“Hot Spot” Detection of Ischemic Myocardium In Vivo by Gamma Camera Imaging
Background—99mTc-HL91 is a new hypoxia imaging agent that demonstrates increased uptake and retention in globally hypoxic myocardium in vitro. The purpose of this study was to determine whether 99mTc-HL91 could detect regional ischemia in vivo by gamma camera imaging.
Methods and Results—Eight open-chest dogs with left circumflex (LCx) stenoses were studied. Injection of 5 mCi of 99mTc-HL91 and microspheres was followed by imaging over 4 hours. Heart slices were imaged, then stained with triphenyltetrazolium chloride (TTC), and tissues were well-counted. TTC staining demonstrated no injury. Mean LCx blood flow was 0.32±0.04 mL · min−1 · g−1, and mean left anterior descending coronary artery (LAD) flow was 0.96±0.02 mL · min−1 · g−1 (ratio, 0.33). “Hot spots” were detected in 8 of 8 experiments in vivo within 60 minutes and improved over 4 hours. Region of interest analysis of LCx/LAD activity ratios demonstrated significant increases within 30 minutes (final ratio, 3.0; P<0.05). LCx and LAD washout curves demonstrated significant differences within 15 minutes. Washout curves were biexponential over 1 hour, followed by linear retention from 1 to 4 hours. Four-hour fractional retention was 0.12 for LAD and 0.44 for LCx (P<0.01). Myocardial flow versus tracer uptake demonstrated 2 phases: phase 1 (flow, 0.05 to 0.7 mL · min−1 · g−1) had an inverse linear correlation (r=−0.80); phase 2, (flow, >0.7 mL · min−1 · g−1) had no correlation. Ischemic heart/liver ratios remained near 1.0 for 4 hours.
Conclusions—99mTc-HL91 positively identifies regional myocardial ischemia in a canine model using 99mTc imaging. Quantitative techniques allowed identification of ischemic myocardium within 15 minutes of tracer administration.
An imaging agent demonstrating increased uptake and long-term retention in ischemic myocardium would be an important addition to currently available agents. Synthesis of a new 99mTc-labeled imaging agent, 99mTc-HL91, has recently been reported.1 Initial studies demonstrated increased uptake in tumors with a hypoxic fraction and also low hepatic uptake.1 Using a reperfused, isolated, perfused-heart model, our laboratory subsequently described increased myocardial 99mTc-HL91 uptake and retention in ischemic and hypoxic hearts.2 We also found that there was no increased uptake in acutely and irreversibly injured myocardium early after reperfusion.3 The purpose of the present study in an intact canine model with a partial coronary stenosis at rest was to (1) determine whether these properties could be used to detect regional ischemic viable myocardium in vivo with gamma camera imaging, (2) describe uptake and clearance kinetics and their time course to determine the optimal imaging period, (3) describe microsphere-determined flow versus tracer activity relationships, (4) describe target-to-nontarget ratios over time (lung/liver/gallbladder), and (5) determine blood clearance over time.
Eight adult mongrel dogs with LCx coronary clamp (mean weight, 23 kg; range, 20 to 25 kg) were prepared as previously reported.4 99mTc-HL91 was prepared and radiochemical purity determined as reported.2
Gamma Camera Imaging
Our methods for acquiring gamma camera images have been reported.4 Images were qualitatively assessed without background subtraction.
Image quality was assessed quantitatively by use of target-to-background pixel-count ratios. Images were also quantified by calculation of pixel count ratios in ROIs in the control (LAD) and ischemic (LCx) myocardial zones. ROIs in the LCx region were defined by the size of the “hot spot.” In the LAD region, ROI size was equal to the entire septal wall. Average counts per pixel in these ROIs were then obtained and ratios calculated for each experiment. Images were background-subtracted and corrected for physical decay of 99mTc. Background subtraction was performed by placing an ROI on the thoracic wall in an area that avoided overlap with lung, heart, great vessels, and liver. The size of the ROI varied slightly between experiments because of anatomic variations (153.5±8.5 pixels, mean±SEM) but was constant within an experiment. The location of the ROI was unchanged across images within an experiment. Mean background counts (cpm/pixel) for each image were then subtracted from the organ counts within that image. LCx/lung and LCx/liver ratios were calculated for each experiment.
