Myocardial Rubidium-82 Tissue Kinetics Assessed by Dynamic Positron Emission Tomography as a Marker of Myocardial Cell Membrane Integrity and Viability
Background Recent reports have demonstrated the clinical use of rubidium-82 chloride (Rb-82) in combination with positron emission tomography (PET) not only as a tracer of myocardial blood flow but also as a marker of cell membrane integrity using static imaging early and late after tracer injection. The purpose of this study was to compare myocardial Rb-82 kinetics assessed by dynamic PET imaging as a marker for tissue viability with regional fluorine-18 fluorodeoxyglucose (FDG) uptake in patients with coronary artery disease.
Methods and Results Twenty-seven patients with angiographically proven coronary artery disease and 5 subjects with a low likelihood for coronary artery disease underwent dynamic PET imaging under resting conditions using Rb-82 and FDG. Both image sequences served as input data for a semiautomated regional analysis program. This program generated polar maps representing Rb-82 tissue half-life and FDG utilization assessed by Patlak’s approach. Myocardial tissue viability was visually determined from static Rb-82 and FDG images. Regions were categorized as normal, ischemically compromised, and scar tissue. Their coordinates were subsequently copied to the functional polar maps for further analyses. In normal subjects, Rb-82 tissue half-life was homogeneous throughout the left ventricle (90±11 seconds). In coronary patients, differences between Rb-82 tissue half-lives in normal and scar tissue were highly significant (95±10 and 57±15 seconds, respectively; P<.0001). FDG uptake in these two tissue groups was 78±12% and 40±13%, respectively (P<.0001). Ischemically compromised tissue with reduced perfusion but maintained FDG uptake displayed an Rb-82 half-life of 75±9 seconds, indicating active cellular tracer retention, which was significantly different from scar tissue. Overall agreement of tissue categorization as either viable or scar was 86% between Rb-82 kinetics and FDG utilization. In a subgroup of 11 patients with all three tissue types within one image set, Rb-82 tissue half-life discriminated between normal, ischemic, and scar tissue (97±9, 75±9, and 60±15 seconds, respectively; P<.01).
Conclusions This study demonstrated a significant relationship between cell membrane integrity as assessed by dynamic Rb-82 PET imaging and myocardial glucose utilization as a marker for tissue viability. In regions with reduced perfusion, Rb-82 kinetics was different in compromised but metabolically active and irreversibly injured myocardium. The predictive value of this approach must be evaluated in follow-up studies.
Positron emission tomography (PET) in combination with tracers of myocardial perfusion including nitrogen-13 ammonia, oxygen-15 water, or rubidium-82 chloride (Rb-82) and of myocardial metabolism such as carbon-11 acetate or fluorine-18 fluorodeoxyglucose (FDG) has been used as a sensitive imaging technique to assess myocardial tissue viability in patients with coronary artery disease.1 2 3 4 5 6 Preserved or absent FDG uptake in regions with reduced perfusion has proven to predict functional outcome after coronary revascularization.7 8 9 10 11 12 13 14 15 Furthermore, recent reports indicate that a mismatch of flow and metabolism is associated with a high risk for subsequent cardiac events if revascularization is not performed.16 17 18
PET imaging with Rb-82 has proven to be highly accurate in the detection and functional assessment of coronary artery stenoses and infarct size imaging.19 20 21 22 23 24 25 26 27 Experimental and first clinical data indicate the usefulness of Rb-82 in combination with PET to assess myocardial viability. Animal models with coronary occlusion and reperfusion demonstrated rapid Rb-82 clearance from irreversibly injured tissue, whereas the tracer was retained in reversibly damaged and viable tissue.28 29 Based on these experimental data, Gould et al30 subsequently proposed Rb-82 imaging early and late after tracer injection as a new approach for definition of tissue viability in patients with previous myocardial infarction. Recently, the clinical long-term follow-up of patients studied by this approach has been reported and was shown to be of diagnostic value to identify patients with poor prognosis and those who might benefit from coronary revascularization.31
The purpose of this study was to develop a method for quantitative assessment of myocardial Rb-82 kinetics and to compare the results with regional FDG and Rb-82 tracer uptake measurements as independent markers of tissue viability.
Twenty-seven patients (18 men, 9 women; age, 58±13 years) with angiographically proven coronary artery disease were studied at the University of Michigan Medical Center. All were referred for assessment of myocardial viability in regions with impaired function for further treatment planning. Twenty-six patients (96%) had a history of previous myocardial infarction. The median time between infarction and PET imaging was 35 days (range, 4 days to 6 months). Twelve patients had a recent myocardial infarction and 9 patients had received thrombolytic or interventional therapy during the evolving infarction. Global left ventricular ejection fraction was <50% in all patients, with an average of 42±11% (range, 25% to 47%). Twenty-one patients (78%) had angiographic evidence of multivessel disease.
