Myocardial Uptake of 7′-(Z)-[¹²³I]Iodorotenone During Vasodilator Stress in Dogs With Critical Coronary Stenoses

Abstract

Background

There is a well-recognized need for a new generation of single photon emission computed tomography (SPECT) perfusion tracers with improved myocardial extraction over a wide flow range. Radiotracers that target complex I of the mitochondrial electron transport chain have been proposed as a new class of myocardial perfusion imaging agents. 7-(Z)-[¹²⁵I]iodorotenone (¹²⁵I-ZIROT) has demonstrated superior myocardial extraction and retention characteristics in rats and in isolated perfused rabbit hearts. We sought to fully characterize the biodistribution and myocardial extraction versus flow relationship of ¹²³I-ZIROT in an intact large-animal model.

Methods and Results

The ¹²³I-ZIROT was administered during adenosine A_2A agonist-induced hyperemia in 5 anesthetized dogs with critical left anterior descending (LAD) stenoses. When left circumflex (LCx) flow was maximal, ¹²³I-ZIROT and microspheres were coinjected and the dogs were euthanized 5 minutes later. ¹²³I-ZIROT biodistribution was evaluated in 2 additional dogs by in vivo planar imaging. At ¹²³I-ZIROT injection, transmural LAD flow was unchanged from baseline (mean±SEM, 0.90±0.22 versus 0.87±0.11 mL/[min · g]; P=0.92), whereas LCx zone flow increased significantly (mean±SEM, 3.25±0.51 versus 1.00±0.17 mL/[min · g]; P<0.05). Myocardial ¹²³I-ZIROT extraction tracked regional myocardial flow better than either thallium-201 or ^99mTc-sestamibi from previous studies using a similar model. Furthermore, the ¹²³I-ZIROT LAD/LCx activity ratios by ex vivo imaging or well counting (mean±SEM, 0.42±0.08 and 0.45±0.1, respectively) only slightly underestimated the LAD/LCx microsphere flow ratio (0.32±0.09).

Conclusions

The ability of ¹²³I-ZIROT to more linearly track blood flow over a wide range makes it a promising new SPECT myocardial perfusion imaging agent with potential for improved coronary artery disease detection and better quantitative estimation of the severity of flow impairment.

There is a well-recognized need for a new generation of myocardial perfusion imaging (MPI) agents with improved myocardial extraction over a wide flow range.^1–4 The “roll-off” phenomenon that occurs for most of the clinically used perfusion tracers results in underestimation of perfusion abnormalities, which can be a limitation in the detection of mild-to-moderate stenosis.^5,6 In addition, more linear extraction of the tracer over a wide range of flow is a requirement for quantitative estimation of absolute blood flow.

It has long been recognized that cationic MPI agents, including the current clinical tracers ^99mTc-sestamibi and ^99mTc-tetrofosmin, accumulate in cardiac cells based on the mitochondrial membrane potential.^7–10 Cationic tracers are nonspecifically trapped within the mitochondria because of their charge,^11,12 and their accumulation correlates well with the 20% to 30% mitochondrial mass in the heart.¹³ More recently, radiotracers that specifically target complex I of the mitochondrial electron transport chain (mitochondrial complex I [MC-I] and nicotinamide adenine dinucleotide:ubiquinone oxidoreductase) have been proposed as a new class of MPI agents.^14–18 These tracers are structural analogs of rotenone,^14–16 fenazaquin,¹⁸ or pyridaben,^17,19 3 well-known potent MC-I inhibitors.

Rotenone, a neutral, lipophilic, natural product, and its derivatives have been radiolabeled with either iodine^14,15 or carbon and fluorine¹⁶ for single proton emission computed tomography (SPECT) and positron emission tomography (PET) imaging. In Sprague-Dawley rats, 7′-(Z)[¹²⁵I]iodorotenone (¹²⁵I-ZIROT) demonstrated myocardial uptake similar to ^99mTc-sestamibi at 1 and 2 hours after injection, with significantly superior heart:lung ratios.¹⁴ Marshall et al¹⁵ demonstrated that ¹²⁵I-ZIROT extraction and retention were greater than those of ^99mTc-sestamibi in ex vivo–isolated, erythrocyte and albumin–perfused rabbit hearts. The objective of the present study was to evaluate the distribution of ¹²³I-ZIROT and to fully characterize the myocardial flow/extraction relationship during vasodilator stress in an open-chest canine model of coronary stenosis.