Figure 1⇓ illustrates the experimental protocol. Baseline hemodynamic measurements were recorded during a 20-minute period after surgical preparation and instrumentation. During the baseline period, microspheres were injected to determine regional myocardial blood flows as previously described.4 Then, the occluder on the LCx was partially tightened to provide a 90% reduction in baseline epicardial resting blood flow. Thirty minutes later, 5 mCi of 99mTc-HL91 was injected into the left femoral vein via the Teflon catheter. A microsphere blood flow determination was made simultaneously. The microspheres were labeled with either 113Sn or 103Ru. The order in which the microspheres were injected was randomized. Serial gamma camera images were acquired every 30 minutes between 30 minutes and 4 hours after injection. No splanchnic shielding was used.
To measure blood 99mTc activity over the 4-hour study period, 1.0-mL serial arterial blood samples were collected at 30-second intervals during the first 2 minutes and then at 2, 4, 6, 8, 10, 20, 30, 60, 120, 180, and 240 minutes after injection. The dogs were then euthanized.
After death, the LCx was completely occluded, and hearts were infused with Evans blue dye to delineate the area at risk. Hearts were immediately excised, and the precise location of the LCx was marked with a suture. The entire heart in each experiment was then placed in plastic wrap on a plastic board and imaged for 5 minutes on the face of the gamma camera. After the right ventricle and great vessels were removed, hearts were then sliced into four 1-cm-thick rings from apex to base. The 4 rings were placed on the face of the gamma camera and imaged together. In a subset of 4 experiments, the entire heart, liver, and gallbladder were sectioned and assayed in a dose calibrator for 99mTc activity immediately after the end of the experiment.
The heart slices were then weighed, and the myocardial rings were stained in a 2% solution of TTC for 15 minutes at 37°C. Any unstained areas were measured as a percentage of the total left ventricle, and infarct size was quantified by computerized digital planimetry of slice photographs. Forty-eight tissue samples, 24 samples representing the normally perfused LAD zone and 24 representing the occluded LCx zone, were assayed in the gamma well counter for 99mTc and microsphere activity. These samples were taken from the middle 2 slices. The apex and base slices tended not to have an ischemic zone. Inclusion of these slices would have disproportionately increased the number of normal pieces.
Ten 1-mL samples were pipetted from different, randomly chosen locations within the stirred TTC solution after removal of the heart slices. These samples were subsequently assayed in a gamma well counter for 99mTc activity. Count data were then background-subtracted and decay-corrected, and counts were averaged for the 10 samples. The average value was then multiplied to equal the final volume of the TTC solution. Activity calculated in this manner was compared with resident heart activity and was found not to exceed 1% of resident heart activity in any experiment.
Within 12 hours of collection, the 99mTc serial blood samples and the myocardial tissue samples were counted for 5 minutes each in a gamma well counter as previously described,4 and myocardial blood flow was calculated.4
Data Analysis and Statistical Methods
For calculation of 99mTc-HL91 blood clearance, the first blood sample (t=0) in each experiment was discarded, because the counts in this sample were lower than that of the 30-second sample as a result of inadequate time for complete mixing of 99mTc-HL91 in the blood pool. Then, background-subtracted and decay-corrected serial blood sample counts were normalized as a percentage of the counts at 30 seconds (peak activity), and the individual activity versus time curves were fitted by a nonlinear regression analysis (Tablecurve 2D for Windows, Jandel Scientific).
The first minute of myocardial clearance data from images was omitted from further analysis because of the potential for added activity from the blood pool. The count data were then normalized to the percent activity at 2 minutes. The clearance data from both zones (LCx and LAD) for each experiment were fitted by nonlinear and linear regression techniques as described above to compare parameters of myocardial clearance kinetics.
Fractional myocardial clearance from the normal and stenosed zones was defined as the difference between the initial and final counts divided by the initial counts. Clearance was calculated beginning 2 minutes after 99mTc-HL91 injection from the background and decay-corrected, normalized data.
Percent injected dose calculations were made for heart, liver, and gallbladder in 4 experiments by dividing the decay-corrected assayed activity in these organs by the initial injected activity and expressing this as a percentage.