For comparison, 5 subjects (3 men, 2 women; age, 50±20 years) with a low likelihood of hemodynamically relevant coronary artery disease by history, ECG criteria, echocardiography, and negative stress thallium-201 scintigraphy were studied to evaluate the homogeneity of Rb-82 clearance parameters in normal myocardium.
Positron Emission Tomography
Patients had a light breakfast in the morning and received either 50 g dextrose orally or, in the presence of diabetes mellitus, small doses of insulin to titrate plasma glucose levels at approximately 120 mg/100 mL before the imaging procedure.32
After positioning of the patient in the scanner gantry, a 20-minute transmission scan for attenuation correction was performed using a retractable germanium-68 ring source. Consistency in patient positioning was achieved by marking the chest with washable ink and aligning the marks with a low-power laser light beam from the tomograph. Patients were not moved out of the scanner during the study.
All studies were performed under resting conditions using a two-ring, multislice, whole-body PET scanner (ECAT 931, CTI/Siemens), allowing simultaneous imaging of 15 transaxial slices, with a slice thickness of 6.75 mm encompassing the entire heart. All studies were obtained using the following reconstruction parameters: matrix size, 128×128 pixels with a pixel size of 2.35 mm and Hanning reconstruction filter with a cutoff frequency of 0.30.
Rb-82 delivered from a strontium-82/Rb-82 generator (Squibb) was injected using an automated infusion system (Harvard model 975, Harvard Apparatus) capable of accurate delivery of Rb-82, with a maximum volume of 100 mL per infusion. A dose of 60 mCi Rb-82 was diluted and infused over 30 seconds. Dynamic data acquisition was initiated at the beginning of the infusion and continued for 8 minutes, with a total of 15 frames (12×10 seconds and 3×120 seconds).
After Rb-82 decay, 10 mCi FDG was injected as a slow intravenous bolus over 30 seconds, and dynamic PET images were obtained for 60 minutes (12×5 minutes).
Data Processing and Analysis
Three-dimensional reorientation. A reorientation algorithm provided by the scanner software system was used to generate 12 short-axis cardiac planes for each time point obtained from dynamic Rb-82 or FDG protocols.33 The short-axis cardiac images served as input data for a semiautomated regional analysis program developed at our institution, which generates quantitative polar maps of regional myocardial Rb-82 clearance rates and of myocardial FDG utilization.
Topographic definition of myocardial regions for quantitative analysis. An automated region definition program that had been developed for N-13 ammonia studies at our institution was used for the calculation of regional time-activity curves.34 Epicardial and endocardial borders were visually defined by the operator for each short-axis plane in the time frame with the highest ratio of myocardial to background tracer activity. For the Rb-82 study, this was usually frame 13 (2 to 4 minutes). In this frame, tracer activity had already cleared from the blood pool, but significant decay of myocardial Rb-82 had not yet occurred. In the case of FDG study with progressive myocardial tracer accumulation over time, the final frame (55 to 60 minutes) was chosen.
Activity maxima within the two concentric borders were defined along 36 circumferential radii, beginning at the posterior intersection of the left and right ventricular walls. Thirty-six regions of 3×3 pixels in each of the 8 short-axis planes were created along these radii, centered at the point of maximal activity of each radius. Subsequently, the locations of these regions were copied to all time frames. Time-activity curves were finally determined for 36 circumferential regions in each short-axis cardiac plane for further analysis.
Blood pool definition. For both tracers, definition of the region representing the blood pool activity was confined to the two most basal planes to minimize resolution distortion. For Rb-82 kinetics, the blood pool region covering ≥9 pixels was defined in the left ventricular chamber using the frame showing peak Rb-82 activity.35 For the FDG studies, the blood pool region covering ≥9 pixels was placed in the right ventricular chamber, with sufficient distance to the ventricular septum to minimize myocardial tracer contribution at late imaging time frames.
Calculation of Rb-82 tracer kinetics. A monoexponential, least squares curve fit was applied to non–decay-corrected regional time-activity curves. Dependent on the individual shape of the tissue curve, the start time point for the curve fit was chosen interactively by the operator so that the rapid component of the blood clearance function was completed and only the slow component remained (Fig 1⇓). This time point was defined in most patients approximately 90 seconds after the start of tracer infusion. All following data points until completion of the study were included into the fitting procedure. Regional time-activity curves were fitted to the formula
where CRb(t) represents the total tissue activity predicted by the model at time t and Ca(t) represents the actual blood pool activity at time t. The two estimated parameters were A and kt. Parameter A  represents Rb-82 uptake at the start time point of curve fit, and parameter kt [min−1] characterizes the combined physical tracer decay and tissue clearance rate of Rb-82. The constant TBV represents the total blood volume and spillover fraction within the defined region. Based on previous studies, TBV was preset to a value of 0.3 (30%).34
Upon completion of the fitting procedure, the resulting values for parameters A and kt were presented in a polar map display. To prevent interstudy variation, the polar map representing parameter A was normalized to the individual maximum and expressed as a percentage.
