Physiological Basis of Myocardial Contrast Enhancement in Fast Magnetic Resonance Images of 2-Day-Old Reperfused Canine Infarcts
Background Contrast-enhanced fast magnetic resonance (MR) images of acute, reperfused human infarcts demonstrate regions of hypoenhancement and hyperenhancement. The relations between the spatial extent and time course of these enhancement patterns to myocardial risk, infarct, and no-reflow regions have not been well characterized.
Methods and Results The proximal left anterior descending coronary artery was occluded in 11 closed-chest dogs for 90 minutes followed by 2 days of reperfusion. Regional blood flow was determined by use of radioactive microspheres. The animals were studied at the 2-day time point with contrast-enhanced fast MRI (Signa 1.5 T, General Electric). Thioflavin-S was administered to demarcate no-reflow regions. The hearts were then excised, sectioned into five base-to-apex slices, stained with 2,3,5-triphenyltetrazolium chloride (TTC), and photographed under room light (for TTC) and ultraviolet light (for thioflavin). The spatial extents of thioflavin-negative, TTC-negative, and risk regions were compared planimetrically with MRI hypoenhanced and hyperenhanced regions. The spatial locations of subendocardial hypoenhancement in MR images correlated closely with those of thioflavin-negative regions. Microsphere blood flow in these regions was significantly reduced compared with remote regions (0.37±0.09 versus 0.88±0.10 mL/min per gram, respectively, P<.001) and with baseline (0.37±0.09 versus 0.87±0.15 mL/min per gram, P<.01). The spatial extent of hyperenhancement was smaller than the risk region (r=.64, slope=0.48, P<.001) but highly correlated with TTC-negative regions and were, on average, 12% larger (r=.93, slope=1.12, P=.035).
Conclusions In contrast-enhanced MR images of 2-day-old reperfused canine infarcts, myocardial regions of hypoenhancement are related to the no-reflow phenomenon. Approximately 90% of the myocardium within hyperenhanced regions is nonviable.
Reestablishing perfusion to regions of acute MI is an important determinant of long-term prognosis.1 2 3 The success of attempts to reperfuse via thrombolytics, angioplasty, or bypass surgery typically has been evaluated with angiography,4 but this technique addresses only the issue of large-vessel patency. The question of whether or not reperfusion has been reestablished at the microvascular level has usually been addressed by use of nuclear medicine techniques such as thallium scintigraphy.5 6
Over the past several years, a number of experimental studies have suggested that contrast-enhanced MR imaging could be used to characterize microvascular perfusion.7 8 9 10 11 12 13 14 Early studies using spin-echo imaging required several minutes to obtain each image of the heart.10 11 12 13 14 More recently, several groups have used fast MR imaging to characterize the myocardial first pass of MR contrast agents after bolus administration.8 15 Our group recently reported the use of fast MR to examine myocardial enhancement over a time scale of minutes in acute, reperfused human infarcts.16 These studies have demonstrated that when the infarct-related artery remains occluded, myocardial image intensity in the area perfused by that artery is reduced compared with normal myocardium after contrast administration.14 15 17 Conversely, in those infarcts that are successfully reperfused, myocardial image intensity is increased compared with normal myocardium.10 13 14 16 18
Ultimately, the potential of contrast MR to provide information about microvascular perfusion will be determined by the relation of these altered enhancement patterns to the underlying pathophysiology. Many key issues concerning the relation of flow to hypoenhancement and hyperenhancement in contrast-enhanced MR images of the heart remain poorly understood. In the present study, we examined the relation between regional myocardial blood flow and MR contrast-enhancement patterns in a closed-chest canine model of acute, reperfused MI. This canine model has been well characterized by our group19 and results in well-defined regions of ischemic myocardium at risk, infarcted myocardium, and regions of no-reflow that can be delineated by established methods.19 The imaging methods used in the present study were identical to those of our clinical study16 to facilitate comparison with the images obtained from patients.