The myocardial blood flow versus 99mTc-HL91 activity data for each individual experiment were normalized to a flow of 1.0 mL · min−1 · g−1, then normalized to the tissue piece with the highest activity, which was arbitrarily assigned as 100%. Data points for flows <0.05 mL · min−1 · g−1 were not included in the regression analysis, although they are demonstrated in Figure 2⇓. Previous studies using other related agents have demonstrated markedly reduced tracer uptake at these very low flows, probably due to poor tracer delivery or irreversible myocardial injury. The remaining data were then divided into 2 phases. The first phase ranged from a flow of 0.05 mL · min−1 · g−1 to a flow of <0.7 mL · min−1 · g−1. The cutoff value for flow of 0.7 mL/min was computer-derived. The flat second phase was for flows ≥0.7 mL · min−1 · g−1. These individual phases were fitted by linear regression analysis. The individual experimental results were then combined to demonstrate mean regression lines.
All results were expressed as mean±SEM. The significance of mean differences among groups was assessed with a one-way repeated-measures ANOVA. Post hoc comparisons were made by use of t tests with correction for multiple comparisons made by the Bonferroni procedure (Crunch Software Corp). Temporal comparisons were made by a paired t test. Values of P<0.05 were considered significant. The goodness of fit for linear and nonlinear regression analyses was calculated with the fit standard error statistic. Pearson r correlation coefficients were also reported.
All experimental animals were handled in accordance with the guiding principles of the American Physiological Society and by the Institutional Animal Care and Use Committee of the University of Oklahoma Health Sciences Center.
Table 1⇓ lists the hemodynamic data for the 8 experiments. After the stenosis was applied, relative LCx epicardial blood flow fell markedly, from 100% to 10% (P<0.05), and mean arterial pressure fell slightly, from 114.2±3.0 to 100.7±3.3 mm Hg (P<0.05). During the stenosis period, mean heart rate fell slightly, from 102.1±9.2 to 80.9±8.8 bpm. Mean left atrial pressure and cardiac output did not change significantly after stenosis.
Table 2⇓ lists the individual and mean microsphere-determined blood flows for the 8 experiments. Before occlusion, the mean LCx flow was 0.94±0.03 mL · min−1 · g−1, the mean LAD flow was 1.00±0.01 mL · min−1 · g−1, and the mean ratio was 0.93 (LCx/LAD zone). After stenosis, the mean LCx flow was 0.33±0.05 mL · min−1 · g−1, the mean LAD flow was 0.96±0.02 mL · min−1 · g−1, and the mean ratio was 0.33 (LCx/LAD zone).
TTC staining demonstrated no myocardial injury. Furthermore, after the myocardial tissue was removed from the staining solution, the solution was found to have no significant activity above background when counted in a well counter. This analysis indicates that TTC did not extract 99mTc-HL91 from myocardial tissue in this study.
99mTc-HL91 Myocardial Uptake Versus Flow
Figure 2⇓ demonstrates the myocardial 99mTc-HL91 uptake versus microsphere-determined flow curves for the 8 individual experiments. Each curve was divided into 2 phases and analyzed (Table 3⇓). Phase 1 consisted of data points between flows >0.05 and <0.7 mL · min−1 · g−1. This phase had a linear regression equation of y=116.8−125.2x (r=0.80). Phase 2 consisted of data points for flows ≥0.7 mL · min−1 · g−1. This phase had a linear regression equation of y=24.3+0.01x (r=0.06). Because this phase was flat, the linear correlation was near zero, as expected. However, the goodness of fit was excellent (fit standard error, 1.5). The mean regression lines for the 2 phases are demonstrated in Figure 3⇓.
99mTc-HL91 In Vivo Gamma Camera Images
Figure 4⇓ demonstrates representative 99mTc-HL91 in vivo gamma camera images. Lung activity was not prominent. Liver and gallbladder activities were present but did not significantly affect myocardial image quality. Images demonstrated increased 99mTc-HL91 uptake in the stenosis zone (LCx) at 60 minutes after tracer administration, although there was also significant uptake in the nonstenosis zone (LAD). However, by 120 minutes after tracer administration and thereafter, there was progressively less activity in the normal zone (LAD) compared with the stenosis zone (LCx). Qualitative analysis of in vivo images demonstrated hot spots in the stenosis zone (LCx) in all 8 experiments.