Analysis of FDG time-activity curves. Blood pool and myocardial FDG time-activity curves were used to calculate regional values for the Patlak analysis as described previously from our laboratory.32 36
The analysis describes the correlation between the ratio Ct/Cb and the expression ∫Cb(0-t)/Cb, where Ct is the decay-corrected myocardial activity and Cb is the decay-corrected blood pool activity at any given time t. ∫Cb(0-t) represents the integral of blood pool activity from time zero to time t and serves as an index of the arterial FDG input function. Since FDG–6-phosphate accumulates in tissue as a function of uptake and phosphorylation of FDG, this relationship becomes linear after equilibration of free tissue FDG. The linear part of this plot was determined with the use of a derivative approach.
Regional Patlak slopes derived from all 36 time-activity curves were displayed in polar map format. Results of each region were presented as a percentage of individual maximum. This relative expression was chosen to avoid additional scatter in the data caused by intraindividual and interindividual variation of absolute metabolic values.14
Regional analysis of data. Based on the time-activity curves and analysis scheme, the software package generated three coordinate polar maps for each individual study: (1) percentage of maximal Rb-82 uptake (RbUpt%): average noncorrected Rb-82 tissue activity between 2 and 4 minutes after start of Rb-82 infusion, expressed as a percentage of the individual maximum; (2) Rb-82 tissue half-life (RbT1/2) derived from the washout rate constant kt converted into tracer half-lives (RbT1/2=[ln|0.5|/kt]×60): a time of 76 seconds represents physical tracer decay; higher values indicate cellular tracer accumulation and lower values express accelerated clearance from tissue; and (3) percentage of maximal FDG Patlak slope (FDGSlo%): regional Patlak slopes were normalized to the individual maximum within each polar map and expressed as a percentage.
Visual definition of regional tissue viability. Myocardial viability was defined by visual analysis of conventional static Rb-82 and FDG PET images (short axis and horizontal and vertical long axis). These static images were created from the dynamic studies by summing up frames acquired from several consecutive time frames. In the case of Rb-82 studies, frames 9 through 15 (80 seconds to 8 minutes) were combined. For creation of static FDG images, summation was carried out from frames 7 through 12 (30 to 60 minutes). From these images, tissue regions representing three different tissue types were categorized visually as previously described from our laboratory15 : normal, regions with normal or highest Rb-82 uptake; ischemically compromised, regions with reduced Rb-82 uptake but maintained or elevated F-18 uptake (mild match and mismatch); and scar, regions with concordant reduced Rb-82 and FDG uptake (match).
Based on the well-validated match or mismatch pattern, this approach was used to optimally separate into discrete regions the various tissue types that each heart contained in varying extent and mixture.
This visual analysis allowed localization of the regions of interest on the polar maps for further quantitative analysis. One region in each individual patient was defined to represent normal tissue. Additionally, in 21 patients one region was defined in each as representing ischemically compromised tissue with maintained viability. In another 17 patients one region in each was categorized as scar tissue with markedly reduced perfusion and concordant reduction of FDG utilization. Thus, each individual patient could have a maximum of three regions with different tissue categories.
A region definition procedure similar to that described for the dynamic protocol was applied to the static images and polar maps representing Rb-82 distribution and FDG activity. On these static polar maps, the regions of interest for the individual tissue categories were placed according to the previous visual localization. The coordinates of the so-defined regions were subsequently copied to the polar maps of the dynamic studies representing RbUpt%, RbT1/2, and FDGSlo%, as described above.
Analysis of Regional Homogeneity of Rb-82 Kinetics
In the control population, the left ventricular myocardial wall was divided into four quadrants (anterior, lateral, inferior, and septal). These quadrants were divided into basal and mid ventricular parts. The apex was defined as a separate region. Overall, nine myocardial regions were defined in each subject to study the regional homogeneity of tracer uptake and clearance. Analysis of regional Rb-82 kinetics was performed as described for patient studies.
Values are given as mean±SD. For statistical analysis, the software package StatView 4.02 (Abacus Concepts Inc) was used. Comparison between two groups was performed with the use of Student’s t test for either unpaired or paired data sets. Linear regression analysis was applied to study relationships between different parameters. A value of P<.05 was considered statistically significant.
Regional Homogeneity of Rb-82 Tissue Half-life in Normal Myocardium
In 5 control subjects without evidence for coronary artery disease, mean tissue half-life of Rb-82 was 90±11 seconds (range, 70 to 127 seconds) and was homogeneous throughout the left ventricle without significant regional differences (Table 1⇓).
Rb-82 Kinetics and FDG Uptake in Different Tissue Types
The results of quantitative analyses of regional Rb-82 uptake, Rb-82 half-life, and FDG utilization are given in Table 2⇓. Rb-82 uptake as well as Rb-82 half-life was significantly different between normal and scar tissue (P<.0001).
