Assessment of Myocardial Viability by Use of 11C-Acetate and Positron Emission Tomography
Threshold Criteria of Reversible Dysfunction
Background Dual positron emission tomography (PET) imaging with a perfusion tracer and 18F-fluorodeoxyglucose (FDG) can detect myocardial viability. This approach may be replaced by a single 11C-acetate study, which enables quantification of both regional blood flow and oxidative metabolism. The significance of acetate-derived indexes for myocardial viability is examined.
Methods and Results Thirty postinfarct patients with akinetic ventricular segments, a mean ejection fraction of 42±11%, and high-grade coronary obstructions were studied with serial 11C-acetate PET scanning before and 7±5 months after coronary revascularization. Acetate PET was tested against FDG and serial assessments of segmental wall motion. Sixty of 155 severely dysfunctional LV segments improved postoperatively, and regional blood flow increased. Flow estimates after revascularization suggested little fibrosis in reversible segments. At baseline, blood flows differed between normal myocardium, reversible dysfunction, and irreversible dysfunction (1.04±0.27, 0.73±0.18, and 0.43±0.18 mL·min−1·g−1, respectively; P<.001). Oxidative metabolic rates were reduced only in irreversibly injured LV segments. Multivariate analysis identified the acetate perfusion index as the only independent predictor of postoperative recovery. Its predictive accuracy was similar to that of FDG imaging but superior to indexes of flow-metabolic mismatch or oxidative metabolism.
Conclusions After myocardial infarction, quantitative indexes of perfusion and oxidative metabolism from acetate PET indicate a critical threshold beyond which tissue is irreversibly injured. Findings support the use of blood flow as a marker of myocardial viability in chronic postinfarct patients with modestly reduced ejection fractions.
Myocardial viability is an important clinical issue in patients with chronic myocardial infarction and LV dysfunction who are considered for coronary revascularization. Among the radionuclide imaging techniques developed for the detection of reversible dysfunction, single-photon tomography with a flow tracer and PET with 18F-FDG are most widely used.1 Because glucose is the preferred substrate under ischemic conditions, dual-tracer PET imaging of glucose uptake and myocardial perfusion is considered the standard of reference for the detection of residual metabolic activity in ischemic myocardium.
Even with the concept of flow-metabolism mismatch, however, predictive accuracies have been reported to vary between 60% and 90%, probably due to differences in patient selection, processing of data, and interventional success at follow-up.2 3 4 5 Despite its clinical success, limitations of the FDG approach accrue from its substrate dependency and from other complexities of metabolic control.6 11C-Acetate has been proposed as an alternative marker with minimal substrate dependency. Apart from possible advantages in diabetic patients and early after acute myocardial infarction, it is attractive as a single-tracer technique that yields quantitative data on flow and metabolism and may help curtail the use of double-tracer studies.5 7 8 It remains unclear, however, which marker is most reliable for the assessment of viability.
It was the objective of the present study to compare 18F-FDG and 11C-acetate as markers of reversible dysfunction in a group of patients with subacute or chronic myocardial infarctions, severe regional wall-motion abnormalities, and obstructed but collateralized infarct vessels. Adequate revascularization was documented in all patients by follow-up angiography or PET. In contrast to current clinical practice, all PET data were processed quantitatively by use of tracer kinetic modeling. Using repeated PET and wall-motion studies before and after revascularization, we attempted to define threshold criteria of MBF and metabolism for accurate prediction of reversible dysfunction.
Patients and Study Protocol
The study population was assembled during a 6-month period from a larger group of patients with chronic myocardial infarctions who were considered for coronary revascularization at our institution. Patients were eligible if large, severe wall-motion abnormalities were present in the supply area of ≥1 totally or subtotally occluded coronary arteries and if there was angiographic evidence of some residual or collateral flow. Patients with an acute myocardial infarction or with insulin-dependent diabetes mellitus were excluded. Limited PET examination capacities further contributed to patient selection. Forty-one consecutive patients were initially enrolled in the study, of whom 1 eventually received a heart transplant, 4 were treated medically, 3 had recurrent coronary syndromes or reocclusions before completion of the study, and 3 refused postoperative reexaminations. The final study population therefore consisted of 30 patients (aged 55±18 years) who were successfully revascularized and had a complete postoperative follow-up. This included serial wall-motion studies and PET metabolic and perfusion studies before and after coronary revascularization. The adequacy of revascularization was verified by follow-up angiography in 20 patients and by follow-up PET (restored blood flow) in the remaining 10. All patients were stable at enrollment and had an uneventful course until completion of the study. In 23 patients, myocardial infarction had occurred between 4 weeks and several years before enrollment; 7 patients were studied within the first 4 weeks after infarction. Nine patients had intravenous thrombolysis. ECG infarct locations were anterior in 13 patients, lateral in 3, inferior in 7, and anterior plus lateral or inferior in 7. One-, two-, or three-vessel disease (>50% stenosis) was present in 14, 10, and 6 patients, respectively.