Quantitative Analysis of In Vivo Myocardial Gamma Camera Imaging
Figure 5⇑ demonstrates the in vivo myocardial washout curve from scans. The 240-minute fractional myocardial 99mTc-HL91 retention was 0.44±0.06 for the stenosis zone (LCx) and 0.12±0.01 for the normal zone (LAD) (P<0.001). The results of nonlinear and linear regression fits are listed in Table 4⇓. The stenosis zone (LCx) washout curve was biexponential during the first hour of clearance (r2=0.99), followed by a linear retention from 1 to 4 hours (fit standard error, 0.71). The first exponential phase was rapid, with t1/2=2.4±0.8 minutes. The second exponential was slow, with t1/2=161.6±73.6 minutes. The late linear retention appeared to have a slightly positive slope, but this was not significant. The control zone (LAD) washout curve was biexponential during the first hour of clearance (r2=0.99), followed by a linear retention from 1 to 4 hours (fit standard error, 3.11). The first exponential phase for the normal zone (LAD) was rapid and not significantly different from the stenosis zone (LCx), with t1/2=2.2 minutes (P=NS versus LCx). The second exponential phase was faster than the stenosed LCx, with t1/2=32.8 minutes (P<0.05 versus LCx). The late linear retention had a slope of 0.01±0.05 (P=NS versus LCx).
Figure 6⇑ demonstrates the mean hot spot–to–normal myocardial (LCx/LAD), LCx/lung, and LCx/liver activity ratios over 240 minutes. The LCx/LAD myocardial activity ratio progressively increased over time (P<0.05). The final LCx/LAD myocardial activity ratio was 3.0±0.4 at 4 hours. The LCx/lung and LCx/liver activity ratios did not change significantly over time. The LCx/lung activity ratio was 3.4±0.7 at the end of the 4-hour experiment. The LCx/liver activity ratio was 1.0±0.2 at the end of the 4-hour experiment.
Ex Vivo Gamma Camera Imaging
Figure 7⇑ demonstrates ex vivo images of myocardial slices from a representative dog. Qualitative analysis of these images demonstrated increased 99mTc-HL91 activity in the stenosed LCx zone in all 8 experiments.
Quantitative analysis of the ex vivo gamma camera images is demonstrated in Figure 8⇑. For the whole heart, the mean LCx/LAD activity ratio was 2.8±0.2. The mean LCx/LAD activity ratio for the individual slices ranged from 2.2±0.2 for the base to 3.9±0.7 for slice 2.
The mean percent injected 99mTc-HL91 dose was 1.1±0.1% in the heart, 5.8±0.7% in the liver, and 0.6±0.1% in the gallbladder 240 minutes after tracer administration. After completion of the protocol scans, additional scans of the urinary bladder indicated that this is a major route of excretion for this agent.
99mTc-HL91 Blood Clearance
Figure 9⇑ demonstrates the 99mTc-HL91 blood activity clearance curve. Clearance was triexponential (r2=0.99). The best-fit equation was y=a exp(−x/b)+c exp(−x/d)+e exp(−x/f), where a=444.1, b=0.5, c=43.0, d=9.7, e=9.2, and f=609.5.
Nitroimidazoles were initially used as markers of tissue hypoxia in tumor cells.5 Initial cardiac studies in cultured myocytes, perfused hearts, and dogs detected hypoxic myocardial tissue by use of 18F-labeled misonidazole and positron imaging.6 7 This technology was limited by the need for a cyclotron and the expense of a positron camera. Studies using an 123I-labeled analogue were described, thus allowing gamma camera imaging.8 However, high liver uptake and the relative unavailability of 123I limited the usefulness of this agent. More recently, a 99mTc-labeled nitroimidazole has been reported.9 10 11 12 13 This agent had the advantages of using a widely available molybdenum-generated radiolabel that could be detected by standard gamma cameras. However, high hepatic activity on canine images was described.
More recently, several new nitroimidazole-containing ligands have been synthesized for labeling with technetium. The 99mTc complex of one of these compounds, 99mTc-HL91M, demonstrated improved hypoxic/normal count ratios compared with other compounds. Furthermore, when the 2-nitroimidazole moiety was removed from the core ligand of HL91M, the resulting technetium-labeled compound (99mTc-HL91) demonstrated significantly improved hypoxic/normal myocardial uptake and reduced liver activity compared with other compounds.1 Our laboratory used a perfused-heart model to demonstrate increased 99mTc-HL91 uptake and retention in ischemic and hypoxic myocardium.2 Both models had increased uptake and retention compared with controls. Similar results were noted in Krebs-Henseleit–perfused and red blood cell–perfused hearts. Subsequent studies in our laboratory demonstrated no increased 99mTc-HL91 uptake in reperfused nonviable myocardium.3 The purpose of the present study was to determine whether increased uptake and retention of 99mTc-HL91 could be demonstrated with gamma camera imaging in an intact canine model. A model of coronary stenosis–induced regional ischemia was used to compare ischemic with normal myocardium.
