Single-Shot Inversion Recovery TrueFISP for Assessment of Myocardial Infarction

Transcription

1 TrueFISP of Myocardial Infarction Cardiac Imaging Original Research C D E M N E U T R Y L I M C I G O F I N G rmin Huber 1 Stefan O. Schoenberg 1 enedikt Spannagl 1 Johannes Rieber 2 Isabelle Erhard 2 Volker Klauss 2 Maximilian F. Reiser 1 Huber, Schoenberg SO, Spannagl, et al. Keywords: ischemia, heart, MR contrast agents, MRI, myocardial infarction DOI:.2214/JR Received May, 4; accepted after revision December 22, 4. 1 Institut für Klinische Radiologie, Klinikum Groβhadern, Marchioninistrasse 15, Munich 81377, Germany. ddress correspondence to. Huber. 2 Department of Cardiology, Klinikum Innenstadt, Munich, Germany. JR 6; 186: X/6/ merican Roentgen Ray Society Single-Shot Inversion Recovery TrueFISP for ssessment of Myocardial Infarction OJECTIVE. The aim of the study was to assess the diagnostic accuracy of imaging the myocardium with a fast multislice inversion recovery 2D single-shot true fast imaging with steady-state precession (truefisp) sequence during a single breath-hold in comparison with an established segmented inversion recovery turbo fast low-angle shot (turboflsh) sequence. SUJECTS ND METHODS. Forty-three patients with myocardial infarction were examined on a 1.5-T MR system min after administration of contrast material (gadodiamide,.2 mmol/kg) with a single-shot 2D multislice technique (single-shot inversion recovery true- FISP) that allows one to image the entire short axis during one breath-hold (18 heartbeats) and with a segmented 2D single-slice technique (inversion recovery turboflsh) that requires one breath-hold per slice (12 heartbeats). Signal intensity was determined in normal myocardium, in infarcted myocardium, and in the left ventricle. The contrast-to-noise ratio (CNR) of normal and infarcted myocardium was determined. The areas of hyperintense infarctions on selected slices and the entire volumes were compared for both sequence techniques. RESULTS. The inversion recovery truefisp sequence has a lower CNR than the inversion recovery turboflsh sequence (mean values,. vs 12.9, respectively; p =.5) for differentiation of viable from nonviable myocardium. The CNR of injured myocardium and blood in the left ventricular cavity also has a lower value for the multislice technique compared with the single-slice technique (.6 vs 1.2, respectively; p =.45). ssessment of the area of infarction within one slice (r =.97, p <.2) and of the volume of the entire infarction (r =.96, p <.3) is possible with excellent correlation of both techniques. CONCLUSION. Despite having a lower CNR, the inversion recovery 2D single-shot truefisp sequence allows fast and accurate identification of the area and volume of infarction with high spatial resolution within a single breath-hold. im et al. [1] showed in an animal K experiment that MR images acquired with a delay after contrast material administration allow visualization of acute and chronic infarctions with high spatial resolution and accurate spatial correlation in comparison with triphenyltetrazolium chloride (TTC)- stained histopathologic slices. The infarcted area shows delayed enhancement. Recovery of contractility in hypokinetic segments of the myocardium, caused by ischemia, can be predicted by high-resolution imaging of the infarction before surgical or interventional revascularization is performed or drug therapy is started [2 4]. The superior spatial resolution of MRI is especially advantageous compared with 18 F-FDG PET resulting in a more accurate assessment of transmural extent of the infarction. For cardiac MRI, high-end MR systems with strong gradient systems, dedicated surface coil techniques, and advanced pulse sequence techniques are necessary. The reason is that both high spatial and temporal resolution and a high signal-to-noise ratio have to be achieved, and artifacts caused by respiratory motion and cardiac contraction have to be suppressed [5]. Shorter scan times are advantageous to achieve better patient compliance; to avoid contrast change in the myocardium over time; and to shorten the examination time for a comprehensive examination of function, perfusion, and viability during one examination. For imaging viability, single-slice MR techniques are widely used. The most established well-documented single-slice technique in the literature is an inversion recovery turbo fast low-angle shot (turboflsh) technique for the detection of myocardial infarc- JR:186, March 6 627