Coronary revascularization was performed by bypass grafting in 21 patients and by coronary angioplasty in 9. Segments that were considered nonviable by the initial PET study were not necessarily excluded from revascularization, especially in multivessel disease. Unsuccessfully revascularized segments were later excluded from follow-up analysis. Segmental LV function, blood flow, and metabolism were studied before and 6.9±5.1 months after revascularization (range, 2 to 21 months). The assessment of regional LV function was based on preoperative and postoperative contrast angiograms in 20 patients who had a follow-up catheterization and on paired echocardiograms in the remaining 10 patients. MBF and oxidative metabolism were assessed preoperatively and postoperatively by dynamic 11C-acetate–PET scanning. Exogenous glucose utilization was assessed by 18F-FDG in the preoperative study. PET and wall-motion studies were performed within a few days of one another. The local ethics committee reviewed and approved the protocol, and informed consent was obtained from each patient.
Coronary Arteriography and Evaluation of LV Function
Routine coronary arteriography was performed with a nonionic contrast agent on a biplane Philips DCI system after intracoronary injection of nitroglycerin. LVEF was calculated from the angiogram with the use of the standard Simpson’s rule. Premature and postextrasystolic beats were excluded. The evaluation of regional LV performance before and after coronary revascularization was based on paired biplane contrast angiograms or on two-dimensional echocardiograms with the use of four- and two-chamber long-axis and short-axis views. For the purpose of analysis, the preoperative and postoperative studies were displayed in parallel and the LV wall was segmented into 12 regions (anterior, inferior, lateral, and septal, each divided into a basal, mid, and apical portion from the 30° right anterior oblique and 60° left anterior oblique angiogram). Segmental wall motion was visually scored on a four-point scale3 9 (normal, mildly hypokinetic, severely hypokinetic, and akinetic or dyskinetic) by one experienced observer who was blinded to the PET data.
PET Image Acquisition
18F-FDG and 11C-acetate were prepared as previously described.10 Radionuclide imaging was performed with the Hannover Medical School ECAT 951/31 PET scanner (Siemens/CTI). The resolution was ≈9 to 10 mm full width of half maximum (31 slices, plane separation 3.4 mm, 128×128 matrix, Hann filter cutoff 0.3). After positioning of the patient, a 20-minute transmission scan was acquired. Thereafter, 1.1 gBq 11C-acetate was injected intravenously as a slow bolus and dynamic scanning was started (10×10 seconds, 1×60 seconds, 5×100 seconds, 3×180 seconds, and 4×300 seconds). In the preoperative study, 370 MBq 18F-FDG was injected 1 hour later and 12 5-minute frames were recorded. Subjects were studied after an oral glucose load (50 g). Venous glucose levels, arterial blood pressures, and heart rates were determined twice during each study to ensure maintenance of a steady state.
PET Image Analysis
The PET studies were analyzed with the use of Siemens/CTI software. The tomographic data were corrected for attenuation and radioactive decay and the ventricles reorientated into six short-axis slices, which were later combined to three. The tomographic slices of the preoperative and postoperative PET studies were carefully matched to avoid misalignment between individual studies. The tracer input functions were determined from a region inside the LV. Twelve myocardial ROIs were generated by dividing each of the short-axis slices into quarters, which were identical to the 12 segments used for wall-motion analysis. Indexes of segmental flow and metabolism were computed from these regions. At least 2 remote segments per patient were defined as having maximal flow, normal wall motion, and a normal coronary angiogram.