In the present study, epicardial LCx blood flow fell significantly, as expected, with the placement of the coronary stenosis. This resulted in a slight fall in mean arterial blood pressure but no significant change in heart rate, left atrial pressure, or cardiac output. During the subsequent 4 hours, heart rate fell slightly, but epicardial LCx blood flow, mean arterial blood pressure, left atrial pressure, and cardiac output did not change significantly. The fall in heart rate is probably due to compensatory adaptation to a severe LCx flow reduction. Thus, no hemodynamic changes could account for the results reported below. It should be noted that although myocardial blood flow was reduced and ischemia was assumed to be present, there was no direct measurement of tissue hypoxia.
Blood Flow as Determined by Microspheres
Mean microsphere myocardial blood flows (mL · min−1 · g−1) before the stenosis were 0.94 for the LCx and 1.00 for the LAD, indicating no differences at baseline. After the stenosis was stable, mean microsphere blood flow for the LCx zone was 0.33 versus 0.96 mL · min−1 · g−1 for the LAD zone (ratio, 0.33), consistent with a significant resting flow reduction in the LCx zone, whereas normal flow was maintained in the LAD zone.
TTC demonstrated no significant myocardial injury. Furthermore, after the myocardial tissue was removed from the TTC solution, samples of the solution were found to have no significant activity above background in a well counter. This would indicate that 99mTc-HL91 is not leached significantly from myocardial tissue by the TTC salts, as has been previously described with sestamibi and thallium.14
In Vivo Gamma Camera Imaging
In the present study, qualitative assessment of in vivo gamma camera images demonstrated increased LCx zone activity compared with LAD zone activity at 60 minutes. However, there was still significant LAD zone activity at 60 minutes. By 90 to 120 minutes, differential 99mTc-HL91 clearance had significantly decreased LAD zone activity; consequently, the LCx increased activity was easily discernable as a hot spot. The resolution of this hot spot continued to improve over the remaining 4-hour imaging period. All 8 experiments qualitatively demonstrated increased LCx zone activity at 1, 2, 3, and 4 hours. Hot spot identification was possible without splanchnic (ie, hepatic) shielding, both qualitatively and quantitatively. Thus, 99mTc-HL91 can identify acutely ischemic viable myocardium on planar scans.
Quantitative analysis demonstrated differences between the LCx and LAD zones in 2 ways: First, the LCx/LAD activity ratio determined by ROI analysis of the serial images progressively increased over 4 hours. This increasing ratio was statistically significant as early as 30 minutes after tracer administration. By 2 hours after tracer administration, the LCx/LAD activity ratio was ≈90% of the 4-hour value. The final 4-hour in vivo LCx/LAD zone activity ratio was 3.0. Second, LCx and LAD zone time-activity curves determined by ROI analysis of serial images demonstrated significantly greater retention in the LCx zone than in the LAD zone. This differential retention was statistically significant 15 minutes after tracer administration. At the end of the 4-hour period, LCx zone fractional retention was 0.44 versus 0.12 for the LAD zone (P<0.001). Therefore, use of quantitative methods allows early detection of ischemic territories.
The results of the present study are consistent with previously published perfused-heart results from our laboratory.2 The present canine study demonstrated ischemic zone 1-hour fractional retention of 0.44 and normal zone fractional retention of 0.12. The corresponding perfused-heart 1-hour fractional retention was 0.31 for the ischemic zone and 0.13 for the normal zone. It should be noted that the 4-hour LCx/LAD activity ratio was 3.0 in the present study, compared with a 1-hour normal/ischemic ratio of 13.6 in the perfused-heart studies.2 This difference is probably due to uniformly severe ischemia in the globally perfused hearts compared with a more heterogeneous pattern of ischemia in an intact dog heart with a regional stenosis and with the presence of recirculating blood activity in the intact-animal study.
In the present study, the normal and ischemic zone myocardial 99mTc-HL91 clearance curves were biexponential over the first hour of clearance. This agrees with our previous results in isolated perfused hearts.2 In the present study, however, we observed 99mTc-HL91 myocardial clearance for an additional 3 hours after tracer administration. Clearance during this time demonstrated linear retention with a minimal positive slope for the ischemic zones and a minimal negative slope for the normal zones (P=NS). These results indicate a late myocardial 99mTc-HL91 component that is relatively tightly sequestered. Biexponential clearance implies an early component of unsequestered fractions related to flow and a later component related to a slower process perhaps due to reversible binding or enzymatic reconstitution. Linear retention very late implies a tightly sequestered fraction. Although the LCx late retention phase appeared to have a slightly positive slope, this was not significant. A slightly positive slope could be explained by delayed uptake due to residual blood activity.