Methods

The animal studies were conducted in compliance with the position of the American Heart Association on the use of research animals in accordance with the Principles of Laboratory Animal Care (National Institutes of Health publication 86 to 23, revised 1985). The animal protocols were approved by the Institutional Animal Care and Use Committees at both Lawrence Berkeley National Laboratory and the University of Virginia.

Preparation of ¹²³I-ZIROT

Preparation of ¹²³I-ZIROT was similar to the previously described synthesis of the iodine-125–labeled analog.¹⁴ Iodine-123 (0.1N in NaOH; MDS Nordion, Ottawa, Canada) was loaded onto an AG 4-X4 anion exchange resin column (Bio-Rad acetate form; 0.57×2 cm in a 1-mL syringe). The column was rinsed 5 times with 18 MΩ deionized water. The iodine-123 was eluted from the column with 4 mol/L NH₄OH (0.4-mL fractions). The selected fraction was transferred into a 1-mL v-vial. The liquid was evaporated under a gentle stream of nitrogen at 115°C. The 7′-(Z)-tributylstannylrotenone¹⁴ (20 μg in 100 μL of dry acetonitrile) was added to the cooled vial equipped with a magnetic stir vane. NaOAc buffer (40 μL; 0.1 mol/L NaOAc and 1.3 mol/L HOAc, pH 3.5) and peracetic acid (10 μL) were added, and the reaction was stirred for 60 min at room temperature. The reaction was quenched with 10 μL of 0.1% N₂S₂O₅ (aqueous), diluted with 350 μL of water, and injected onto a Whatman Partisil 10 ODS-3 high pressure liquid chromatography column (9.4 mm×50 cm; 60/40 MeOH/H₂O, 6 mL/min). The product fractions were pooled, diluted 3 times with water, and passed through a Waters C₁₈ Sep-Pak Light cartridge. The trapped product was eluted from the cartridge with 1 mL of United States Pharmacopeia ethanol through a Millex FG (22-μm) filter. The Sep-Pakfilter assembly was rinsed with 10 mL of sterile saline, providing a 9% ethanol/saline solution for injection. Radio thin-layer chromatographic evaluation of ZIROT was undertaken before injection of the tracer (1:1 ethyl acetate/hexane; R_f=0.44).

Surgical Preparation

Five fasted adult mongrel dogs (mean±SEM weight, 26.9±4.4 kg) were anesthetized with sodium pentobarbital (30 mg/kg IV), intubated, and ventilated with room air on a respirator with 5 cm of positive end-expiratory pressure. Anesthesia was maintained using an intravenous infusion of pentobarbital (2.5 mg/kg per hour). Saturation of peripheral oxygen and expiratory end-tidal CO₂ were continuously monitored using a portable multiparameter vital signs monitor (Surgivet; Waukesha, WI). In addition, arterial blood gases were monitored throughout the experiment using a blood gas analyzer (ABL5; Radiometer America, Inc., Westlake, OH). SpO₂,_expCO₂, pH, pO₂, pCO₂, and HCO₃ were maintained at physiological levels. Lead II of the ECG was monitored continuously. Both femoral veins and the right jugular vein were isolated and cannulated using 8F polyethylene catheters for pentobarbital, fluid, ¹²³I-ZIROT, and vasodilator administration. Both femoral arteries were isolated and cannulated with 8F polypropylene catheters for arterial blood sampling and radiolabeled microsphere reference blood withdrawal. A 7F catheter was inserted in the right carotid artery for continuous monitoring of the mean systemic arterial pressure.

A left thoracotomy was performed at the level of the fifth intercostal space, and the heart was suspended in a pericardial cradle. The heart was instrumented as previously described.^5,6 A snare occluder and ultrasonic transducer were placed proximal to the first diagonal branch of the left anterior descending (LAD) coronary artery. A second flow transducer was placed on the left circumflex (LCx) coronary artery. The LAD and LCx flow, left atrial pressure, mean and end-diastolic left ventricular pressures, heart rate, and positive first derivative of left ventricular pressure with respect to time were continuously monitored and digitally recorded using LDS Ponemah software (DSI; St Paul, MN).