Regional MBFs and oxidative metabolic rates were derived from 11C-acetate uptake and clearance rate constants (K1 and k2, respectively), as described elsewhere in detail.11 12 Acetate kinetics were evaluated during the first 20 minutes by use of a one-tissue-compartment model, which corrects for tracer recirculation, fractional blood volume, and spillover activity from blood pool to tissue (see “Appendix A” and “Appendix B”). The glucose metabolic rate (MRGlc) was assessed by Patlak’s plot procedure using frames 5 to 12 of the FDG scan and a lump constant of 0.67. This method does not correct for partial volume effects. In three patients, the quality of the FDG images was insufficient for quantification. MBF is given in milliliters per minute and gram, oxidative metabolism as rate constant per minute, and glucose utilization in micromoles per minute and gram. Metabolic rate to blood flow ratios were calculated relative to the remote segments. Two different definitions of a flow-glucose mismatch pattern were considered, namely, the ratio and the difference of the MRGlc and blood flow.
All data were expressed as mean±SD. Differences between means were analyzed by use of Student’s t test, if applicable. The relationships between PET indexes and the degree of segmental dysfunction at baseline were analyzed by use of ANOVA, and the relationships between these indexes and the reversibility of dysfunction were analyzed by use of a logistic regression analysis. Probability values <.05 were considered significant.
LV function was studied before and >2 months (mean, 6.9 months) after revascularization by repeat contrast angiograms or two-dimensional echo. LVEF averaged 42.5±11.5% preoperatively and increased by 5.8±7.2% (range, −4% to 23%) postoperatively in patients who had a second contrast angiogram. Preoperative wall motion could be assessed in 351 of 360 segments and follow-up wall motion in 337 segments. Wall motion was rated as normal, mildly hypokinetic, severely hypokinetic, and akinetic or dyskinetic in 151, 45, 60, and 95 segments, respectively. After revascularization, wall motion improved by 1 grade in 34 segments with severe hypokinesis or akinesia at baseline and by ≥2 grades in 26 segments. Improvement in ≥2 segments was accompanied by a 9.5±7.0% increase in LVEF (versus 0.4±2.8% in the remaining patients; P<.01). Increases in LVEF were related to the number of improved segments, being ≈5% in patients with 2 improved segments and 10% in those with 4 improved segments (linear correlation r=.67). At the time of infarction, peak levels of serum creatine kinase were lower in patients with reversible dysfunction than in those with an unchanged function (P<.05; Table 1⇓). Q-wave infarctions were present in all patients with unchanged function and in 13 of the 17 patients with reversible dysfunction; 4 had anterior non–Q-wave infarctions. The locations of functional improvement corresponded with that of ECG infarction in all patients. They were anterior in 11 patients (7 Q-wave, 4 non–Q-wave), inferior in 1, anterior and inferior in 2, and inferolateral in 3 (all Q-wave). The other clinical variables, such as the number of diseased vessels, LVEF, and pressure-rate product, were similar in both groups by multiple regression analysis.
Regional MBF and Metabolism
MBFs and metabolic rates were assessed on a segmental basis parallel to wall motion (Table 2⇓). Remote segments with maximal flow were used for intraindividual normalization of absolute measurements. Remote blood flows, oxidative metabolic rates, and MRGlc at baseline were 1.16±0.27 mL·min−1·g−1, 0.122±0.038 per minute, and 0.29±0.10 μmol·min−1·g−1, respectively (at serum glucose levels of 6.3±1.7 mmol/L). Intraindividual SDs were 9% for blood flow, 7% for the oxidative metabolic rate, and 8% for the MRGlc.
As anticipated, flow and metabolic parameters were related to the segmental function at baseline. They gradually decreased with the degree of dysfunction. However, quantitative expressions of the mismatch pattern were only weakly correlated with baseline wall motion (ie, MRGlc/MBF, MRGlc−MBF). Corresponding to the ECG sites of infarcts, the PET images revealed concordant perfusion FDG defects in all patients.
