardial wall (stenosed) and the inferior myocardial wall (remote). Vertical lines illustrate the period that helical multidetector computed tomography scanning took place during this study. Helical scanning was triggered when bolus tracking detected a signal density of 180 HU in the ascending aorta. LV ⫽ left ventricle.

analysis of the myocardium provides an accurate surrogate for MBF. Microvascular physiology and its implications for MDCT perfusion imaging. Autoregulation of MBF governed by vasoconstriction/vasodilation of arterioles and venules, in addition to capillary recruitment and derecruitment, is essential for myocyte homeostasis (15–17). In the setting of a normal epicardial coronary artery and myocardial hyperemia secondary to adenosine receptor agonists, arteriole and venule vasodilation decreases the resistance across the coronary microcirculation, increasing MBF (18). Additionally, studies using EBCT and myocardial contrast echocardiography show increased myocardial blood volume (MBV) in myocardium supplied by a normal coronary artery during hyperemia (19,20). Conversely, in the setting of a coronary stenosis, hyperemic MBF begins to decrease with a diameter stenosis ⬎40% to 50%, and capillary derecruitment results in decreased MBV in myocardial territories supplied by a stenosed artery (18,21). Using X-ray-computed tomography in “dynamic” or “flow” mode, Wang et al. (22) have studied the relationships between MBF and MBV in the coronary microcirculation. In summary, they calculated MBV using the ratio of the area under the curve for contrast-enhanced myocardium and the aortic blood pool. Then MBF was derived using the ratio of MBV and the contrast bolus transit time. Using this method, they accurately estimated MBF compared with microsphere-derived measurements. Furthermore, they demonstrated that MBV increases with increased MBF. In another study that examined MBF and MBV under several hemodynamic conditions, including adenosine infusion, they again concluded that MBV has a predictable relationship to MBF (23). More recently, Möhlenkamp et al. (19), using EBCT and intracoronary Doppler ultrasound, further confirmed that there is a strong correlation between EBCT-

derived MBF and MBV and that both increase in response to adenosine. Our study uses the ratio of myocardial SD and LV blood pool SD, which mathematically more closely resembles the measurement of MBV in these previous studies. While this ratio may not directly measure MBF and more likely represents a measurement of MBV, previous studies show a tight correlation between MBF and MBV and support their use as a surrogate measure of myocardial perfusion. Clinical implications. Multidetector computed tomography has significant unique advantages over other tomographic imaging modalities including unsurpassed spatial resolution, short scanning times, and the ability to image patients with implantable electronic devices. Faster scanning times with 64-detector systems make it possible to assess perfusion during a helical MDCT scan because there is less slice-to-slice variability in the Z axis when scanning from base to apex. Since the introduction of MDCT, studies have documented improved accuracy for the detection of coronary stenoses ⬎50% (24 –29). Multidetector computed tomography MPI could offer a second “lens” of interpretation through which a cardiac MDCT coronary angiogram could be evaluated. An algorithm that determines not only the degree of percent stenosis, but also whether a perfusion defect is present or absent in the corresponding myocardial region would be a valuable tool in the non-invasive diagnosis of CAD. Our study demonstrates that differences in regional and slice-by-slice myocardial SD reflect differences in myocardial perfusion. While this technique falls short of absolute MBF quantification, the myocardial SD ratio is a robust semiquantitative metric that reflects differences in both MBF and MBV within the clinically important range of flows (up to 8 ml/g/min) as previously defined by positron emission tomography, magnetic resonance, and EBCT at

George et al. MDCT Perfusion Imaging

JACC Vol. 48, No. 1, 2006 July 4, 2006:153–60

159

multiple stenoses making it difficult to define perfusion deficits indexed or normalized to normal myocardium. Human studies may require either a rest and stress scan or, alternatively, may require defining normal values for contrast-enhanced myocardium normalized to the LV blood pool SD. Lastly, beam-hardening artifacts can have an MDCT signal appearance that closely resembles a perfusion deficit. In the current study, beam hardening artifacts were easily distinguishable from perfusion deficits because they had several unique characteristics including a triangular, transmural, and hypoenhanced appearance emanating from a structure with a high SD, and they did not follow the distribution of a coronary vascular bed. Conclusions. Adenosine-augmented MDCT MPI can be successfully performed in a canine model of LAD stenosis using first-pass, contrast-enhanced helical MDCT imaging. Perfusion deficits can be detected visually, and the myocardial SD ratio provides a semiquantitative metric that correlates well with microsphere-derived MBF. The ability to perform MPI during MDCT coronary imaging could have significant implications in the diagnosis and treatment of patients with CAD.

