Several flow tracers have been introduced to assess myocardial perfusion under rest and stress conditions (Table 30.1). Potassium 43 (⁴³K) and ²⁰¹Tl were introduced because of their attractive physiological characteristics: high first-pass extraction, rapid blood clearance, and slow tissue washout (5). ²⁰¹Tl became a very successful clinical flow tracer, but its low photon energy (80 to 90 KeV) and long physical half-life (73 hours) affect the image quality and limit the applicable dose with tolerable radiation exposure. Pohost et al. first described the use of ²⁰¹Tl redistribution as marker of viable tissue (6). Based on the tissue kinetics of ²⁰¹Tl, the myocardium can be characterized as normal (normal uptake, normal washout), ischemic (low uptake, delayed washout), or scar (low uptake, fast washout). Comparing early tracer retention at 15 to 30 minutes after injection with that of later timepoints (3 to 4 hours) allows for the separation of left ventricular segments with and without tracer redistribution. Some investigators proposed to monitor the redistribution process over prolonged time periods. Gutman et al. reported that in the presence of severe stenoses, defects that appeared fixed at 3 to 5 hours often displayed redistribution after 24 hours (7). Kiat et al. demonstrated subsequently that the redistribution pattern at 24 hours is more predictive for tissue recovery after revascularization than 3- to 4-hour data (8).

Although the presence of redistribution has been shown to serve as specific marker of viability, the absence of redistribution at 3 to 4 hours after tracer injection does not rule out tissue viability (8,9,10). In order to accelerate the redistribution of relative activity in the myocardium, a second ²⁰¹Tl injection has been proposed (11). An injection of ²⁰¹Tl at peak exercise with early imaging is followed by an injection
at rest and imaging after 3 to 4 hours. Based on this protocol, ²⁰¹Tl scintigraphy has gained wide clinical acceptance as the most appropriate viability test using SPECT imaging.

The ^99mTc flow tracers have more suitable physical characteristics than ²⁰¹Tl. A photon energy of 140 KeV and a half-life of 6 hours provide better count statistics because of less attenuation and higher allowable dose application. However, the first-pass extraction fraction of ^99mTc[sestamibi] and [^99mTc]tetrofosmin is lower than that of ²⁰¹Tl (12,13,14). This may be a limitation for the assessment of perfusion at high flow rates but does not affect resting tracer distribution necessary for viability studies. Both ^99mTc-labeled tracers are retained in viable myocardium for prolonged time periods. [^99mTc]sestamibi is taken up in the mitochondria and retained by an electrochemical gradient. The relative distribution of both tracers at rest reflect perfusion and tissue viability (15).

Blood flow tracers can be classified based on their physiological behavior (see Table 30.1). Oxygen 15 [¹⁵O]water, for example, represents a freely diffusible tracer that washes in and out of myocardial tissue as a function of blood flow. The first-pass extraction of [¹⁵O]water in the heart is not diffusion limited, nor is [¹⁵O]water tissue extraction affected by any metabolic pathways (Fig. 30.1) (16,17,18). The second group of flow markers are radiotracers retained in myocardial tissue proportional to myocardial blood flow. For these radiopharmaceuticals, the initial tracer extraction (first-pass extraction) and their tissue retention are important factors defining their suitability as blood flow tracers. Nitrogen 13 (¹³N) ammonia is highly extracted by myocardial tissue in the form of [¹³N]ammonia (19,20,21). Within the tissue, the tracer can either back-diffuse into the vascular space or be trapped in the form of [¹³N]glutamine (Fig. 30.1). Ionic tracers such as rubidium 82 (⁸²Rb), rubidium 81 (⁸¹Rb), and potassium 38 (³⁸K) display tracer kinetics similar to those of ²⁰¹Tl (12,22,23). Initial extraction of these compounds ranges from 50% to 70%. For both [¹³N]ammonia retention and ionic tracer extraction, a nonlinear relationship exists between blood flow and tissue tracer extraction (22,24). For methodological details of myocardial flow measurements by PET, please refer to an excellent recent review by Kaufmann and Camici (25).