2 tion and accurate assessment of viability [2, 4, 6]. This pulse sequence technique served as the gold standard in our study for the evaluation of a new fast multislice technique: an inversion recovery 2D single-shot true fast imaging with steady-state precession (truefisp) sequence. This sequence allows imaging of nine slices during one breath-hold. The purpose of the study was to investigate the contrast-to-noise ratio (CNR) of infarction and normal myocardium for both pulse sequence techniques, to determine diagnostic accuracy of the multislice technique in detecting myocardial infarction (MI), and to compare the area and volume of infarction with both pulse sequence techniques. The single-shot inversion recovery truefisp sequence acquires one slice during one R-R cycle every second R-R cycle. In addition to the shorter scanning time, the single-shot technique may help to avoid respiratory motion artifacts in uncooperative patients compared with a segmented pulse sequence technique. recent study using a similar MR technique showed promising initial results in a small patient population [7]. In that study, a clear decrease of CNR was found. The current study with a large patient population can show if the change of CNR has an impact on diagnostic accuracy or if CNR is still high enough to use the technique in the future with high-spatial-resolution and parallel imaging techniques. Subjects and Methods Patient Population Forty-three patients who had suffered an MI were prospectively enrolled in the study (inclusion dates, June, 3 November 15, 3). The age of the infarctions ranged from 2 weeks to 3 months (mean, 24.3 ± 16 [SD] days). Informed consent was obtained from each patient for cardiac MRI. The mean age of the 43 patients (36 men, seven women) was 56 ± 12.3 (SD) years (age range, 76 years). The patients did not have unstable angina, New York Heart ssociation class III or IV, or contraindications to MRI. The diagnosis of MI was based on pathologic ECGs and biochemical evidence. The institutional ethics committee approved the study. MRI ll MR images were obtained with a 1.5-T unit (Magnetom Sonata, Maestro Class, Siemens Medical Solutions) with 8 channels and a dedicated 12- element phased-array body coil. ECGs and heart rate were monitored with the physiologic monitor of the imager (MR-compatible monitoring system). ll patients were examined in the supine position. -gauge cannula was placed in an antecubital vein for injection of contrast material. Scout images were obtained to determine the exact position and standard orientations: the long-axis view, the fourchamber view, and the short-axis views of the left ventricle. Gadodiamide (Omniscan, mersham Health) was injected at a dosage of.2 mmol/kg of body weight, followed by ml of saline. The examination of viability started after a -min delay. Optimization of Inversion Time (TI) Ten minutes after contrast material administration, a segmented inversion recovery cine true- FISP pulse sequence (TI scout) acquisition was performed at a midventricular short-axis location. This acquisition was used to determine the TI at which the signal of normal myocardium is null for the subsequent delayed enhancement acquisition. The details of the sequence, based on an approach proposed by Scheffler and Henning [8], are shown schematically in Figure 1. Thus, after a nonselective hyperbolic secant pulse applied after the R wave trigger, multiple cardiac phases were acquired using a segmented truefisp readout. The resulting 23 cine images show the recovery of the longitudinal magnetization at multiple time intervals in increments given by the temporal resolution after inversion. In our implementation, 7 k- space lines were acquired per segment per cardiac phase, resulting in a temporal resolution of 15 msec (TR/TE, 2.2/1.1; flip angle, 5 ) and a breath-hold duration of 14 cardiac cycles. The spatial resolution was typically mm 3, the field of view was cm, and the matrix was From the resulting images IR pulse t 1 t 2 Phase 1 t m Phase 2 t m+1 Phase m TI with TI values from 185 to 515 msec, the optimum TI value was determined by visual assessment in agreement by two operators. The optimum TI was determined as the image in which the signal intensity of normal myocardium was near null and the signal intensity of the left ventricular cavity was lower compared with that of the infarcted region. There was positive contrast between infarction and normal myocardium [3]. The TI scout sequence is available as a standard product. From the resulting images, the optimum TI value could be determined. The TI was optimized to null the signal of normal myocardium, which means blood in the left ventricular cavity had lower signal intensity than infarcted myocardium and there was positive contrast between infarction and normal myocardium [6]. Imaging Delayed Enhancement segmented inversion recovery turboflsh sequence was used to cover the entire left ventricle in the short-axis view. One slice was acquired