Experimental Protocol

The experimental protocol used in 5 dogs is depicted in Figure 1. After a brief stabilization period, 15-μM microspheres (Sr-85, Nb-95, or Sc-46; Perkin-Elmer Life Sciences, Downer’s Grove, Ill) were injected into the left atrial catheter to measure baseline resting blood flow. The microsphere technique is the “gold standard” for measuring regional myocardial blood flow, has been well validated,²⁰ and is routinely used in our laboratory.^5,6,21 Next, a critical LAD stenosis was produced by tightening the snare occluder, as previously described in detail.^5,6,21 A critical stenosis was defined as that which produced no change in resting coronary flow but with the hyperemic response to a brief 10-s LAD occlusion completely abolished. Once the stenosis was established, a second set of microspheres was administered to document myocardial blood flow with the stenosis. Afterward, an intravenous infusion of the A_2A adenosine receptor agonist ATL-146e (0.3 μg/kg per minute; PGx-Health, LLC) was initiated.²² When LCx flow was maximal, a mean±SEM of 109±49 MBq (3±1 mCi) of ¹²³I-ZIROT and a third set of microspheres were coinjected. Five minutes later, the dogs were euthanized with an overdose of sodium pentobarbital.

Figure 1.

Experimental protocol. LAD, left anterior descending coronary artery; mic, radiolabeled microsphere injections.

Ex Vivo Heart Slice Imaging and Image Quantification

Immediately after euthanasia, the heart was removed and sliced into four 1-cm short-axis slices from apex to base. The heart slices were placed on a thin piece of cardboard, covered tightly with plastic wrap, and imaged directly on a low-energy all-purpose parallel hole collimator of a standard clinical γ camera (Siemens Orbiter). Images were collected using a 128×128 count matrix with a 20% window centered on the 159-keV iodine-123 photopeak for a total of 1×10⁶ counts. The heart slice images were quantified by regions of interest (ROIs) drawn on the LAD defect area of the anteroseptal wall and on the normal LCx-perfused posterior wall, as previously described.^5,6

Determination of Tracer Activities and Regional Myocardial Blood Flow

Each myocardial slice was then subdivided as previously described, resulting in 72 myocardial samples for each dog.^5,6 γ Well counting was performed on these samples and on the blood reference samples to determine iodine-123 and microsphere activities using a γ well counter (MINAXI 5550; Perkin-Elmer Life Sciences). The tissue counts were corrected for background, decay, and isotope spillover; regional myocardial blood flow was calculated with specialized computer software (PCGERDA; Scientific Computing Solutions, LLC, Charlottesville, Va). Transmural activities and flow values were calculated as the weighted mean of the corresponding epicardial, midwall, and endocardial samples. The 5 transmural sections (15 segments) with the lowest flow values were classified as the stenotic LAD zone, and the 5 transmural sections with the highest flow values were classified as the normal LCx zone. For each of the 5 individual animals, raw data plots of ZIROT activities versus microsphere-derived flow at injection during vasodilator stress were created with all 72 myocardial tissue segments from each dog. A plot of data pooled from the 5 dogs was created by normalizing the activity data from each individual dog to the activity of a segment in the same dog, with a normal flow of ≈1.0 mL/(min · g), similar to previous studies^5,6 from our laboratory.

Whole Body Planar Imaging

Two fasted adult mongrel dogs (weight, 20–25 kg) were sedated with sodium pentothal (30 mg/kg IV). The animals were not pretreated with potassium iodide to block thyroid uptake of free iodine-123. The dogs were intubated and maintained using inhaled isoflurane gas (2%). The ¹²³I-ZIROT (148–222 MBq) was injected in a femoral vein. Serial planar images were acquired with a dual-head clinical γ camera (GE Millenium Hawkeye) using a low-energy, high-resolution, parallel-hole collimator and a 20% window centered on the 159-keV iodine-123 photopeak. The dynamic images were acquired from both heads in the left and right lateral views for 4 minutes each during the first hour and for 8 minutes each between 1 and 2 hours. In 1 dog, serial arterial blood samples (2 mL) were acquired at 5-s intervals for the first minute and then at 1.5, 2, 2.5, 3, 4, 5, 10, 15, 20, 30, and 60 min after ¹²³I-ZIROT injection. The blood samples were counted in a γ well counter, and the blood clearance time constant was calculated by fitting the time-activity data with an exponential function.

