Utility of Cardiovascular Magnetic Resonance to Predict Left Ventricular Recovery After Primary Percutaneous Coronary Intervention for Patients Presenting With Acute ST-Segment Elevation Myocardial Infarction

Utility of Cardiovascular Magnetic Resonance to Predict Left Ventricular Recovery After Primary Percutaneous Coronary Intervention for Patients Presenting With Acute ST-Segment Elevation Myocardial Infarction Michael D. Shapiro, DO*, Koen Nieman, MD, Khurram Nasir, MD, MPH, Cesar H. Nomura, MD, Ammar Sarwar, MD, Maros Ferencik, MD, PhD, Suhny Abbara, MD, Udo Hoffman, MD, MPH, Herman K. Gold, MD, Ik-Kyung Jang, MD, PhD, Thomas J. Brady, MD, and Ricardo C. Cury, MD Cardiac magnetic resonance (CMR) has been shown to predict left ventricular (LV) recovery in patients after acute ST-segment elevation myocardial infarction. The purpose of this investigation was to determine the relative values of infarct transmurality and microvascular obstruction (MVO) using delayed enhancement CMR to predict LV recovery. We studied 17 patients (mean age 60 ⴞ 10 years, 14 men) presenting with first acute ST-segment elevation myocardial infarction treated with primary percutaneous coronary intervention who underwent CMR within 6 days after presentation and again at 6 months. In total 680 myocardial segments were evaluated, of which 267 (39%) demonstrated delayed hyperenhancement (DHE) and 116 (18%) demonstrated MVO. Unadjusted odds ratio (OR) for any improvement in regional LV function with increasing DHE category (<50%, 51% to 75%, >75% transmurality) was 0.20 (95% confidence interval [CI] 0.13 to 0.30, p <0.0001), whereas it was 0.40 (95% CO 0.28 to 0.55, p <0.0001) with increasing MVO category (0, <50th, >50th percentile). However, when coadjusted together, the relation remained robust with regard to degree of transmurality of DHE (OR 0.21, 95% CI 0.13 to 0.36, p <0.0001), but the relation was lost for MVO (OR 0.90, 95% CI 0.58 to 1.40, p ⴝ 0.64). In conclusion, when using the delayed enhancement technique for assessment of DHE and MVO, degree of infarct transmurality appears to be a more powerful predictor of LV recovery by CMR. © 2007 Elsevier Inc. All rights reserved. (Am J Cardiol 2007;100: 211–216)

The ability of cardiovascular magnetic resonance (CMR) to assess the transmural extent of myocardial necrosis and viability is well documented and generally regarded as the clinical standard.1,2 A large volume of data supports the use of single photon emission computed tomography for infarct quantification; however, accurate assessment of infarct transmurality is limited by its modest spatial resolution.3 Previous work has shown that CMR has the ability to predict left ventricular (LV) recovery after revascularization in the setting of chronic4 and acute5–7 myocardial infarction. Because of its excellent spatial resolution, CMR can evaluate microvascular obstruction (MVO) infarct and transmurality. Several studies have characterized the individual utilities of delayed hyperenhancement (DHE) and MVO after revascularized myocardial infarction for predicting subsequent LV recovery as assessed by improved reCardiac MR PET CT Program, Department of Radiology, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts. Manuscript received January 15, 2007; revised manuscript received and accepted February 19, 2007. Dr. Shapiro, Dr. Nasir, and Dr. Ferencik received support from Grant 1T32 HL076136-02 from the National Institutes of Health, Bethesda, Maryland. *Corresponding author: Tel: 617-726-3745; fax: 617-724-4152. E-mail address: [email protected] (M.D. Shapiro). 0002-9149/07/$ – see front matter © 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.amjcard.2007.02.079