per breath-hold. This sequence technique was used as standard of reference, because it is well established and was investigated in the current literature [2, 4, 6]. For the inversion recovery turboflsh sequence, the TR was 11 msec; TE, 4 msec; flip angle, 25 ; and bandwidth, 1 Hz/pixel. Thirty-two k-space lines were acquired during one cardiac cycle. With data acquisition every second heart beat, imaging of one slice was possible during nine cardiac cycles. Data acquisition was performed at mid diastole at a time window of minimal cardiac motion during 352 msec. T1 relaxation of normal myocardium Image at Phase 1 collected with TI = t 1 Phase 2 collected with TI = t 2 Phase m collected with TI = t m Optimum TI = Trigger time (TT) of the cine phase image when normal myocardium is dark Fig. 1 Diagram shows timing of cine inversion recovery (IR) true fast imaging with steady-state precession (truefisp) pulse sequence. Segments acquired with identical inversion time (TI) over all cardiac cycles during one breath-hold are used to reconstruct one image. For each delay time after inversion pulse, an image with typical contrast for defined TI is reconstructed. 628 JR:186, March 6

3 TrueFISP of Myocardial Infarction The inversion recovery turboflsh sequence was compared with a new pulse sequence technique: a single-shot 2D inversion recovery true- FISP technique. That new technique allows imaging of nine slices during one breath-hold with data acquisition every second heart beat. The single-shot 2D inversion recovery truefisp sequence was performed first, just after the TI scout, because the acquisition takes only a single breath-hold. Thus, no relevant change of optimal TI can be expected before the inversion recovery turboflsh images are acquired. For the single-shot inversion recovery truefisp sequence, the TE was 1.1 msec; TR, 2.2 msec; flip angle, ; and the bandwidth, 1,2 Hz/pixel. The acquisition window was 352 msec for 1 k-space lines. With data acquisition every second heart beat, imaging of one slice was possible during one cardiac cycle. Seventeen cardiac cycles were necessary to image nine slices during one single breath-hold. The single-shot inversion recovery truefisp sequence [6] involves constant radiofrequency pulsing and gradient refocusing of transverse magnetization. The nonselective inversion recovery pulse is used to improve the T1 contrast compared with the T2 contrast. For both pulse sequences, the previously determined optimal TI was used. Typical values for the field of view were mm; for the spatial resolution, mm 3 ; and for the matrix size, Identical values were used for both pulse sequences. Data Evaluation Regions of interests were defined for the hyperenhanced and infarcted areas in the normal myocardium and in the left ventricular cavity with rgus software at a Leonardo workstation (Siemens Medical Solutions). The regions of interest provide information about signal intensity, contrast, and area of infarction. For calculation of CNR, additional regions of interest were positioned in the background that provided information about the SD of background noise. The maximum transmural extent of the infarction was determined in relation to the thickness of the myocardium. The signal intensity (SI) values and the CNR values were determined for both sequence types. The CNR value for the infarcted myocardium in comparison with the normal myocardium was determined as follows: CNR = SI infarction SI normal myocardium / SD noise.(1) The CNR value for the infarcted myocardium in comparison with the blood in the left ventricular cavity was calculated as follows: CNR = SI infarction SI left ventricular cavity / SD noise.(2) The results of the CNR values were compared for both pulse sequence techniques with the Student s t test. The area of hyperenhanced myocardium was determined for each short-axis slice on inversion recovery turboflsh images and on the single-shot inversion recovery truefisp images. First, one representative slice was selected for comparison of both sequence types. Second, the single values of the area of infarction were summarized for each patient and each sequence type to determine the entire infarct volume. The results of the two pulse sequence types were compared. Paired Student s t tests were calculated for the values of the hyperenhanced area on a representative slice (in centimeters squared) and of the volume (in milliliters) of the entire infarction for a significance level of p =.1 (H : µ d =), where H means null hypothesis, µ d= µ 1 µ 2, µ 1 is the mean value of population 1, and µ 2 is the mean value of population 2. The correlation coefficient and the equation of regression were determined for the infarct volume and the area of the infarction on a