Eleven mongrel dogs weighing 20 to 25 kg were studied. The care and treatment of all animals involved in this study were in accordance with institutional guidelines. On day 1, the animals were anesthetized with sodium pentobarbital (35 mg/kg IV), intubated, and mechanically ventilated. A catheter sheath was placed into the right femoral artery and used to introduce a 6F pigtail catheter into the LV under fluoroscopic guidance. This catheter was used for microsphere administration and to monitor LV pressure. The sheath of this catheter was used for microsphere reference sampling from the femoral artery. A second catheter sheath was then placed in the left femoral artery and used to introduce a 7F left Judkins 2.5 guiding catheter into the left main coronary artery. A 3.5F angioplasty balloon catheter was passed through the guiding catheter and slid into the LAD over a 0.014-in guide wire that was positioned in the proximal LAD. The balloon was then inflated for 90 minutes to produce myocardial infarction and then deflated to allow reperfusion of the infarcted myocardium. Regional flow was measured immediately before LAD occlusion, 80 minutes after LAD occlusion, and 3.5 hours after reperfusion (Fig 1⇓) by LV injection of ≈2 million sonicated microspheres19 (15±1 μm in diameter) labeled with one of several radionuclides (153Gd, 113Sn, 103Ru, 95Nb, or 46Sc; Du Pont). The animals were then allowed to recover.
All images were acquired after 48 h of reperfusion (Fig 1⇑). The dogs were again anesthetized and mechanically ventilated, and a pigtail catheter was placed in the LV under fluoroscopic guidance. A fourth set of microspheres was administered, and the animals were transported to the MR facility for scanning in a whole-body 1.5-T Signa scanner (General Electric). The animals were placed in the right antecubital position, and a flexible radiofrequency receiver coil was placed around the chest. Scout images were obtained to locate the long axis, from which four (n=6) or six (n=5) parallel short-axis slices were prescribed. ECG-gated mid-diastolic fast MR images were then acquired by use of a pulse sequence that provides a linear relation between image intensity and contrast concentration over a wide range and that has been described in detail elsewhere.20 21 Briefly, magnetization is driven to steady state before image acquisition using a train of dummy radiofrequency pulses, resulting in strongly T1-weighted images with little or no influence of T2 and/or T2*. Imaging parameters were image matrix 256×96, α=45°, TR=6.5 ms, TE=2.3 ms, field of view 32 cm, and slice thickness=10 mm. Voxel dimensions were 1.2×3.3×10 mm.
The time line for MR imaging is shown in Fig 2⇓. The effects of cardiac motion were minimized by acquiring only one fourth of each image (24 phase encodes) in each cardiac cycle. All image data were acquired during breath hold, which was achieved by turning off the respirator at end expiration. Typically, breath-hold duration was 32 cardiac cycles, during which each of the four (or six) base-apex locations were imaged twice (Fig 2⇓).
After baseline precontrast images were acquired, 0.1 mmol/kg of a nonionic clinically approved MR contrast agent (gadoteridol, ProHance, Squibb) was injected by hand as a bolus into a femoral vein. Images were then acquired beginning 10 seconds after contrast injection. After the 32 cardiac cycles required for image acquisition, the animal was ventilated for ≈30 seconds, after which the second cycle of image acquisition began (Fig 2⇑). This cycle was repeated for 15 minutes. If the 10 seconds to begin imaging and 10 seconds to turn the respirator on and off are accounted for, the effective temporal resolution between any two images of the same base-apex location was approximately 30 seconds.
Measurement of Risk, Infarct, and No-Reflow Regions
Immediately after MR imaging, 20 mL of the fluorescent dye thioflavin-S (2% solution) was injected into the LV via the pigtail catheter to define the extent of the no-reflow region.19 One minute after thioflavin injection, the hearts were arrested with potassium chloride and excised. The atria, epicardial fat, valvular tissue, and right ventricular free wall were trimmed away. The LV was then sectioned into five short-axis slices, each of which was then incubated in a 2% solution of TTC for 20 minutes at 37°C. Regions that failed to demonstrate brick-red staining were considered to represent infarcted myocardium.22 Each slice was then photographed under ultraviolet light (for thioflavin) and room light (for TTC). For each slice, shallow cuts were made to divide ≈50% of the LV circumference centered in the infarcted region into 8 to 12 pie-shaped sections. An additional pie-shaped section was identified on the side of the ventricle opposite the infarcted region for measurement of blood flow in a remote region. The exact locations of these cuts were recorded before the slices were photographed to relate thioflavin and TTC regions to the location from which tissue samples were taken for microsphere blood flow analysis.