Baseline levels of blood flow and metabolism were also related to the occurrence of later improvement of wall motion by ≥1 grade at follow-up. This was observed in the clinically important subgroup of severely hypokinetic or akinetic segments that were successfully revascularized (Table 3⇓). There was a 35% reduction in baseline segmental blood flow (to 0.73±0.18 mL·min−1·g−1) in segments with later improvement, compared with a 60% reduction (to 0.43±0.18 mL·min−1·g−1) in those without recovery. A small subset of segments with postoperatively normalized wall motion had a 25% reduction in baseline flow and normal glucose uptake. The acetate-derived metabolic rate was almost normal in reversible dysfunction (0.105±0.024 per minute). Baseline variables were not different if wall motion improved by just 1 or by >1 grade at follow up. In general, regional metabolic rates and blood flows were more markedly depressed in segments with persistent dysfunction by univariate analysis. Stepwise multivariate testing revealed segmental blood flow at baseline to be the only independent determinant of functional recovery after revascularization.
The relationship between regional metabolic rates and blood flows is shown in Figs 1⇓ and 2⇓. There was no correlation between glucose utilization and blood flow except in irreversibly injured segments, in which both were reduced proportionately (r=.54). The majority of segments with later improvement lay above this regression line. Accordingly, their MRGlc relative to blood flow (MRGlc/MBF) tended to be higher but overlapped with those of the irreversibly injured segments (1.49±1.17 versus 1.16±0.49). The overall relationship of k2 and regional blood flows was sigmoidal in shape (Fig 2⇓). The correlation was steep and almost linear for blood flows <0.5 and >1 mL·min−1·g−1. The low-flow range included most of the irreversibly injured segments. The correlation was less steep in the intermediate-flow range, which included most of the segments with reversible dysfunction.
Follow-up PET studies were available in all patients. At follow-up, there were minor decreases of flow and metabolism in normal segments (Table 2⇑). However, there were increases of blood flow in dysfunctional segments by ≈0.2 mL·min−1·g−1 (P<.00001). Increases were present in segments with and without functional recovery but led to almost normal perfusion levels only in segments with reversible dysfunction. Values displayed in Table 3⇑ remained essentially unchanged irrespective of the inclusion of septal segments.
Predictive Accuracy of PET Indexes
Fig 3⇓ displays paired cumulative incidences of severely dysfunctional segments with and without functional improvement as continuous functions of PET indexes of perfusion and metabolism. The selectivity of an index for predicting postoperative outcome is represented by the distance between curves for reversible and irreversible dysfunction. Cutoff values with maximal discriminative power have maximal ordinate differences between a pair of curves. When 0.5 mL·min−1·g−1 (flow) and 50% (FDG) were used as cutoff values, the positive and negative predictive accuracies were 79% and 90% for blood flow and 78% and 85% for FDG uptake, respectively (Table 4⇓). Lower positive and negative predictive accuracies were calculated for k2 (62% and 65%, cutoff at 0.09 per minute), the differential mismatch pattern MRGlc−MBF (74% and 58%), and the ratio of FDG uptake to blood flow (56% and 55%).
MBF and the Return of Function
Based on a more qualitative image analysis, the mismatch pattern of glucose uptake and blood flow has been used for detection of reversibly injured myocardium. Because quantification of viable myocardium is important,1 the present study compared several quantitative indexes of cardiac metabolism and blood flow. Estimates of absolute perfusion from 11C-acetate kinetics were identified as the only independent predictors of reversible dysfunction in patients with chronic myocardial infarctions and modest reductions in ejection fraction.
Our data show that a postoperative return of function is unlikely if transmural blood flow is reduced below 0.5 mL·min−1·g−1. In a pig model of short-term hibernation,13 no myocardial infarction developed if microsphere blood flow was >0.34 mL·min−1·g−1. A clinical study by Gewirtz et al14 reported a similar threshold level for ammonia-derived flow, although no follow-up data were presented. A higher cutoff blood flow (0.64 mL·min−1·g−1) was more recently reported in patients with ischemic anterior wall dysfunction.15 The latter study reported a similar prognostic relevance for estimates of absolute blood flow and glucose uptake. Differences between threshold levels are probably due to different flow tracers and handling of finite resolution effects. These were treated by a priori corrections for regional wall thickening in one study15 and by a kinetic modeling approach in ours.