Ex Vivo Gamma Camera Imaging
In the present study, images of 8 of 8 ex vivo hearts and heart slices demonstrated increased 99mTc-HL91 activity in the LCx zone when assessed qualitatively and quantitatively. After 4 hours, the LCx/LAD ratios ranged from 2.2 at the base to 3.9 in the second slice. These values compare well in magnitude with the LCx/LAD activity ratio of 3.0 obtained at 4 hours in vivo with gamma camera imaging. This indicates that other organ and blood activities did not contribute significantly to values obtained from in vivo scans of the heart.
Regional Myocardial Blood Flow Versus 99mTc-HL91 Tissue Activity
Analysis of the 8 individual graphs of regional myocardial blood flow versus 99mTc-HL91 myocardial activity demonstrated 2 phases. During the first phase, decreases in flow from 0.7 to 0.05 mL · min−1 · g−1 resulted in marked increases in tracer activity. This phase was described by an ascending straight line with r=0.80. Thus, within this phase, increasing severity of ischemia led to increasing 99mTc-HL91 uptake in a direct negative linear relationship. The second phase consisted of flat tracer activity for flows >0.7 mL · min−1 · g−1. This phase was described by a flat line with an r value near zero, as expected (r=0.06). The goodness of fit statistic (fit standard error, 0.71) was excellent. Thus, it appears that as blood flow is decreased below normal resting values (<1 mL · min−1 · g−1), there is no increased uptake of 99mTc-HL91. At 70% of normal resting flows, 99mTc-HL91 begins to accumulate in direct linear proportion to the severity of flow reduction.
99mTc-HL91 uptake was markedly reduced at flows <0.05 mL · min−1 · g−1. Previous studies using 18F-misonidazole7 and BMS18132111 also demonstrated reduced tracer uptake at very low flows. This reduced uptake could be a result of poor delivery of the tracer due to very low flow. Alternatively, reduced uptake at very low flows could be due to irreversible myocardial injury. Perfused-heart studies in our laboratory have demonstrated no increased uptake of 99mTc-HL91 in irreversibly injured reperfused myocardium. Although the present study demonstrated no gross injury by TTC staining, microscopic islands of injury cannot be excluded.
It should be noted that 99mTc-HL91 tissue activity at 4 hours was compared with microsphere blood flow at the time of tracer injection. Epicardial blood flow was kept constant during this period.
Blood, Lung, Liver, and Gallbladder Activity
In the present study, 99mTc-HL91 blood activity levels fell rapidly. Clearance was triexponential, and the fit was excellent (fit standard error, 0.45). Thus, background activity from blood pools probably did not affect image quality beyond a few minutes.
The hot spot/lung activity ratio remained relatively constant throughout the 4-hour experiment, ending at 3.4 after 4 hours. Thus, lung activity was essentially background and did not degrade myocardial image quality. Gallbladder activity was observed on the images but did not significantly affect myocardial image quality.
The hot spot/liver activity ratio remained relatively constant throughout the 4-hour experiment, ending at 1.0 after 4 hours. Unshielded images were thought to be of excellent quality. Thus, this agent is the first to demonstrate excellent ischemic/normal uptake and retention ratios while maintaining reduced liver uptake compared with previously described agents.
Nitroimidazole Trapping Mechanism
In normal tissue, it was thought that previously described nitroimidazoles enter cells by way of their lipophilicity and exit by back-diffusion without further reaction. When a nitroimidazole undergoes a single-electron reduction of the nitro group, a free radical is formed. Free radicals are unstable and highly reactive species that are unlikely to remain in this state for long. Nitroimidazoles may undergo 1 reduction reaction and subsequently gain an electron, restoring the original molecular configuration and allowing diffusion out of the cell. In this situation, retention is not prolonged in normal tissue. In hypoxic tissue, nitro free radicals are less likely to gain an electron, and cellular retention is prolonged. 99mTc-HL91 is actually produced by removal of the 2-nitro- imidazole moiety from the parent nitroimidazole. Thus, in hypoxic tissue, nitroimidazoles may actually undergo a sequence of reduction reactions with intracellular enzymes to form more reactive products that are able to bind to cellular components. Although this sequence of reactions is not completely understood, further biochemical modifications to the original molecule are possible through additional reduction reactions that may result in binding to cellular elements. The net effect of these structural alterations is to reduce the membrane permeability of the compound so that cellular retention is prolonged.