In regions defined as ischemically compromised, Rb-82 uptake was significantly different from normal and scar tissue. Rb-82 tissue half-life was 75±9 seconds in these regions, which approximates physical tracer decay and indicates active cellular tracer retention. In contrast, Rb-82 half-life was significantly shorter in scar tissue, with 57±15 seconds (P<.0001) indicating accelerated tracer clearance from myocardium. FDG uptake in ischemically compromised tissue was 68±15% and significantly higher compared with scar tissue (40±13%, P<.0001).
Fig 2⇓ illustrates the individual Rb-82 tissue half-lives in all 27 patients. In regions defined as normal, 26 of 27 regions (96%) had RbT1/2≥76 seconds (physical tracer decay). In ischemically compromised regions, 9 of 17 regions (53%) displayed RbT1/2>76 seconds. In contrast, 16 of 17 (94%) of the regions defined as scar tissue demonstrated accelerated Rb-82 washout (RbT1/2<76 seconds).
Ten regions displayed a perfusion/metabolism mismatch with FDG utilization >70%. Rb-82 uptake was 47±10%, and FDG utilization was 81±7%. Myocardial Rb-82 half-life averaged 78±9 seconds (range, 68 to 88), with 7 of 10 regions displaying RbT1/2≥76 seconds.
Eleven patients displayed all three different tissue categories (normal, ischemically compromised, and scar) within one image set. The comparison of different tissue types in these 11 patients is presented in Table 3⇓. As for all patients, differences for Rb-82 uptake and Rb-82 tissue half-life between normal and scar tissue were highly significant (P<.0001). Again, in regions with reduced Rb-82 uptake, RbT1/2 differentiated between ischemically compromised and necrotic tissue. However, Rb-82 uptake in these two groups also was different. FDG utilization in regions defined as either normal or compromised was not significantly different and confirmed the previous visual definition for ischemically jeopardized regions.
Relationship Between Rb-82 Uptake and Rb-82 Tissue Half-life
There was a significant linear relationship between Rb-82 uptake and Rb-82 tissue half-life (r=.81, P<.0001). Fig 3⇓ illustrates the correlation between those two parameters in all regions with reduced perfusion. Inclusion of normal regions did not change the relationship (r=.86, P<.0001).
Comparison of Rb-82 Uptake and Rb-82 Tissue Half-life as Indicators of Absent or Maintained Tissue Viability
The following two criteria were chosen to compare the diagnostic value of Rb-82 uptake and of Rb-82 tissue half-life for evaluation of myocardial viability. The thresholds for preserved viability were RbUpt%≥50% of individual maximum30 and RbT1/2≥normal mean−2 SD.
These two criteria were applied to regions that had been classified by FDG utilization as either viable (>50%) or nonviable (≤50%).
Agreement between the different approaches is presented in Table 4⇓. For all regions, maintained viability as defined by FDG uptake was identified correctly in 76% by Rb-82 uptake and in 92% by Rb-82 tissue half-life. Consistent categorization of scar tissue was achieved in 100% and 71%, respectively. Overall agreement for viability categorization by FDG uptake was 86% with Rb-82 kinetics and 81% with Rb-82 uptake.
Rb-82 uptake identified 48% of regions with preserved FDG utilization when only regions with reduced perfusion were analyzed. In contrast, agreement between Rb-82 tissue half-life and FDG utilization was 86% for identification of maintained viability.
The data presented indicate that regional myocardial Rb-82 kinetics assessed by dynamic PET imaging allow differentiation between normal and irreversibly injured myocardium. Furthermore, in regions with reduced resting perfusion as defined from static Rb-82 uptake images, Rb-82 kinetics can distinguish ischemically compromised but metabolically active and thus potentially salvageable myocardium from scar tissue.
Rb-82 kinetics correlated with regional FDG uptake measurements in regions with reduced perfusion. This finding suggests a relationship between cell membrane function as assessed by Rb-82 tissue kinetics and preservation of intracellular glucose metabolism as independent indicators of maintained myocardial viability in the presence of reduced blood flow.
Using FDG uptake as the “gold standard” for tissue viability, Rb-82 agreed with this parameter in 86% for definition of scar tissue and viable myocardium.
This study was not designed to define the predictive accuracy of dynamic Rb-82 PET imaging for functional outcome after coronary revascularization but to evaluate the feasibility to assess tissue tracer kinetics compared with the established FDG method. Future studies in patients with severe left ventricular dysfunction must elucidate the predictive value of this approach for functional outcome.
