- Email: [email protected]
Functional imaging in the assessment of myocardial infarction: MR imaging vs. MDCT vs. SPECT
European Journal of Radiology 71 (2009) 480–485
Functional imaging in the assessment of myocardial infarction: MR imaging vs. MDCT vs. SPECT Andreas H. Mahnken a,b,∗ , Philipp Bruners a,b , Sven Stanzel c , Ralf Koos d , Georg Mühlenbruch a , Rolf W. Günther a , Patrick Reinartz e,f a Department of Diagnostic Radiology, RWTH Aachen University, Germany Applied Medical Engineering, Helmholtz Institute, RWTH Aachen University, Germany c Institute of Medical Statistics, RWTH Aachen University, Germany d Medical Clinic I, RWTH Aachen University, Germany e Department of Nuclear Medicine, RWTH Aachen University, Germany f Radios Center of Diagnostic Radiology and Nuclear Medicine, Duesseldorf, Germany
b
Received 4 January 2008; received in revised form 20 May 2008; accepted 4 June 2008
Abstract Purpose: To intraindividually compare magnetic resonance (MR) imaging, ECG-gated multi-detector spiral computed tomography (MDCT) and gated single photon emission computed tomography (SPECT) for the evaluation of global and regional myocardial function and the identification of myocardial perfusion abnormalities. Materials and methods: Nine patients (8 men; 55.1 ± 8.9 years) with a history of myocardial infarction (MI) were included in this retrospective study. All patients had undergone segmented k-space steady state free precession MR imaging, 99m Tc-MIBI gated myocardial perfusion SPECT and contrast enhanced ECG-gated 16-MDCT. Ventricular volumes and ejection fraction (EF) were calculated. Left ventricular (LV) wall motion at rest was analyzed. For SPECT and arterial phase MDCT perfusion abnormalities were assessed. Data was compared with Lin’s concordance–correlation coefficient (ρc ), Bland–Altman plots and kappa statistics. Results: For EF, there was an excellent concordance and correlation (ρc = 0.99) between SPECT (EF = 41.7 ± 10.4%), MDCT (EF = 42.2 ± 11.1%), and MR imaging (EF = 41.9 ± 11.4%). Considering MR imaging as standard of reference, MDCT (κ = 0.86) is superior to SPECT (κ = 0.51) for the assessment of the regional wall motion at rest. There was a good agreement between SPECT and MDCT regarding the detection of perfusion abnormalities (κ = 0.62). Conclusion: MDCT, MR imaging, and SPECT allow for the reliable assessment of global and regional left ventricular function in patients with a history of MI. MDCT also allows to some extent for the detection of perfusion abnormalities. With its potential to assess both, the coronary arteries as well as the myocardium, MDCT a promising modality for the comprehensive diagnostic work-up in patients with suspected myocardial ischemia. © 2008 Elsevier Ireland Ltd. All rights reserved. Keywords: MDCT; MR imaging; SPECT; Heart; Myocardial ischemia
1. Introduction For an efficient management of patients with myocardial ischemia, assessment of left ventricular (LV) function and the identification of myocardial perfusion defects are of particular interest to provide optimal treatment. Various imaging ∗ Corresponding author at: Department of Diagnostic Radiology, Aachen University of Technology, Pauwelsstrasse 30, D-52074 Aachen, Germany. Tel.: +49 241 8088332; fax: +49 241 8082499. E-mail address: [email protected] (A.H. Mahnken).
0720-048X/$ – see front matter © 2008 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.ejrad.2008.06.003
techniques provide information on different aspects of LVfunction and perfusion. Among others, these include magnetic resonance (MR) imaging, gated single photon emission computed tomography (SPECT), and retrospectively ECG-gated multi-detector spiral computed tomography (MDCT). While MR imaging is considered the reference standard for assessing global and regional LV-function [1], SPECT proved its value for clinical decision making [2]. While MR imaging and SPECT both focus on myocardium, MDCT has emerged as a reliable tool for the non-invasive assessment of the coronary arteries [3].
A.H. Mahnken et al. / European Journal of Radiology 71 (2009) 480–485
Although each technique is typically used to solve a particular diagnostic problem, all of them are known to provide objective data on global and regional LV-function. Therefore, these imaging approaches permit complex examination strategies for assessing LV-function and perfusion. Various studies compared two of the above-mentioned imaging modalities for assessing myocardial function or perfusion [4–7]. Some authors also reported the combination of SPECT perfusion imaging with coronary MDCT [8,9]. However, there is no data on an intraindividual face-to-face comparison of all of these imaging techniques for the comprehensive work-up of patients with myocardial ischemia. The purpose of this study was to compare cardiac MR imaging, MDCT, and gated 99m Tc-MIBI SPECT for the evaluation of global and regional LV-function and the identification of myocardial perfusion abnormalities.
2. Materials and methods For this retrospective study, data of all patients who had undergone MR imaging, MDCT, and gated 99m Tc-MIBI SPECT of the heart within a 4-week period between 1999 and 2007 were retrieved from the hospital information system. Patients with cardiac events or change of medication during this period were excluded from further data analysis. The study was approved by the institutional review board. MR imaging was performed on a 1.5-Tesla whole body MR-scanner (Intera, Philips Medical Systems, Best, The Netherlands) with a five-element cardiac synergy coil during end-expiratory breath-hold. The patients were positioned supine. 8 mm double oblique short-axis, 2-, 3-, and 4-chamber orientated images without interslice gap were acquired. A segmented k-space steady state free precession (SSFP) sequence using a prospective ECG trigger technique with retrospective adjustment of the heart phases acquiring 25 heart phases was applied. Acquisition time per image was 2.8–3.2 s, depending on the individual patient’s heart rate. Imaging parameters were as follows: TR (repetition time) = 3.1 ms, TE (echo time) = 1.6 ms, flipangle = 60◦ , field-of-view = 350 × 270 mm2 , 256 × 152 matrix, 10 phase-encoding steps per time frame. Temporal resolution at a heart rate of 65 beats per minute (bpm) was 37 ms. Retrospectively ECG-gated 16-slice CT (Sensation 16, Siemens, Forchheim, Germany) was performed during endexpiratory breath-hold with patients in supine position. A standardized examination protocol with 16 × 0.75 mm collimation, 120 kV tube voltage, 550 mA seff , 3.4 mm table feed/rotation and a tube rotation time of 420 ms was applied. For contrast enhancement 30 ml of non-ionic contrast material (Ultravist370, Bayer–Schering Healthcare, Berlin, Germany) was injected at 4 ml/s followed by 50 ml at 3 ml/s. Contrast material was followed by a 50 ml saline chaser bolus injected at 3 ml/s. The start delay was determined using the bolus tracking technique applying a threshold of 140 Hounsfield units (HU) and an additional delay of 6 s. No beta-blockers were administered prior to the examination.