Image Quantification

The dynamic image data sets acquired from the dual heads were summed at each point and were corrected for background and decay. The ROIs were drawn on the heart and liver, and the activities within each region were expressed as a percentage of the initial activity. Heart/liver ratios were calculated by dividing the mean counts in the heart ROI by the mean counts in the liver ROI.

Radiation Dosimetry

The ROIs were drawn in areas of minimal organ overlap. Residence times were calculated from the integral of the time activity curves for the following organs: brain, thyroid, heart, stomach, gallbladder, upper small and large intestines, kidney, liver, and urinary bladder. Radiation dose estimates were calculated using MIRDOSE3.1.²³

Statistical Analysis

All statistical computations were performed with SYSTAT software (Systat Software, Inc, San Jose, Ca). The results were expressed as mean±SEM. Differences between means were assessed with either 1- or 2-way repeated-measures ANOVA, where appropriate, with Bonferroni correction for multiple comparisons, or a paired t test (ischemic/normal ratios), with values of P<0.05 considered significant. The myocardial flow-extraction relationship was analyzed by nonlinear regression using the following equation: y=a0*(1−exp[−a1/x]), according to the kinetic transport model of Gosselin and Stibitz.²⁴

Results

Hemodynamic Parameters

The mean hemodynamic parameters are summarized in the Table. As expected, adenosine receptor agonist (ATL146e) infusion caused a slight transient decrease in mean arterial pressure and left ventricular pressure, with no significant changes in heart rate or left atrial pressure. Furthermore, the vasodilator resulted in no increase in mean ultrasonically measured LAD flow because of the presence of the critical stenosis on that vessel, but there was a significant increase in LCx flow from 37±8 to 121±29 (P<0.001 versus stenosis), resulting in a >3:1 flow disparity between the stenotic LAD and normal LCx coronary artery.

Absolute Myocardial Blood Flow

Figure 2 summarizes the microsphere-derived transmural regional myocardial blood flows. As can be seen, resting flow did not change in the LAD zone on setting of the critical stenosis. During peak vasodilator stress, LAD zone flow remained unchanged; however, there was a 3.3-fold increase in normal LCx zone flow, resulting in a nearly 4:1 mean flow disparity between the 2 zones when ¹²³I-ZIROT was injected.

Figure 2.

Average transmural myocardial blood flow in the LAD and left circumflex (LCx) territories determined using microsphere injections at baseline, after setting the critical LAD stenosis, and during peak vasodilator stress at ¹²³I-ZIROT injection. * P<0.001 vs baseline, and †P<0.001 vs LAD. n=5.

Myocardial ZIROT Uptake as a Function of Regional Blood Flow

Figure 3 depicts the myocardial flow-extraction relationship obtained by γ well counting of the 359 myocardial segments obtained from all 5 animals. The thallium-201 and ^99mTc-sestamibi flow-extraction curves from our previous studies using a similar canine model are shown on the same graphic.⁶ Myocardial ¹²³I-ZIROT extraction tracked regional myocardial blood flow, as determined by the gold standard microspheres, over a wide range of flow, more linearly than either thallium-201 or ^99mTc-sestamibi in our previous studies. As shown in Figure 3, ¹²³I-ZIROT begins to underestimate true flow, as determined by the microsphere measurements, at >2 mL/(min · g). This is better than the current clinical flow tracers that deviate from unity between 1 and 1.5 mL/(min · g).

Figure 3.

Flow-extraction relationship. This graphic represents the pooled results from 359 myocardial segments obtained from the 5 experiments. ¹²³I-ZIROT activity and flow in each myocardial segment were normalized to the area of normal flow (≈1 mL/[min · g]). The curve fit through the experimental data were of the form y=a0*(1−exp[−a1/x]) according to the kinetic transport model of Gosselin and Stibitz.²⁴ The results previously obtained in our laboratory using the same animal model for thallium-201 and ^99mTc-sestamibi are represented on the same graphic.⁶ r² is the coefficient of determination.

