Noninvasive Quantification of Myocardial Blood Flow in Humans
A Direct Comparison of the [13N]Ammonia and the [15O]Water Techniques
Background [13N]Ammonia has been validated in dog studies as a myocardial blood flow tracer. Estimates of myocardial blood flow by [13N]ammonia were highly linearly correlated to those by the microsphere and blood sample techniques. However, estimates of myocardial blood flow with [13N]ammonia in humans have not yet been compared with those by an independent technique. This study therefore tested the hypothesis that the [13N]ammonia positron emission tomographic technique in humans gives estimates of myocardial blood flow comparable to those obtained with the [15O]water technique.
Methods and Results A total of 30 pairs of positron emission tomographic flow measurements were performed in 30 healthy volunteers; 15 volunteers were studied at rest and 15 during adenosine-induced hyperemia. Estimates of average and of regional myocardial blood flow by the [13N]ammonia and the [15O]water approaches correlated well (y=0.02+1.02x, r=.99, P<.001, SEE=0.023 for average and y=0.06+1.00x, r=.97, P<.001, SEE=0.025 for regional values) over a flow range of 0.45 to 4.74 mL·min−1·g−1. At rest, mean myocardial blood flow was 0.64±0.09 mL·min−1·g−1 for [13N]ammonia and 0.66±0.12 mL·min−1·g−1 for [15O]water (P=NS). For adenosine-induced hyperemia, mean myocardial blood flow was 2.63±0.75 mL·min−1·g−1 for [13N]ammonia and 2.73±0.77 mL·min−1·g−1 for [15O]water (P=NS). The coefficient of variation as an index of the observed heterogeneity of myocardial blood flow averaged, for [13N]ammonia, 9±4% at rest and 12±7% during stress and, for [15O]water, 14±11% at rest and 16±9% during stress. The coefficients of variation for [15O]water were significantly higher than those for [13N]ammonia (P=.004 at rest and P=.03 during stress).
Conclusions The two approaches yield comparable estimates of myocardial blood flow in humans, which supports the validity of the [13N]ammonia method in human myocardium previously shown only in animals. However, the [15O]water approach reveals a greater heterogeneity (presumably method-related), which might limit the accuracy of sectorial myocardial blood flow estimates in humans.
The arterial reference sample technique used with radiolabeled microspheres has been considered the “gold standard” of myocardial blood flow measurements. Yet, because of its invasiveness, the approach is unsuitable for measurements in humans. PET affords the noninvasive measurement of regional myocardial blood flow in humans. Various PET approaches are available, including the [15O]water and the [13N]ammonia techniques.1 2 Estimates of regional myocardial blood flow by both PET approaches in animal experiments have been found to correlate linearly with microsphere measurements of regional myocardial blood flow.3 4
The metabolically inert [15O]water freely diffuses across the capillary and cellular membranes and thus rapidly equilibrates between the vascular and extravascular spaces. Achievement of such equilibrium is subsequently referred to (by definition) as the first-pass extraction fraction, which in the case of [15O]water approaches unity and is independent of blood flow. Thus, the net extraction as the product of the first-pass extraction fraction and blood flow correlates linearly with blood flow. The 2-minute physical half-life of the 15O isotope affords repetitive flow measurements at <10- to 15-minute intervals. On the other hand, a shortcoming of the [15O]water approach includes correction for the high 15O activity in the blood pool. This is done by subtraction of blood pool activity, labeled with [15O]carbon monoxide. Such correction adds complexity to the acquisition and processing of the images and data analysis and increases the error sensitivity because of possible image misalignments and low counts due to the subtraction and the short physical half-life of 15O.5 6
[13N]Ammonia offers an image quality superior to that of [15O]water because of its prolonged retention in myocardium, the longer half-life of the 13N isotope (9.8 minutes), and its preferential distribution into the myocardium.7 8 However, the myocardial net extraction of [13N]ammonia is related nonlinearly to myocardial blood flow because the first-pass tracer extraction fraction declines with increasing myocardial blood flow.9 Moreover, because it is trapped metabolically in the myocardium, questions have been raised regarding the effects of metabolic changes and abnormalities on the myocardial 13N retention.10
Both techniques were compared with independent measurements of myocardial blood flow with microspheres in the same dog study. Estimates of myocardial blood flow by [15O]water and by [13N]ammonia were correlated linearly to those by the microsphere technique.11 However, the two approaches have not yet been compared directly in humans. Such comparison is important because the relationship between the myocardial net extraction of [13N]ammonia and myocardial blood flow was derived in canine myocardium but is applied to human myocardium. One might argue that species-related differences in [13N]ammonia trapping and metabolism could alter the flow-extraction relationship. Because of the importance of this relationship for flow measurements, the validity of the [13N]ammonia approach has yet to be demonstrated in humans. On the other hand, [15O]water is metabolically inert, so this approach should be independent of metabolic alterations. Therefore, it was the purpose of this study to answer the question of whether estimates of regional myocardial blood flow in humans by [13N]ammonia are comparable to those by [15O]water.