After the slices were photographed, myocardial samples (0.1 to 0.5 g) were obtained for microsphere counting. For each slice, each of the pie-shaped sections was divided into five approximately equal regions from epicardium to endocardium, yielding a total of ≈250 myocardial samples from each heart. Each of the samples was then weighed and counted in a gamma emission well spectrometer (model 5986, Hewlett-Packard) along with the reference blood samples at appropriate energy windows. Regional myocardial blood flow (mL/min per 100 g) was then calculated by standard methods.19
Subsequently, clear acetate sheets were placed over projections of the photographs, and the infarcted regions (TTC negative) and no-reflow regions (absence of thioflavin) were outlined by hand. The risk regions were defined as areas in which myocardial blood flow (based on the microspheres) was reduced by at least 50% during the occlusion compared with remote (nonischemic) regions. With this definition, the location of each of the myocardial tissue samples within the risk region was transferred to the acetate sheets. This process was guided by knowledge of the locations of the cuts recorded before photography and by the fact that the epicardial and endocardial samples were of equal size. The areas of the risk, infarct, and no-reflow regions were then measured by planimetry of the acetate sheets and combined with the weight of the slices to calculate the size of these regions as a percentage of LV mass.
Regional Blood Flow
To compare blood flow within different myocardial regions, within each animal the myocardial samples were grouped into four regions: remote, at risk but not infarcted, infarcted but not no-reflow, and no-reflow. For each animal, blood flow in the remote region was defined as the transmural average of blood flow in normal myocardium. Blood flow in the “at risk but not infarcted” region was determined by averaging all samples within the risk region but outside the TTC-negative zones. Blood flow in the “infarcted but not no-reflow” region was determined by averaging all samples within the TTC-negative zones but outside the thioflavin-negative zones. Blood flow in the “no-reflow” regions was determined by averaging all samples within the thioflavin-negative zones. Samples that were partially located in more than one zone were excluded from the blood flow calculations.
Measurement of Hypoenhanced and Hyperenhanced Regions
The surface areas of hypoenhanced and hyperenhanced regions in the MR images were measured planimetrically by two independent observers blinded to the histological results. These regions were identified for each slice directly from the images on a Macintosh using the software package nih image. For each slice, the delineation of regions of hypoenhancement was improved by averaging four images acquired within the first 2 minutes after contrast arrival in normal myocardium. The delineation of hyperenhanced regions was facilitated by averaging eight images acquired 6 to 14 minutes after contrast. The spatial extents of hypoenhanced and hyperenhanced regions were expressed as percent LV.
Myocardial MR image time-intensity curves were generated within regions of interest using nih image. Regions of interest were placed within the hypoenhanced and hyperenhanced regions, as well as within the LV cavity and within normal myocardium (remote). Care was taken to define the regions of interest several pixels from the epicardial and endocardial surfaces to avoid partial-volume effects. The image time-intensity curves were normalized by two different methods. With the first normalization method, image intensities within the four regions (blood, remote, hypoenhanced, and hyperenhanced) were expressed as the percent increase in image intensity compared with baseline by use of the following equation:
With the second normalization method, the percent change in myocardial image intensity was divided by the percent change in blood image intensity at each time point. We term this second normalization method the MBSI.
The MBSI was used to test the hypothesis that the kinetics of contrast enhancement in the remote, hypoenhanced, and hyperenhanced regions were the same as those in the blood pool.
Repeated-measures ANOVA23 was used to test the hypotheses that microsphere blood flow varied with time in different regions and that image intensities varied with time in different regions. Differences between specific regions were isolated with Bonferroni-corrected t tests. The spatial extents of hypoenhanced and hyperenhanced regions were then compared with those of thioflavin-negative, risk, and TTC-negative regions by paired t tests. Differences were considered significant at the 5% level. Linear regression analyses were used to compare hypoenhanced regions with thioflavin-negative regions, hyperenhanced regions with risk regions, and hyperenhanced regions with TTC-negative regions. All results are expressed as mean±SEM.
The Table⇓ summarizes the hemodynamic results before, during, and after occlusion of the proximal LAD. The hemodynamic changes were typical of those observed during myocardial infarction and reperfusion.