481
From raw data 8 mm images without slice gap were reconstructed along the short axis as well as the 2-, 3-, and 4-chamber views every 5% of the RR-interval (3D-Recon, Siemens). A field-of-view of 180 × 180 mm2 with a 512 × 512 matrix and a medium smooth convolution kernel (B30f) were chosen. Gated myocardial perfusion scintigraphy was acquired using a triple-head ␥-camera (Siemens MultiSPECT 3, Erlangen, Germany) 60 min after intravenous application of 412 ± 16 MBq 99m Tc-MIBI. A 120◦ rotation per head was done in 20 steps of 30 s each. The zoom factor was 1.23, the matrix size 64 × 64. Reconstruction of the data sets was done by filtered backprojection with a third-order Butterworth filter and a cutoff frequency at 0.5. Each cardiac cycle was divided into eight intervals of equal length. For further analysis all MR and MDCT images were transferred to a workstation (Leonardo, Siemens) equipped with a dedicated cardiac software package (ARGUS, VA60B, Siemens). SPECT data were transferred to an ICON workstation (Siemens), where they were reoriented on the transversal planes, first parallel to the septum and then parallel to the inferior wall. The workstation was equipped with software for quantitative analysis of LV-function and perfusion (QPS/QGS, version 3.0, Cedars Sinai, Los Angeles, CA, USA) [10]. For retrospective data re-evaluation MDCT, MR, and SPECT images were each assessed by different readers who were blinded to the diagnosis as well as to the results of the other modalities. Slices from the apex to the base of the heart were analyzed. In contrast to SPECT, papillary muscles were included within the ventricular lumen by MR and MDCT. For MDCT and MR imaging, endocardial borders of the left ventricle were traced manually for all cardiac phases. SPECT data analysis was performed automatically without user interaction. End-diastolic volume (EDV), end-systolic volume (ESV), stroke volume (SV), and ejection fraction (EF) were calculated. For MDCT and MR imaging wall motion analysis cine loops in the short axis as well as the 2-, 3-, and 4-chamber orientations were evaluated. For SPECT regional wall motion analysis the QGS software package was used applying a three-dimensional surface display. Segmental wall motion was assessed applying a 17-segment model of the left ventricle [11]. Regional wall motion was scored separately: normal, hypokinetic (decreased endocardial excursion and systolic wall thickening), akinetic (absence of endocardial excursion and systolic wall thickening) and dyskinetic (paradoxical outward movement in systole). Perfusion deficits were analyzed from short axis views using the same 17-segment model of the left ventricle. From MDCT images attenuation values were determined for each myocardial segment using individually adapted regions of interest. A segment was defined as hypoperfused if at least one third of the myocardium showed a decrease of attenuation ≥20 HU [12]. SPECT perfusion analysis was performed automatically applying the QPS algorithm. Regarding SPECT, a segment was considered as hypoperfused if the percentaged tracer uptake was at least two standard deviations below the reference value taken from a database of healthy patients [13]. As only functional MR imaging was available there are no data on MR perfusion imaging.
482
A.H. Mahnken et al. / European Journal of Radiology 71 (2009) 480–485
Continuous data are given as mean value ± standard deviation. Agreement in global LV-function parameters between the different imaging modalities is determined with Bland–Altman plots and Lin’s concordance–correlation coefficient (ρc ). Wall motion and perfusion abnormalities were compared using the kappa coefficient (κ). The latter were valued as follows: 0–0.2 low, 0.21–0.4 moderate, 0.41–0.6 substantial, 0.61–0.8 good, >0.81 excellent agreement. Summary statistics and Bland–Altman plots were obtained with Medcalc 9.2 (MedCalc Software, Mariakerke, Belgium). Kappa coefficients and Lin’s concordance–correlation coefficients were computed with the SAS statistical analysis software (The SAS System for Windows, Release 9.1; SAS Institute Inc., Cary, NC, USA). 3. Results A total of 10 patients were identified from the hospital information system. All examinations were performed between 2004 and 2007, while there were no patients identified for the period from 1999 to 2003. Nine patients (8 men) with a mean age of 55.1 ± 8.9 (range: 46–72) years fulfilled the inclusion criteria, one patient suffered a second myocardial infarction (MI) within the 4-week period. All of these patients had a history of ST-segment elevation MI and coronary stent placement for revascularization of the infarct related coronary artery 59–183 days prior to MDCT, MR imaging or SPECT. Patients underwent MDCT and MR imaging (SPECT) within 2.1 ± 2.8 (22.7 ± 8.7) days. The mean interval between MR imaging and SPECT was 23.4 ± 7.5 days. All examinations were suited for analysis. Calculated mean effective radiation exposure for MDCT was 9.71 ± 2.21 mSv. The corresponding value for SPECT was 0.62 ± 0.02 mSv. For EDV, ESV, and SV there were relevant differences of the mean values between SPECT and MDCT as well as SPECT and MR imaging. SPECT showed systematically lower values for the three parameters (Table 1). Although Bland–Altman plots revealed a systematic deviation of the mean between MR imaging and SPECT for EDV and ESV, the concordance–correlation coefficient was in comparable ranges for all modalities. Best agreement between the different imaging modalities was found for EF with a concordance–correlation coefficient of ρc = 0.99 each (Fig. 1, Table 1). Using MR imaging, wall motion abnormalities were observed in 75/153 (49.0%) segments. The corresponding values for MDCT and SPECT are 75/153 (49.0%) and 67/153 (43.8%), respectively. Wall motion
Table 2 Frequency tables for wall motion analysis comparing MDCT and MR imaging (A), SPECT and MDCT (B) and SPECT and MR imaging (C) MDCT
MR imaging Normal
Hypokinetic
Dyskinetic
Akinetic
Normal Hypokinetic Dyskinetic Akinetic
77 1 0 0
1 44 1 5
0 5 2 0
0 0 0 17
SPECT
MDCT Normal
Hypokinetic
Dyskinetic
Akinetic
Normal Hypokinetic Dyskinetic Akinetic
68 10 0 0
17 30 1 2
0 3 0 0
1 7 2 12
SPECT
MR imaging Normal
Hypokinetic
Dyskinetic
Akinetic
69 9 0 0
15 31 0 5
0 2 0 0
2 8 3 9
Normal Hypokinetic Dyskinetic Akinetic
The best agreement was found when comparing MDCT and MR imaging (κ = 0.86). Table 3 Frequency table comparing SPECT and arterial phase MDCT for the detection of perfusion abnormalities SPECT
MDCT
Normal Abnormal
Normal
Abnormal
84 20
7 42
analysis showed an excellent agreement between MDCT and MR imaging (κ = 0.86). The agreement between SPECT and MDCT (κ = 0.53) as well as between SPECT and MR imaging (κ = 0.51; Table 2) turned out to be moderate. Under rest conditions, SPECT identified perfusion abnormalities in 62/153 (40.5%) myocardial segments. In MDCT, perfusion abnormalities were observed in 49/153 (32.0%) segments. The kappa coefficient showed a good agreement between both modalities for the detection of perfusion abnormalities (κ = 0.62; Fig. 2, Table 3).