The study was approved by the UCLA Human Subject Protection Committee. Each study participant signed the informed consent form after the investigative nature of the study, its risks, and its merits had been explained.
Thirty healthy human volunteers (mean age, 26±9 years; range, 19 to 57 years; 13 women, 17 men) were studied. To limit the radiation burden to each study participant, pairs of [15O]water and [13N]ammonia PET studies were performed either at rest or during adenosine-induced hyperemia. Consequently, 15 volunteers were studied at rest (mean age, 28±12 years; range, 20 to 57 years) and 15 during adenosine-induced hyperemia (mean age, 24±4 years; range, 19 to 31 years). None of the participants had a history of cardiovascular disease or smoking. Entrance criteria included normal heart rate, blood pressure, and resting and stress ECGs and a low probability for coronary artery disease.12 In addition, all volunteers were carefully instructed to refrain from intake of caffeine-containing beverages within 12 hours before the study.
All volunteers were injected with 30 mCi [15O]water and 20 mCi [13N]ammonia into a peripheral vein over a 30-second period while acquisition of the serial transaxial tomographic images of the heart was started. Both radiotracers were produced and synthesized as previously reported.9 13
Adenosine (Adenoscan, kindly supplied by Medco Research, Inc; 140 μg·kg−1·min−1) was infused over 6 minutes. Three minutes after the adenosine infusion was started, the radiotracer was injected while acquisition of the PET data began.
PET Study Protocol
All images were acquired on a Siemens/CTI model 931/08-12 tomograph. This device records 15 image planes simultaneously. The axial field of view is 10.8 cm. A 30-minute blank scan was recorded as part of the daily routine procedures. To minimize patient movement within the tomograph, a Velcro strap was wrapped across the chest. Correct positioning of the volunteer’s heart within the axial field of view of the tomograph was ascertained on a 4-minute rectilinear transmission scan. Then, a 20-minute transmission image for photon attenuation correction was obtained; this was followed by the [15O]water and [13N]ammonia myocardial blood flow measurements. Beginning with the intravenous administration of [15O]water, twelve 10-second, four 30-second, and one 60-second frames were acquired. Beginning with the intravenous administration of [13N]ammonia, twelve 10-second, two 30-second, one 60-second, and one 900-second frames were recorded. Measurements of myocardial blood flow by [15O]water were separated by 15 minutes from those by [13N]ammonia to allow for 15O decay (t1/2=2 minutes). ECGs were monitored continuously throughout all studies. Heart rate and blood pressure were recorded at 15- and 60-second intervals, respectively.