Perfusion estimates, which are derived from tracer uptake (such as acetate or ammonia), represent average flow of fibrotic and viable myocardium, the relative contributions of which cannot be distinguished within a single measurement. In contrast to the above-cited literature, however, our study also included perfusion measurements after revascularization. Postoperatively, ischemia has been relieved and the residual flow deficit is likely to reflect only the fraction of irreversible fibrosis, which can thus be assessed retrospectively. In functionally improved segments, relative blood flows were higher at follow-up (87±21% of remote segments) than in unchanged segments (63±29%), with a cutoff near 75%. If the persistent relative flow deficit after revascularization does reflect transmural fibrosis, these data suggest potential functional reversibility in segments with ≤25% fibrosis. Such values are in good accordance with values reported for the water-perfusible tissue index16 or with the amount of fibrosis seen in biopsy samples of PET-viable tissue with and without return of function.17 18
The perfusion state of the nonfibrotic tissue fraction has been investigated selectively by use of H215O PET,19 which measures flow in perfusible tissue, and by ammonia PET in noninfarcted but dysfunctional myocardium.20 In both trials, absolute flows were within normal limits, suggesting that reduced function was not a consequence of chronic hypoperfusion. Again, myocardial perfusion was not restudied after revascularization. Using serial PET scanning, however, we were able to observe a highly significant increase of absolute resting blood flow in dysfunctional myocardium after revascularization that almost reached normal postoperative levels in reversible segments. This would suggest chronic hypoperfusion in at least part of the dysfunctional segments, adding to the reduced uptake of flow markers due to (mild) fibrosis. Clearly, our findings are at variance with the above literature, which may possibly be due to differences in stenosis severity and tracer kinetic methods. Reduced resting flow levels may reflect a more critical stenosis severity. In addition, the low sensitivity of clearance estimates (as of H215O) to partial volume effects may contribute to the differences, leading to higher than average estimates in heterogeneously perfused tissue (see “Appendix C”).
The finding that baseline blood flow and metabolic rate constants were significantly higher in reversible dysfunction supports the observation that functional recovery depends on the preservation of MBF at a level sufficient to maintain oxidative metabolism.21 22 However, outcome was quite variable in moderately hypoperfused segments. It is hypothesized that dysfunctional segments that did not improve despite a transmural flow >0.5 mL·min−1·g−1 may have contained viable but substantially altered cells. Biopsy studies have shown that severe subcellular alterations may prevent contractile recovery despite the absence of significant fibrosis in long-term collateral-dependent myocardium.17
Relationship of MBF and Metabolism
The relationships between regional blood flows and metabolic rate constants support the above threshold values. Segments classified as nonviable typically showed close correlations between glucose or oxidative metabolic rates and blood flows. However, viable segments had a variable relationship between metabolism and flow. In the case of glucose, this is likely to be due to a preferential uptake in reversibly ischemic myocardium, which was most prominent in the small subset of segments that had entirely normalized wall motion postoperatively (Table 3⇑).
Oxidative metabolism and perfusion are known to be closely coupled in normal myocardium. However, the present data suggest that oxidative metabolism was less markedly reduced than blood flow in moderately hypoperfused myocardium. This confirms an observation by Czernin and coworkers9 but differs from the linear correlation seen by others.23 Possible explanations for the observed divergence of metabolism and perfusion include an increased oxygen extraction in ischemic tissue, as was shown by experimental data.24 However, effects cannot be explained by oxygen extraction alone. The fact that flow estimates (from tracer uptake) are more sensitive to partial volume effects than metabolic estimates (from tracer clearance) is likely to play an additional role, as shown in “Appendix C.” This effect may also account, at least in part, for the lack of predictive value for estimates of oxidative metabolism. But whatever the relative contributions of methodological and physiological factors may be, the observed nonlinearity of the metabolism-flow relationship appears to indicate the presence of heterogenous, ie, nontransmural, injury.
There are several limitations to the present study. Some are related to the highly selected study group. First, it cannot be ruled out that the initial PET study influenced the surgical strategy in some patients. This and the exclusion of segments with nonfunctional bypass grafts may explain the high prevalence of postinterventional improvement. Second, postoperative improvement of wall motion has been attributed mainly to myocardial “hibernation,” but effects of stunning cannot be excluded. Such effects, however, are thought to be of minor importance >4 weeks after infarction, as shown by experimental and clinical studies.25 26 Third, although all patients had akinetic segments, global LV function was only mildly impaired (LVEF of 42% on average). Whether our results are applicable in patients with lower ejection fractions will be subject to future studies. A preliminary report has suggested that in such patients, metabolic criteria may be advantageous over blood flow if a glucose clamp technique is used.27
In addition, technical problems must be mentioned. These include the possibility of image misalignments between serial PET and wall-motion studies. To minimize this source of error and to facilitate the definition of segments, the PET images were reorientated with the use of standard software. Second, regional wall motion was analyzed semiquantitatively. Although clearly less suitable for interindividual analysis, this approach is considered useful for comparisons between serial studies of the same patients analyzed in parallel. Third, accurate quantification of tracer concentrations by PET is impeded by variable LV wall thickness and cardiac motion in ungated studies. Such difficulties cannot be completely resolved with the present equipment.