The first radiopharmaceutical used in combination with PET for the assessment of regional cardiac metabolism was carbon 11 [¹¹C]palmitate. [¹¹C]palmitate is avidly taken up by the myocardium, with an extraction fraction of about 50%. The myocardial ¹¹C activity clears in a biexponential pattern, including a rapid early and slow late phase. Experimental studies changing cardiac workload or cardiac substrate availability demonstrated changes in the relative contribution of the early and late phase of [¹¹C]palmitate kinetics (41). Based on these studies, a relation was established between the early clearance phase and the relative contribution of β-oxidation to long-chain fatty acid metabolism (42,43). The contribution of the early clearance phase was correlated with myocardial oxygen consumption (41,43). Myocardial ischemia leads to impaired [¹¹C]palmitate oxidation, as demonstrated by tracer kinetics (44,45). Experimental work correlating [¹¹C]palmitate kinetics with ultrastructural tissue analysis in animals with ischemic injury revealed good correlation between lipid droplets and increased deposition of [¹¹C]palmitate into a slow turnover pool (27). Quantification of mitochondrial function using [¹¹C]palmitate is limited because of the influence of plasma substrate levels on the metabolic fate of exogenous fatty acids in terms of their incorporation into triglycerides or immediate oxidation. To avoid the complexity of substrate interaction defining the relative contribution of long-chain fatty acids and carbohydrates to overall oxidative metabolism, [¹¹C]acetate has been proposed as an alternative probe to describe oxidative metabolism. [¹¹C]acetate is converted to [¹¹C]acetyl-CoA in the mitochondria and enters the TCA cycle. ¹¹C activity equilibrates within TCA cycle intermediates, and ¹¹C activity clears from myocardium in the form of [¹¹C]CO₂ (Fig. 30.3). Several studies have indicated that [¹¹C]acetate kinetics, as assessed by dynamic PET imaging,
correlate closely with myocardial oxygen consumption, as assessed directly in the animal laboratory or by hemodynamic parameters in the clinical setting. Kinetics of [¹¹C]acetate are only sparsely affected by substrate interactions and thus allow quantification of myocardial oxygen consumption (46,47,48,49,50). Tracer kinetic models have been introduced to quantitate [¹¹C]acetate kinetics (51,52). However, most clinical studies use simple determination of clearance rates by using monoexponential curve fitting of regional myocardial ¹¹C myocardial time–activity curves (47).

The use of [¹¹C]palmitate and [¹¹C]acetate in healthy volunteers indicates homogeneous uptake of these tracers within the normal left ventricular myocardium (47,53,54). The high initial extraction of [¹¹C]palmitate and [¹¹C]acetate allows the qualitative assessment of regional myocardial perfusion (41,55,56). The comparison of [¹¹C]acetate uptake and [¹³N]ammonia blood flow measurements also indicates close correlation of results in patients with CAD, which points to the opportunity to assess flow and metabolism with a single tracer injection (see Fig. 30.3).

Fluorine 18 deoxyglucose (FDG) traces transmembranous transport as well as phosphorylation of exogenous glucose (57,58). FDG does not enter any further metabolic pathways but accumulates in myocardium proportional to glucose transport and phosphorylation (59,60). Very little dephosphorylation of FDG occurs in the myocardium. Experimental studies in isolated rabbit septum as well as canine models demonstrated a close relation of FDG uptake and exogenous glucose metabolism, measured by the Fick principle (60,61). However, FDG molecules display a different affinity for glucose transport and phosphor-ylation than does the glucose molecule itself (62). To correct for this discrepancy, a correction term is necessary for the quantification of exogenous glucose utilization of the myocardium by FDG. This correction term (lumped constant) is assumed to be constant under physiologic and most pathophysiologic conditions (63,64). Recent data, however, indicate that under rapidly changing conditions, the lumped constant may actually change as a function of altered affinities for glucose transport as well as for the hexokinase reaction (65,66). Exogenous glucose utilization can be quantified by using a simple fitting procedure of FDG myocardial kinetics and parametric display of regional metabolic data (67,68,69) (see Chapter 13).