gional LV function over time.6 –12 However, there are limited data comparing the relative values of these 2 CMR findings and debate remains as to which of these 2 parameters is more powerful in predicting LV recovery. Although a single report5 found degree of transmurality to be a better predictor of LV recovery in patients revascularized after acute ST-segment elevation myocardial infarction, MVO was assessed using a perfusion pulse sequence. The aim of the present study was to clarify the value of these CMR parameters for prediction of LV recovery in patients presenting with acute ST-segment elevation myocardial infarction using the delayed enhancement technique. Methods We prospectively and consecutively enrolled 22 patients hospitalized for a first acute ST-segment elevation myocardial infarction, which was diagnosed by history, diagnostic electrocardiographic changes, and cardiac enzyme abnormalities in accordance with consensus guidelines.13 To be included, patients had to undergo primary percutaneous coronary intervention as part of their treatment for acute myocardial infarction. Patients with clinical heart failure (intravenous inotropic therapy, intra-aortic balloon pump), ongoing ischemia, or significant arrhythmia were excluded. www.AJConline.org

212

The American Journal of Cardiology (www.AJConline.org)

The institutional review board approved the study and written informed consent was obtained from all participants. Of the 22 patients who underwent baseline magnetic resonance imaging (MRI), 5 patients discontinued participation and refused to undergo a second scan, so a total of 17 patients underwent serial CMR examinations. Baseline CMR was performed 2 to 6 days after primary percutaneous coronary intervention. Follow-up CMR was performed at an average of 6 months later to evaluate global and regional LV recovery. No patient had evidence of a new myocardial infarction between the first and second scan. MRI was performed with a 1.5-T system (GE Signa CV/I, General Electric Medical Systems, Waukesha, Wisconsin) using an 8-channel phased array cardiac coil. Baseline and follow-up CMR protocols consisted of cine MRI, first-pass perfusion, and delayed enhancement imaging. Cine images were acquired using an electrocardiographically gated, segmented k-space, steady-state free precession (FIESTA, General Electric Medical Systems) technique in the short-axis view of the left ventricle covering the entire heart from base to apex. Image parameters were 3.5-ms repetition time, 1.4-ms echo time, 192 ⫻ 192 image matrix, 34- ⫻ 34-cm field of view, 8-mm slice thickness, 2-mm spacing, and 45° flip angle. First-pass perfusion images were acquired during administration of gadopentetate dimeglumine 0.1 mmol/kg (Schering AG, Berlin, Germany) at a rate of 5 ml/s by means of a power injector (Medrad, Indianapolis, Indiana) followed by saline 20 ml at 5 ml/s. These first-pass images were obtained during a breath hold using an electrocardiographically gated saturation recovery interleaved hybrid gradient echo planar imaging sequence. Eight short-axis slices through the left ventricle were obtained every other heartbeat and repeated 30 to 50 times. Imaging parameters were 6.7-ms repetition time, 1.4-ms echo time, 4 echocardiographic train length, 128 ⫻ 128 image matrix, 20° flip angle, 90° saturation pulse, 125-kHz bandwidth, 34- ⫻ 34-cm field of view, and 8-mm slice thickness. After completion of first-pass images, a second dose of gadopentetate dimeglumine 0.1 mmol/kg was administered. Ten minutes after the second dose of gadopentetate dimeglumine, an electrocardiographically gated, breath-hold, segmented k-space, inversion recovery prepared fast gradient echo pulse sequence was performed. Imaging was performed with 7.1-ms repetition time, 3.1-ms echo time, 256 ⫻ 192 image matrix, 20° flip angle, 180° inversion pulse, and 150- to 300-ms inversion time. Inversion time was chosen dynamically to null normal myocardium. Images were transferred to a dedicated CMR image processing workstation (Advantage, General Electric, Milwaukee, Wisconsin). Cine, first-pass perfusion, and DHE images were matched for position with anatomic landmarks, including the papillary muscles and insertion of the right ventricle. Five representative slices were matched between cine and delayed enhancement in the baseline and follow-up examinations for assessment of regional LV function. Each slice was divided in 8 segments, resulting in 40 segments per patient (5 slices, 8 segments per slice). Regional LV function was assessed using a semiquantitative scale (4 ⫽ normal, 3 ⫽ mild to moderate hypokinesis, 2 ⫽ severe hypokinesis, 1 ⫽ akinesis, 0 ⫽ dyskinesis). Coregistered