Predictive Accuracy of PET Viability Markers
Because glucose metabolism predominates in ischemic myocardium, the so-called mismatch pattern between FDG uptake and perfusion has been regarded as a hallmark of residual viability. In view of the clinical success of the concept, its predictive power was lower than expected in the present study, ranging at the lower end of reported values.2 3 4 5 In fact, estimated glucose uptake had a higher predictive accuracy than measures of mismatch, which we suspect reflects the importance of absolute residual flow. A more prominent mismatch pattern was observed only in a small subset with entirely normalized wall motion after surgery, presumably due to the absence of significant fibrosis.
To the best of our knowledge, this is the first study that has tested the predictive accuracy of acetate-derived measures of absolute blood flow and oxidative metabolism as viability markers against serial assessments of LV wall motion before and after coronary revascularization. Using a more qualitative approach without tracer kinetic modeling, Gropler et al5 reported an acceptable accuracy only for estimates of oxidative metabolism. Our data, however, suggest that 11C-acetate provides viability information predominantly as a marker of blood flow with little additive information from oxidative rate constants, which is consistent with the known coupling of perfusion and metabolism. Indeed, perfusion estimates, which per se are not unique to 11C-acetate, had a predictive accuracy in the present study similar to FDG imaging but were clearly superior to indexes of an FDG-perfusion mismatch or the oxidative metabolic rate.
Flow as a measure of viability is documented in the present study for patients with only modest reductions in ejection fraction, reflecting less severe or a smaller extent of ischemic damage or reflecting a mixture of necrotic with hibernating and stunned myocardium. Other reports favoring metabolic indexes of viability over flow measurement may reflect selection of patients with more severe impairment or selection of patients with predominantly hibernating myocardium and very little fibrosis. The present study does not contravene metabolic imaging for assessing viability but rather provides complementary data on effective alternatives within the spectrum of moderate to severe LV impairment and the variable mix of scar, hibernating, and stunned myocardium characterizing many, if not most, postinfarction patients.
Selected Abbreviations and Acronyms
|K 1||=||11C-acetate uptake rate constant|
|k 2||=||11C-acetate clearance rate constant|
|LV||=||left ventricle, left ventricular|
|LVEF||=||left ventricular ejection fraction|
|MBF||=||myocardial blood flow|
|MRGlc||=||metabolic rate of glucose uptake|
|PET||=||positron emission tomography|
|ROI||=||region of interest|
Perfusion Estimates by 11C-Acetate
Myocardial 11C-acetate clearance rate constants have been validated as measures of myocardial oxidative metabolism.28 29 In addition, it has been shown that regional 11C-acetate activity early after injection reflects the relative distribution of perfusion.30 31 We used a simplified one-tissue-compartment model11 12 and converted the uptake rate constant (K1) into flow by use of an experimentally determined extraction-flow relationship.28
The flow estimates from 11C-acetate were compared with 13N-ammonia (as reference tracer) in nine postinfarct patients, three of whom were studied after administration of adenosine 140 μg·kg−1·min−1 IV. The ammonia scan was performed 1 hour after the acetate scan; the ammonia kinetic model followed that of Hutchins et al.32 Flow estimates from ROIs placed identically in the acetate and ammonia images were linearly correlated over a wide range of flows (Fig 4⇓).