Cross-sectional images of the heart were reconstructed by use of a Shepp-Logan filter with a cutoff frequency of 0.96 cycles per centimeter, yielding a final in-plane spatial resolution at the center of the plane of ≈10 mm full-width at half-maximum. During myocardial blood flow measurements at rest, little if any subject movement occurs. However, side effects of intravenous adenosine may cause subject motion. Such movement can artifactually alter the activity distribution throughout the myocardium. This potential source of error might substantially affect estimates of myocardial blood flow. To minimize effects of subject motion, the images acquired during adenosine hyperemia were realigned to the transmission image as previously reported.14
The dynamically acquired sets of 15 transaxial images each for both measurements were reoriented into 6 short-axis images.4 ROIs were then assigned to the reoriented [13N]ammonia images recorded ≈3 minutes after tracer injection. Sectorial, 70° arc ROIs were assigned to three mid–left ventricular short-axis images. The regions corresponded to the distributions of the left anterior descending, the left circumflex, and the right coronary arteries (Fig 1⇓). For [13N]ammonia studies, assignment of the ROIs to each myocardial vascular territory used the “geometric ROI strategy” as proposed by Hutchins et al15 to overcome partial-volume effects. Further, an elliptical ROI with an area of about 30 mm2 was placed into the blood pool of the left ventricle. All ROIs were then copied to the dynamically acquired and reoriented [15O]water and [13N]ammonia short-axis image sets.
Myocardial and blood pool time-activity curves were generated from the first 12 dynamic frames spanning the first 2 minutes after tracer injection and were corrected for radioisotope decay. To minimize statistical fluctuations of the PET data introduced by errors related to the count rate of the images and the size of the ROI,4 one time-activity curve for each sectorial myocardial territory was derived by averaging the individual sector time-activity curves obtained from the three mid–left ventricular short-axis images. Accordingly, blood flow was assessed in about 1.5 g myocardium for each vascular territory.
Correction for Effects of Partial Volume and Spillover
Partial-volume effects result in <100% recovery of tissue activity in structures measuring less than about twice the full-width at half-maximum spatial resolution value.16 Depending on the performance of the system, such activity loss can amount to 30% for myocardium.17 Methods to correct for these activity losses have recently been reported for the [15O]water technique by Iida and coworkers18 and for the [13N]ammonia technique by Hutchins and coworkers.15 In brief, an additional term, the “myocardial blood volume,” was added to the operational model equation (for details, see “Appendix”).
[13N]Ammonia Metabolite Contamination of the Arterial Input Function
For [13N]ammonia measurements, the left ventricular input function was corrected for the time-dependent distribution of [13N] label between ammonia and its metabolites as determined previously. For studies at rest, the mean metabolite fraction at 60 seconds after injection was 1.1% and at 120 seconds after injection, 9.5%. For adenosine hyperemia, the mean metabolite fraction was 2.1% at 60 seconds after injection and 21.8% at 120 seconds after injection (P<.05 versus rest).19
Estimates of Myocardial Blood Flow
Myocardial blood flow was estimated by model fitting of the first 2 minutes of the corrected blood pool and myocardial time-activity curves. The PET-measured time-activity curves were fitted with a one-compartment model for [15O]water18 and a two-compartment model for [13N]ammonia.4
Regional Heterogeneity of Myocardial Blood Flow
The CVs (=SD/mean myocardial blood flow) were calculated for each volunteer.
Homogeneity of Myocardial Blood Flow at Rest and During Stress
For each volunteer, myocardial blood flow polar maps based on [13N]ammonia cardiac PET imaging were generated and compared with a database of normal volunteers.20 This approach was chosen to ascertain that the volunteers were indeed free of significant coronary artery disease, because coronary angiography in human volunteers without signs or symptoms for coronary artery disease was found to be unjustified.
Mean values are given with their SDs. Linear least-squares regression analysis was performed to correlate estimates of myocardial blood flow by [13N]ammonia to those by [15O]water. For the evaluation of mean differences in heart rate, mean arterial pressure, and rate-pressure product between [15O]water and [13N]ammonia measurements as well as mean differences between estimates of myocardial blood flow by [15O]water and [13N]ammonia, the paired t test was used. A value of P<.05 was considered statistically significant.
Homogeneity of Myocardial Blood Flow at Rest and During Stress
Comparison of the polar maps constructed from the rest and hyperemic [13N]ammonia images with a database of normal subjects revealed homogeneous myocardial blood flow for all volunteers both at rest and during adenosine-induced hyperemia. Also, all volunteers had a normal ECG, and none had chest pain during or after adenosine hyperemia.