Table 1 Baseline characteristics (n ⫽ 17) Age (yrs) Men Body mass index (kg/m2) Dyslipidemia (currently on lipid-lowering medication) Type 2 diabetes mellitus Hypertension Current smokers Door to balloon time (min) Time between percutaneous coronary intervention and CMR 1 (days) Time between CMR 1 and 2 (d) Characteristics of myocardial infarction Maximum troponin-1 (␮g/L) Maximum creatine phosphokinase (U/L) Maximum creatine phosphokinase-MB (␮g/L) Myocardial infarction territory Anterior Inferior Lateral Obstructive noninfarct-related artery coronary artery disease

60 ⫾ 10 14 (82%) 26.7 ⫾ 2.3 14 (82%) 3 (17%) 11 (65%) 3 (18%) 81 ⫾ 34 3⫾1 185 ⫾ 48 6.4 ⫾ 3.9 2,117 ⫾ 1,397 233 ⫾ 190 3 (18%) 11 (65%) 3 (18%) 11 (65%)

Values are means ⫾ SDs or numbers of patients (percentages).

segments from MRI 2 were evaluated for change in LV function compared with MRI 1 and an increase ⱖ1 in the segmental wall motion score was considered evidence of LV recovery.7 In addition, to quantify end-diastolic wall thickness, end-systolic wall thickness, LV volumes, and LV mass, epicardial and endocardial contours were detected automatically and corrected manually on the stack of shortaxis cine MRI images using the centerline method (Mass Analysis; Medis, Leiden, The Netherlands). Segmental wall thickening was calculated as (end-systolic wall thickening ⫺ end-diastolic wall thickening)/end-diastolic wall thickening ⫻ 100. Myocardial perfusion was evaluated on first-pass perfusion images. The perfusion defect was manually planimetered in each affected slice to obtain a perfusion defect volume. This volume was then multiplied by 1.05 g/ml to obtain the mass of the myocardial perfusion defect as percent total myocardial mass. Infarct mass was quantified with manual planimetry of regions of DHE from consecutive 2-dimensional slices, similar to the method for quantifying the size of perfusion defects. MVO was identified as an area of signal void within the area of DHE and its mass was obtained as previously described. In addition, transmurality of DHE and MVO per segment (e.g., DHE mass/mass of segment; MVO mass/ mass of segment, respectively) was determined for each of the 40 segments per patient and graded as 0% (grade 0), 1% to 25% (grade 1), 26% to 50% (grade 2), 51% to 75% (grade 3), and 76% to 100% (grade 4). Continuous variables are presented as mean ⫾ SD. Paired Student’s t test or analysis of variance was used to assess differences in continuous variables when appropriate; chi-square testing was applied for assessing differences for categorical variables. Linear and logistic regression analyses were used to assess the relation of LV recovery with DHE and MVO. All tests were 2-sided and a p value ⬍0.05

Coronary Artery Disease/CMR to Predict LV Recovery After MI

213

Table 2 Baseline and follow-up cardiac magnetic resonance characteristics CMR Characteristic

Baseline CMR

Followup CMR

p Value

End-diastolic volume (ml) 113 ⫾ 15 126 ⫾ 34 0.24 End-systolic volume (ml) 52 ⫾ 12 59 ⫾ 21 0.26 Stroke volume (ml) 62 ⫾ 9 67 ⫾ 18 0.38 Ejection fraction (%) 55 ⫾ 6 54 ⫾ 8 0.90 Myocardial mass (g) 125 ⫾ 27 103 ⫾ 15 0.006 First-pass perfusion defect (g) 11 ⫾ 6 4 ⫾ 2 ⬍0.001 DHE (g) 25 ⫾ 11 15 ⫾ 7 ⬍0.001 Microvascular obstruction (g) 4⫾3 0 ⬍0.001 Segmental end-diastolic thickness (mm) 9⫾2 8 ⫾ 2 ⬍0.001 Segmental end-systolic thickness (mm) 13 ⫾ 2 12 ⫾ 2 ⬍0.001

Figure 1. Change in infarct size over time.