Microspheric Blood Flow
Fig 3⇓ shows microspheric blood flow immediately before occlusion (baseline), 80 minutes after onset of occlusion, and after 3.5 and 48 hours of reperfusion in the four different myocardial regions: (1) remote, (2) at risk but not infarcted, (3) infarcted but not no-reflow, and (4) no-reflow. At baseline, blood flow averaged across all regions was 0.71 mL/min per gram, and differences between regions were not statistically significant. During the occlusion, a highly significant decrease in flow was observed in regions 2 through 4 compared with baseline (P<.001) and remote regions (P<.001), as would be expected, with virtually no flow in region 4 (the region that was thioflavin negative at 48 hours). After 3.5 hours of reperfusion, blood flow in all regions increased to values similar to those at baseline (P=NS for flow at 3.5 hours compared with baseline), demonstrating that the myocardium had been successfully reperfused. After 48 hours of reperfusion, blood flow in regions 2 (risk) and 3 (infarcted) was slightly lower than in remote regions (P=NS) but higher than in region 4 (no-reflow, P<.05). Blood flow in the no-reflow region was less than half that in remote regions (0.37±0.09 versus 0.88±0.10 mL/min per gram, respectively, P<.001).
Fig 4⇓ depicts typical MR images acquired in one dog before contrast administration and at various times after contrast. Fig 5A⇓ depicts average image time-intensity curves obtained in all 11 dogs in blood and in three myocardial regions. Fig 5B⇓ shows the same myocardial data as Fig 5A⇓ on an expanded scale. Before contrast administration, image intensity in blood and in all myocardial regions was homogeneous and low, consistent with the design of the imaging pulse sequence20 21 and our observations in humans.16 21 After contrast administration, three distinct myocardial enhancement patterns were apparent. The first pattern, observed in normal myocardial tissue (remote), involved a homogeneous increase in myocardial image intensity within 1 minute, followed by a decline in intensity for the remaining 14 minutes (P<.001). The second pattern, hypoenhancement, was observed in 8 of the 11 animals and involved the central subendocardial regions of the infarct zones. These regions of hypoenhancement were most apparent during the first 2 minutes after contrast administration. Image intensity within these regions increased slowly throughout the 15-minute imaging period (P<.001). The third pattern, hyperenhancement, was observed in 10 of the 11 animals. These hyperenhanced regions appeared similar to normal myocardium during the first 2 minutes after contrast but then became progressively better delineated about 5 minutes after contrast and remained so for the duration of the 15-minute imaging period. Image intensity in these regions did not change after 2 minutes postcontrast (P=NS).
Fig 6⇓ shows the MBSI for the same data as shown in Fig 5⇑. In normal myocardium, the MBSI reached a value of ≈0.2 within the first 2 minutes after contrast and remained constant thereafter. No statistically significant changes in the MBSI were observed after 2 minutes (P=NS), suggesting that contrast kinetics in normal myocardium were essentially the same as those of the blood pool. In hypoenhanced myocardium, the MBSI increased slowly throughout the imaging period (P<.001). In hyperenhanced myocardium, the MBSI also increased throughout the imaging period (P<.001). The increases in MBSI over time in hypoenhanced and hyperenhanced regions strongly suggest that contrast kinetics in these regions, unlike normal myocardium, are dissociated from those of the blood pool.
Relationship of Hypoenhancement to No-Reflow Regions
Fig 7⇓ compares typical MR images obtained in two different dogs within the first 3 minutes after contrast to the photographs of thioflavin deposition obtained postmortem. Qualitatively, a strong similarity was observed between the location of the hypoenhanced regions in the MR images and the absence of thioflavin deposition. Similar results were observed in all eight of the dogs in which hypoenhancement was present. In the remaining three dogs, no regions of hypoenhancement were detected. In each of these three animals, the infarct was very small (<2.5% of ventricle based on TTC), and the photographs of thioflavin deposition revealed very small (<3 mm) subendocardial no-reflow regions in two animals and none in the other animal.
Fig 8⇓ is a plot of the spatial extent of the hypoenhanced regions, expressed as a percentage of the total LV mass, against the extent of the no-reflow. Hypoenhanced regions were smaller than thioflavin-negative regions (0.98±0.26% versus 2.50±0.79%, P<.05; ranges, 0% to 2.3% and 0% to 7.9%, respectively). On average, the size of the hypoenhanced regions was 25% of the size of the thioflavin-negative regions (r=.75, slope=0.25).