Comparison of 99mTc-HL91 With Other Agents
Initial cardiac studies in vitro and in vivo detected hypoxic myocardial tissue by use of 18F-labeled misonidazole and positron imaging.6 7 Shelton and associates6 studied 18F-misonidazole in a canine model of coronary occlusion. They reported a 23% retention in the ischemic zone and a 2% retention in the normal zone. Martin and associates7 also studied 18F-misonidazole in a canine model of coronary occlusion. They reported an ischemic/normal zone activity ratio of 3.0 after 4 hours. This is the same ratio as the hot spot/LAD myocardial activity ratio determined from gamma camera images at 4 hours in the present study.
One other 99mTc-labeled hypoxia-avid imaging agent has been studied extensively to date. This agent is BMS181321.9 In perfused-heart experiments, serial increases in BMS181321 myocardial uptake have correlated with both serial decreases in tissue oxygen and serial decreases in blood flow.10 12 Stone and associates13 studied this agent in a swine model of ischemia with extracorporeal circulation. They demonstrated increased tracer uptake and retention in ischemic myocardium. The correlation of flow to tracer uptake was r=0.68 (compared with r=0.80 in the present study with 99mTc-HL91). They reported an ischemic/normal myocardial BMS181321 activity ratio of 1.7 (compared with 3.0 for 99mTc-HL91 in the present study), a heart/liver ratio of 0.58 (1.0 in the present study), and a heart/lung ratio of 3.1 (3.4 in the present study). Shi and associates11 also studied BMS181321 in a canine model of coronary stenosis with pacing. There was an inverse correlation of myocardial blood flow to tracer uptake, with r=0.67 with a second-order polynomial fit. The ischemic/normal BMS181321 activity ratio was 1.61 at 60 minutes. The hepatic/ischemic myocardium BMS181321 ratio was 4.2 at 60 minutes. Although BMS181321-increased ischemic zone uptake was found in all ex vivo and in vivo studies, the authors cautioned that “an unfavorable heart to liver ratio was observed with in vivo planar imaging which may limit its use in clinical myocardial imaging.” Preliminary reports have described BMS194796, a modification of BMS181321, which may have reduced hepatic tracer uptake and faster ischemic tissue tracer uptake.15
The present study using 99mTc-HL91 demonstrates the ability to positively identify regionally ischemic myocardium by use of 99mTc imaging with standard gamma camera equipment. By qualitative analysis, identification of regional ischemia is possible within 60 minutes of tracer administration. Image quality subsequently improves from 60 to 240 minutes. By use of quantitative analysis of LCx/LAD activity ratios and LCx and LAD zone washout rates, identification of regional ischemia may be possible as early as 15 minutes after tracer administration.
An assessment of the severity of myocardial ischemia may be theoretically possible with 99mTc-HL91, because there is a progressive linear increase in activity for flows between 0.7 and 0.05 mL · min−1 · g−1. It should be noted that this study used an open-chest model. Clinically, it may be more difficult to quantify uptake and clearance. Furthermore, clinically, one may not have a severe ischemia reference point, although normal tissue uptake should provide a baseline for comparison. Possible clinical situations in which the use of a hot spot ischemia–detecting agent could be useful include (1) the detection of “hibernating” myocardium before cardiac transplantation or coronary bypass surgery; (2) the detection of salvageable myocardium before reperfusion therapy during acute myocardial infarction. This potential application would be feasible only if an assessment could be made rapidly by quantitative techniques; and (3) the detection of exercise-induced transient myocardial ischemia in the detection of chronic coronary artery disease. This potential application may be problematic, given the short period of ischemia.
1. This study demonstrates the ability of 99mTc-HL91 to detect regional myocardial ischemia in vivo in an intact canine model with a coronary stenosis by use of gamma camera imaging. Hot spot detection is possible qualitatively within 60 minutes and improves until 4 hours after tracer administration. Quantitative ROI analysis of in vivo LCx/LAD activity ratios demonstrated significant increases within 30 minutes after tracer administration, with a final value of 3.0 at 4 hours. Quantitative ROI analysis of LCx and LAD zone washout curves demonstrated significant differences within 15 minutes of tracer administration.
