Quantitative Measurements of Cardiac Phosphorus Metabolites in Coronary Artery Disease by 31P Magnetic Resonance Spectroscopy
Background 31P metabolite measurements in the human heart by magnetic resonance spectroscopy (MRS) have been reported previously. By use of a method in which metabolite content was quantified with reference to a standard located outside the chest, it has become possible to measure the content of phosphocreatine (PCr) and ATP in vivo in the human heart. In this study, PCr and ATP contents were measured by 31P MRS and compared in human myocardium with reversible ischemia or scar diagnosed by exercise thallium scintigraphy.
Methods and Results Forty-one subjects with stenosis of the left anterior descending coronary artery (>50%) and 11 healthy control subjects (C) composed the present study group. Patients were divided into two groups on the basis of exercise 201Tl scintigraphy: a reversible 201Tl defect group (RD[+], n=29) who demonstrated redistribution at late image and a fixed 201Tl defect group (RD[−], n=12). While the subjects lay supine within the magnet, 31P MR spectra were obtained from the anterior and apical regions of the left ventricle by slice-selected one-dimensional chemical shift imaging. For metabolite quantification, a standard was placed at the center of the surface coil. ANOVA revealed significant differences among the three groups with respect to the mean (±SD) PCr at rest (C, 12.14±4.25 >RD[+], 7.64±3.00 >RD[−], 3.94±2.21 μmol/g wet heart tissue, P<.05) as well as a significant decrease in ATP in the RD(−) group (C, 7.72±2.97; RD[+], 6.35±3.17 >RD[−], 4.35±1.52 μmol/g wet heart tissue, P<.05).
Conclusions Compared with healthy control subjects, PCr content decreased significantly in patients with both reversible and fixed 201Tl defects, and ATP content decreased significantly in subjects with fixed thallium defects. These results suggest that the measurement of ATP content in the human heart by 31P MRS is a clinically important method for the evaluation of myocardial viability.
31P magnetic resonance spectroscopy (31P MRS) has been shown to be an important technique for studying the influence of coronary artery diseases (CAD) by monitoring myocardial metabolism.1 2 3 4 5 6 The development of MRS has made it possible to directly measure metabolically important compounds noninvasively. Until now, the ratio of phosphocreatine (PCr) to ATP has been used as an index of 31P MRS evaluation. However, the evaluation by PCr/ATP ratio cannot detect simultaneous proportional reductions of both PCr and ATP.
Recently, absolute quantification of PCr and ATP has been attempted by use of 31P MRS.7 8 9 Bottomley et al7 reported that when 31P MRS by three-dimensional chemical shift imaging (3D CSI) was used, metabolite content could be quantified with reference to a standard located outside the chest, yielding normal in vivo human heart content of PCr and ATP of 11.0±2.7 and 6.9±1.6 μmol/g wet heart tissue, respectively. In addition, the same researchers demonstrated that high-energy phosphate content did not vary significantly with location in the normal myocardium, although 2,3-diphosphoglycerate (2,3-DPG) signals from blood varied with subject and location.
Since we are not aware of any reports on myocardial high-energy phosphate levels in human CAD, the main purpose of this study was to investigate whether phosphorus metabolite content measurements by 31P MRS, compared with exercise 201Tl scintigraphy, could serve as a useful tool for evaluating CAD.
Patients with stenosis of the left anterior descending coronary artery were divided into two groups by exercise 201Tl scintigraphy: a reversible 201Tl defect group and a fixed 201Tl defect group. After 31P MRS by slice-selected one-dimensional (1D) CSI was performed on all patients and healthy subjects, PCr and ATP contents were quantified with reference to a standard located against the chest.
Forty-one patients with stenosis (≥50%) of the left anterior descending coronary artery (LAD) and 11 healthy volunteers with no clinical evidence of cardiac disease composed the present study group. Coronary angiograms were analyzed by two independent, blinded, experienced angiographers. Luminal diameter stenosis was measured by videodensitometry with a Vanguard Coronary Analyzer System (Vanguard Instrument Corp).
Patients were divided into two groups by exercise 201Tl scintigraphy with 3-hour postexercise study. The reversible 201Tl defect group (RD[+]) consisted of 22 men and 7 women 44 to 77 years old (mean±SD, 62±8 years). The fixed 201Tl defect group (RD[−]) consisted of 8 men and 4 women 36 to 77 years old (mean±SD, 64±12 years) (Table⇓).