representative slice for the two pulse sequences. land-ltman s plot was calculated for the volume and area of infarction. Results Eighteen infarctions were located in the anteroseptal segments, 16 were located in the inferoseptal segments, and nine in the inferolateral segments. Twenty-six infarctions showed a complete transmural extent; 15, 5 75% of the thickness of the myocardium; and two, 25 5% of the thickness of the myocardium. The inversion recovery truefisp sequence revealed a lower CNR than the inversion recovery turboflsh sequence (mean values,. vs 12.9, respectively; p =.5) for viable and nonviable myocardium, calculated using equation 1. The CNR of injured myocardium and blood in the left ventricular cavity, calculated using equation 2, was also lower for the multislice technique compared with the single-slice technique (.6 vs 1.2, respectively; p =.45). The mean values of the signal intensity of normal myocardium and infarcted myocardium and the SD of background noise were 12.2, 54.5, and 4.3 for the single-shot inversion recovery truefisp sequence and., 83.6, and 4.9 for the inversion recovery turbo- FLSH, respectively. The mean values for the signal-to-noise ratio (SNR) of infarcted myocardium were 12.6 and 17.1, respectively. The mean imaging time for the inversion recovery single-shot truefisp sequence was 16 ± 3.2 (SD) sec. The mean imaging time for the segmented inversion recovery turboflsh sequence was min 13 sec ± 45 sec. The mean volume of the injured myocardium per patient was 26.3 ± 13.4 ml with the inversion recovery turboflsh sequence and 26.1 ± 13.5 ml with the single-shot inversion recovery truefisp sequence. The Student s t test showed that there is no difference for values of infarct volume for a significance level of p =.1. The assessment of the volume of the infarction was possible with excellent correlation of both techniques (r =.96, p <.3). The regression equation was y = x, where dependent y is the volume determined with inversion recovery truefisp, and independent x is the volume determined with inversion recovery turboflsh. The value for comparison of the area of infarction on a single slice was r =.97 (p <.2). The regression equation was y = x, where dependent y is the area determined with inversion recovery truefisp, and independent x is the area determined with inversion recovery turboflsh. The Student s t test showed that there is no difference for values of infarct area on a single slice for a significance level of p =.1. Figure 2 reveals a comparison of MR images of a 63-year-old patient with occlusion of the left anterior descending artery. Figure 3 shows an anteroseptal myocardial infarction visualized with the inversion recovery truefisp sequence (Fig. 3) and with the inversion recovery turboflsh sequence (Fig. 3). Figure 4 shows scattergrams with the infarction volumes and areas determined with both pulse sequence techniques. The scattergrams include the regression lines. Figure 5 shows land-ltman s plots of the volume and area of infarction. Figure 6 shows nine MR images acquired with the single-shot inversion recovery truefisp technique during a single breath-hold. Discussion Kim et al. [1] investigated acute and chronic infarction in an animal experiment with dogs. MR images acquired min after contrast material application showed hyperenhancement with excellent spatial correlation with the histopathologic examination of necrotic areas after staining of the entire left ventricle with TTC. However, transient ischemia with reperfusion caused no hyperenhancement. JR:186, March 6 629

4 In contrast to 18 F-FDG PET, contrast-enhanced MRI is able to image the myocardium with superior spatial resolution [2]. Thus, Fig year-old woman with myocardial infarction in anteroseptal and inferoseptal segments after occlusion of left anterior descending artery., Inversion recovery true fast imaging with steady-state precession (truefisp) image shows myocardial infarction (arrow) with hyperintense signal intensity. Infarction has complete transmural extent., Inversion recovery turbo fast low-angle shot (turboflsh) image shows myocardial infarction (arrow) with hyperintense signal intensity. Infarction has complete transmural extent. rea of infarction is identical for both pulse sequence techniques. Fig year-old man with myocardial infarction in anteroseptal and inferoseptal segments after occlusion of left anterior descending artery., Inversion recovery turbo fast low-angle shot (turboflsh) image shows myocardial infarction with hyperintense signal intensity. Infarction has complete transmural extent in anteroseptal segment (arrows). Extent of infarction in a part of inferoseptal segment is 5% of thickness of myocardium., Inversion recovery true fast imaging