Table 1 Summary of global LV-functional parameters as determined from MDCT, MR imaging, and SPECT MDCT
ESV (ml) EDV (ml) SV (ml) EF (%)
96.5 164.8 68.3 42.2
± ± ± ±
MRI
35.1 44.7 20.4 11.1
101.8 170.9 69.1 41.9
SPECT
± ± ± ±
40.1 46.7 17.6 11.4
85.0 143.0 58.0 41.7
± ± ± ±
MDCT vs. MRI
35.6 5.1 15.5 10.4
MDCT vs. SPECT
MRI vs. SPECT
Bland–Altman
ρc
Bland–Altman
ρc
Bland–Altman
ρc
−5.4 (–45.6; 34.9) −6.1 (–67.1; 55.0) −0.7 (–22.1; 20.7) 0.3 (–2.0; 2.6)
0.84 0.76 0.83 0.99
11.5 (–25.5; 48.4) 21.8 (–42.4; 86.0) 10.3 (–18.9; 39.6) 0.6 (–2.2; 3.4)
0.82 0.66 0.57 0.99
16.8 (2.4; 31.3) 27.9 (3.7; 52.1) 11.1 (–5.1; 27.3) 0.3 (–3.2; 3.7)
0.89 0.81 0.71 0.99
Results of the Bland–Altman plots are presented as mean difference with the upper and lower limits of agreement given in brackets. There is an excellent concordance and correlation (ρc ) for determining EF.
A.H. Mahnken et al. / European Journal of Radiology 71 (2009) 480–485
483
Fig. 1. Bland–Altman plots for EDV (A–C) and EF (D–F). While there is an excellent agreement between MR imaging, MDCT, and SPECT for the assessment of the EF (D–F), a relevant deviation of the mean was found between SPECT on the one hand and MDCT (B) as well as MR imaging on the other hand for EDV (C). A similar finding was observed for ESV (graphs not shown).
4. Discussion Myocardial ischemia is one of the most relevant socioeconomic challenges in medicine. Each day, about 2400
Americans die of cardiovascular disease. The estimated annual incidence of new and recurrent MI is 865,000 in the US alone. In 2007 heart disease is estimated to cause total costs of US$ 277.1 billion [14]. Therefore efficient examination strategies are
484
A.H. Mahnken et al. / European Journal of Radiology 71 (2009) 480–485
needed for the workup of patients with suspected myocardial ischemia. This is particularly true for patients with MI, as the spatial extent and degree of ischemic injury determine the individual patient’s outcome [15]. Abnormalities of the ventricular wall motion are important markers of myocardial ischemia [16], whereas global LV-function reflects the hemodynamics and the prognosis in patients with coronary artery disease [17]. Several cross-sectional imaging techniques provide the potential for assessing LV-function including ventricular volumes, regional wall motion, and myocardial perfusion. Among these techniques, MR imaging is an accepted reference standard for the evaluation of ventricular function [1], while SPECT is established for the assessment of myocardial perfusion [2]. Global LV function agreed well between MDCT and MR imaging. This finding is in accordance with previous reports
Fig. 2. 54-year-old man with a history of ST-segment elevation myocardial infarction. Side-by-side comparison of short axis MDCT (A) and SPECT images (B) acquired under resting conditions shows corresponding areas of reduced perfusion (arrows).
[6,7,18]. When compared to SPECT, there were relevant differences with respect to the LV-volumes while the agreement of the EF was nearly perfect. The differences in the LV-volumes are supposed to be due to differences in the identification of the LV-base, which was better defined on MDCT and MR images. This causes a systematic difference of the LV-volumes that will be cancelled out when calculating the EF. Differences in the spatial resolution may contribute to this effect, as 8 mm slice thickness as used for MR imaging and for MDCT is considerably below the spatial resolution of SPECT (10 mm). However, by defining the LV-base in the SPECT studies manually instead of using the automated algorithm, this problem can be overcome even with differences of the spatial resolution [19]. The use of a Butterworth may also contribute to these differences as it generally results in smaller LV-volumes, when compared to a Metz filter [20]. Another substantial source of error is the temporal resolution, which was best in MR imaging. In previous studies, however, this was of minor importance [7,19]. Nevertheless, although all of these techniques agree well they should not be used interchangeably. MDCT and MR imaging showed an excellent agreement for wall motion at rest, as it has been shown in previous reports [6]. However, the work-up of patients with ischemia requires stress testing to define risk levels for coronary events [21]. Although there is promising preliminary data on MDCT stress imaging [4], it is out of bounds for clinical routine use. This technique requires repeated radiation and contrast exposure at a level several times above SPECT. Moreover, image quality degrades due to motion artifacts caused by elevated heart rates. The latter might be overcome by introduction of dual-source CT, but stays unjustified as imaging techniques with very low or even no radiation exposure like SPECT, echocardiography or MR imaging are readily available. The unique advantage of MDCT however, is its potential to assess the coronary arteries. In severely hypoperfused areas SPECT assessment of wall motion might be hampered by the potentially inadequate count rate, which limits visualization of the myocardium in the affected segment. This effect may also influence quantification of the LV-volumes. Consequently, the evaluation of LV-function can be compromised in these low count regions. This effect is thought to contribute to the observed differences [22]. In accordance with previous reports, hypoperfused myocardium displayed decreased attenuation during arterial phase MDCT [23,24]. This finding is unspecific and represents various types of perfusion abnormalities, including MI [25]. In MI the extent of disease typically exceeds beyond the areas of decreased attenuation seen in arterial phase MDCT. SPECT has to be regarded as more sensitive to perfusion abnormalities [26]. This is reflected by the greater number of myocardial segments with perfusion abnormalities identified with SPECT when compared to MDCT. Nevertheless, the good agreement between both techniques supports the hypothesis that areas of reduced attenuation on arterial phase CT truly represent areas of reduced perfusion. There are several limitations of this study. First, only few patients were included. With all of these patients having a history of MI there will be a selection bias. Therefore, the presented data
A.H. Mahnken et al. / European Journal of Radiology 71 (2009) 480–485