In the RD(+) group, 18 patients showed evidence of prior myocardial infarction. In the RD(−) group, all patients demonstrated prior myocardial infarction and had extensive anterior wall 201Tl defects at rest, as shown in the Table⇑, and severe hypokinesis, akinesis, or dyskinesis in the anterior, septal, and apical regions by left ventriculography.
All patients continued to receive conventional anti-ischemic medication at the time of the study (nitrates, β-blocking agents, calcium channel blockers, anticoagulants, etc). The control group (C) consisted of 11 healthy subjects (7 men and 4 women) 27 to 67 years old (mean±SD, 49±13 years). All subjects in the control group were free of any previous clinical history of heart disease. Informed consent was obtained from all subjects.
Exercise Single Photon Emission Computed Tomography 201Tl Imaging
Exercise 201Tl imaging was performed within 1 month of the MRS study. No significant clinical events (acute myocardial infarction or unstable angina) intervened between the exercise test and the MRS study. All patients underwent multistage treadmill exercise testing according to the Bruce protocol. Heart rate, blood pressure, and 12-lead ECG were monitored during exercise. Incremental exercise testing was continued until the onset of either anginal chest pain, ST-segment depression of ≥2 mm, fatigue, or target heart rate. At peak exercise, a dose of 111 MBq of 201Tl was injected intravenously, after which exercise was continued for an additional 30 seconds. Single photon emission computed tomography (SPECT) imaging was obtained within 3 minutes after the cessation of exercise and was repeated 3 hours later. Images were acquired with a large-field-of-view rotating gamma camera (GCA-901, Toshiba Medical) equipped with a low-energy, general-purpose collimator interfaced to a dedicated computer (GMS-550U, Toshiba Medical). Early and delayed (3 hours later) 201Tl images were obtained over a 180° arc (32 images per study; 40 seconds per image), spanning from a 45° right anterior oblique to a 45° left posterior oblique view.
SPECT 201Tl Image Interpretation
A five-point scoring system by visual interpretation (0, normal; 1, equivocal; 2, moderate; 3, severe reduction of radioisotope uptake; and 4, absence of detectable tracer in a segment) for 20 segments of the left ventricular myocardium was used for the assessment of stress perfusion, according to the methods of Berman et al.10 The intensity of each image was normalized to the highest pixel value in that image.10 11 12 Segments were scored by the consensus of two experienced observers who were unaware of the clinical history, results of coronary angiography, and MRS testing results. Three short-axis slices on the apical, midventricular, and basal regions were divided into six evenly spaced segments each, and the vertical long-axis slice was scored for two apical segments (segments 1, 7, and 13, anterior; 2, 8, and 14, anteroseptal; 3, 9, and 15, inferoseptal; 4, 10, and 16, inferior; 5, 11, and 17, inferolateral; 6, 12, and 18, anterolateral; and 19 and 20, apical).10 Assignments of individual segments to LAD territories were determined on segments 1, 2, 3, 7, 8, 13, 14, 19, and 20 as previously described.10
A segment with a score ≥2 was considered to have a defect. A reversible perfusion defect was defined as one in which a stress defect was associated with a rest score ≤1 or a stress defect score of 4 with a rest score of 2. Nonreversible segments with stress scores of 4 or 3 were subcategorized as showing no reversibility (scores 4/4 or 3/3) or minimal reversibility (scores 4/3 or 3/2).
31P MR experiments were performed with a 1.5-T, 1-m–bore, whole-body MR imaging system (Signa; General Electric Medical Systems). A 20-cm transmitter diameter, 12.5-cm receiver diameter, circular 1H surface coil and a 31P surface coil with the same size transmitter and receiver were used. The 31P surface coil transmitter and receiver were coaxial and coplanar. Fig 1⇓ shows 1H transaxial and coronal images taken with the 1H surface coil in place to show the variation in intensity across the region of interest (ROI), including extension into surrounding tissues. The 1H coronal image suggests that the heart was favorably selected by the surface coil.
Patients were examined in the supine position. Conventional 1H MR imaging was performed by the instrument’s body coil before 31P MRS to confirm and guide the placement of the surface coil over the anteroapical region of the left ventricle. Images were acquired in the transaxial and coronal planes. Subjects were kept in a constant position on the examination table during the acquisition of 31P MRS data, and the table was maintained at an established center position during the acquisitions. The subjects were removed twice from the magnet during the examination: for positioning of the 1H coils precisely over the heart for shimming and replacing them with 31P coils for spectroscopy.