Figure 9. First-pass, adenosine-augmented multidetector computed tomography (MDCT) myocardial perfusion imaging in a patient referred for invasive angiography after single-photon emission computed tomography showed a fixed perfusion deficit in the inferior and inferolateral territories. Panels A and C demonstrate an inferior and inferolateral subendocardial perfusion deficit in the mid and distal left ventricle, respectively (arrows). Using semi-automated function/perfusion software, myocardium meeting the perfusion deficit signal density threshold of one standard deviation below the remote myocardial signal density is designated in blue in panels B and D. Invasive angiography shows a chronically occluded distal right coronary artery with left to right collaterals filling the posterior descending (arrows) and posterolateral branches (E). (F) 17-segment polar plot of MDCT-derived myocardial signal densities. Note the hypoperfused inferior and inferolateral regions displayed in blue.

baseline and during hyperemia (30 –32). Clinical studies currently underway at our institution will determine if MDCT can reliably predict the physiological significance of coronary stenoses in patients (Fig. 9). Study limitations. A potential limitation of using adenosine for stress MDCT MPI is tachycardia and arrhythmias that can significantly degrade the quality of MDCT imaging. To offset the tachycardic effect, beta-blockers were used, which, in turn, have been shown to increase hyperemic MBF, and this may be more pronounced in ischemic segments (33,34). Although this effect predisposed our hypothesis to failure, significant differences between ischemic and remote myocardial segments were still detected. Our study demonstrates the feasibility of evaluating myocardial perfusion in a single scan in the setting of an experimental stenosis in the LAD territory only. It is unclear whether we would obtain the same results in patients with stenoses in other territories or who may have

Acknowledgment The authors thank Jorge Guzman for his technical assistance in this project. Reprint requests and correspondence: Dr. Albert C. Lardo, Johns Hopkins School of Medicine, Division of Cardiology, 720 Rutland Avenue, 1042 Ross Building, Baltimore, Maryland 21205. E-mail: [email protected].

REFERENCES 1. Klocke FJ, Baird MG, Lorell BH, et al. ACC/AHA/ASNC guidelines for the clinical use of cardiac radionuclide imaging— executive summary: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (ACC/AHA/ ASNC Committee to Revise the 1995 Guidelines for the Clinical Use of Cardiac Radionuclide Imaging). J Am Coll Cardiol 2003; 42:1318 –33. 2. Nguyen T, Heo J, Ogilby JD, Iskandrian AS. Single photon emission computed tomography with thallium-201 during adenosine-induced coronary hyperemia: correlation with coronary arteriography, exercise thallium imaging and two-dimensional echocardiography. J Am Coll Cardiol 1990;16:1375– 83. 3. Parodi O, Marcassa C, Casucci R, et al. Accuracy and safety of technetium-99m hexakis 2-methoxy-2-isobutyl isonitrile (sestamibi) myocardial scintigraphy with high dose dipyridamole test in patients with effort angina pectoris: a multicenter study. Italian Group of Nuclear Cardiology. J Am Coll Cardiol 1991;18:1439 – 44. 4. Coyne EP, Belvedere DA, Vande Streek PR, Weiland FL, Evans RB, Spaccavento LJ. Thallium-201 scintigraphy after intravenous infusion of adenosine compared with exercise thallium testing in the diagnosis of coronary artery disease. J Am Coll Cardiol 1991;17:1289 –94. 5. Nishimura S, Mahmarian JJ, Boyce TM, Verani MS. Equivalence between adenosine and exercise thallium-201 myocardial tomography: a multicenter, prospective, crossover trial. J Am Coll Cardiol 1992;20: 265–75. 6. Gupta NC, Esterbrooks DJ, Hilleman DE, Mohiuddin SM. Comparison of adenosine and exercise thallium-201 single-photon emission computed tomography (SPECT) myocardial perfusion imaging. The

160

7.

8. 9.

10.

11.

12.

13.

14. 15. 16. 17. 18. 19.

20.