does not provide information on the comparability of the evaluated imaging modalities for the assessment of the functional relevance of coronary artery stenoses. Moreover, all examinations were performed at rest. Thus, there is no information on the assessment of functional reserve or exercise induced ischemia. However, cardiac MDCT is typically performed under rest conditions. Thus, only data at rest is eligible for this side-by-side comparison. MDCT and MR imaging were analyzed manually, whereas SPECT data were assessed automatically. This may contribute to the differences between the different imaging modalities, particularly when assessing global ventricular function. Finally, the transmural extent of the ischemic injury was not assessed because all of imaging techniques compared in this study are known to be unreliable for assessing transmurality of MI and none of them is considered the standard of reference for assessing MI. Although the data has to be interpreted with care, this approach realistically reflects data handling in clinical routine. 5. Conclusion MDCT, MR imaging, and SPECT permit the assessment of global and regional LV-function. Considering MR imaging as standard of reference, MDCT is superior to SPECT for the assessment of LV-volumes and regional wall motion at rest. With SPECT as a standard of reference arterial phase MDCT also allows to some extent for the detection of LV perfusion abnormalities. With its potential to reliably assess the coronaries and the myocardium, MDCT is a promising modality for the comprehensive diagnostic work-up in patients with suspected myocardial ischemia. Nonetheless, the assessment of functional parameters including perfusion abnormalities remains an adjunct to MDCT coronary angiography. References [1] Rerkpattanapipat P, Mazur W, Link KM, Hundley WG. Assessment of cardiac function with MR imaging. Magn Reson Imaging Clin N Am 2003;11:67–80. [2] Hachamovitch R, Berman DS. The use of nuclear cardiology in clinical decision making. Semin Nucl Med 2005;35:62–72. [3] Hamon M, Morello R, Riddell JW, Hamon M. Coronary arteries: diagnostic performance of 16- versus 64-section spiral CT compared with invasive coronary angiography-meta-analysis. Radiology 2007;245:720–31. [4] Kurata A, Mochizuki T, Koyama Y, et al. Myocardial perfusion imaging using adenosine triphosphate stress multi-slice spiral computed tomography: alternative to stress myocardial perfusion scintigraphy. Circ J 2005;69:550–7. [5] Schepis T, Gaemperli O, Koepfli P, et al. Comparison of 64-slice CT with gated SPECT for evaluation of left ventricular function. J Nucl Med 2006;47:1288–94. [6] Mahnken AH, Muhlenbruch G, Koos R, et al. Automated vs. manual assessment of left ventricular function in cardiac multidetector row computed tomography: comparison with magnetic resonance imaging. Eur Radiol 2006;16:1416–23. [7] Dewey M, Muller M, Eddicks S, et al. Evaluation of global and regional left ventricular function with 16-slice computed tomography, biplane cineventriculography, and two-dimensional transthoracic echocardiography: comparison with magnetic resonance imaging. J Am Coll Cardiol 2006;48:2034–44.
485
[8] Rispler S, Keidar Z, Ghersin E, et al. Integrated single-photon emission computed tomography and computed tomography coronary angiography for the assessment of hemodynamically significant coronary artery lesions. J Am Coll Cardiol 2007;49:1059–67. [9] Hacker M, Jakobs T, Matthiesen F, et al. Comparison of spiral multidetector CT angiography and myocardial perfusion imaging in the noninvasive detection of functionally relevant coronary artery lesions: first clinical experiences. J Nucl Med 2005;46:1294–300. [10] Germano G, Kavanagh PB, Berman DS. An automatic approach to the analysis, quantitation and review of perfusion and function from myocardial perfusion SPECT images. Int J Card Imaging 1997;13:337–46. [11] Cerqueira MD, Weissman NJ, Dilsizian V, et al. Standardized myocardial segmentation and nomenclature for tomographic imaging of the heart: a statement for healthcare professionals from the Cardiac Imaging Committee of the Council on Clinical Cardiology of the American Heart Association. Circulation 2002;105:539–42. [12] Mahnken AH, Koos R, Katoh M, et al. Assessment of myocardial viability in reperfused acute myocardial infarction using 16-slice computed tomography in comparison to magnetic resonance imaging. J Am Coll Cardiol 2005;45:2042–7. [13] Sharir T, Germano G, Waechter P, et al. A new algorithm for the quantitation of myocardial perfusion SPECT. II: validation and diagnostic yield. J Nucl Med 2000;41:720–7. [14] Rosamond W, Flegal K, Friday G, et al. Heart disease and stroke statistics-2007 update: a report from the American Heart Association Statistics Committee and Stroke Statistics Subcommittee. Circulation 2007;115:e69–171. [15] Gersh BJ, Anderson JL. Thrombolysis and myocardial salvage. Results of clinical trials and the animal paradigm-paradoxic or predictable? Circulation 1993;88:296–306. [16] Lieberman AN, Weiss JL, Jugdutt BI, et al. Two-dimensional echocardiography and infarct size: relationship of regional wall motion and wall thickening to the extent of myocardial infarction in the dog. Circulation 1981;63:739–46. [17] Emond M, Mock MB, Davis KBLD, et al. Long-term survival of medically treated patients in the Coronary Artery Surgery Study (CASS) Registry. Circulation 1994;90:2645–57. [18] Juergens KU, Seifarth H, Maintz D, et al. MDCT determination of volume and function of the left ventricle: are short-axis image reformations necessary? AJR Am J Roentgenol 2006;186(6 Suppl. 2):S371– 378. [19] Schaefer WM, Lipke CS, Standke D, et al. Quantification of left ventricular volumes and ejection fraction from gated 99mTc-MIBI SPECT: MRI validation and comparison of the Emory Cardiac Tool Box with QGS and 4D-MSPECT. J Nucl Med 2005;46:1256–63. [20] Vakhtangandze T, Hall DO, Zananiri FV, Rees MR. The effect of Butterworth and Metz reconstruction filters on volume and ejection fraction calculations with 99Tcm gated myocardial SPECT. Br J Radiol 2005;78:733–6. [21] Kuijpers D, van Dijkman PR, Janssen CH, Vliegenthart R, Zijlstra F, Oudkerk M. Dobutamine stress MRI. Part II. Risk stratification with dobutamine cardiovascular magnetic resonance in patients suspected of myocardial ischemia. Eur Radiol 2004;14:2046–52. [22] Manrique A, Faraggi M, Vera P, et al. 201Tl and 99mTc-MIBI gated SPECT in patients with large perfusion defects and left ventricular dysfunction: comparison with equilibrium radionuclide angiography. J Nucl Med 1999;40:805–9. [23] Gosalia A, Haramati LB, Sheth MP, Spindola-Franco H. CT detection of acute myocardial infarction. AJR Am J Roentgenol 2004;182:1563–6. [24] Hoffmann U, Millea R, Enzweiler C, et al. Acute myocardial infarction: contrast-enhanced multi-detector row CT in a porcine model. Radiology 2004;231:697–701. [25] Francone M, Carbone I, Danti M, et al. ECG-gated multi-detector row spiral CT in the assessment of myocardial infarction: correlation with noninvasive angiographic findings. Eur Radiol 2006;16:15–24. [26] Tamaki N, Takahashi N, Kawamoto M, et al. Myocardial tomography using technetium-99m-tetrofosmin to evaluate coronary artery disease. J Nucl Med 1994;35:594–600.