MR spectra were spatially localized by 1D CSI with slice selection in the sagittal direction (see below) (Fig 2⇓). Before each measurement, shimming by the proton signal of water in the ROI was performed over the sagittal slice by applying a gradient and using a slice-selective pulse. The typical line widths were between 0.4 and 0.8 ppm. Phase-encoding gradients were applied in the coronal direction for 500 μs, with 16 phase-encoding views acquired over a 32-cm field of view (Fig 3⇓). Each phase-encoding step required an average of 32 free induction decays, which were acquired during every other heartbeat at end systole (350 ms after the R wave of the ECG). The 31P sweep width was 2000 Hz, and the data block size was 512. The radiofrequency pulse power was kept constant for all subjects.
31P MRS Localization Technique
Signal localization to myocardial tissue was achieved by slice-selected 1D CSI with ECG gating to systole. PCr resonance sometimes split into two peaks in ordinary one-dimensional approaches for localization such as depth-resolved surface coil spectroscopy (DRESS) or 1D CSI. The splitting of PCr resonance is considered to be caused by the contamination of the chest muscle (mainly left chest wall muscle).9 We added a sagittal slice selection for the left ventricle to the 1D CSI sequence (Fig 4A⇓), which resulted in a favorable decrease of chest muscle contamination (Fig 4B⇓).
The slice thickness of the selection in the sagittal direction was between 60 and 80 mm, and the slice was positioned so that much of it was filled by the left ventricle. Sixteen phase-encoding steps were applied over a 32-cm field of view. Each resulting spectrum was thus representative of a 2-cm-thick section in the coronal plane. Radiofrequency excitation was achieved with a selective sinc pulse of 1-ms duration, and pulse power was kept constant between patients. The power of the surface coil was not adjusted between patients because coil loading was almost constant between subjects. Phantom studies demonstrated that at an average depth of 0.0 and 5.5 cm, the coil deposited approximately 145° and 68° flip angles, respectively.
The distance between the surface coil and the center of the slice, which included anterior myocardium, ranged from 4.6 to 7.2 cm, with a mean±SD of 5.8±0.7 cm. Phantom studies have shown that <15% of “signal bleed” occurs from adjacent sections. Total scanning time ranged from 12 to 15 minutes per set of spectra. The total MR examination time ranged from 50 to 60 minutes. The integrated areas of the resonances of phosphocreatine (PCr) and β-phosphate of ATP were measured after a 15-Hz line-broadening exponential filter was applied and the peaks were fitted to a lorentzian line with a home-built automatic data processing station using the simplex technique (Fig 5⇓).13
For metabolite quantification, a 0.5-mL reference vial of 5.75 mol/L hexamethylphosphoric triamide (HMPT) solution was placed on the axis of the surface detection coil against the chest. The sensitivity profile in the transverse field of our 31P 12.5-cm receiver coil is shown in Fig 6⇓. Signal intensities for 1 mL of 17.2 mol/L phosphoric acid (inorganic phosphate [Pi], spin-lattice relaxation time [T1]=1.31 seconds) were measured at different points under the surface coil with the same transmit attenuation. This sample was used in conjunction with a phantom that loaded the coil to the same extent as the subjects. This phantom yielded the same Q factor (54.9) as the subjects. Shifting of the chemical shift maps was not used to superimpose them on the intensity map. This phantom study revealed that the excitation field was comparatively uniform over the ROI for the anterior myocardium in the sensitive volume of the 12.5-cm coil.
The content of metabolite was determined from the formula of Bottomley et al.7 The volume of the anterior myocardium in the ROI was estimated from the transaxial 1H MR image, assuming that the volume was enclosed by two spherical surfaces, namely, epicardial and endocardial, and that the specific gravity of the myocardial tissue was 1.05 (Fig 7⇓). Since no transaxial slicing was performed, the accurate transaxial dimension of the slice was unknown. 1H images taken with the 1H surface coil suggest that the transaxial dimension of the slice seems to be almost the same as the dimension of the 31P detection coil. Thus, we estimated the dimensions of the slices as 2.0×(6.0 to 8.0)×12.5 cm. The ratio of the detection coil sensitivities at the sample and at the reference was derived from phantom studies. Corrections for signal loss due to phase variations and missing data points were not made for our series.