Hemodynamic parameters remained constant during [15O]water and [13N]ammonia myocardial blood flow measurements at rest as well as during adenosine hyperemia. For measurements at rest, the mean heart rate averaged 63±7 bpm during the [13N]ammonia and 63±6 bpm during the [15O]water study (P=.63). The mean arterial pressure was 84±9 mm Hg during [13N]ammonia and 84±8 mm Hg during [15O]water measurements (P=.62) (Fig 2⇓). For adenosine hyperemia, the mean heart rate was 94.5±13 bpm during [13N]ammonia and 94±11 bpm during [15O]water measurements (P=.24). The mean arterial pressure was 77.8±8 mm Hg for [13N]ammonia and 76.5±8 mm Hg for [15O]water measurements (P=.18) (Fig 3⇓).
Myocardial Blood Flow
Estimates of mean myocardial blood flow by [13N]ammonia were linearly correlated to those obtained by [15O]water by y=0.02+1.02x, r=.99, P<.001, SEE=0.023 (Fig 4A⇓). Furthermore, estimates of regional myocardial blood flow by both tracers also correlated linearly (y=0.06+1.00x, r=.97, P<.001, SEE=0.025) (Fig 4B⇓). At rest, myocardial blood flow by [13N]ammonia averaged 0.64±0.09 mL·min−1·g−1 and by [15O]water, 0.66±0.12 mL·min−1·g−1. The mean difference was 0.03±0.11 mL·min−1·g−1, P=.33, for mean (n=15) and 0.03±0.15 mL·min−1·g−1, P=.15, for regional (n=45) estimates of myocardial blood flow. For adenosine-induced hyperemia, myocardial blood flow averaged 2.63±0.75 mL·min−1·g−1 for [13N]ammonia and 2.73±0.77 mL·min−1·g−1 for [15O]water. The difference between the two measurements averaged 0.09±0.18 mL·min−1·g−1, P=.16, for mean (n=15) and 0.09±0.37 mL·min−1·g−1, P=.11, for regional (n=45) estimates of myocardial blood flow. The differences between the two approaches were not systematic but rather distributed randomly, as is also indicated by the slopes of the regression lines, which for both the mean and regional flow measurements approached unity.
Myocardial Blood Flow and Cardiac Work
At rest, there was a linear correlation between the rate-pressure product as an index of cardiac work and the estimates of mean myocardial blood flow by [13N]ammonia (y=0.20+0.00006x, r=.61, P=.017) and by [15O]water (y=0.24+0.00006x, r=.58, P=.035). During adenosine-induced hyperemia, no significant correlation between blood flow and cardiac work was observed.
Heterogeneity of Myocardial Blood Flow
For measurements at rest, the mean CV was 9±4% for [13N]ammonia and 14±11% for [15O]water (P=.004). For adenosine-induced hyperemia, the mean CV was 12±7% for [13N]ammonia and 16±9% for [15O]water (P=.03).
The results of this study indicate that the [13N]ammonia dynamic PET approach yields estimates of myocardial blood flow that are comparable to those by [15O]water. This applies to both the rest and hyperemic measurements. In addition, the good agreement between estimates of mean and regional myocardial blood flow by both PET approaches further supports the validity of the [13N]ammonia PET technique in humans. It implies the absence of a possible species-related difference in [13N]ammonia trapping between canine and human myocardium. However, the [15O]water approach yielded somewhat more heterogeneous estimates of regional myocardial blood flow than the [13N]ammonia approach.
Limitations in the Present Study
Several limitations related to the volunteers and/or PET technique might have influenced the results of this study. An essential requirement for comparing two different measurement techniques of myocardial blood flow in the same individual is that blood flow remain constant. Whether this was in fact the case remains unknown. However, since blood flow at rest and, to some extent, during hyperemia depends on cardiac work, heart rate, and arterial blood pressure and since all indexes of cardiac work and all hemodynamic parameters were virtually identical for both studies in each volunteer, it seems that true myocardial blood flow was indeed constant during both the [15O]water and the [13N]ammonia measurements.