Functional imaging in the assessment of myocardial infarction: MR imaging vs. MDCT vs. SPECT Andreas H. Mahnken a,b,∗ , Philipp Bruners a,b , Sven Stanzel c , Ralf Koos d , Georg Mühlenbruch a , Rolf W. Günther a , Patrick Reinartz e,f a Department of Diagnostic Radiology, RWTH Aachen University, Germany Applied Medical Engineering, Helmholtz Institute, RWTH Aachen University, Germany c Institute of Medical Statistics, RWTH Aachen University, Germany d Medical Clinic I, RWTH Aachen University, Germany e Department of Nuclear Medicine, RWTH Aachen University, Germany f Radios Center of Diagnostic Radiology and Nuclear Medicine, Duesseldorf, Germany
b
Received 4 January 2008; received in revised form 20 May 2008; accepted 4 June 2008
Abstract Purpose: To intraindividually compare magnetic resonance (MR) imaging, ECG-gated multi-detector spiral computed tomography (MDCT) and gated single photon emission computed tomography (SPECT) for the evaluation of global and regional myocardial function and the identification of myocardial perfusion abnormalities. Materials and methods: Nine patients (8 men; 55.1 ± 8.9 years) with a history of myocardial infarction (MI) were included in this retrospective study. All patients had undergone segmented k-space steady state free precession MR imaging, 99m Tc-MIBI gated myocardial perfusion SPECT and contrast enhanced ECG-gated 16-MDCT. Ventricular volumes and ejection fraction (EF) were calculated. Left ventricular (LV) wall motion at rest was analyzed. For SPECT and arterial phase MDCT perfusion abnormalities were assessed. Data was compared with Lin’s concordance–correlation coefficient (ρc ), Bland–Altman plots and kappa statistics. Results: For EF, there was an excellent concordance and correlation (ρc = 0.99) between SPECT (EF = 41.7 ± 10.4%), MDCT (EF = 42.2 ± 11.1%), and MR imaging (EF = 41.9 ± 11.4%). Considering MR imaging as standard of reference, MDCT (κ = 0.86) is superior to SPECT (κ = 0.51) for the assessment of the regional wall motion at rest. There was a good agreement between SPECT and MDCT regarding the detection of perfusion abnormalities (κ = 0.62). Conclusion: MDCT, MR imaging, and SPECT allow for the reliable assessment of global and regional left ventricular function in patients with a history of MI. MDCT also allows to some extent for the detection of perfusion abnormalities. With its potential to assess both, the coronary arteries as well as the myocardium, MDCT a promising modality for the comprehensive diagnostic work-up in patients with suspected myocardial ischemia. © 2008 Elsevier Ireland Ltd. All rights reserved. Keywords: MDCT; MR imaging; SPECT; Heart; Myocardial ischemia
1. Introduction For an efficient management of patients with myocardial ischemia, assessment of left ventricular (LV) function and the identification of myocardial perfusion defects are of particular interest to provide optimal treatment. Various imaging ∗ Corresponding author at: Department of Diagnostic Radiology, Aachen University of Technology, Pauwelsstrasse 30, D-52074 Aachen, Germany. Tel.: +49 241 8088332; fax: +49 241 8082499. E-mail address: [email protected] (A.H. Mahnken).
0720-048X/$ – see front matter © 2008 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.ejrad.2008.06.003
techniques provide information on different aspects of LVfunction and perfusion. Among others, these include magnetic resonance (MR) imaging, gated single photon emission computed tomography (SPECT), and retrospectively ECG-gated multi-detector spiral computed tomography (MDCT). While MR imaging is considered the reference standard for assessing global and regional LV-function [1], SPECT proved its value for clinical decision making [2]. While MR imaging and SPECT both focus on myocardium, MDCT has emerged as a reliable tool for the non-invasive assessment of the coronary arteries [3].
A.H. Mahnken et al. / European Journal of Radiology 71 (2009) 480–485
Although each technique is typically used to solve a particular diagnostic problem, all of them are known to provide objective data on global and regional LV-function. Therefore, these imaging approaches permit complex examination strategies for assessing LV-function and perfusion. Various studies compared two of the above-mentioned imaging modalities for assessing myocardial function or perfusion [4–7]. Some authors also reported the combination of SPECT perfusion imaging with coronary MDCT [8,9]. However, there is no data on an intraindividual face-to-face comparison of all of these imaging techniques for the comprehensive work-up of patients with myocardial ischemia. The purpose of this study was to compare cardiac MR imaging, MDCT, and gated 99m Tc-MIBI SPECT for the evaluation of global and regional LV-function and the identification of myocardial perfusion abnormalities.
2. Materials and methods For this retrospective study, data of all patients who had undergone MR imaging, MDCT, and gated 99m Tc-MIBI SPECT of the heart within a 4-week period between 1999 and 2007 were retrieved from the hospital information system. Patients with cardiac events or change of medication during this period were excluded from further data analysis. The study was approved by the institutional review board. MR imaging was performed on a 1.5-Tesla whole body MR-scanner (Intera, Philips Medical Systems, Best, The Netherlands) with a five-element cardiac synergy coil during end-expiratory breath-hold. The patients were positioned supine. 8 mm double oblique short-axis, 2-, 3-, and 4-chamber orientated images without interslice gap were acquired. A segmented k-space steady state free precession (SSFP) sequence using a prospective ECG trigger technique with retrospective adjustment of the heart phases acquiring 25 heart phases was applied. Acquisition time per image was 2.8–3.2 s, depending on the individual patient’s heart rate. Imaging parameters were as follows: TR (repetition time) = 3.1 ms, TE (echo time) = 1.6 ms, flipangle = 60◦ , field-of-view = 350 × 270 mm2 , 256 × 152 matrix, 10 phase-encoding steps per time frame. Temporal resolution at a heart rate of 65 beats per minute (bpm) was 37 ms. Retrospectively ECG-gated 16-slice CT (Sensation 16, Siemens, Forchheim, Germany) was performed during endexpiratory breath-hold with patients in supine position. A standardized examination protocol with 16 × 0.75 mm collimation, 120 kV tube voltage, 550 mA seff , 3.4 mm table feed/rotation and a tube rotation time of 420 ms was applied. For contrast enhancement 30 ml of non-ionic contrast material (Ultravist370, Bayer–Schering Healthcare, Berlin, Germany) was injected at 4 ml/s followed by 50 ml at 3 ml/s. Contrast material was followed by a 50 ml saline chaser bolus injected at 3 ml/s. The start delay was determined using the bolus tracking technique applying a threshold of 140 Hounsfield units (HU) and an additional delay of 6 s. No beta-blockers were administered prior to the examination.