The saturation factors (SF) were determined by the following formula14 :
where TR is repetition time.
The flip angle, α, for the subject was estimated by use of a phantom that yielded the same Q factor as the subject. The T1 values of cardiac PCr and β-phosphate of ATP were estimated from the spectra of four healthy volunteers acquired at TRs of 1, 2, 3, and 15 seconds with no ECG gating. The T1 values of PCr and β-phosphate of ATP were 4.2 and 1.7 seconds, respectively. We used the same T1 value for all the subjects. This correction assumes that normal and infarcted myocardium have the same spin-lattice relaxation times. The T1 value of HMPT was 11.5 seconds.
Levels of ATP were represented by the β-phosphate of ATP, since this does not overlap with the resonances of other compounds. Contamination of blood ATP from the section intersecting the ventricular cavity was corrected by subtraction of a blood ATP signal corresponding to 15% of the total integrated 2,3-DPG signal from the β-phosphate of the ATP signal.15
Metabolite content in unit tissue volume was converted to content per gram wet weight by use of a specific gravity of 1.05 for adult heart tissue.
When the same subjects were studied twice on different days, the percent change variation obtained in each case was <10% each for PCr and ATP.
Data are presented as mean±SD. Scheffé’s F test for multiple contrasts was applied to detect significant differences as defined by ANOVA among the three groups (RD[+], RD[−], and control subjects). A probability value of P<.05 was considered to be significant.
There was no significant difference in heart rate among the three groups (66±13, 74±15, and 66±9 beats per minute for RD[+], RD[−], and C, respectively, P=NS by ANOVA).
There was no significant difference in left ventricular mass in the ROI among the three groups (39.35±8.45, 38.67±6.83, and 38.64±8.09 g wet tissue for RD[+], RD[−], and C, respectively, P=NS by ANOVA).
The 31P MR spectra of typical patients in the three groups are shown in Fig 8⇓. The results of one-way ANOVA demonstrated that subjects in the RD(−) group had significantly lower myocardial PCr content than those in the RD(+) (P<.05) and C (P<.01) groups. In addition, subjects in the RD(+) group had lower PCr content than the C group (P<.05) (RD[+], 7.64±3.00; RD[−], 3.94±2.21; and C, 12.14±4.25 μmol/g wet tissue) (Fig 9⇓).
Subjects in the RD(−) group had significantly lower myocardial ATP content than those in the C group (P<.05). However, no significant differences existed between the RD(+) and C groups (RD[+], 6.35±3.17; RD[−], 4.35±1.52; and C, 7.72±2.97 μmol/g wet tissue) (Fig 10⇓).
Subjects in the RD(−) group had significantly lower myocardial PCr/ATP ratios than those in the C group (P<.05). However, no significant differences existed between the RD(+) and C groups (RD[+], 1.37±0.57; RD[−], 0.94±0.41; and C, 1.80±1.03) (Fig 11⇓).
MRS has been established as an important technique for studying the effects of CAD on myocardial metabolism.16 17 Recently, stress exercise testing and intravenous dobutamine loading have also been used effectively to elicit myocardial ischemia, resulting in abnormalities of high-energy phosphate metabolism in patients studied with MRS.1 6 18 19 Before the development of MRS, assessment of myocardial biochemical alterations associated with CAD was limited to the analysis of biopsy specimens or via indirect analysis by coronary sinus blood sampling for compounds such as lactate and succinate.20 21 The development of MRS has made it possible to directly measure metabolically important compounds noninvasively.
MRS research concerning absolute quantification of human myocardial PCr and ATP has been limited to only a few preliminary studies.7 8 9 Evaluation by PCr/ATP ratio cannot detect proportional increases or decreases of both PCr and ATP. To obtain further insight into myocardial metabolism, quantification of PCr and ATP content is essential. Accurate measurement of Pi is difficult to perform in humans because of contamination of cavity blood. Accordingly, only PCr and ATP were measured in this study by this method.
Bottomley et al7 stated that mean values from the heart in healthy control subjects are 11.7±2.5 μmol/g for PCr and 7.2±1.2 μmol/g for ATP by 3D CSI. Okada et al9 reported that mean values from the heart in healthy control subjects are 13.0±3.0 μmol/g for PCr and 7.8±2.6 μmol/g for ATP by the DRESS technique. These findings are compatible with the results obtained in our study.