The comparison of [¹¹C]acetate uptake and clearance with regional FDG kinetics in normal volunteers demonstrated an inhomogeneity of regional glucose utilization in the heart (70). Regional FDG uptake is increased in the lateral wall of the left ventricle and slightly decreased in the area of intraventricular septum (71). Combined [¹¹C]acetate and FDA myocardial substrate metabolism can be used for the quantification and determination of the relative contribution of glucose to overall oxidative metabolism (70,72). Such measurements have been performed to evaluate cardiac glucose metabolism in patients with insulin-dependent and non-insulin-dependent diabetes mellitus. In patients with a history of insulin-dependent diabetes, there was no significant difference in overall glucose utilization as compared with a control population if insulin was replaced by a euglycemic insulin clamp (73,74,75).

There are numerous methods proposed to assess myocardial viability with ²⁰¹Tl. The most commonly used imaging protocol with ²⁰¹Tl includes stress injection of the tracer followed by reinjection and redistribution imaging 3 to 4 after tracer application. Bax et al. analyzed 33 studies with over 800 patients in a meta-analysis and found that ²⁰¹Tl imaging using rest and stress protocols had a sensitivity of 86% and a specificity of 59% in predicting functional recovery following revascularization (76). The relatively low specificity may reflect the criteria used to classify viability as well as the protocol of data acquisition. Several investigators observed that the redistribution process following ²⁰¹Tl injection may be prolonged in patients with severe coronary stenosis (7,8). Gutman et al. reported that in the presence of severe stenosis perfusion defects that appeared fixed at 3 to 6 hours after injection often showed redistribution at 24 hrs (7). Kiat et al. supported the observation that late imaging can be used to identify ²⁰¹Tl redistribution in areas that were labeled fixed at 4 hours after injection (8). The investigation of Liu et al. following patients with single-vessel disease before and after coronary angioplasty confirm the notion that a fixed ²⁰¹Tl defect does not exclude tissue viability (9). In this cohort of patients, 72% of patients with fixed defects also had evidence of myocardial viability after revascularization. This is further supported by studies comparing ²⁰¹Tl kinetics with FDG-PET images. Brunken et al. demonstrated that patients who had fixed ²⁰¹Tl defects displayed ¹⁸F-FDG uptake in the area of fixed defects (10). In this cohort, 58% of fixed ²⁰¹Tl defects had evidence of viability based on metabolic imaging. Subsequently ²⁰¹Tl reinjection was introduced to enhance the redistribution process. Dilsizian et al. reported that in patients with a fixed ²⁰¹Tl defect on redistribution images, reinjection caused normalization of the defect in 49% of cases (77). Further studies by the same group indicated that ²⁰¹Tl reinjection imaging has a much greater concordance with FDG-PET imaging than does conventional ²⁰¹Tl redistribution, indicating a higher sensitivity for tissue viability by the reinjection procedure. In this study, 88% of viable segments by ²⁰¹Tl reinjection protocols were also viable based on the FDG-PET studies (78).

Stress perfusion imaging may not be necessary in patients with known coronary artery disease who are scheduled for viability studies. Therefore, a number of investigators evaluated the value of rest-delayed ²⁰¹Tl imaging as the preferred method for assessing regional tissue viability. In a group of 26 patients, Iskandrian found that 12 of 16 patients with normal or transient ²⁰¹Tl defects showed improved ventricular function after surgery, whereas 2 out of 10 patients with fixed defects demonstrated some degree of functional recovery (79).