Stenotic/Normal Count Ratios for Flow Versus Tracer Activities

Representative ZIROT ex vivo heart slice images are displayed in Figure 4A. Large perfusion defects were easily observed in the anteroseptal wall (slices 1–3). The average defect:normal count ratio for the 3 slices with defects in this dog was 0.20. The bar graphs in Figure 4B compare the mean stenotic LAD/normal LCx ratios for flow versus ¹²³I-ZIROT activities determined by either γ well counting of transmural myocardial segments or ex vivo image quantification. The lower the bar, the more severe the flow disparity between the stenotic and normal zones. As shown, using 2 separate and independent analyses of tracer deposition, ¹²³I-ZIROT only slightly underestimated the true microsphere flow ratio. The ¹²³I-ZIROT LAD/LCx flow ratio was 0.32±0.09, compared with LAD/LCX activity ratios of 0.45±0.10 by γ well counting (P<0.05 versus flow ratio) and 0.42±0.08 by ex vivo imaging (P<0.05 versus flow ratio). Also shown in Figure 4B are previous data evaluating ^99mTc-sestamibi in the same model with comparable LAD-LCx flow disparity at injection (0.43±0.04).

Figure 4.

A, Representative example of ¹²³I-ZIROT ex vivo imaging. Note the presence of a large defect on the anteroseptal segments of the most apical short-axis slices (slices 1–3). The defect:normal count ratio in this dog was 0.20. B, Comparison between the average transmural flow defect magnitude and the average ¹²³I-ZIROT defect magnitude, as determined by GWC (γ well counting) or ex vivo imaging (n=5). For comparison, our previous results using the same animal model for ^99mTc-sestamibi with comparable flow at injection are represented on the same graphic (n=9).⁶ *P<0.05 vs flow ratio. P indicates posterior; A, anterior; S, septal; L, lateral.

Determination of Biodistribution, Organ Kinetics, and Radiation Dosimetry Using Whole Body Dynamic Planar Imaging

Planar images of ¹²³I-ZIROT at 15, 30, 60, and 120 min after injection are shown in Figure 5. The heart is easily identifiable by planar imaging as early as 15 min after injection. The low blood pool activity allows early visual delineation of both the left and right ventricle walls. This observation was supported by the serial blood sampling data in this dog that revealed that ¹²³I-ZIROT rapidly cleared from the blood pool (t_1/2=17 s), with <2% of the initial dose present in the blood pool at 15 min after injection. Limited uptake of free iodine-123 was seen in the thyroid during the 2-hour duration of the imaging study.

Figure 5.

Representative whole body planar images of ¹²³I-ZIROT at 15, 30, 60, and 120 min after injection. H indicates heart; Li, liver; K, kidney; GB, gall-bladder; T, thyroid; SG, salivary glands.

Both myocardial and liver uptake values were maximal at the 2-min point. Myocardial and liver clearance values were then moderate, with t_1/2 values of ≈30 and 11 minutes, respectively (Figure 6A). At 15 min after ¹²³I-ZIROT injection, the heart/liver activity ratio reached a plateau of 1.32±0.14 that was maintained for at least 2 hours (Figure 6B); there was progressive accumulation of the tracer in the gallbladder. The ¹²³I-ZIROT uptake in the lungs was minimal, and the lungs were not visible by planar imaging at any point.

Figure 6.

Results obtained from quantification of the in vivo planar images (n=2). The regions of interest (ROIs) were drawn around the heart and liver to determine the heart and liver kinetics (A) and the heart/liver ratio (B) over time.

Preliminary radiation dose estimates from the planar images were determined on selected organs, including the gallbladder wall, stomach, heart wall, kidneys, liver, thyroid, and salivary gland. The organ with the highest radiation dose was the gallbladder wall, with 1.14 rad/mCi injected. This was followed by the salivary gland, stomach, and thyroid, with doses of 0.66, 0.46, and 0.26 rad/mCi, respectively. The heart wall (0.03 rad/mCi) and the clearance organs, liver and kidney (0.04 rad/mCi), showed the least radiation dose. The whole body dose was 0.011 rad/mCi.