This study was not a comparison of 201Tl methods with MR techniques, but rather a study of the use of nuclear methods to establish experimental group criteria. We demonstrated that subjects in the RD(−) group had lower myocardial PCr than those in both the RD(+) and C groups and that patients in the RD(+) group had lower ratios than those in the C group. These results suggest that the degree of CAD is inversely proportional to the PCr content.
We successfully revealed that subjects in the RD(−) group had significantly lower myocardial ATP content than those in the C group, whereas no significant differences existed between the RD(+) and C groups. The myocardium in CAD without scarred tissue, as in the RD(+) group, had myocardial ATP content similar to that of healthy control subjects. These results suggest that ATP content measurement in the human heart is a clinically important method for the evaluation of myocardial viability. Since myocardial scar tissue presumably contains little or almost no ATP, the acquired signal in the RD(−) group most likely came from the residual viable myocardium, which was often remodeled to maintain adequate global left ventricular function. Myocardial infarction, particularly one that is large and transmural, can produce alterations in the topography of both the infarcted and noninfarcted regions of the ventricle. These abnormalities of remodeled myocardium might have altered the metabolism in the surviving muscle.
We clearly showed that subjects in the RD(−) group had significantly lower myocardial PCr/ATP ratios than those in the C group. Patients in the RD(−) group had not only low myocardial ATP content but also much lower myocardial PCr, and thus lower PCr/ATP ratios, than those in the C group. Weiss et al1 noted that the resting PCr/ATP ratio was slightly lower in patients with coronary heart disease than in control subjects and in patients with nonischemic heart disease. Hardy et al17 found that reduced PCr/ATP ratios in patients with congestive heart failure resulted from severe multivessel CAD. Luney et al5 pointed out a reduction in PCr/ATP in myocardium identified as scar by 201Tl scintigraphy. Using the DRESS technique, we recently demonstrated that subjects in the RD(−) group had lower myocardial PCr/ATP ratios at rest than those in both the RD(+) and C groups and that subjects in the RD(+) group had lower ratios than those in the C group.6 These findings are in agreement with the results of the present research. These results (the lack of more normal PCr/ATP levels in the residual myocytes in the infarcted area) suggest that the metabolism of residual viable myocytes may have been changed by the mechanisms of remodeling. Abnormal transmural distribution of high-energy phosphate compounds is evident in remodeled myocardium.22 This abnormality may be related in part to the mismatch of oxygen delivery and demand.22 23 The 31P MRS method may be more heavily influenced by perfusion than by the extent of tissue viability.
Further studies on a larger number of patients, evaluated before and after revascularization, are clearly needed to establish more broadly applicable and reliable diagnostic algorithms by use of 201Tl SPECT,10 11 12 24 25 26 27 28 fluorodeoxyglucose positron emission tomography (PET),10 12 29 30 31 or dobutamine echocardiography32 33 in both hibernating and stunned myocardium. Although PET is superior to SPECT, it is very expensive, requires cyclotron technology, and is not readily available. In contrast, thallium is easily accessible for clinical use and provides a less expensive alternative to PET assessment of regional metabolic activity. We used exercise 201Tl SPECT imaging as an index of myocardial ischemia or scar in patients with CAD. Exercise 201Tl scintigraphy has also been used to assess myocardial viability in patients with CAD.24 Reversible 201Tl defects are associated with viable ischemic myocardium, whereas fixed 201Tl defects are usually regarded as representing nonviable infarcted regions. However, recent studies have demonstrated that fixed 201Tl defects observed on 3- or 4-hour postexercise images can become reversible when patients are reimaged at 24 hours after exercise.12 25 Additionally, fixed 201Tl defects have been reported to disappear after revascularization.27 28 More recently, thallium reinjection after stress redistribution imaging has been proposed for identifying viable myocardium.
Thus, it is possible that perhaps certain patients with both necrotic and injured, but still viable, myocardium were present in our RD(−) group. In the future, MRS may possess the ability to adequately differentiate conditions in which perfusion is abnormal but no evidence of ischemia exists.19 The term “hibernating” myocardium has been applied to chronic left ventricular dysfunction without angina or ischemic ECG changes in patients with CAD that is reversed by therapy that increases myocardial blood flow. Subjects in the RD(−) group, who showed normal ATP content, might have hibernating myocardium, and intervention for revascularization may be indicated in these cases.
Measuring absolute content is difficult but not impossible, although accurate knowledge of the tissue volume in the voxel is needed.7 Metabolite imaging and noninvasive quantification could be quite useful in studying the heterogeneity of human CAD.