Despite careful instruction, it is possible that not all volunteers refrained from intake of caffeine. However, this would affect both measurements in the same way but might account for the variability of the hyperemic response. For the pairs of hyperemic measurements, adenosine was administered twice. This raises the question as to whether tachyphylaxis to adenosine may account for a submaximal vasodilation achieved during the second administration. A recent study in our laboratory demonstrated that hyperemic flows achieved during the second adenosine hyperemia were identical to those achieved during the first adenosine study,21 which implies that similar degrees of hyperemia were achieved during the two adenosine infusions.
Myocardial blood flow depends on numerous factors.22 For example, it is determined at rest largely by oxygen demand, which in turn is a function of cardiac work.23 24 25 Therefore, myocardial blood flow at rest would be expected to be related to the rate-pressure product as an index of cardiac work, which in fact was observed in this study.23
Accurate noninvasive quantification of regional myocardial blood flow is important for a correct assessment of blood flow in normal myocardium and in coronary artery disease as well as its consequences for myocardial function. Given the availability of various approaches, this then raises the question of whether there is an optimum technique. For [13N]ammonia, the results of the study rule out possible species-related differences in metabolic trapping of the tracer in the myocardium and thus lend further support to the validity of this approach in humans. On the other hand, metabolites of [13N]ammonia may contaminate the arterial tracer input function and thus lead to an underestimation of myocardial blood flow. Correction of the input function for 13N-labeled metabolites as performed in this study largely eliminated this contamination. Furthermore, the excellent agreement of estimates of myocardial blood flow by [13N]ammonia and the metabolically inert [15O]water indicates that such contamination would be very small, at least within the first 2 minutes after tracer injection. In addition, extraction of metabolites by the myocardium would, to some extent, offset this underestimation. With regard to tracer kinetic modeling for [13N]ammonia, the measured myocardial time-activity curves by PET were fitted to a previously validated two-compartment model.4 This approach has subsequently been shown by others to yield the most accurate estimates of myocardial blood flow compared with other proposed modeling approaches.26 27 Metabolic trapping of [13N]ammonia has been found to be reduced in the posterolateral and lateral walls of the left ventricular myocardium in humans.20 Nevertheless, estimates of blood flow from the initial uptake data recorded during the first 2 minutes after tracer injection have not been affected by this reduced metabolic trapping and indeed revealed homogeneous blood flow.23 28
For [15O]water, as with [13N]ammonia, some assumptions were made to allow measurements of human myocardial blood flow. For example, the tissue-blood partition coefficient of water is constant (and therefore fixed) and is the same in each individual. The distribution of the freely diffusible water in myocardium is always uniform. In addition, venous and tissue activities are treated as a single compartment, because the volume of distribution of [15O]water in myocardium is unity. Finally, a potential disadvantage of the [15O]water technique is the need for blood volume correction for each individual myocardial blood flow measurement.5 This requires either an additional imaging procedure with its inherent compounding of errors or inclusion of the myocardial blood volume term in the operational model equation. The latter method as used in this study was found by others to be an acceptable alternative to the blood pool subtraction method in terms of both accuracy and precision6 and has been validated in experimental animals against independent measurements of blood flow by the microsphere technique.
Effects of partial volume were corrected by addition of a myocardial blood volume term to the operational model equations. This method allowed us to overcome errors introduced by the limited PET resolution in the same way for both measurements. Interestingly, a recent study has demonstrated that a misalignment of 0.5 cm between transmission and emission images results in a significant change in radioactivity distribution throughout the myocardium.29 In contrast to this study, recent studies did not account for minimization of error propagation by correcting for image misalignment for hyperemic myocardial blood flow measurements.