To examine whether the smaller size of the hypoenhanced regions was due to the level of blood flow required for detection, we computed the percent LV with microsphere flows less than 0.1, 0.2, 0.3, and 0.4 mL/min per gram and compared them with the size of the hypoenhanced and thioflavin-negative regions. The results are shown in Fig 9⇓. The mean size of the hypoenhanced regions was similar to the mean size of myocardium receiving flow <0.1 mL/min per gram. Conversely, the mean size of thioflavin-negative regions was closer to the mean size of myocardium receiving flow <0.2 mL/min per gram. The results shown in Fig 9⇓ suggest that hypoenhanced regions are smaller than thioflavin-negative regions because the MR contrast agent molecule (gadoteridol) can be delivered in sufficient amounts to be detected at lower flows than the thioflavin-S molecule. Because gadoteridol and thioflavin-S have similar molecular weights (≈600 D for both), this difference may be due to a requirement for higher concentrations of thioflavin before the myocardium appears stained in ultraviolet photographs.
Relationship of Hyperenhancement to Infarct and Risk Regions
Fig 10⇓ compares typical MR images obtained in 2 different dogs ≈10 minutes after contrast administration with the photographs of the TTC-stained myocardium obtained postmortem. Qualitatively, the location of the hyperenhanced regions was always concordant with the location of the infarct. Similar results were observed in 10 of the 11 dogs. The 1 animal in which hyperenhancement was not observed had a very small infarct (2.6% of ventricle).
The sizes of hyperenhanced regions (8.4±2.2%; range, 0% to 24%) and risk regions (23.5±2.9%; range, 2.4% to 38%) were only modestly correlated and were approximately half as large (r=.64, slope=0.48, intercept=−3.0). The differences were statistically significant (P<.001).
The sizes of hyperenhanced regions were strongly correlated with TTC-negative regions (6.4±1.8%; range, 0.7% to 20%). Fig 11⇓ is a plot of the sizes of the hyperenhanced regions compared with the sizes of the TTC-negative regions. A straight line fit through these data had an r value of .93 and coefficients of 1.12 (slope) and 1.17 (intercept). However, the hyperenhanced regions were slightly larger than TTC-negative regions (12%), and this difference was statistically significant (P=.035).
This study demonstrates that there is a relation between MR contrast-enhancement patterns and the extent and type of myocardial injury. Comparison of contrast-enhanced MR images to histology showed that in hearts subjected to 90 minutes of LAD occlusion followed by 48 hours of reperfusion, hypoenhanced regions are related to the no-reflow phenomenon and that ≈90% of the myocardium within hyperenhanced regions is nonviable.
The initial slopes of the time-intensity curves for the remote, hypoenhanced, and hyperenhanced regions (Fig 5⇑) appeared to reflect regional variations in myocardial blood flow as measured by microspheres (Fig 3⇑). Flow at 48 hours (Fig 3⇑) was lowest in the thioflavin-negative regions, which also had the smallest initial slope in the time-intensity curves (Fig 5⇑). These observations are similar to those of Wilke et al,15 who studied pharmacologically vasodilated myocardium in which LAD flow was reduced to various levels via a hydraulic occluder in dogs, and to results reported by Manning et al,8 who compared normal myocardium with regions distal to angiographically occluded arteries in humans. In addition to regional flow, however, other factors such as contrast extravasation24 25 26 27 probably play an important role in the initial slope of these time-intensity curves. More importantly, however, since image intensity is derived from the proton signal rather than from the contrast agent itself, image intensity may depend on the number of water protons that are exposed to the agent and therefore on the compartment (intravascular, interstitial, or cellular) in which the agent is distributed. Recent studies by Wendland et al28 and our group29 suggest that under first-pass conditions, a limitation of myocardial water exchange, most likely between the intravascular and interstitial spaces, introduces a strong nonlinearity between myocardial contrast concentration and image intensity in T1-weighted images. Thus, although the initial slope of image intensity correlates with flow, it is unclear whether this correlation is due to flow itself, contrast extravasation, intercompartmental water exchange, or some combination of these factors.