It is common to both tracers that the quantification of myocardial uptake and hence flow is severely affected by the limited spatial resolution of the scanner and by variations in wall thickness. There are two possible approaches to correct for limited recovery and spillover of activity from the LV cavity to the myocardium. The first is to correct the PET data of a myocardial ROI for partial volume and spillover effects before evaluation by a kinetic model.23 However, such a strategy requires knowledge of acquisition parameters, true object size, and averaging effects in ungated studies. Alternatively, we did not correct the PET data a priori for limited recovery but treated total blood volume (tbv) formally as fractional blood volume, even though spillover does not correspond to a fraction of the observed tissue volume, by use of the following equation: The parameter of the total blood volume (tbv) is given by tbv=fr×fbv+fs, where fr is fractional recovery in tissue ROI, fbv is fractional blood volume, fs is fractional spillover into tissue ROI, and c is tracer activity. Because fr×fbv is significantly smaller than fs, the fitted value of tbv yields approximately the recovery coefficient fr≈1−tbv, which is thus automatically included in the equation as a correction factor multiplying ctissue. Error estimates were obtained from simulations performed for a resolution of 9 mm and a wall thickness of 10 mm. On average, recovery was 69% and spillover was 20%. The average error was ≈10% of the true tracer uptake rate or flow. Additional a priori corrections for limited recovery at this point would have caused significant overestimation. Similar results were reported in a recent paper by Nuyts et al.33 The authors concluded that inclusion of total blood volume in the kinetic modeling is a sufficient means of accounting for limited recovery.
Evaluation of PET studies is based on the assumption of a homogeneous response of the region under investigation. To assess the influence of tissue heterogeneity on quantitative estimates with 11C-acetate, the tissue-response curves of two regions representing normal and infarcted tissue were calculated from the analytical solution of the model equations. A typical input function from the present study was used for the calculations. Various weighted averages of the response curves (representing ROIs composed of the two tissue types) were then fitted, and the relation between the resulting parameters and the true parameters underlying the simulation was determined. The main result is the following.
The fitted uptake parameter K1 (and thus flow) is nearly always equal to the weighted average of the individual uptake parameters. In contrast, the fitted washout rate k2 is not reduced in proportion to K1 if flow in the low-flow component is <40% of normal values. Nevertheless, the decreased washout rate (k2) from a heterogeneous ROI was still detectable even if flow and metabolism in the low-flow component were reduced by a factor of 10. The simulations shown in Fig 5⇓ reproduce the nonlinear shape of the clinically observed relation between oxidative rate constants (k2) and regional blood flows <1 mL·min−1·g−1.
- Received July 1, 1996.
- Revision received October 24, 1996.
- Accepted November 18, 1996.
- Copyright © 1997 by American Heart Association
Dilsizian V, Arrighi JA, Diodati JG, Quyyumi AA, Alavi K, Bacharach SL, Marin-Neto JA, Katsiyiannis PT, Bonow RO. Myocardial viability in patients with chronic artery disease: comparison of 99mTc-sestamibi with thallium reinjection and [18F]fluorodeoxyglucose. Circulation. 1994;89:578-587.
Marwick TH, MacIntyre WJ, Lafont A, Nemec JJ, Salcedo EE. Metabolic responses of hibernating and infarcted myocardium to revascularization: a follow-up study of regional perfusion, function, and metabolism. 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.
Hariharan R, Bray M, Ganim R, Doenst T, Goodwin GW, Taegtmeyer H. Fundamental limitations of [18F]2-deoxy-2-fluoro-d-glucose for assessing myocardial glucose uptake. Circulation. 1995;91:2435-2444.
Schwaiger M, Wolpers HG. Advances in the assessment of myocardial metabolism by positron emission tomography. Coron Artery Dis. 1990;1:547-555.
Czernin J, Porenta G, Brunken R, Krivokapich J, Chen K, Bennett R, Hage A, Fung C, Tillisch J, Phelps ME, Schelbert HR. Regional blood flow, oxidative metabolism, and glucose utilization in patients with recent myocardial infarction. Circulation. 1993;88:884-895.
Meyer GJ, Günter K, Matzke KH, Harms T, Hundeshagen H. A modified preparation method for 11C-acetate preventing liquid phase extraction steps. J Labelled Comp Radiopharm. 1993;32:182-183.
Buck A, Wolpers HG, Hutchins GD, Savas V, Mangner TJ, Nguyen N, Schwaiger M. Effect of C-11-acetate recirculation on estimates of myocardial oxygen consumption by PET. J Nucl Med. 1991;32:1950-1957.
van den Hoff J, Burchert W, Wolpers HG, Meyer GJ, Hundeshagen H. A kinetic model for cardiac PET with [1-carbon-11]-acetate. J Nucl Med. 1996;37:521-529.