One significant difficulty is the current inability to obtain MR data from regions other than the anterior wall. Several CSI methods have been proposed since the inception of MR imaging. CSI with slice selection in the sagittal direction was applied to the subjects in our study. There might have been contributions from surrounding tissue, such as signals from ATP in the liver, that occur in the absence of a transaxial slice selection. Localized human in vivo 31P MR spectra of metabolites are almost invariably acquired under conditions of partial signal saturation to optimize the signal-to-noise ratio per unit time. In addition, this technique requires delays of approximately 1.5 ms between the center of the excitation pulse and the commencement of data acquisition to accomplish spatial encoding and rephasing after slice selection. Owing to these limitations, systematic underestimation of ATP content, but not PCr, might have been observed compared with undelayed acquisition because 31P spin–spin relaxation time (T2) of ATP is shorter than that of PCr.34 35 Variabilities in saturation conditions and/or acquisition delays are likely to be major sources of systematic error.36
In this study, spectral data were analyzed with the saturation factor calculated on the assumption that T1 values of phosphorus metabolite in the heart for all subjects were the same as in healthy control subjects. The effects of aging and various heart diseases on the T1 values of cardiac phosphorus metabolite should be examined in future studies.
As in conventional MR imaging, discrete Fourier transform reconstruction in spectroscopic imaging can theoretically result in variable artifactual subtractions and additions from a neighboring volume element, depending on the location and distribution of the signal sources in each element. Varying underestimations or overestimations of concentrations and ratios may inevitably occur under conditions in which metabolite distribution is heterogeneous, as is the case for PCr in the chest wall and myocardium. These effects are difficult to identify conclusively and are neither easily corrected nor confidently modeled by simple phantom studies. Furthermore, inclusion of chamber blood in the volume will mask or falsely elevate the Pi resonance by signals from 2,3-DPG. Therefore, Pi could not be reliably measured in this study. Contamination of blood ATP from the section intersecting the ventricular cavity was corrected by subtraction of a blood ATP signal corresponding to 15% of the total integrated 2,3-DPG signal from the β-phosphate of ATP signal. However, the contribution that ATP in the blood makes to the intensity of the β-phosphate of ATP must be variable and have an effect on the metabolite content calculations.
The bin displayed in Fig 6⇑, which corresponds to a typical location of anterior myocardium, is a region with variations in intensity of about 50% (1449 to 1000). Such a wide range might have significantly affected the content calculations. We used the same sensitivity value in the region occupied by the anterior myocardium for each subject. When the field map of Fig 6⇑ was aligned with the 1H images showing the position of the myocardium in the slice from which metabolite content measurements were made for five subjects, the percent variation of the weighted average of sensitivity over the ROI in each case compared with the value used in this study was <8%.
31P MRS using 1D CSI with slice selection in the sagittal direction was performed in patients with CAD (RD[+] and RD[−]) and healthy control subjects. Patients with fixed perfusion defects by thallium scintigraphy had lower myocardial PCr content relative to those patients with reversible defects (P<.05) or control subjects (P<.01). In addition, patients with reversible defects had lower PCr content than control subjects (P<.05). However, ATP content decreased significantly in only the RD(−) group (C >RD(−), P<.05). These results suggest that ATP content measurement in the human heart by 31P MRS is a clinically important method for the evaluation of myocardial viability.
This work was supported in part by a research grant for cardiovascular diseases from the Ministry of Health and Welfare (4C-1). The authors would like to thank Drs M. Okada and S. Morikawa for their expert technical assistance and valuable discussion.
Reprint requests to Takahiro Yabe, MD, First Department of Internal Medicine, Shiga University of Medical Science, Tsukinowa-cho, Seta, Otsu 520-21, Japan.
- Received October 17, 1994.
- Revision received December 21, 1994.
- Accepted January 2, 1995.
- Copyright © 1995 by American Heart Association
Ingwall JS. Phosphorus nuclear magnetic resonance spectroscopy of cardiac and skeletal muscle. Am J Physiol. 1982;242(Heart Circ Physiol 11):H729-H744.
Neubauer S, Krahe T, Schindler R, Horn M, Hillenbrand H, Entzeroth C, Mader H, Kromer EP, Riegger GAJ, Lackner K, Ertl G. 31P magnetic resonance spectroscopy in dilated cardiomyopathy and coronary artery disease: altered cardiac high-energy phosphate metabolism in heart failure. Circulation. 1992;86:1810-1818.