Contrary to the results of the initial slopes of the image time-intensity curves, the interpretation of myocardial enhancement a few minutes after contrast administration may be less complex for several reasons. First, the influence of vascular-interstitial water exchange is probably less important after the first pass because contrast concentrations are much lower and change slowly over time.30 Under these conditions, contrast concentration appears to be linearly related to the change in 1/T1 in both blood31 and myocardium.32 33 34 Second, the pulse sequence used in this study was specifically designed to produce a linear relationship between image intensity and 1/T1,20 21 suggesting that our image intensity data obtained after 2 minutes postcontrast (ie, well after the first pass) are linearly related to contrast concentration. Third, equilibrium between intravascular and interstitial contrast concentrations most likely occurs much faster24 25 27 than the time required to clear the agent from the blood pool.30 Thus, in normal myocardium it is likely that after the initial transients that follow the bolus, contrast concentrations in the intravascular and interstitial spaces are essentially always equal to that of the blood, and consequently the ratio of myocardial to blood contrast concentrations would remain constant. This is precisely what was observed experimentally in normal tissue, as shown in Fig 6⇑ (MBSI is constant over time).
Infarct Size Versus Hyperenhancement
Our study is the first to compare infarct sizes with hyperenhancement after 2 days of reperfusion. Several investigators, however, have studied hyperenhancement of myocardial tissue in acutely reperfused infarcts.10 11 13 18 In agreement with our findings (Fig 10⇑), Saeed et al18 showed that in rats subjected to 2 hours of occlusion followed by 3±0.5 hours of reperfusion, the spatial extent of hyperenhanced regions in T1-weighted images strongly correlates with the spatial extent of TTC-negative zones. Schaeffer et al13 studied myocardial hyperenhancement in a canine model of reperfused myocardial infarction and found that the spatial extent of hyperenhancement was dependent on when the contrast agent was administered after reperfusion. When the contrast agent was administered early in the reperfusion period (5 minutes), hyperenhancement correlated with the size of the risk region, whereas when the contrast agent was administered later in the reperfusion period (90 minutes), hyperenhanced regions were much smaller despite similar infarct sizes.13
The finding that hyperenhanced regions closely correlate with TTC-negative regions and are only slightly (12%) larger is of potentially great importance because it strongly suggests that in infarcts reperfused for several days, as would be the case if the method of this paper was applied clinically, regions that become hyperenhanced several minutes after contrast administration are mostly nonviable. Consistent with this observation, we found, in another study, that in patients with reperfused infarcts examined within the first week, the subgroup of patients in whom the MBSI of hyperenhanced regions increased over time developed larger scars 6 months after infarct.35
Physiological Basis of Hyperenhancement
Schaeffer et al13 measured myocardial Gd-DTPA concentrations postmortem with radiolabeled Gd and found higher Gd concentrations in infarcted, reperfused canine myocardium than in normal myocardium within 1 hour after administration of contrast. This finding is consistent with the results of Geschwind et al,36 who measured Gd-DTPA concentrations using inductively coupled plasma mass spectroscopy postmortem in rats and found higher Gd concentrations in infarcted, reperfused myocardium. The results of these studies strongly suggest that hyperenhancement of infarcted, reperfused myocardium is at least in part due to higher Gd concentrations.13
At least two mechanisms could explain the increased contrast concentrations in the hyperenhanced regions. First, the volume of distribution for the contrast molecule within the voxel may increase. This may be secondary to interstitial edema and/or eventual disruption of the myocyte membrane after prolonged ischemia,37 which may allow the contrast molecule to enter the myocyte intracellular space. Second, the rate at which the contrast molecule washes in and out of the myocardial tissue may be slower in infarcted, reperfused regions. Our finding that image intensity in the hyperenhanced regions did not change over time more than 2 minutes after contrast (Fig 5B⇑) supports the hypothesis that washout of the contrast molecule is impaired.
The MBSI (Fig 6⇑) provides additional insight into these two potential mechanisms. We found that in normal myocardium, the MBSI remains constant over time, whereas the MBSI of the hyperenhanced regions increased significantly over time. If the volume of distribution for the contrast agent in hyperenhanced regions was increased but contrast agent concentrations in the interstitial space (and/or cell space) remained equal to that of blood, the MBSI would shift to a higher level than that of normal tissue but would remain constant over time. The observation that the MBSI in hyperenhanced regions increases over time therefore strongly suggests that the rapid communication between the interstitial space and the blood space observed in normal myocardium no longer exists in hyperenhanced regions.