481
From raw data 8 mm images without slice gap were reconstructed along the short axis as well as the 2-, 3-, and 4-chamber views every 5% of the RR-interval (3D-Recon, Siemens). A field-of-view of 180 × 180 mm2 with a 512 × 512 matrix and a medium smooth convolution kernel (B30f) were chosen. Gated myocardial perfusion scintigraphy was acquired using a triple-head ␥-camera (Siemens MultiSPECT 3, Erlangen, Germany) 60 min after intravenous application of 412 ± 16 MBq 99m Tc-MIBI. A 120◦ rotation per head was done in 20 steps of 30 s each. The zoom factor was 1.23, the matrix size 64 × 64. Reconstruction of the data sets was done by filtered backprojection with a third-order Butterworth filter and a cutoff frequency at 0.5. Each cardiac cycle was divided into eight intervals of equal length. For further analysis all MR and MDCT images were transferred to a workstation (Leonardo, Siemens) equipped with a dedicated cardiac software package (ARGUS, VA60B, Siemens). SPECT data were transferred to an ICON workstation (Siemens), where they were reoriented on the transversal planes, first parallel to the septum and then parallel to the inferior wall. The workstation was equipped with software for quantitative analysis of LV-function and perfusion (QPS/QGS, version 3.0, Cedars Sinai, Los Angeles, CA, USA) [10]. For retrospective data re-evaluation MDCT, MR, and SPECT images were each assessed by different readers who were blinded to the diagnosis as well as to the results of the other modalities. Slices from the apex to the base of the heart were analyzed. In contrast to SPECT, papillary muscles were included within the ventricular lumen by MR and MDCT. For MDCT and MR imaging, endocardial borders of the left ventricle were traced manually for all cardiac phases. SPECT data analysis was performed automatically without user interaction. End-diastolic volume (EDV), end-systolic volume (ESV), stroke volume (SV), and ejection fraction (EF) were calculated. For MDCT and MR imaging wall motion analysis cine loops in the short axis as well as the 2-, 3-, and 4-chamber orientations were evaluated. For SPECT regional wall motion analysis the QGS software package was used applying a three-dimensional surface display. Segmental wall motion was assessed applying a 17-segment model of the left ventricle [11]. Regional wall motion was scored separately: normal, hypokinetic (decreased endocardial excursion and systolic wall thickening), akinetic (absence of endocardial excursion and systolic wall thickening) and dyskinetic (paradoxical outward movement in systole). Perfusion deficits were analyzed from short axis views using the same 17-segment model of the left ventricle. From MDCT images attenuation values were determined for each myocardial segment using individually adapted regions of interest. A segment was defined as hypoperfused if at least one third of the myocardium showed a decrease of attenuation ≥20 HU [12]. SPECT perfusion analysis was performed automatically applying the QPS algorithm. Regarding SPECT, a segment was considered as hypoperfused if the percentaged tracer uptake was at least two standard deviations below the reference value taken from a database of healthy patients [13]. As only functional MR imaging was available there are no data on MR perfusion imaging.
482
A.H. Mahnken et al. / European Journal of Radiology 71 (2009) 480–485
Continuous data are given as mean value ± standard deviation. Agreement in global LV-function parameters between the different imaging modalities is determined with Bland–Altman plots and Lin’s concordance–correlation coefficient (ρc ). Wall motion and perfusion abnormalities were compared using the kappa coefficient (κ). The latter were valued as follows: 0–0.2 low, 0.21–0.4 moderate, 0.41–0.6 substantial, 0.61–0.8 good, >0.81 excellent agreement. Summary statistics and Bland–Altman plots were obtained with Medcalc 9.2 (MedCalc Software, Mariakerke, Belgium). Kappa coefficients and Lin’s concordance–correlation coefficients were computed with the SAS statistical analysis software (The SAS System for Windows, Release 9.1; SAS Institute Inc., Cary, NC, USA). 3. Results A total of 10 patients were identified from the hospital information system. All examinations were performed between 2004 and 2007, while there were no patients identified for the period from 1999 to 2003. Nine patients (8 men) with a mean age of 55.1 ± 8.9 (range: 46–72) years fulfilled the inclusion criteria, one patient suffered a second myocardial infarction (MI) within the 4-week period. All of these patients had a history of ST-segment elevation MI and coronary stent placement for revascularization of the infarct related coronary artery 59–183 days prior to MDCT, MR imaging or SPECT. Patients underwent MDCT and MR imaging (SPECT) within 2.1 ± 2.8 (22.7 ± 8.7) days. The mean interval between MR imaging and SPECT was 23.4 ± 7.5 days. All examinations were suited for analysis. Calculated mean effective radiation exposure for MDCT was 9.71 ± 2.21 mSv. The corresponding value for SPECT was 0.62 ± 0.02 mSv. For EDV, ESV, and SV there were relevant differences of the mean values between SPECT and MDCT as well as SPECT and MR imaging. SPECT showed systematically lower values for the three parameters (Table 1). Although Bland–Altman plots revealed a systematic deviation of the mean between MR imaging and SPECT for EDV and ESV, the concordance–correlation coefficient was in comparable ranges for all modalities. Best agreement between the different imaging modalities was found for EF with a concordance–correlation coefficient of ρc = 0.99 each (Fig. 1, Table 1). Using MR imaging, wall motion abnormalities were observed in 75/153 (49.0%) segments. The corresponding values for MDCT and SPECT are 75/153 (49.0%) and 67/153 (43.8%), respectively. Wall motion
Table 2 Frequency tables for wall motion analysis comparing MDCT and MR imaging (A), SPECT and MDCT (B) and SPECT and MR imaging (C) MDCT
MR imaging Normal
Hypokinetic
Dyskinetic
Akinetic
Normal Hypokinetic Dyskinetic Akinetic
77 1 0 0
1 44 1 5
0 5 2 0
0 0 0 17
SPECT
MDCT Normal
Hypokinetic
Dyskinetic
Akinetic
Normal Hypokinetic Dyskinetic Akinetic
68 10 0 0
17 30 1 2
0 3 0 0
1 7 2 12
SPECT
MR imaging Normal
Hypokinetic
Dyskinetic
Akinetic
69 9 0 0
15 31 0 5
0 2 0 0
2 8 3 9
Normal Hypokinetic Dyskinetic Akinetic
The best agreement was found when comparing MDCT and MR imaging (κ = 0.86). Table 3 Frequency table comparing SPECT and arterial phase MDCT for the detection of perfusion abnormalities SPECT
MDCT
Normal Abnormal
Normal
Abnormal
84 20
7 42
analysis showed an excellent agreement between MDCT and MR imaging (κ = 0.86). The agreement between SPECT and MDCT (κ = 0.53) as well as between SPECT and MR imaging (κ = 0.51; Table 2) turned out to be moderate. Under rest conditions, SPECT identified perfusion abnormalities in 62/153 (40.5%) myocardial segments. In MDCT, perfusion abnormalities were observed in 49/153 (32.0%) segments. The kappa coefficient showed a good agreement between both modalities for the detection of perfusion abnormalities (κ = 0.62; Fig. 2, Table 3).