Similar to other potassium analogues such as thallium-201, Rb-82 is rapidly extracted by the myocardium. Its uptake reflects the equilibration within the potassium pool with little activity clearance in the presence of intact cell membranes. Based on these physiological characteristics, Rb-82 is currently used as a flow tracer for the detection of regionally impaired coronary flow reserve in patients with suspected or known coronary artery disease.19 20 21 22 23 24 25 26 27
Since the intracellular retention of potassium and of Rb-82 requires active maintenance of ionic gradients across cell membranes, Rb-82 tissue kinetics allows study of the integrity of membrane function. The accelerated leakage of potassium from myocardium in the presence of coronary occlusion has been experimentally documented and correlated with early impaired cellular membrane function as well as with tissue necrosis in a canine model.37 Experimental studies by Goldstein28 29 evaluated Rb-82 as a marker of potassium fluxes. These investigations recorded regional tracer kinetics of Rb-82 using β-detectors placed on the epicardial surface of the heart in acutely ischemic animals after coronary occlusion and reperfusion. Rb-82 tissue activity increased over time in viable myocardium reflecting intracellular retention, while the tracer cleared rapidly from tissue defined as necrotic by histochemical staining methods.
Based on these observations, Gould et al30 proposed the clinical use of Rb-82 as an indicator of cell membrane integrity and myocardial viability. They used static data acquisition early and late after tracer injection to estimate the regional Rb-82 uptake. Infarct size based on these Rb-82 uptake and retention images correlated closely with infarct size and location on FDG images. The results suggested that loss of cell membrane integrity as indicated by accelerated Rb-82 clearance parallels loss of intracellular glucose metabolism.
These observations, obtained by comparison of two static images, are supported and further elucidated by our findings from dynamic PET imaging. In our study, the tissue half-life of Rb-82 exceeded the physical tracer half-life of 76 seconds as evidence for active tissue tracer retention in 96% of regions previously defined as normal. Regions with evidence of irreversible injury by FDG criteria displayed accelerated tracer clearance with tissue half-life times <76 seconds in 94% of regions.
However, few of our study patients were studied in the early subacute phase of myocardial infarction, which would most closely reflect the above-cited experimental condition. The Rb-82 kinetics in patients with chronic coronary artery disease and infarctions may reflect primarily the retention capacity of the tracer within a given region. Since blood flow to infarcted tissue may recover, as demonstrated in animal experiments using microspheres, the clearance rate of Rb-82 in these segments may correspond to the number of viable cells that can retain the tracer. Therefore, the Rb-82 kinetics may be best described by the distribution volume of the tracer in tissue as a marker for the available potassium space. However, the calculation of the distribution volume requires tracer kinetic modeling defining tracer delivery (K1) and clearance (K2). The PET camera system in this study did not allow for the fit of the entire dynamic study because of dead-time considerations, limiting the analysis to the estimates of tracer clearance rates only. The hypothesis that Rb-82 clearance rates may reflect the number of viable cells in one segment is supported by the good agreement with FDG data in regions with reduced perfusion.
An additional technical aspect is the fact that tissue clearance parameters are less affected by partial volume effects, which may be especially important in areas with infarcted myocardium. This may even offer advantages for the Rb kinetic analyses over the assessment of tissue FDG uptake.
The observed overlap of individual clearance rates, as illustrated in Fig 2⇑, is most likely a result of our approach to choose large regions for analysis. It cannot be excluded that ischemically compromised regions predominantly presented “hibernating” myocardium but that islets of necrotic tissue were also included in those regions and influenced the resulting signal (mild match). This explanation is supported by the finding that myocardial FDG utilization of all these regions was slightly but not significantly lower compared with normal regions. Because of the limited spatial resolution of the PET system used in this study, it was impossible to distinguish if intermediate values predominantly reflected truly accelerated washout (compared with normal regions) of not irreversibly damaged myocardium or if this pattern predominantly reflected a transmural mixture of different tissue types in the presence of a nontransmural infarction. However, despite this technical limitation, the data provide evidence that Rb kinetics in myocardium with reduced perfusion but maintained FDG utilization differs from those observed in irreversibly damaged myocardium and that this approach may be helpful in the diagnostic workup of patients with ischemic heart disease and reduced ventricular function. Future studies with advanced scanner systems and higher spatial resolution might provide further insights into the Rb-82 kinetics of hibernating myocardium.