with steady-state precession (truefisp) image shows myocardial infarction with hyperintense signal intensity. Infarction has complete transmural extent in anteroseptal segment (arrows). Extent of infarction in part of inferoseptal segment is 5% of thickness of myocardium. complete transmural infarctions can be distinguished from nontransmural infarctions in at least 25% steps of the thickness of the left ventricular myocardium. Viable myocardium has the potential to improve its contractility after revascularization even if it has a transmural extent of 5%, whereas complete transmural infarctions have a low probability to improve their contractility. Especially in patients with a reduced left ventricular ejection fraction, the prediction of possible improvement of regional wall thickening is important for their outcome after surgical revascularization. The spatial resolution we used in our study had a maximum in-plane voxel length of 1.8 mm. Thus, a subdivision of the thickness of the myocardium in four 25% steps for assessment of the transmurality of the infarction was possible for both pulse sequence techniques, because the identical spatial resolution was used for the segmented single-slice technique and the multislice technique. It has been known for several years that regions of myocardial infarction exhibit higher signal intensity than normal myocardium on T2-weighted images [9, ] and on T1- weighted images after administration of extravascular contrast agents [2, 4, 6]. Since the initial experience, many studies have been performed using a variety of pulse sequences to differentiate injured from normal myocardium. Kim et al. [2] and Simonetti et al. [6] implemented a breath-hold inversion recovery segmented turboflsh sequence for T1- weighted postcontrast imaging of infarction that produces strongly T1-weighted images because of the inversion pulse. Simonetti et al. compared pulse sequences to acquire contrast material enhanced images after myocardial infarction and found the greatest differences in regional myocardial signal intensity for the breath-hold inversion recovery turboflsh sequence. In the current literature, this sequence is the most investigated and established one [1 4, 6, 7], which is the reason we chose to use it as a standard of reference in our study. The new pulse sequence technique (inversion recovery truefisp) allows imaging nine slices during one breath-hold (Fig. 6). Thus, the entire left ventricular myocardium can be covered. Our results show a close agreement of the area and volume of infarction for the reference sequence (inversion recovery turbo- FLSH) and the inversion recovery truefisp sequence. mixed T1 to T2 contrast is known for truefisp sequences [6]. Therefore, the area of infarction could have been overestimated in patients with subacute and acute myocardial infarctions because the edema of the infarction was larger than the delayed 6 JR:186, March 6

5 TrueFISP of Myocardial Infarction IR TurboFLSH (ml) 7 5 IR TrueFISP (ml) 5 7 Fig. 4 Scattergrams reveal volumes and areas of myocardial infarctions., Scattergram reveals values of volumes of myocardial infarctions for two pulse sequence techniques: inversion recovery (IR) true fast imaging with steady-state precession (truefisp) and IR turbo fast low-angle shot (turboflsh). Values for volume of myocardial infarction show excellent correlation., Scattergram reveals values of area of myocardial infarctions on selected slice. Two pulse sequence techniques, IR truefisp and IR turboflsh, are compared. IR TurboFlash (cm 2 ) IR TrueFISP (cm 2 ) verage of IR TrueFISP and IR TurboFLSH Fig. 5 land-ltman s plots on selected slice of two pulse sequence techniques: inversion recovery true fast imaging with steady-state precession (IR truefisp) and inversion recovery turbo fast low-angle shot (IR turboflsh)., land-ltman s plot of infarct volumes. One data point is above threshold, as defined by SD., land-ltman s plot of infarct areas. One data point is above threshold and one data point below threshold, as defined by SD (IR TrueFISP IR TurboFLSH) / verage % (IR TrueFISP IR TurboFLSH) / verage % SD 45.4 Mean SD SD 41.6 Mean SD verage of IR TrueFISP and IR TurboFLSH enhancing nonviable myocardium [9, ]. The results of our study show that the area of infarction is not overestimated on the inversion recovery truefisp sequence when compared with the inversion recovery turbo- FLSH sequence. The first reason may be the use of the inversion pulse. The second reason may be the use of the individual sequence parameters, a short TR and TE, and a flip angle of. The inversion pulse and the sequence parameters support the T1 contrast compared with the T2 contrast. However, in spite of the optimization of the T1 contrast by the pulse sequence, strong signal was observed in the fluid in patients with pericardial effusion or pleural effusion, indicating that a relevant T2 contrast is still present. Therefore, the main reason for accurate detection of the volume of infarction may be the weak T2 signal and contrast caused by the edema in the infarcted muscle and the surrounding area compared with fluid or with contrast enhancement of the infarcted myocardium. Determination of the SNR values JR:186, March 6 631