2. 99mTc-HL91 normal and ischemic zone washout curves were biexponential over the first hour of clearance, followed by a flat linear retention from 1 to 4 hours. The 4-hour fractional retention was 0.12 for the normal zone and 0.44 for the ischemic zone.
3. Myocardial blood flow versus 99mTc-HL91 retention curves demonstrated 2 phases. The first phase was gradually increasing tracer activity for flows between 0.7 and 0.05 mL · min−1 · g−1. This phase was described by an ascending straight line with r=0.80. The second phase demonstrated no increase in tracer activity for flows ≥0.7 mL · min−1 · g−1, indicating no flow dependence at normal and high flows.
4. Ischemic myocardial/liver activity ratios remained near 1.0 for 4 hours after tracer administration.
5. Blood clearance was relatively rapid and did not interfere with early myocardial imaging.
Selected Abbreviations and Acronyms
|LAD||=||left anterior descending coronary artery|
|ROI||=||region of interest|
- Received February 4, 1997.
- Revision received January 15, 1998.
- Accepted January 23, 1998.
- Copyright © 1998 by American Heart Association
Archer CM, Edwards B, Kelly JD, King AC, Burke JF, Riley ALM. Technetium labelled agents for imaging tissue hypoxia in vivo. In: Nicolini M, Bandoli G, Mazzi U, eds. Technetium and Rhenium in Chemistry and Nuclear Medicine. Padova, Italy: SGE Ditoriali Publishers; 1995:535–539.
Okada RD, Nguyen KM, Edwards B, Archer CM, Kelly JD, Johnson G III. HL91-technetium-99m: effects of low flow and hypoxia on a new ischemia-avid myocardial imaging agent. Circulation. 1997;95:1892–1899.
Johnson G III, Nguyen KN, Edwards B, Archer CM, Okada RD. HL91-technetium-99m: uptake and retention of a new hypoxia avid imaging agent can differentiate ischemic viable from ischemic nonviable myocardium. Circulation. 1995;92(suppl I):I-788. Abstract.
Johnson G, Glover DL, Herbert CB, Okada RD. Early myocardial clearance kinetics of Tc-99m-Teboroxime differentiate normal and flow restricted canine myocardium at rest. J Nucl Med. 1993;34:630–636.
Shelton ME, Dence CS, Hwang D-R, Herrero P, Welch MJ, Bergmann SR. In vivo delineation of myocardial hypoxia during coronary occlusion using fluorine-18 fluoromisonidazole and positron emission tomography: a potential approach for identification of jeopardized myocardium. J Am Coll Cardiol. 1990;16:477–485.
Martin GV, Caldwell JH, Rasey JS, Grunbaum Z, Cerqueira M, Krohn KA. Enhanced binding of the hypoxic cell marker [3H] fluoromisonidazole in ischemic myocardium. J Nucl Med. 1989;30:194–201.
Martin GV, Biskupiak JE, Caldwell JH, Rasey JS, Krohn KA. Characterization of iodovinylmisonidazole as a marker for myocardial hypoxia. J Nucl Med. 1993;34:918–924.
Rumsey WL, Patel B, Linder K. Effects of graded hypoxia on retention of technetium-99m-nitroheterocycle in perfused rat heart. J Nucl Med. 1995;36:632–636.
Shi CQ, Sinusas AT, Diane DP, Singer MJ, Young LH, Heller EN, Rinker BP, Wackers FJT, Zaret BL. Technetium-99m nitroimidazole (BMS 181321): a positive imaging agent for detecting myocardial ischemia. J Nucl Med. 1995;36:1078–1086.
Stone CK, Mulnix T, Nickles RJ, Renstrom B, Nellis SH, Liedtke J, Nunn AD, Kuczynski BL, Rumsey WL. Myocardial kinetics of a putative hypoxic tissue marker, 99mTc-labeled nitroimidazole (BMS 181321), after regional ischemia and reperfusion. Circulation. 1995;92:1246–1253.
Reinhardt CP, Weinstein H, Wironen J, Leppo J. Effect of triphenyl tetrazolium chloride staining on the distribution of radio labeled pharmaceuticals. J Nucl Med. 1993;34:1722–1727.
Rumsey WL, Patel B, Kuczynski B, Hood C, Linder E, Nunn A, Strauss HW. Comparison of two novel technetium agents for imaging ischemic myocardium. Circulation. 1995;92(suppl I):I-181.