Heterogeneity of Myocardial Blood Flow
The concept of heterogeneity of myocardial blood flow at any given time in small adjacent areas of the myocardium under a variety of conditions is supported by observations in animal studies.30 31 32 33 34 Several factors may account for such flow heterogeneity. For example, “temporal heterogeneities” may cause changing variations between regions.35 Further, “spatial heterogeneity” in regional myocardial blood flow may be linked to arteriolar or intratissue Po2 over a broad range, from 40 to 200 mm Hg in dogs.36 Moreover, a linear relationship between tissue norepinephrine content and coronary blood flow distribution might exist.37 Another explanation for the flow heterogeneity could be “twinkling,” which is a moment-to-moment variation in regional myocardial blood flow.38 Despite these physiological phenomena, which account for the temporal and spatial heterogeneities of myocardial blood flow estimates, the major contributor to this heterogeneity was the “method-related” heterogeneity in the present study. This heterogeneity depends on cardiac motion, regional partial volume effects, assumptions made in tracer kinetic modeling, the size of the ROIs,39 40 and flow result errors produced by the least-squares nonlinear regression algorithm that computes the flow estimates. However, the greater heterogeneity of myocardial blood flow at rest and during adenosine-induced hyperemia observed for [15O]water is presumably method related, because the short physical half-life of the 15O isotope causes a rapid decline in count rates, which result in considerable statistical noise.
Implications of Findings
The two tracer approaches yield comparable estimates of myocardial blood flow in humans. Thus, they are equally suited for the quantification of myocardial blood flow. To explore short-term responses to physiological and/or pharmacological interventions, the short physical half-life of [15O]water offers the opportunity for multiple flow measurements at short time intervals. Conversely, the somewhat higher method-related heterogeneity might limit the accuracy of measurements of regional blood flow. Also, static images of the relative distribution of myocardial blood flow are frequently of suboptimal diagnostic quality. [13N]Ammonia, conversely, yields diagnostically better images of the distribution of myocardial blood flow, which at the same time facilitates accurate placement of ROIs.
Selected Abbreviations and Acronyms
|CV||=||coefficient of variation|
|PET||=||positron emission tomography|
|ROI||=||region of interest|
Tracer Kinetic Modeling for the Quantification of Myocardial Blood Flow
The theory for the measurement of blood flow by use of the inert, freely diffusable [15O]water was described for myocardium by Bergmann et al3 and Iida et al.18 The general principles were proposed by Kety.41 In this study, a one-compartment model was used for estimating myocardial blood flow by [15O]water. The following differential equation describes the rate of change of [15O]within the myocardium: where Ct is the [15O]water activity in the ROI in counts per pixel per minute, MBF is myocardial blood flow, and V is the volume of myocardium within the ROI (in milliliters). The constant ρ is determined by the tissue/blood partition coefficient (mL/g), the specific gravity of the myocardium, and the extraction fraction of [15O]water in the myocardium. Solving Equation 1 for Ct yields Simplifying Equation 2 by expressing one derives where the symbol ⊗ indicates the mathematical operation of convolution. Integrating Equation 4 for the scan time duration ti to ti+1 yields ∫titi+1Ctdt is measured by PET for t=ti to ti+1 for i=1. . . 12 and ∫titi+1 Ca(t)⊗ ρke−ktdt is calculated by use of the one-compartment model and the left ventricular input function for a given ρ and MBF. Regional MBF was then estimated by use of Equation 5 to fit the measured myocardial time-activity curves.