6 Fig year-old man with transmural myocardial infarction after occlusion of circumflex artery. Nine MR images, acquired with inversion recovery true fast imaging with steady-state precession (truefisp) during a single breath-hold, reveal transmural infarction (arrow) as hyperenhanced region in inferolateral segments. showed a higher mean value for the truefisp sequence than for the turboflsh sequence. Thus, reduction of CNR is caused by higher signal intensity of the normal myocardium. There may be two possible reasons: either the TI value determined by the TI scout by visual evaluation needs to be corrected by a certain value compared with the inversion recovery turboflsh sequence or very low contrast uptake in normal myocardium after min causes a higher signal on the truefisp sequence than on the turboflsh sequence. Usually, the TI is adapted individually to null the signal of the normal myocardium and improve contrast between infarcted and viable myocardium. The performance of inversion recovery delayed hyperenhancement is highly sensitive to the TI selected using magnitude images. suboptimal TI can cause a reduction in contrast or even destroy the contrast between infarcted and normal myocardium completely [6]. Different methods with which to optimize the TI value are available. Different TI values can be used, and the best one can be chosen by trial and error. nother method is to use a TI scout sequence, which acquires data sets with different TI values and reduced spatial resolution during one breathhold [8, 11, 12]. From the complete image set at one slice position, the TI value of the image with the optimal contrast can be selected for imaging the entire myocardium. The CNR of infarcted myocardium versus viable myocardium was significantly lower (22%) for the multislice single-shot inversion recovery truefisp sequence than for the segmented inversion recovery turboflsh sequence. However, in our study the difference of CNR between the inversion recovery true- FISP sequence and the inversion recovery turboflsh sequence for infarction and normal myocardium was lower compared with that reported in a recent study with a small number of patients [7]. Thus, it may be possible to increase spatial resolution by factor of 2 or 3 by combining the single-shot inversion recovery truefisp pulse sequence with parallel imaging. It may also be possible to use turboflsh techniques as single-shot sequences when they are combined with parallel imaging for example, a sensitivity-encoding (SENSE) factor of 4 or 8 with an 8-channel system. However, further studies are necessary to show whether CNR is still high enough with the use of higher acceleration factors. The contrast between injured myocardium and the left ventricular cavity is relatively low for both pulse sequence types and lower for the multislice technique. However, a positive contrast that allows accurate determination of the area and volume of infarction is still present. To increase contrast between the left ventricular cavity and injured myocardium, using a longer time interval between contrast material application and imaging viability could help. n advantage of the single-shot technique compared with the inversion recovery turboflsh technique is that one slice is acquired during one R-R cycle, not one slice in a segmented way over several R-R cycles. Therefore, the single-shot technique may help to reduce respiratory artifacts in uncooperative patients and cardiac motion artifacts in patients with arrhythmias. With a short time interval for imaging the entire left ventricle, contrast change in the myocardium over time can be avoided [13]. Various methods of nuclear medicine are available to distinguish viable from nonviable myocardium. 18 F-FDG PET is known as the most reliable method [14, 15]. The thallium SPECT examination has a lower diagnostic accuracy for differentiation of viable from nonviable myocardium than 18 F-FDG PET. The reason is that infarcted and viable myocardium can be hypoperfused at rest and under pharmacologic stress or exercise [16 19]. n important disadvantage of the 18 F-FDG PET examination is a lower spatial resolution compared with delayed enhancement MRI. Furthermore, the examination of glucose metabolism can be difficult, especially in diabetic patients who are among the group of patients with elevated risk of coronary heart disease [18, 19]. In some cases, it is possible that 18 F-FDG PET shows falsepositive results in normokinetic segments of the myocardium [15]. 632 JR:186, March 6