Schulz R, Rose J, Martin C, Brodde OE, Heusch G. Development of short-term myocardial hibernation. Circulation. 1993;88:684-695.
Gewirtz H, Fischman AJ, Abraham S, Gilson M, Strauss HW, Alpert NM. Positron emission tomographic measurements of absolute regional myocardial blood flow permits identification of nonviable myocardium in patients with chronic myocardial infarction. J Am Coll Cardiol. 1994;23:851-859.
Grandin C, Wijns W, Melin JA, Bol A, Robert AR, Heyndrickx GR, Michel C, Vanoverschelde JJ. Delineation of myocardial viability with PET. J Nucl Med. 1995;36:1543-1552.
Yamamoto Y, DeSilva R, Rhodes CG, Araujo LI, Iida H, Rechavia E, Nihoyannopohlos P, Hackett D, Galassi AR, Taylor CJV, Lammertsma AA, Jones T, Maseri A. A new strategy for the assessment of viable and regional myocardial blood flow using 15O-water and dynamic positron emission tomography. Circulation. 1992;86:167-178.
Flameng W, Shivalkar B, Borgers M. Myocardial viability: stunning and hibernation. In: van der Wall EE, Blanksma PK, Niemeyer MG, Paana AMJ, eds. Cardiac Positron Emission Tomography. Dordrecht, Netherlands: Kluwer Academic Publishers; 1995:15-24.
Depre C, Vanoverschelde JJ, Melin JA, Borgers M, Bol A, Ausma J, Dion R, Wijns W. Structural and metabolic correlates of the reversibility of chronic left ventricular ischemic dysfunction in humans. Am J Physiol. 1995;268:H1265-H1275.
Marinho NVS, Keogh BE, Costa DC, Lammertsma AA, Ell PJ, Camici PG. Pathophysiology of chronic left ventricular dysfunction: new insights from the measurement of absolute myocardial blood flow and glucose utilization. Circulation. 1996;93:737-744.
Vanoverschelde JL, Wijns W, Depre C, Essamri B, Heyndrickx GR, Borgers M, Bol A, Melin JA. Mechanisms of chronic regional postischemic dysfunction in humans: new insights from the study of noninfarcted collateral-dependent myocardium. Circulation. 1993;87:1513-1523.
Buxton DB, Mody FV, Krivokapich J, Phelps ME, Schelbert HR. Quantitative assessment of prolonged metabolic abnormalities in reperfused canine myocardium. Circulation. 1992;85:1842-1856.
Vanoverschelde JL, Melin JA, Bol A, Vanbutsele R, Cogneau M, Labar D, Robert A, Michel C, Wijns W. Regional oxidative metabolism in patients after recovery from reperfused anterior myocardial infarction. Circulation. 1992;85:9-21.
Gerber BL, Vanoverschelde JL, Bol A, Michel C, Wijns W, Robert A, Melin JA. Relative FDG imaging during euglycemic hyperinsulinemic glucose clamp accurately identifies viable myocardium in patients with ischemic left ventricular dysfunction. J Nucl Med. 1994;35:39P. Abstract.
Armbrecht JJ, Buxton DB, Schelbert HR. Validation of 11C-acetate as a tracer for noninvasive assessment of oxidative metabolism with positron emission tomography in normal, ischemic, postischemic, and hyperemic canine myocardium. Circulation. 1990;81:1594-1605.
Gropler RJ, Siegel BA, Geltman EM. Myocardial uptake of carbon-11-acetate as an indirect estimate of regional myocardial blood flow. J Nucl Med. 1991;32:245-251.
Chan SY, Brunken RC, Phelps ME, Schelbert HR. Use of the metabolic tracer carbon-11-acetate for evaluation of regional myocardial perfusion. J Nucl Med. 1991;32:665-672.
Nuyts J, Maes A, Vrolix M, Schiepers C, Schelbert H, Kuhle W, Bormans G, Poppe G, Buxton D, Suetens P, De Geest H, Mortelmans L. Three-dimensional correction for spillover and recovery of myocardial PET images. J Nucl Med. 1996;37:767-774.