George et al. MDCT Perfusion Imaging

GE SPECT Multicenter Adenosine Study Group. J Am Coll Cardiol 1992;19:248 –57. Miller DD, Donohue TJ, Younis LT, et al. Correlation of pharmacological 99mTc-sestamibi myocardial perfusion imaging with poststenotic coronary flow reserve in patients with angiographically intermediate coronary artery stenoses. Circulation 1994;89:2150 – 60. Hoffmann U, Millea R, Enzweiler C, et al. Acute myocardial infarction: contrast-enhanced multi-detector row CT in a porcine model. Radiology 2004;231:697–701. Mahnken AH, Koos R, Katoh M, et al. Assessment of myocardial viability in reperfused acute myocardial infarction using 16-slice computed tomography in comparison to magnetic resonance imaging. J Am Coll Cardiol 2005;45:2042–7. Lardo AC, Cordeiro MA, Silva C, et al. Contrast enhanced multidetector computed tomography viability imaging following myocardial infarction: characterization of myocyte death, microvascular obstruction and chronic scar. Circulation 2006;113:394 – 404. Hacker M, Jakobs T, Matthiesen F, et al. Comparison of spiral multidetector CT angiography and myocardial perfusion imaging in the noninvasive detection of functionally relevant coronary artery lesions: first clinical experiences. J Nucl Med 2005;46:1294 –300. Reinhardt CP, Dalhberg S, Tries MA, Marcel R, Leppo JA. Stable labeled microspheres to measure perfusion: validation of a neutron activation assay technique. Am J Physiol Heart Circ Physiol 2001;280: H108 –16. Cerqueira MD, Weissman NJ, Dilsizian V, et al. Standardized myocardial segmentation and nomenclature for tomographic imaging of the heart: a statement for healthcare professionals from the Cardiac Imaging Committee of the Council on Clinical Cardiology of the American Heart Association. Circulation 2002;105:539 – 42. Bowman FD, Waller LA. Modelling of cardiac imaging data with spatial correlation. Stat Med 2004;23:965– 85. Habazettl H, Vollmar B, Christ M, Baier H, Conzen PF, Peter K. Heterogeneous microvascular coronary vasodilation by adenosine and nitroglycerin in dogs. J Appl Physiol 1994;76:1951– 60. House SD, Johnson PC. Diameter and blood flow of skeletal muscle venules during local flow regulation. Am J Physiol 1986;250: H828 –37. Jarhult J, Mellander S. Autoregulation of capillary hydrostatic pressure in skeletal muscle during regional arterial hypo- and hypertension. Acta Physiol Scand 1974;91:32– 41. Jayaweera AR, Wei K, Coggins M, Bin JP, Goodman C, Kaul S. Role of capillaries in determining CBF reserve: new insights using myocardial contrast echocardiography. Am J Physiol 1999;277:H2363–72. Möhlenkamp S, Behrenbeck TR, Lerman A, et al. Coronary microvascular functional reserve: quantification of long-term changes with electron-beam CT preliminary results in a porcine model. Radiology 2001;221:229 –36. Bin JP, Le E, Pelberg RA, Coggins MP, Wei K, Kaul S. Mechanism of inducible regional dysfunction during dipyridamole stress. Circulation 2002;106:112–7.

JACC Vol. 48, No. 1, 2006 July 4, 2006:153–60 21. Uren NG, Melin JA, De Bruyne B, Wijns W, Baudhuin T, Camici PG. Relation between myocardial blood flow and the severity of coronary-artery stenosis. N Engl J Med 1994;330:1782– 8. 22. Wang T, Wu X-S, Chung N, et al. Myocardial blood flow estimated by synchronous multislice, high-speed computed tomography. IEEE Trans Med Imaging 1989;8:70 –7. 23. Wu X-S, Ewert DL, Liu YH, Ritman EL. In vivo relation of intramyocardial blood volume to myocardial perfusion. Evidence supporting microvascular site for autoregulation. Circulation 1992;85: 730 –7. 24. Kuettner A, Beck T, Drosch T, et al. Diagnostic accuracy of noninvasive coronary imaging using 16-detector slice spiral computed tomography with 188 ms temporal resolution. J Am Coll Cardiol 2005;45:123–7. 25. Nieman K, Cademartiri F, Lemos PA, Raaijmakers R, Pattynama PM, de Feyter PJ. Reliable noninvasive coronary angiography with fast submillimeter multislice spiral computed tomography. Circulation 2002;106:2051– 4. 26. Ropers D, Baum U, Pohle K, et al. Detection of coronary artery stenoses with thin-slice multi-detector row spiral computed tomography and multiplanar reconstruction. Circulation 2003;107:664 – 6. 27. Leber AW, Knez A, von Ziegler F, et al. Quantification of obstructive and nonobstructive coronary lesions by 64-slice computed tomography: a comparative study with quantitative coronary angiography and intravascular ultrasound. J Am Coll Cardiol 2005;46:147–54. 28. Leschka S, Alkadhi H, Plass A, et al. Accuracy of MSCT coronary angiography with 64-slice technology: first experience. Eur Heart J 2005;26:1482–7. 29. Raff GL, Gallagher MJ, O’Neill WW, Goldstein JA. Diagnostic accuracy of noninvasive coronary angiography using 64-slice spiral computed tomography. J Am Coll Cardiol 2005;46:552–7. 30. Bell MR, Lerman LO, Rumberger JA. Validation of minimally invasive measurement of myocardial perfusion using electron beam computed tomography and application in human volunteers. Heart 1999;81:628 –35. 31. Muehling OM, Jerosch-Herold M, Panse P, et al. Regional heterogeneity of myocardial perfusion in healthy human myocardium: assessment with magnetic resonance perfusion imaging. J Cardiovasc Magn Reson 2004;6:499 –507. 32. Krivokapich J, Smith GT, Huang SC, et al. 13N ammonia myocardial imaging at rest and with exercise in normal volunteers. Quantification of absolute myocardial perfusion with dynamic positron emission tomography. Circulation 1989;80:1328 –37. 33. Bottcher M, Czernin J, Sun K, Phelps ME, Schelbert HR. Effect of beta 1 adrenergic receptor blockade on myocardial blood flow and vasodilatory capacity. J Nucl Med 1997;38:442– 6. 34. Koepfli P, Wyss CA, Namdar M, et al. Beta-adrenergic blockade and myocardial perfusion in coronary artery disease: differential effects in stenotic versus remote myocardial segments. J Nucl Med 2004;45: 1626 –31.