7 TrueFISP of Myocardial Infarction In contrast to the methods of nuclear medicine, delayed enhancement MRI with a multislice approach can be performed during a shorter scanning time with less extensive preparation of the patient and contrast material. With short examination times, MRI has the potential to become a more available and more economic examination than 18 F-FDG PET []. The MR examination of myocardial viability can be combined with an examination of function and perfusion at rest and under pharmacologic stress [21]. In conclusion, single-shot inversion recovery truefisp allows reliable detection of infarction for an accurate assessment of the area and volume of infarction and an examination with high spatial resolution. This new multislice MR technique seems to be able to replace the established single-slice technique without loss of relevant diagnostic information. References 1. Kim RJ, Fieno DS, Parrish T, et al. Relationship of MRI delayed contrast enhancement to irreversible injury, infarct age, and contractile function. Circulation 1999; 9: Kim RJ, Wu E, Rafael, et al. The use of contrastenhanced magnetic resonance imaging to identify reversible myocardial dysfunction. N Engl J Med ; 343: ello D, Shah DJ, Farah GM, et al. Gadolinium cardiovascular magnetic resonance predicts reversible myocardial dysfunction and remodeling in patients with heart failure undergoing beta-blocker therapy. Circulation 3; 8: Jimenez orreguero LJ, Ruiz-Salmeron R. ssessment of myocardial viability in patients before revascularization. Rev Esp Cardiol 3; 56: Nguyen TD, Nuval, Mulukutla S, Wang Y. Direct monitoring of coronary artery motion with cardiac fat navigator echoes. Magn Reson Med 3; 5: Simonetti OP, Kim RJ, Fieno DS, et al. n improved MR imaging technique for the visualization of myocardial infarction. Radiology 1; 218: Huber, Schonberg SO, Spannagl, Rieber J, Klauss V, Reiser MF. Determining myocardial vitality in myocardial infarction: comparison of single and multislice MRI techniques with TurboFLSH and TrueFISP sequences [in German]. Radiologe 4; 44: Scheffler K, Henning JT. Quantification with inversion recovery TrueFISP. Magn Reson Med 1; 45: Chung YC, Lee VS, Laub G, Simonetti OP. Inversion recovery cine TrueFISP. ISMRM Tenth nnual Meeting May, 2, Honolulu, HI. erkeley, C: International Society for Magnetic Resonance in Imaging, 2:219. Spuentrup E, uecker, Karassimos E, Gunther RW, Stuber M. Navigator-gated and real-time motion corrected free-breathing MR imaging of myocardial late enhancement. Rofo 2; 174: Nilsson JC, Nielsn G, Groenning, et al. Sustained postinfarction myocardial oedema in humans visualised by magnetic resonance imaging. Heart 1; 85: Schulz-Menger J, Gross M, Messroghli D, Uhlich F, Dietz R, Friedrich M. Cardiovascular magnetic resonance of acute myocardial infarction at a very early stage. J m Coll Cardiol 1; 42: Oshinski JN, Yang Z, Jones JR, Mata JF, French. Imaging time after Gd-DTP injection is critical in using delayed enhancement to determine infarct size accurately with magnetic resonance imaging. Circulation 1; 4: Chiu CW, So NM, Lam WW, Chan KY, Sanderson JE. Combined first-pass perfusion and viability study at MR imaging in patients with non-st segment-elevation acute coronary syndromes: feasibility study. Radiology 3; 226: Marshall RC, Tillisch JH, Phelps ME, et al. Identification and differentiation of resting myocardial ischemia and infarction in men with positron computed tomography: 18 F-labeled fluorodeoxyglucose and N-12 ammonia. Circulation 1983; 67: Klein C, Nekolla SG, engel F, et al. ssessment of myocardial viability with contrast-enhanced magnetic resonance imaging: comparison with positron emission tomography. Circulation 1; 5: Kloner R, Przyklenk K. Hibernation and stunning of the myocardium. N Engl J Med 1991; 325: Rahimtoola SH. The hibernating myocardium. m Heart J 1989; 117: Vanoverschelde JL, Wijns W, orgers M, et al. Chronic myocardial hibernation in humans: from bedside to bench. Circulation 1997; 95: Zamorano J, Delgado J, lmeria C, et al. Reasons for discrepancies in identifying myocardial viability by thallium-1 redistribution, magnetic resonance imaging, and dobutamine echocardiography. m J Cardiol 2; 9: Kuhl HP, eek M, van der Weerdt P, et al. Myocardial viability in chronic ischemic heart disease: comparison of contrast-enhanced magnetic resonance imaging with (18)F-fluorodeoxyglucose positron emission tomography. J m Coll Cardiol 3; 41: JR:186, March 6 633