No-Reflow Size Versus Hypoenhancement
Although zones of hypoenhancement within infarcted myocardium have been observed by other groups in the absence of reperfusion,14 15 17 to our knowledge the only study documenting such hypoenhanced regions in reperfused infarcts is our recent clinical study.16 Our finding that these regions of hypoenhancement show a high degree of correlation with the spatial location of the no-reflow zones by thioflavin (Fig 7⇑) provides strong evidence that hypoenhancement is related to the no-reflow phenomenon described by us19 and others.6 38 Consistent with previous observations, we found that myocardial blood flow in thioflavin-negative regions at the time of MR imaging (48 hours, Fig 3⇑) was much lower than in the remote regions (0.37±0.09 versus 0.88±0.10 mL/min per 100 grams, respectively, P<.001). This provides additional strong evidence that hypoenhanced regions represent infarcted regions with extremely low blood flow.
Physiological Basis of Hypoenhancement
We found that before contrast administration, image intensity in the hypoenhanced regions was not significantly lower than intensity in the other two myocardial regions of interest (hypo, 17.1±1.6; hyper, 16.7±1.3; remote, 19.8±1.5; P=NS). This suggests that hypoenhancement is not due to some baseline abnormality but rather is related to delayed contrast penetration in these regions. Delayed contrast penetration might be explained by widespread microvascular damage during the reperfusion period, a feature characteristic of no-reflow regions.19 This widespread microvascular damage decreases regional flow secondary to decreases in functional capillary density. In addition, the decrease in functional capillary density would result in a large increase in the time required for the contrast molecule to diffuse into the extravascular space (on a voxel scale) due to increased diffusion distances and the fact that diffusion time increases with the square of distance.
In a previous study that used the same canine model of reperfused infarction as the present study, we compared thioflavin-negative with thioflavin-positive regions by light and electron microscopy.19 Thioflavin-negative regions exhibited widespread microvascular damage characterized by microvascular obstruction due to neutrophil plugging, microvascular thrombosis, endothelial cell swelling, extravascular compression by tissue edema, or a combination of these factors. Thioflavin-positive areas within the risk region demonstrated moderate tissue injury, with a mixture of coagulation and contraction band necrosis.19 Contraction band necrosis within thioflavin-negative regions was widespread after 3.5 hours of reperfusion19 but was not prominent after shorter reperfusion times,19 38 suggesting that in addition to “true” (ie, immediate) no-reflow found at the time of reperfusion, a substantial portion of postischemic myocardium undergoes progressive vascular obstruction during reperfusion. This suggests that the presence and spatial extent of no-reflow regions observed with MR may depend on postinfarct reperfusion times.
Although the presence of hypoenhanced and hyperenhanced regions is related to regional myocardial blood flow, other factors, such as the rate of contrast clearance from the blood, the rate of diffusion of the contrast molecule out of the vascular space, and the functional capillary density, probably play important roles. Thus, the regions of hypoenhancement and hyperenhancement observed in this study appear to represent disruptions in the normal rapid diffusion of the contrast molecule in and out of the myocardial interstitium. These enhancement patterns might be better described as providing information about the delivery of nutrients to cardiac myocytes rather than perfusion through the microvasculature.
Selected Abbreviations and Acronyms
|LAD||=||left anterior descending coronary artery|
|MBSI||=||myocardium-to-blood signal intensity ratio|
This study was supported by the Frank T. McClure Fellowship in Cardiovascular Research (Dr Judd), NIH-NHLBI grant HL-17655 (Specialized Center of Research in Ischemic Heart Disease, Dr Becker), and NIH-NHLBI grant HL-45090 (Dr Zerhouni). The authors thank Anthony DiPaula for his work with the experimental preparation and data analysis, Kenneth Rent for his work with the MR imaging, and Ann Capriotti for the illustrations and figures.
- Received February 7, 1995.
- Revision received April 3, 1995.
- Accepted April 21, 1995.
- Copyright © 1995 by American Heart Association
Gruppo Italiano per lo Studio della Streptochinasi nell’Infarto Miocardico (GISSI). Effectiveness of intravenous thrombolytic treatment in acute myocardial infarction. Lancet. 1986;1:871-874.
Chesebro JH, Knatterud G, Roberts R, Borer J, Cohen LS, Dalen J, Dodge HT, Francis CK, Hillis D, Ludbrook P, Markis JE, Mueller H, Passamani ER, Powers ER, Rao AK, Robertson T, Ross A, Ryan TJ, Sobel BE, Wellerson J, Williams DO, Waret BL, Braunwald E. Thrombolysis in Myocardial Infarction (TIMI) Trial, phase I: a comparison between intravenous tissue plasminogen activator and intravenous streptokinase. Circulation. 1987;76:142-154.