Several methods are available to correct for intravascular activity within the myocardial ROI before finally estimating myocardial blood flow. One uses blood pool images obtained with [15O] or [11C]carbon monoxide and subtracting these images from the [15O]water images. In the present study, a third parameter for calculation of the blood pool activity (vascular blood pool in myocardium [MBV] observed in the analyzed ROI) was added to the [15O]water model to correct for intravascular activity within tissue ROIs. This approach was chosen because it was found by other investigators to be more accurate than direct measurements of myocardial intravascular volume with carbon monoxide myocardial blood pool imaging.6 The latter would have required an additional imaging procedure, with its inherent compounding of errors. Incorporating MBV in Equation 5 yields Expressing Equation 6 according to the proposed strategy for quantification of myocardial blood flow to overcome the partial volume effect by Iida et al18 yields For [13N]ammonia, the myocardial and blood pool time-activity curves were fitted to a previously established two-compartment model.4 In brief, this model consists of the functional compartments C1, ie, the freely diffusible [13N]ammonia distributed in the volume V (vascular and interstitial space, 0.8 mL/g), and C2, representing 13N activity trapped into myocardium. At any time t after tracer injection, 13N activity concentration in myocardium [Qi(t)], measured by PET for a given myocardial ROI, equals the sum of radioactivity of free [13N]ammonia in the freely diffusible pool Qf(t), the radioactivity of [13N]ammonia metabolites in the bound pool Qt(t), and the spillover of activity from blood pool to myocardium SP(BM), and the arterial concentration of 13N activity AB(t). This relationship is expressed by The rate constants describing the 13N activity exchange between the two functional compartments are for forward (eg, Qf to Qt) transport K1 (mL·min−1·g−1) and for backward (eg, Qt to Qf) transport K2 (min−1). The latter has been shown to be negligible and is therefore fixed to zero.4 K1 is related to MBF by as determined previously in dog experiments.9 The two parameters estimated by model fitting are myocardial blood flow in mL·min−1·g−1 and the spillover fraction from the blood pool into the myocardium. Expressing Equation 8 to accomplish the proposed minimization of resolution distortions for quantification of myocardial blood flow based on [13N]ammonia by Hutchins et al15 yields
This study was supported by the US Department of Energy by contract DE-FC03-87ER60615; by the Director of the Office of Energy Research, Office of Health and Environmental Research, Washington, DC; by research grants HL-29845 and HL-33177 from the National Institutes of Health, Bethesda, Md; and by an Investigative Group Award by the Greater Los Angeles (Calif) Affiliate of the American Heart Association. The authors thank Ronald Sumida, Lawrence Pang, Francine Aquilar, Gloria Stocks, Der-Jenn Liu, and Mark Hulgan for their technical assistance; the UCLA Medical Cyclotron staff for producing the isotopes and radiopharmaceuticals; and Eileen Rosenfeld for her secretarial assistance in preparing this manuscript.
- Received August 28, 1995.
- Revision received November 13, 1995.
- Accepted November 19, 1995.
- Copyright © 1996 by American Heart Association
Bergmann SR, Fox KAA, Rand AL, McElvany KD, Welch MJ, Markham J, Sobel BE. Quantification of regional myocardial blood flow in vivo with H215O. Circulation. 1984;70:724-733.
Kuhle WG, Porenta G, Huang SC, Buxton D, Gambhir SS, Hansen H, Phelps ME, Schelbert HR. Quantification of regional myocardial blood flow using 13N-ammonia and reoriented dynamic positron emission tomographic imaging. Circulation. 1992;86:1004-1017.
Huang SC, Schwaiger M, Carson RE, Carson J, Hansen H, Selin C, Hoffman EJ, MacDonald N, Schelbert HR, Phelps ME. Quantitative measurement of myocardial blood flow with oxygen-15 water positron computed tomography: an assessment of potential and problems. J Nucl Med. 1985;26:616-625.
Lammertsma AA, De Silva R, Araujo LI, Jones T. Measurement of regional myocardial blood flow using C15O2 and positron emission tomography: comparison of tracer models. Clin Phys Physiol Meas. 1992;13:1-20.
Schelbert HR, Czernin J, Huang SC. Quantification of regional myocardial blood flow: is there an optimum tracer? J Nucl Med. 1990;31:271-273.
Schelbert HR, Phelps ME, Huang SC, McDonald NS, Hansen H, Selin C, Kuhl DE. N-13 ammonia as an indicator of myocardial blood flow. Circulation. 1981;63:1259-1272.
Bergmann SR, Hack S, Tewson T, Welch MJ, Sobel BE. The dependence of accumulation of 13NH3 by myocardium on metabolic factors and its implications for quantitative assessment of perfusion. Circulation. 1980;61:34-43.
Bol A, Melin JA, Vanoverschelde J-L, Baudhuin T, Vogelaers D, De Pauw M, Michel C, Luxen A, Labar D, Cogneau M, Robert A, Heyndrickx GR, Wijns W. Direct comparison of [13N] ammonia and [15O] water estimates of perfusion with quantification of regional myocardial blood flow by microspheres. Circulation. 1993;87:512-525.
Hoh CK, Dahlbom M, Harris G, Choi Y, Hawkins RA, Phelps ME, Maddahi J. Automated iterative three-dimensional registration of positron emission tomography images. J Nucl Med. 1993;34:2009-2018.