Table 1 Summary of global LV-functional parameters as determined from MDCT, MR imaging, and SPECT MDCT
ESV (ml) EDV (ml) SV (ml) EF (%)
96.5 164.8 68.3 42.2
± ± ± ±
MRI
35.1 44.7 20.4 11.1
101.8 170.9 69.1 41.9
SPECT
± ± ± ±
40.1 46.7 17.6 11.4
85.0 143.0 58.0 41.7
± ± ± ±
MDCT vs. MRI
35.6 5.1 15.5 10.4
MDCT vs. SPECT
MRI vs. SPECT
Bland–Altman
ρc
Bland–Altman
ρc
Bland–Altman
ρc
−5.4 (–45.6; 34.9) −6.1 (–67.1; 55.0) −0.7 (–22.1; 20.7) 0.3 (–2.0; 2.6)
0.84 0.76 0.83 0.99
11.5 (–25.5; 48.4) 21.8 (–42.4; 86.0) 10.3 (–18.9; 39.6) 0.6 (–2.2; 3.4)
0.82 0.66 0.57 0.99
16.8 (2.4; 31.3) 27.9 (3.7; 52.1) 11.1 (–5.1; 27.3) 0.3 (–3.2; 3.7)
0.89 0.81 0.71 0.99
Results of the Bland–Altman plots are presented as mean difference with the upper and lower limits of agreement given in brackets. There is an excellent concordance and correlation (ρc ) for determining EF.
A.H. Mahnken et al. / European Journal of Radiology 71 (2009) 480–485
483
Fig. 1. Bland–Altman plots for EDV (A–C) and EF (D–F). While there is an excellent agreement between MR imaging, MDCT, and SPECT for the assessment of the EF (D–F), a relevant deviation of the mean was found between SPECT on the one hand and MDCT (B) as well as MR imaging on the other hand for EDV (C). A similar finding was observed for ESV (graphs not shown).
4. Discussion Myocardial ischemia is one of the most relevant socioeconomic challenges in medicine. Each day, about 2400
Americans die of cardiovascular disease. The estimated annual incidence of new and recurrent MI is 865,000 in the US alone. In 2007 heart disease is estimated to cause total costs of US$ 277.1 billion [14]. Therefore efficient examination strategies are
484
A.H. Mahnken et al. / European Journal of Radiology 71 (2009) 480–485
needed for the workup of patients with suspected myocardial ischemia. This is particularly true for patients with MI, as the spatial extent and degree of ischemic injury determine the individual patient’s outcome [15]. Abnormalities of the ventricular wall motion are important markers of myocardial ischemia [16], whereas global LV-function reflects the hemodynamics and the prognosis in patients with coronary artery disease [17]. Several cross-sectional imaging techniques provide the potential for assessing LV-function including ventricular volumes, regional wall motion, and myocardial perfusion. Among these techniques, MR imaging is an accepted reference standard for the evaluation of ventricular function [1], while SPECT is established for the assessment of myocardial perfusion [2]. Global LV function agreed well between MDCT and MR imaging. This finding is in accordance with previous reports
Fig. 2. 54-year-old man with a history of ST-segment elevation myocardial infarction. Side-by-side comparison of short axis MDCT (A) and SPECT images (B) acquired under resting conditions shows corresponding areas of reduced perfusion (arrows).
[6,7,18]. When compared to SPECT, there were relevant differences with respect to the LV-volumes while the agreement of the EF was nearly perfect. The differences in the LV-volumes are supposed to be due to differences in the identification of the LV-base, which was better defined on MDCT and MR images. This causes a systematic difference of the LV-volumes that will be cancelled out when calculating the EF. Differences in the spatial resolution may contribute to this effect, as 8 mm slice thickness as used for MR imaging and for MDCT is considerably below the spatial resolution of SPECT (10 mm). However, by defining the LV-base in the SPECT studies manually instead of using the automated algorithm, this problem can be overcome even with differences of the spatial resolution [19]. The use of a Butterworth may also contribute to these differences as it generally results in smaller LV-volumes, when compared to a Metz filter [20]. Another substantial source of error is the temporal resolution, which was best in MR imaging. In previous studies, however, this was of minor importance [7,19]. Nevertheless, although all of these techniques agree well they should not be used interchangeably. MDCT and MR imaging showed an excellent agreement for wall motion at rest, as it has been shown in previous reports [6]. However, the work-up of patients with ischemia requires stress testing to define risk levels for coronary events [21]. Although there is promising preliminary data on MDCT stress imaging [4], it is out of bounds for clinical routine use. This technique requires repeated radiation and contrast exposure at a level several times above SPECT. Moreover, image quality degrades due to motion artifacts caused by elevated heart rates. The latter might be overcome by introduction of dual-source CT, but stays unjustified as imaging techniques with very low or even no radiation exposure like SPECT, echocardiography or MR imaging are readily available. The unique advantage of MDCT however, is its potential to assess the coronary arteries. In severely hypoperfused areas SPECT assessment of wall motion might be hampered by the potentially inadequate count rate, which limits visualization of the myocardium in the affected segment. This effect may also influence quantification of the LV-volumes. Consequently, the evaluation of LV-function can be compromised in these low count regions. This effect is thought to contribute to the observed differences [22]. In accordance with previous reports, hypoperfused myocardium displayed decreased attenuation during arterial phase MDCT [23,24]. This finding is unspecific and represents various types of perfusion abnormalities, including MI [25]. In MI the extent of disease typically exceeds beyond the areas of decreased attenuation seen in arterial phase MDCT. SPECT has to be regarded as more sensitive to perfusion abnormalities [26]. This is reflected by the greater number of myocardial segments with perfusion abnormalities identified with SPECT when compared to MDCT. Nevertheless, the good agreement between both techniques supports the hypothesis that areas of reduced attenuation on arterial phase CT truly represent areas of reduced perfusion. There are several limitations of this study. First, only few patients were included. With all of these patients having a history of MI there will be a selection bias. Therefore, the presented data
A.H. Mahnken et al. / European Journal of Radiology 71 (2009) 480–485