The presented approach might be of particular interest for those nuclear facilities without on-site cyclotron, since Rb-82 can be eluted from a convenient generator system. It allows the combined assessment of perfusion by early tracer uptake measurement and assessment of viability by either serial imaging as proposed by Gould et al or dynamic data acquisition as performed in our study with washout analysis after only a single tracer injection. The polar maps, which could be created easily, allowed quantitation of defect sizes or regions with potentially salvageable myocardium. This might be of particular interest in patients with regional wall motion abnormalities to identify those in whom revascularization might result in significant functional improvement. Furthermore, the short half-life of Rb-82 allows repetitive studies in combination with exercise or pharmacological stress testing.24 25 38
Besides the logistic aspects of the Rb-82 generator, the combined assessment of myocardial perfusion and viability with a single tracer infusion offers clinical advantages with respect to patient throughput and possibly costs.39 Currently, the combined assessment of myocardial perfusion and metabolism is the most widely used PET imaging technique for assessment of myocardial viability.6 However, these study protocols need two tracer injections, with higher radiation exposure to the patient and longer imaging times. Furthermore, metabolic conditions such as the patient’s nutritional state or diabetes mellitus, which may limit the interpretation of FDG images, do not influence Rb-82 image quality. This may be of particular consequence for laboratories with high patient throughput, since individualized patient preparation is required for FDG imaging. Finally, the effect of regional heterogeneities of myocardial glucose utilization, limiting the diagnostic accuracy of FDG imaging, may be avoided.36 40 However, other metabolic conditions might influence myocardial Rb uptake and need further studies in patients with coronary artery disease.41
All patients in this study had reduced left ventricular function (ejection fraction ranged from 25% to 47%), and the majority had a history of previous myocardial infarction. Since PET viability imaging is clinically most useful in patients with markedly reduced ventricular function, a subanalysis was performed in 17 patients with left ventricular ejection fractions <40%. Eight of these 17 patients had all three tissue types within their individual PET data set. The results were comparable to the entire patient population, with an Rb-82 tissue half-life of 97±11 seconds in normal regions, 77±12 seconds in compromised regions, and 59±14 seconds in scar tissue regions. Similarly, Rb uptake values and FDG utilization did not differ substantially from the results observed in all patients.
This evaluation of Rb-82 kinetics must be validated in a prospective study in patients undergoing coronary revascularization with assessment of functional outcome at follow-up. Recently, the study by Yoshida and Gould,31 with analysis of the clinical follow-up in the patient population described 2 years earlier by Gould et al,30 indicated the clinical usefulness of Rb-82 in combination with PET imaging for assessment of myocardial viability and of patient prognosis.
The lack of follow-up data on regional or global left ventricular function somewhat limits the clinical conclusions that can be drawn from our results. Confirmation of improved wall motion in ischemic regions after coronary revascularization would certainly serve as the “gold standard” for assessment of myocardial viability. However, analysis of FDG uptake has been validated by follow-up studies in several laboratories as an accurate indicator of myocardial tissue viability.7 8 9 10 11 13 14 15 Therefore, we used the direct comparison between Rb-82 kinetics and FDG imaging to validate this dynamic imaging approach.
Normalization of regional FDG uptake to the individual region with maximal FDG utilization may provide results other than normalization to the region with highest or “normal” perfusion, as proposed by other groups.14 42 However, since normalized FDG utilization values did not differ significantly between normal and ischemically compromised regions, we do not expect that this would substantially change the results of the Rb-82 kinetic calculations obtained in this study.
Because of the count rate limitations of the system, it was not possible to include the first part of the dynamic data set into the fitting procedure. The maximal sensitivity of this PET scanner was approximately 12 000 prompts per second. At this count rate, the error caused by dead-time losses is less than 5%. The typical peak count rate of the Rb-82 study was approximately 15 000 prompts per second after 20 to 30 seconds of Rb-82 infusion in the basal planes of the left ventricle. The count rate fell rapidly after approximately 45 seconds as a result of the decreasing amount of activity infused by the generator and the short physical half-life of Rb-82 (Fig 1⇑). The typical count rate at 60 seconds was 5000 to 7000 prompts per second. Because of the high initial count rate, it was necessary to delay the start of the fitting routine until the blood activity had cleared from the myocardium. Consequently, it was not possible to measure the initial tracer delivery to tissue. Hence, an analytic approach with the use of a compartmental model such as the one proposed by Herrero et al43 44 could not be used, since this approach requires knowledge of blood and tissue time-activity curves during the initial part of the study.
The physical decay of Rb-82 is very fast compared with the duration of the imaging procedure. At the end of the 8-minute protocol, less than 3% of the administered activity is still present. As a result, large correction factors (factors of 10 to 100) are necessary to correct for physical decay. Such correction magnifies the noise contribution as a function of time. To reduce this influence, Rb-82 time-activity curves were not decay corrected. The calculated rate constant kt therefore represented both physical decay and tissue clearance rate and was subsequently converted into tracer tissue half-life.
Selection of different tissue types from visual analysis of flow-metabolism patterns for further quantitative analysis procedures was performed to provide pure tissue types—normal, viable but ischemic, and scar—with prognostic impact of functional outcome that have been validated in several clinical studies with functional follow-up evaluation.7 8 9 11 13 15 This approach was believed to represent the best way of comparing the different types of quantitative data (Rb-82 uptake and kinetics and FDG utilization) from the same region of interest selected to minimize mixtures of different tissue types.
The standardized presentation of Rb-82 and FDG data in a polar map display allowed direct comparison of both tracers using the same regions of interest. The software package written in our institution allowed interactive definition of regions in polar map coordinates, their storage and easy manipulation, and overlay between different studies. The advantage of this procedure lies in the convenient display of several polar maps side by side and the fact that regions defined in one of these polar maps can be automatically copied to all other polar maps.