Hutchins GD, Caraher JM, Raylman RR. A region of interest strategy for minimizing resolution distortions in quantitative myocardial PET studies. J Nucl Med. 1992;33:1243-1250.
Schelbert HR, Schwaiger M. PET studies of the heart. In: Phelps M, Mazziotta J, Schelbert H, eds. Positron Emission Tomography and Autoradiography: Principles and Applications for the Brain and Heart. New York, NY: Raven Press; 1986:581-661.
Iida H, Kanno I, Takahashi A, Miura S, Murakami M, Takahashi K, Ono Y, Shishido F, Inugami A, Tomura N, Higano S, Fujita H, Sasaki H, Nakamichi H, Mizusawa S, Kondo Y, Uemura K. Measurement of absolute myocardial blood flow with H215O and dynamic positron-emission tomography: strategy for quantification in relation to the partial-volume effect. Circulation. 1988;78:104-115.
Porenta G, Kuhle W, Czernin J, Ratib O, Brunken RC, Phelps ME, Schelbert HR. Semiquantitative assessment of myocardial blood flow and viability using polar map displays of cardiac PET images. J Nucl Med. 1992;33:1623-1631.
Czernin J, Kim A, Dominquez V, Phelps M, Schelbert H. How reproducible are measurements of myocardial blood flow by N-13 ammonia and PET? J Nucl Med. 1994;35(suppl S1):24P.
Marcus ML. The Coronary Circulation in Health and Disease. New York, NY: McGraw-Hill; 1983:65-92.
Czernin J, Müller P, Chan S, Brunken RC, Porenta G, Krivokapich J, Chen K, Chan A, Phelps ME, Schelbert HR. Influence of age and hemodynamics on myocardial blood flow and flow reserve. Circulation. 1993;88:62-69.
Nichols WW, O’Rourke MF, Avolio AP, Yaginuma T, Murgo JP. Effects of age on ventricular vascular coupling. Am J Cardiol. 1984;55:1179-1184.
Sunderland JJ, Dewald CL, Martin CC, Esterbrooks DJ, Gupta NC. Comparison of six N-13 ammonia myocardial blood flow measurement techniques. J Nucl Med. 1993;34(suppl):50P. Abstract.
Votaw JR, Delbeke D, Woods R, Holburn G, Sandler MP. Comparison of different techniques for calculating myocardial blood flow. J Nucl Med. 1993;34(suppl):86P-87P. Abstract.
Hutchins GD, Beanlands RSB, Muzik O, Schwaiger M. Quantitative vs semi-quantitative PET myocardial blood flow: influence of regional ammonia kinetics. Circulation. 1992;86(suppl I):I-710. Abstract.
McCord ME, Bacharach SL, Bonow RO, Dilsizian V, Cuocolo A, Freeman N. Misalignment between PET transmission and emission scans: its effect on myocardial imaging. J Nucl Med. 1992;33:1209-1214.
Franzen D, Conway RS, Zhang H, Sonnenblick EH, Eng C. Spatial heterogeneity of local blood flow and metabolite content in dog hearts. Am J Physiol. 1988;254:H344-H353.
Wolpers HG, Hoeft A, Korb H, Lichtlen PR, Hellige G. Heterogeneity of myocardial blood flow under normal conditions and its dependence on arterial PO2. Am J Physiol. 1990;258:H549-H555.
Yipintsoi T, Dobbs WA, Scanlon PD, Knopp TJ, Bassingthwaighte JB. Regional distribution of diffusable tracers and carbonized microspheres in the left ventricle of isolated dog hearts. Circ Res. 1973;33:573-587.
King RB, Bassingthwaighte JB, Hales JRS, Rowell LB. Stability of heterogeneity of myocardial blood flow in normal awake baboons. Circ Res. 1985;57:285-295.
Bassingthwaighte J, Malone M, Moffett T, King R, Little S, Link J, Groth K. Validity of microsphere depositions for regional myocardial flows. Am J Physiol. 1987;253:H184-H193.
Kety S. The theory and applications of the exchange of inert gas at the lungs and tissues. Pharmacol Rev. 1951;3:1-41.