Luney DJE, den Hollander JA, Evanochko WT, Johnson LL, Pohost GM. Spectroscopy of human myocardial scar. Proc Soc Magn Reson Med, 12th Annual Scientific Meeting. 1993;3:1091. Abstract.
Yabe T, Mitsunami K, Okada M, Morikawa S, Inubushi T, Kinoshita M. Detection of myocardial ischemia by 31P magnetic resonance spectroscopy during hand-grip exercise. Circulation. 1994;89:1709-1716.
Okada M, Mitsunami K, Yabe T, Kinoshita M, Morikawa S, Inubushi T. Quantitative comparison of cardiac 31P NMR spectroscopy and 201Tl SPECT imaging in patients with old myocardial infarction. Proc Soc Magn Reson Med, 12th Annual Scientific Meeting. 1993;3:1090. Abstract.
Berman DS, Kiat H, Friedman JD, Wang FP, van Train K, Matzer L, Maddahi J, Germano G. Separate acquisition rest thallium-201/stress technetium-99m sestamibi dual-isotope myocardial perfusion single-photon emission computed tomography: a clinical validation study. J Am Coll Cardiol. 1993;22:1455-1464.
Chua T, Kiat H, Germano G, Maurer G, van Train K, Friedman J, Berman D. Gated technetium-99m sestamibi for simultaneous assessment of stress myocardial perfusion, postexercise regional ventricular function and myocardial viability: correlation with echocardiography and rest thallium-201 scintigraphy. J Am Coll Cardiol. 1994;23:1107-1114.
Brunken RC, Mody FV, Hawkins RA, Nienaber C, Phelps ME, Schelbert HR. Positron emission tomography detects metabolic viability in myocardium with persistent 24-hour single-photon emission computed tomography 201Tl defects. Circulation. 1992;86:1357-1369.
Conway MA, Bristow JD, Blackledge MJ, Rajagopalan B, Radda GK. Cardiac metabolism during exercise in healthy volunteers measured by 31P magnetic resonance spectroscopy. Br Heart J. 1991;65:25-30.
Schaefer S. Clinical application of cardiac spectroscopy. In: Schaefer S, Balaban RS, eds. Cardiovascular Magnetic Resonance Spectroscopy. Boston, Mass: Kluwer Academic Publishers; 1993:215-224.
Parker JO, West RO, Case RB, Chiong MA. Temporal relationships of myocardial lactate metabolism, left ventricular function and S-T segment depression during angina precipitated by exercise. Circulation. 1969;40:97-111.
Mudge GH Jr, Mills RM Jr, Taegtmeyer H, Gorlin R, Lesch M. Alterations of myocardial amino acid metabolism in chronic ischemic heart disease. J Clin Invest. 1976;58:1185-1192.
Schaefer S, Schwartz GG, Wisneski JA, Trocha SD, Christoph I, Steinman SK, Garcia J, Massie BM, Weiner MW. Response of high-energy phosphates and lactate release during prolonged regional ischemia in vivo. Circulation. 1992;85:342-349.
Pohost GM, Zir LM, Moore RH, McKusick KA, Guiney TE, Beller GA. Differentiation of transiently ischemic from infarcted myocardium by serial imaging after a single dose of thallium-201. Circulation. 1977;55:294-302.
Dilsizian V, Bonow RO. Current diagnostic techniques of assessing myocardial viability in patients with hibernating and stunned myocardium. Circulation. 1993;87:1-20.
Schelbert HR, Phelps ME, Huang SC, MacDonald NS, Hansen H, Selin C, Kuhl DE. N-13 ammonia as an indicator of myocardial blood flow. Circulation. 1981;63:1259-1272.
Vanoverschelde JLJ, 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.
Smart SC, Swada S, Ryan T, Segar D, Atherton L, Berkovitz K, Bourdillon PDV, Feigenbaum H. Low-dose dobutamine echocardiography detects reversible dysfunction after thrombolytic therapy of acute myocardial infarction. Circulation. 1993;88:405-415.
Turner CJ, Garlick P. One- and two-dimensional 31P spin-echo studies of myocardial ATP and phosphocreatine. J Magn Reson. 1984;57:221-227.
Balaban RS. NMR spectroscopy of the heart: part 1. Concepts Magn Reson. 1989;1:15-26.