This study demonstrated a significant relationship between cell membrane integrity as assessed by dynamic Rb-82 PET imaging and myocardial viability defined by FDG utilization. Myocardial tracer distribution and tissue half-life of Rb-82 in normal myocardium were homogeneous. In patients with coronary artery disease, Rb-82 kinetics was significantly different and in good agreement with FDG data in compromised but metabolically active and irreversibly injured myocardium. Further investigation with prospective evaluation of wall motion recovery after coronary revascularization is required to determine the prognostic impact of this approach in a clinical setting.
This study was performed during the tenure of an established investigatorship (M. Schwaiger) from the American Heart Association and was supported in part by the National Institutes of Health, Bethesda, Md (RO1-HL-41047-01). J. vom Dahl’s research fellowship at the University of Michigan was supported by Medical Clinic I (Department of Cardiology) at the University of Aachen, Germany. O. Muzik was recipient of a research grant from the Austrian Erwin Schroedinger Foundation (J0473-MED). The authors thank the University of Michigan Cyclotron Facility staff for preparation of F-18 fluorodeoxyglucose. The authors appreciate the excellent technical assistance of E. McKenna, A. Weeden, and J. Rothley in the PET suite.
- Received August 9, 1994.
- Revision received August 1, 1995.
- Accepted August 29, 1995.
- Copyright © 1996 by American Heart Association
Marshall RC, Tillisch JH, Phelps ME, Huang SC, Carson R, Henze E, Schelbert HR. Identification and differentiation of resting myocardial ischemia in man with positron computed tomography, 18F-labeled fluorodeoxyglucose, and N-13 ammonia. Circulation. 1983;67:766-778.
Camici P, Araujo L, Spinks T, Lammertsma AA, Jones T, Maseri A. Myocardial glucose utilization in ischemic heart disease: preliminary results with F18-fluorodeoxyglucose and positron emission tomography. Eur Heart J. 1986;7:19-23.
Schwaiger M, Brunken RC, Krivokapich J, Child JS, Tillisch JH, Phelps ME, Schelbert HR. Beneficial effect of residual anterograde flow on tissue viability as assessed by positron emission tomography in patients with myocardial infarction. Eur Heart J. 1987;8:981-988.
Carrel T, Jenni R, Haubold-Reuter S, von Schulthess G, Pasic M, Turina M. Improvement of severely reduced left ventricular function after surgical revascularization in patients with preoperative myocardial infarction. Eur J Cardiothorac Surg. 1992;6:479-484.
Marwick TH, MacIntyre WJ, Lafont A, Nemec JJ, Salcedo EE. Metabolic responses of hibernating and infarcted myocardium to revascularization. Circulation. 1992;85:1347-1353.
Gropler RJ, Geltman EM, Sampathkumaran K, Perez JE, Schechtman KB, Conversano A, Sobel BE, Bergmann SR, Siegel BA. Comparison of carbon-11-acetate with fluorine-18-fluorodeoxyglucose for delineating viable myocardium by positron emission tomography. J Am Coll Cardiol. 1993;22:1587-1597.
Knuuti MJ, Saraste M, Nuutila P, Haerkoenen R, Wegelius U, Haapanen A, Bergman J, Haaparanta M, Savunen T, Voipio-Pulkki L-M. Myocardial viability: fluorine-18-deoxyglucose positron emission tomography in prediction of wall motion recovery after revascularization. Am Heart J. 1994;127:785-796.
vom Dahl J, Eitzman DT, Al-Aouar ZR, Kanter HL, Hicks RJ, Schwaiger M. Relationship between regional function, perfusion, and metabolism in patients with advanced coronary artery disease undergoing surgical revascularization. Circulation. 1994;90:2355-2366.
Tamaki N, Kawamoto M, Takahashi N, Yonekura Y, Magata Y, Nohara R, Kambara H, Sasayama S, Hirata K, Ban T, Konishi J. Prognostic value of an increase in fluorine-18 deoxyglucose uptake in patients with myocardial infarction: comparison with stress thallium imaging. J Am Coll Cardiol. 1993;22:1621-1627.
DiCarli MF, Davidson M, Little R, Khanna S, Mody FV, Brunken RC, Czernin J, Rokshar S, Stevenson LW, Laks H, Hawkins R, Schelbert HR, Phelps ME, Maddahi J. Value of metabolic imaging with positron emission tomography for evaluating prognosis in patients with coronary artery disease and left ventricular dysfunction. Am J Cardiol. 1994;73:527-533.
Goldstein R, Mullani N, Wong W, Hartz R, Hicks C, Fuentes F, Smalling R, Gould K. Positron imaging of myocardial infarction with rubidium-82. J Nucl Med. 1986;27:1824-1829.
Gould K, Goldstein R, Mullani N, Kirkeeide R, Wong W, Tewson T, Berridge M, Bolomey L, Hartz R, Smalling R, Fuentes F, Nishikawa A. Noninvasive assessment of coronary stenoses by myocardial perfusion imaging during pharmacologic coronary vasodilation, VIII: clinical feasibility of positron cardiac imaging without a cyclotron using generator-produced rubidium-82. J Am Cardiol. 1986;7:775-789.