Discussion

Although SPECT imaging with thallium-201, ^99mTc-sestamibi, and ^99mTc-tetrofosmin has been the workhorse for clinical MPI, there are shortcomings with these agents. Despite the fact that they provide highly relevant diagnostic information about myocardial perfusion and viability, several known limitations have been identified.¹ The roll-off phenomenon remains the principal limitation to the use of these MPI agents. Because of the underestimation of the highest flow values during a stress test, the activity ratio between a myocardial region with a mild impairment in flow reserve and a normal region might be close to unity, thus limiting the detection of an ischemic region.^5,6 Therefore, there is a well-defined need for a new generation of perfusion tracers with improved myocardial extraction over a broader flow range. Moreover, quantitative flow measurement may be much more relevant than relative perfusion imaging for the evaluation of patients with homogeneous perfusion defects, such as in balanced coronary artery disease,²⁵ cardiomyopathy,²⁶ or endothelial dysfunction.²⁷ Therefore, accurately tracking flow over a wide flow range is also 1 of the requirements for the determination of absolute myocardial blood flow.^1,28 Because the prevalence of cardiovascular diseases associated with diffuse perfusion abnormalities, such as diabetes mellitus, is increasing, the need for quantitative flow measurement using either SPECT or PET is increasing.² PET imaging, with better attenuation correction and higher spatial and temporal resolution than SPECT, is the best candidate for quantitative flow measurements in nuclear medicine. Nevertheless, even if the technique is not yet available, it is reasonable to think that the next generation of SPECT cameras, together with better attenuation correction and mathematical modeling, may enable the absolute measurement of myocardial blood flow using a single-photon perfusion tracer.

Rotenone derivatives radiolabeled with either iodine (¹²³I-ZIROT) or fluorine-18 ([¹⁸F]fluorodihydrorotenone) have been proposed as new perfusion imaging agents.^14–16 Rote-none is a natural product that has been extracted from the roots of the tropical plant Derris elliptica and has a nanomolar affinity for the MC-I of the electron transport chain.²⁹ Because mitochondrial dysfunction has been reported in various pathological conditions, such as neurological diseases, rotenone derivatives have been previously radiolabeled with carbon-11 and tritium to evaluate the density of the MC-I in the brain and other tissues.^30,31 Initial evaluations of labeled rotenone derivatives were focused on brain distribution relative to degenerative diseases. Once these compounds were labeled with PET and SPECT radioisotopes, their avidity for tissues rich in mitochondria, especially the heart, was noted.

Using mitochondria as a target for MPI is not new. Both sestamibi and tetrofosmin accumulate in the heart based on the mitochondrial membrane potential.^8–10,12 Prior studies¹⁴ have shown that ¹²⁵I-ZIROT demonstrates a high myocardial uptake in vivo in normal rats (3.71±0.20% ID/g at 1 h post-injection), with modest washout (2.96±0.24% ID/g at 2 h post-injection). Marshall et al¹⁵ evaluated ¹²⁵I-ZIROT in an ex vivo model of isolated retrograde blood-perfused rabbit hearts and showed that both extraction and retention of ¹²⁵I-ZIROT were significantly higher than that of sestamibi. Similarly, in the same model, [¹⁸F]fluorodihydrorotenone tracked flow better than thallium-201.¹⁶ The radiolabeled rotenone analogs have, therefore, demonstrated excellent characteristics as MPI agents. Accordingly, analogs of 2 other MC-I inhibitors, fenazaquin and pyridaben, have been subsequently radiolabeled with fluorine-18 and evaluated as new PET MPI agents.^{17–19,32–34} The results obtained in various animal models using the pyridaben analog ¹⁸F-BMS747158-02 have been encouraging.^{17,19,32–34} The MC-I inhibitor analogs can, therefore, be considered as a new class of MPI agents with excellent characteristics for both SPECT and PET imaging.

Myocardial Flow-Extraction Relationship

In the current study, infusion of the A_2A adenosine receptor agonist ATL-146e²² increased flow by up to ≈6 times in normal zone myocardial segments, whereas flow in the critically stenotic LAD region was unchanged, producing a wide range of flow. Over this wide range, ¹²³I-ZIROT tracked flow more linearly than was previously found using either thallium-201 or ^99mTc-sestamibi.⁶ As shown in Figure 3, ¹²³I-ZIROT begins to underestimate the true flow, as determined by the microsphere measurements, at >2 mL/(min · g). This is better than the current clinical flow tracers that deviate from unity between 1 and 1.5 mL/(min · g). Because of the more linear myocardial extraction, the ¹²³I-ZIROT activity ratio between high- and low-flow regions more closely approximated the gold standard microsphere flow ratio in these same regions. Also, because of the more linear extraction, ¹²³I-ZIROT produced large transmural perfusion defects on images.