Peshock RM, Malloy CR, Buja LM, Nunnally RL, Parkey RW, Willerson JT. Magnetic resonance imaging of acute myocardial infarction: gadolinium diethylenetriamine pentaacetic acid as a marker of reperfusion. Circulation. 1986;74:1434-1440.
Wolfe CL, Moseley ME, Wikstrom MG, Sievers RE, Wendland MR, Dupon JW, Finkbeiner WE, Lipton MJ, Parmley WW, Brasch RC. Assessment of myocardial salvage after ischemia and reperfusion using magnetic resonance imaging and spectroscopy. Circulation. 1989;80:969-982.
Lima JAC, Judd RM, Bazille A, Schulman SP, Atalar E, Zerhouni EA. Regional heterogeneity of human myocardial infarcts demonstrated by contrast-enhanced MRI: potential mechanisms. Circulation. 1995;92:1117-1125.
Ambrosio G, Weisman HF, Mannisi JA, Becker LC. Progressive impairment of regional myocardial perfusion after initial restoration of postischemic blood flow. Circulation. 1989;80:1846-1861.
Judd RM, Atalar E, Reeder SB, McVeigh ER, Zerhouni EA. A pulse sequence with increased dynamic range and linear signal intensity for myocardial perfusion studies. Proc Soc Magn Reson Med. 1993;2:625. Abstract.
Winer BJ. Statistical Principles in Experimental Design. New York, NY: McGraw Hill; 1971:351-359.
Canty JM Jr, Judd RM, Brody AS, Klocke FJ. First-pass entry of nonionic contrast agent into the myocardial extravascular space. Circulation. 1991;84:2071-2078.
Crone C, Levitt DG. Capillary permeability to small solutes. In: Renkin EM, Michel CC, Geiger SR, eds. Handbook of Physiology. Bethesda, Md: American Physiological Society; 1984;4:411-439.
Rumberger JA, Bell MR, Feiring AJ, Behrenbeck T, Marcus ML, Ritman EL. Measurement of myocardial perfusion using fast computed tomography. In: Marcus ML, Skorton DL, Schelbert HR, Wolf GL, eds. Cardiac Imaging. Philadelphia, Pa: WB Saunders Co; 1991:688-702.
Wu XW, Ewert DL, Liu YH, Ritman EL. In vivo relation of intramyocardial blood volume to myocardial perfusion: evidence supporting microvascular site for autoregulation. Circulation. 1992;85:730-737.
Wendland MF, Saeed M, Ku KY, Roberts TPL, Lauerma K, Derugin N, Varadarajan J, Watson AD, Higgins CB. Inversion recovery EPI of bolus transit in rat myocardium using intravascular and extravascular gadolinium-based contrast media: dose effects on peak signal enhancement. Magn Reson Med. 1994;32:319-329.
Wolf GL. Contrast agents for cardiac MRI. In: Marcus ML, Skorton DL, Schelbert HR, Wolf GL, eds. Cardiac Imaging. Philadelphia, Pa: WB Saunders Co; 1991:794-810.
Lima JAC, Judd RM, Lugo-Olivieri CH, Schulman S, Atalar E, Gerstenblith G, Zerhouni EA. Myocardial perfusion pattern by contrast enhanced MRI in the acute post-infarct period can predict myocardial necrosis and eventual extent of scar formation. Circulation. 1994;90(pt 2):I-410. Abstract.
Geschwind JF, Wendland MF, Saeed M, Lauerma K, Derugin N, Higgins CB. Evidence for loss of cell membrane integrity in reperfused myocardial injury using dual mechanisms of contrast-enhanced MR imaging. Proc Soc Magn Reson Med. 1994;2:888. Abstract.
Jennings RB, Murry CE, Steenbergen CJ, Reimer KA. Development of cell injury in sustained acute ischemia. Circulation. 1990;82(suppl II):II-2-II-12.
Kloner RA, Ganote CE, Jennings RB. The “no-reflow” phenomenon after temporary coronary occlusion in the dog. J Clin Invest. 1974;54:1496-1508.