does not provide information on the comparability of the evaluated imaging modalities for the assessment of the functional relevance of coronary artery stenoses. Moreover, all examinations were performed at rest. Thus, there is no information on the assessment of functional reserve or exercise induced ischemia. However, cardiac MDCT is typically performed under rest conditions. Thus, only data at rest is eligible for this side-by-side comparison. MDCT and MR imaging were analyzed manually, whereas SPECT data were assessed automatically. This may contribute to the differences between the different imaging modalities, particularly when assessing global ventricular function. Finally, the transmural extent of the ischemic injury was not assessed because all of imaging techniques compared in this study are known to be unreliable for assessing transmurality of MI and none of them is considered the standard of reference for assessing MI. Although the data has to be interpreted with care, this approach realistically reflects data handling in clinical routine. 5. Conclusion MDCT, MR imaging, and SPECT permit the assessment of global and regional LV-function. Considering MR imaging as standard of reference, MDCT is superior to SPECT for the assessment of LV-volumes and regional wall motion at rest. With SPECT as a standard of reference arterial phase MDCT also allows to some extent for the detection of LV perfusion abnormalities. With its potential to reliably assess the coronaries and the myocardium, MDCT is a promising modality for the comprehensive diagnostic work-up in patients with suspected myocardial ischemia. Nonetheless, the assessment of functional parameters including perfusion abnormalities remains an adjunct to MDCT coronary angiography. References [1] Rerkpattanapipat P, Mazur W, Link KM, Hundley WG. Assessment of cardiac function with MR imaging. Magn Reson Imaging Clin N Am 2003;11:67–80. [2] Hachamovitch R, Berman DS. The use of nuclear cardiology in clinical decision making. Semin Nucl Med 2005;35:62–72. [3] Hamon M, Morello R, Riddell JW, Hamon M. Coronary arteries: diagnostic performance of 16- versus 64-section spiral CT compared with invasive coronary angiography-meta-analysis. Radiology 2007;245:720–31. [4] Kurata A, Mochizuki T, Koyama Y, et al. Myocardial perfusion imaging using adenosine triphosphate stress multi-slice spiral computed tomography: alternative to stress myocardial perfusion scintigraphy. Circ J 2005;69:550–7. [5] Schepis T, Gaemperli O, Koepfli P, et al. Comparison of 64-slice CT with gated SPECT for evaluation of left ventricular function. J Nucl Med 2006;47:1288–94. [6] Mahnken AH, Muhlenbruch G, Koos R, et al. Automated vs. manual assessment of left ventricular function in cardiac multidetector row computed tomography: comparison with magnetic resonance imaging. Eur Radiol 2006;16:1416–23. [7] Dewey M, Muller M, Eddicks S, et al. Evaluation of global and regional left ventricular function with 16-slice computed tomography, biplane cineventriculography, and two-dimensional transthoracic echocardiography: comparison with magnetic resonance imaging. J Am Coll Cardiol 2006;48:2034–44.
485
[8] Rispler S, Keidar Z, Ghersin E, et al. Integrated single-photon emission computed tomography and computed tomography coronary angiography for the assessment of hemodynamically significant coronary artery lesions. J Am Coll Cardiol 2007;49:1059–67. [9] Hacker M, Jakobs T, Matthiesen F, et al. Comparison of spiral multidetector CT angiography and myocardial perfusion imaging in the noninvasive detection of functionally relevant coronary artery lesions: first clinical experiences. J Nucl Med 2005;46:1294–300. [10] Germano G, Kavanagh PB, Berman DS. An automatic approach to the analysis, quantitation and review of perfusion and function from myocardial perfusion SPECT images. Int J Card Imaging 1997;13:337–46. [11] Cerqueira MD, Weissman NJ, Dilsizian V, et al. Standardized myocardial segmentation and nomenclature for tomographic imaging of the heart: a statement for healthcare professionals from the Cardiac Imaging Committee of the Council on Clinical Cardiology of the American Heart Association. Circulation 2002;105:539–42. [12] Mahnken AH, Koos R, Katoh M, et al. Assessment of myocardial viability in reperfused acute myocardial infarction using 16-slice computed tomography in comparison to magnetic resonance imaging. J Am Coll Cardiol 2005;45:2042–7. [13] Sharir T, Germano G, Waechter P, et al. A new algorithm for the quantitation of myocardial perfusion SPECT. II: validation and diagnostic yield. J Nucl Med 2000;41:720–7. [14] Rosamond W, Flegal K, Friday G, et al. Heart disease and stroke statistics-2007 update: a report from the American Heart Association Statistics Committee and Stroke Statistics Subcommittee. Circulation 2007;115:e69–171. [15] Gersh BJ, Anderson JL. Thrombolysis and myocardial salvage. Results of clinical trials and the animal paradigm-paradoxic or predictable? Circulation 1993;88:296–306. [16] Lieberman AN, Weiss JL, Jugdutt BI, et al. Two-dimensional echocardiography and infarct size: relationship of regional wall motion and wall thickening to the extent of myocardial infarction in the dog. Circulation 1981;63:739–46. [17] Emond M, Mock MB, Davis KBLD, et al. Long-term survival of medically treated patients in the Coronary Artery Surgery Study (CASS) Registry. Circulation 1994;90:2645–57. [18] Juergens KU, Seifarth H, Maintz D, et al. MDCT determination of volume and function of the left ventricle: are short-axis image reformations necessary? AJR Am J Roentgenol 2006;186(6 Suppl. 2):S371– 378. [19] Schaefer WM, Lipke CS, Standke D, et al. Quantification of left ventricular volumes and ejection fraction from gated 99mTc-MIBI SPECT: MRI validation and comparison of the Emory Cardiac Tool Box with QGS and 4D-MSPECT. J Nucl Med 2005;46:1256–63. [20] Vakhtangandze T, Hall DO, Zananiri FV, Rees MR. The effect of Butterworth and Metz reconstruction filters on volume and ejection fraction calculations with 99Tcm gated myocardial SPECT. Br J Radiol 2005;78:733–6. [21] Kuijpers D, van Dijkman PR, Janssen CH, Vliegenthart R, Zijlstra F, Oudkerk M. Dobutamine stress MRI. Part II. Risk stratification with dobutamine cardiovascular magnetic resonance in patients suspected of myocardial ischemia. Eur Radiol 2004;14:2046–52. [22] Manrique A, Faraggi M, Vera P, et al. 201Tl and 99mTc-MIBI gated SPECT in patients with large perfusion defects and left ventricular dysfunction: comparison with equilibrium radionuclide angiography. J Nucl Med 1999;40:805–9. [23] Gosalia A, Haramati LB, Sheth MP, Spindola-Franco H. CT detection of acute myocardial infarction. AJR Am J Roentgenol 2004;182:1563–6. [24] Hoffmann U, Millea R, Enzweiler C, et al. Acute myocardial infarction: contrast-enhanced multi-detector row CT in a porcine model. Radiology 2004;231:697–701. [25] Francone M, Carbone I, Danti M, et al. ECG-gated multi-detector row spiral CT in the assessment of myocardial infarction: correlation with noninvasive angiographic findings. Eur Radiol 2006;16:15–24. [26] Tamaki N, Takahashi N, Kawamoto M, et al. Myocardial tomography using technetium-99m-tetrofosmin to evaluate coronary artery disease. J Nucl Med 1994;35:594–600.