Comparison With Other SPECT MPI Agents

Thallium-201 tracks regional myocardial blood flow better than the other clinically used SPECT MPI agents, ^99mTc-sestamibi and ^99mTc-tetrofosmin. Indeed, thallium-201 has higher first-pass extraction and the roll-off phenomenon occurs at higher flow rates than is observed with these ^99mTc-labeled perfusion tracers.^6,35 However, the physical properties of thallium-201 are not ideal because of its relatively long half-life, which limits the injected dose, and its suboptimal energy for SPECT camera detection. Therefore, despite the relatively high liver uptake, ^99mTc-sestamibi and ^99mTc-tetrofosmin are often preferred for SPECT MPI.^36,37

We previously evaluated the thallium-201 and ^99mTc-sestamibi flow-extraction relationship in a canine model with a similar degree of coronary stenosis as that used in the present study, as evidenced by comparable LAD/LCx flow ratios at tracer injection.⁶ The thallium-201 and ^99mTc-sestamibi activity ratios obtained by γ well counting underestimated the degree of flow disparity to a greater extent than ¹²³I-ZIROT (0.62±0.04 and 0.79±0.03 versus 0.45±0.10, respectively). This finding is in agreement with the results obtained by Marshall et al^15,16 in the isolated perfused rabbit heart model.

The ^99mTc-tetrofosmin flow-extraction relationship was first evaluated by Sinusas et al³⁸ in a canine model of permanent LAD occlusion. When injected at peak adenosine or dipyridamole stress, there was good correlation between ^99mTc-tetrofosmin extraction and absolute flow (r=0.84). However, the plateau in tetrofosmin extraction occurred at even lower flow rates than had been previously described for thallium-201 and ^99mTc-sestamibi.⁶ These results were confirmed in our laboratory using a canine critical stenosis model similar to that used in the present study.³⁵

^99mTcN-NOET is a neutral lipophilic compound with a high myocardial first-pass extraction fraction.^39,40 Calnon et al⁴¹ evaluated the ^99mTcN-NOET myocardial flow-extraction relationship using a similar canine model as used in the present study. They reported a stenotic/normal ^99mTcN-NOET activity ratio of 0.55±0.05 by well counting (versus 0.45±0.10 for ¹²³I-ZIROT) and 0.58±0.64 by ex vivo imaging (versus 0.42±0.08 for ¹²³I-ZIROT) in the presence of a stenotic/normal flow ratio similar to that obtained for ¹²³I-ZIROT (0.33±0.04 versus 0.32±0.09, respectively). Despite its favorable characteristics and extensive preclinical and clinical testing, ^99mTcN-NOET was ultimately not approved for MPI in humans.

ZIROT Biodistribution, Organ Clearance Kinetics, and Radiation Dosimetry

Remarkable in vivo stability of ¹²³I-ZIROT was noted in this study. Pretreatment with potassium iodide is typical for studies involving radioiodinated tracers to prevent the accumulation of free radioiodine in the thyroid. In this study, potassium iodide was not administered, yet the accumulation of free radioiodine in the thyroid was surprisingly low, even 2 to 4 h after ¹²³I-ZIROT injection (4-h data not shown). This is strong evidence that metabolic deiodination of ¹²³I-ZIROT is minimal in dogs.

The myocardium was clearly identifiable on in vivo planar images at early points (<15 min), suggesting a favorable myocardial extraction associated with a rapid blood clearance. The ¹²³I-ZIROT cardiac retention was moderate, with a t_1/2 of ≈30 minutes. There are advantages and disadvantages for an MPI agent to be retained in the myocardium after injection.⁴ High myocardial retention allows gated imaging for simultaneous determination of relative perfusion and LV function. On the other hand, myocardial tracer clearance kinetics can be useful for absolute flow quantification using mathematical compartmental modeling.

A preliminary estimation of radiation dosimetry to several key organs and the whole body was determined for ¹²³I-ZIROT. The estimated heart wall, clearance organs, and whole body doses are similar to those estimated for the current clinically approved MPI SPECT tracers.^36,37 However, the gallbladder wall is the organ with the highest dose per mCi of tracer injected in the canine model and may be the organ that limits the injected dose of the tracer in human subjects. The protocols involved in the current studies used injected doses that range from 109 to 222 MBq (3–6 mCi). Suitable uptake and high-quality images were seen with 3 to 4 mCi injected into canine. This translates to 3.42 to 4.56 rad to the gallbladder wall, a dose that is generally recognized as safe and effective by the Food and Drug Administration (Code of Federal Regulations §361.1). The stability of ¹²³I-ZIROT to deiodination, as previously noted, is also seen in the thyroid radiation dose estimate. Thyroid blocking with potassium iodide, a common practice for radioiodinated tracers, may be reduced or unnecessary with ¹²³I-ZIROT. Given the preliminary radiation dose data in the canine, a sufficient ¹²³I-ZIROT dose may be injected without undue dose burden to the patient and without compromising image quality.