Gould KL. Identifying and measuring severity of coronary artery stenosis: quantitative coronary arteriography and positron emission tomography. Circulation. 1988;78:237-245.
Gould K, Goldstein R, Mullani N, Kirdeeide R, Wong WT, Berridge M, Bolomey L, Hartz R, Smalling R, Fuentes F, Nishikawa A. Clinical feasibility, sensitivity and specificity of positron cardiac imaging without a cyclotron using generator produced Rb-82 for the diagnosis of coronary artery disease. J Nucl Med. 1986;27:976. Abstract.
Demer L, Gould K, Goldstein R, Kirkeeide R, Mullani N, Smalling R, Nishikawa A, Merhige M. Assessment of coronary artery disease severity by positron emission tomography: comparison with quantitative arteriography in 193 patients. Circulation. 1989;79:825-835.
Go RT, Marwick TH, MacIntyre WJ, Saha GB, Neumann DR, Underwood DA, Simpfendorfer CC. A prospective comparison of rubidium-82 PET and thallium-201 SPECT myocardial perfusion imaging utilizing a single dipyridamole stress in the diagnosis of coronary artery disease. J Nucl Med. 1990;31:1899-1905.
Goldstein R. Kinetics of rubidium-82 after coronary occlusion and reperfusion: assessment of patency and viability in open-chested dogs. J Clin Invest. 1985;75:1131-1137.
Goldstein RA. Rubidium-82 kinetics after coronary occlusion: temporal relation of net myocardial accumulation and viability in open-chested dogs. J Nucl Med. 1986;27:1456-1461.
Gould KL, Yoshida K, Hess MJ, Haynie M, Mullani N, Smalling RW. Myocardial metabolism of fluorodeoxyglucose compared to cell membrane integrity for the potassium analogue rubidium-82 for assessing infarct size in man by PET. J Nucl Med. 1991;32:1-9.
vom Dahl J, Herman WH, Hicks RJ, Ortiz-Alonso FJ, Lee KS, Allman KC, Wolfe ER, Kalff V, Schwaiger M. Myocardial glucose uptake in patients with insulin-dependent diabetes mellitus assessed quantitatively by dynamic positron emission tomography. Circulation. 1993;88:395-404.
Kotzerke J, Hicks RJ, Wolfe E, Herman WH, Molina E, Kuhl DE, Schwaiger M. Three-dimensional assessment of myocardial oxidative metabolism: a new approach for regional determination of PET-derived carbon-11-acetate kinetics. J Nucl Med. 1990;31:1876-1893.
Muzik O, Beanlands R, Wolfe E, Hutchins G, Schwaiger M. Automated region definition for cardiac nitrogen-13-ammonia PET imaging. J Nucl Med. 1993;34:336-344.
Herrero P, Hartman JJ, Senneff MJ, Bergmann SR. Effects of time discrepancies between input and myocardial time-activity curves on estimates of regional myocardial perfusion with PET. J Nucl Med. 1994;35:558-566.
Hill JL, Gettes LS. Effect of acute coronary artery occlusion on local myocardial extracellular K+ activity in swine. Circulation. 1980;61:768-777.
Camici P, Araujo LI, Spinks T, Lammertsma AA, Kaski JC, Shea MJ, Selwyn AP, Jones T, Maseri A. Increased uptake of F-18 deoxyglucose in postischemic myocardium of patients with exercise-induced angina. Circulation. 1986;74:81-88.
Patterson RE, Eisner RL, Horowitz SF. Comparison of cost-effectiveness and utility of exercise ECG, single photon emission tomography, positron emission tomography, and coronary angiography for diagnosis of coronary artery disease. Circulation. 1995;91:54-65.
Gropler RJ, Siegel BA, Lee KJ, Moerlein SM, Perry DJ, Bergmann SR, Geltman EM. Nonuniformity in myocardial accumulation of F-18 fluorodeoxyglucose in normal fasted humans. J Nucl Med. 1990;31:1749-1756.
Altehoefer C, vom Dahl J, Biedermann M, Sheehan FH, Beilin I, Uebis R, Hanrath P, Buell U. Significance of defect severity in technetium-99m-MIBI SPECT at rest to assess myocardial viability: comparison with fluorine-18-FDG PET. J Nucl Med. 1994;35:569-574.
Herrero P, Markham J, Shelton ME, Weinheimer CJ, Bergmann SR. Noninvasive quantification of regional myocardial perfusion with rubidium-82 and positron emission tomography. Circulation. 1990;82:1377-1386.
Herrero P, Markham J, Shelton ME, Bergmann SR. Implementation and evaluation of a two-compartment model for quantification of myocardial perfusion with rubidium-82 and positron emission tomography. Circ Res. 1992;70:496-507.