Limitations of the Study

In this experimental study, we did not perform in vivo SPECT imaging of the myocardial uptake of ¹²³I-ZIROT. The experimental protocol required euthanizing the animals at 5 min after tracer injection, which did not allow sufficient time to permit SPECT imaging. The reason for selecting the 5-min termination point was to permit comparison with our previous studies using other MPI agents, including Tl-201, under the same experimental conditions. In these earlier studies, the 5-min end point was chosen because it was sufficient to allow enough time for maximal extraction of the tracers being evaluated, based on blood pool availability, but did not allow time for Tl-201 to undergo appreciable redistribution. A second limitation of this study is that we did not perform measurements of myocardial first-pass extraction fraction. The first-pass extraction fraction is an important parameter that needs to be measured in a future study compared with other perfusion tracers. Nevertheless, given the extraction versus flow relationship presented in the current study, the first-pass myocardial extraction fraction for ¹²³I-ZIROT can be inferred as higher than that of any of the currently approved myocardial perfusion radiotracers. Thus, although additional ¹²³I-ZIROT studies appear warranted, other analogs of IROT are being sought with more favorable pharmacokinetics that might reduce the radiation dose burden to the gallbladder wall.

Summary and Conclusions

In a canine model of critical stenosis, ¹²³I-ZIROT tracked regional myocardial blood flow over a wide range. In addition, the myocardium was clearly visible by in vivo planar imaging as early as 15 min after tracer injection. The ability to more linearly track changes in blood flow over a wide range, combined with favorable biodistribution properties, makes ¹²³I-ZIROT a promising new SPECT MPI agent.

Table 1.

Table Hemodynamic Parameters at Baseline, After the Setting of the Stenosis, and During Vasodilator Stress at ¹²³I-ZIROT Injection

Parameters	Rest		Stress
	Baseline	Stenosis
HR, BPM	123±15	121±13	135±19
MAP, mm Hg	102±7	99±6	80±8*†
LAP, mm Hg	5±1	6±2	6±2
LVP, mm Hg	119±8	113±8*	102±10*†
LVEDP, mm Hg	6±2	7±2	6±2
dP/dt, mm Hg/s	2030±313	1779±148	2489±342†

Data are given as mean±SEM. n=5. BPM indicates beats/minute; dP/dt, first derivative of LVP with respect to time; HR, heart rate; LAP, left atrial pressure; LVP, left ventricular pressure; LVEDP, left ventricular end-diastolic pressure; MAP, mean arterial pressure.

^*P<0.05 vs baseline.

^†P<0.05 vs stenosis.

CLINICAL PERSPECTIVE.

The clinical utility of radiotracer imaging of myocardial perfusion depends, in part, on the accuracy of the tracer reflecting myocardial blood flow and the relative biodistribution in cardiac versus noncardiac tissue. Current agents are limited by suboptimal energy characteristics, high subdiaphragmatic activity, or nonlinear uptake at high flow. The accuracy of a new myocardial perfusion radiotracer that targets complex I of the mitochondrial electron transport chain, ¹²³I-ZIROT, was tested in an in vivo dog model of critical coronary artery stenosis with pharmacological stress-induced hyperemia. Myocardial ¹²³I-ZIROT extraction tracked regional myocardial blood flow, as determined by the microspheres, over a wide range of flows, whereas the ¹²³I-ZIROT LAD/LCx activity ratios obtained by ex vivo imaging only slightly underestimated the LAD/LCx microsphere flow ratio. ¹²³I-ZIROT was identified in the myocardium 15 min after injection, with low lung and blood pool activity. These experiments indicate that ¹²³I-ZIROT has advantages in extraction and biodistribution compared with either thallium-201 or ^99mTc-sestamibi and, hence, could be a useful agent to image myocardial blood flow in humans.