was considered statistically significant. All statistical analyses were performed using dedicated software (STATA 8, STATA Corp., Austin, Texas). Results We studied 17 patients (mean age 60 ⫾ 10 years, 14 men) presenting with first acute ST-segment elevation myocardial infarction treated with primary percutaneous coronary intervention. Patient demographics are listed in Table 1. Baseline and follow-up CMR characteristics are presented in Table 2. MVO was detected in 15 of 17 patients. Figure 1 depicts the change in infarct size over time. In total, 680 segments were evaluated. At baseline, 415 (61%), 92 (14%), 89 (13%) and 84 (12%) segments demonstrated normal function, mild hypokinesia, severe hypokinesia, and akinesia, respectively. Of the 680 segments, 267 (39%) had some degree of DHE and 116 (18%) demonstrated some degree of MVO. Figure 2 depicts the relation of the degree of transmurality of DHE to MVO at baseline. Of the 413 segments without DHE, 386 (93%) demonstrated normal LV wall thickening at baseline and remained normal at follow-up CMR; the remaining 27 segments (7%) without DHE were mildly hypokinetic at baseline and 24 of 27 (89%) of these segments recovered to normal function at follow-up CMR. The remaining 3 mildly hypokinetic seg-

Figure 2. Relation of degree of transmurality of DHE and no MVO (white bars), MVO ⬍50th percentile (light gray bars), and MVO ⬎50th percentile (dark gray bars) at baseline.

ments did not demonstrate any deterioration of function at follow-up. Degree of DHE and MVO predicted LV recovery at follow-up. Figure 3 displays baseline and follow-up segmental LV functions according to degree of DHE and MVO. Figure 4 shows change in segmental LV function according to degree of DHE and MVO. Unadjusted odds ratios (OR) for any improvement in function were 0.20 (95% confidence interval [CI] 0.13 to 0.30, p ⬍0.0001) with increasing DHE category and 0.40 (95% CI 0.28 to 0.55, p ⬍0.0001) with increasing MVO category. However, when coadjusted, the OR for LV recovery remained robust for degree of DHE (OR 0.21, 95% CI 0.13 to 0.36, p ⬍0.0001), but the relation with MVO was lost (OR 0.90, 95% CI 0.58 to 1.40, p ⫽ 0.64). Quantitative analysis using systolic wall thickness showed similar results (Table 3). In unadjusted linear regression analysis, change in systolic wall thickness was negatively associated with increasing DHE category (regression coefficient ⫺0.07, 95% CI ⫺0.10 to ⫺0.03, p ⬍0.0001) and MVO (regression coefficient ⫺0.04, 95% CI ⫺0.08 to ⫺0.01, p ⬍0.0001). When coadjusted together, the relation for change in systolic wall thickness remained robust for degree of DHE (regression coefficient ⫺0.09, 95% CI ⫺0.14 to ⫺0.04, p ⫽ 0.001), but the relation with increasing MVO was nonsignificant (regression coefficient 0.009, 95% CI ⫺0.04 to 0.05, p ⫽ 0.70). Discussion Since the introduction of the delayed enhancement technique, important insights have been gained regarding prediction of LV recovery. Specifically, degree of DHE transmurality has been shown to predict LV recovery in the setting of chronic4 and acute5–7 myocardial infarction. Further studies have indicated that MVO also provides important prognostic information. For instance, Gerber et al14 found that extent of MVO had a better ability to predict LV remodeling than did total infarct size after myocardial infarction. Wu et al15 showed that CMR-determined MVO predicts more frequent major adverse cardiac events and remains a strong prognostic marker even after control for infarct size. In this study, we evaluated MVO and DHE using an inversion recovery gradient echocardiographic pulse sequence 10 minutes after infusion of gadolinium.

214

The American Journal of Cardiology (www.AJConline.org)

Figure 3. Baseline and follow-up segmental LV function according to degree of transmurality of DHE (A, B) and MVO (C, D).

Figure 4. Worse (white bars), same (light gray bars), and improved (dark gray bars) segmental LV function according to degree of transmurality of DHE (left) and MVO (right).

However, before introduction of the DHE technique, MVO was typically assessed on imaging in the first few minutes after contrast infusion as the presence of hypoenhanced regions at the core of hyperenhanced infarctions using a hybrid gradient echo planar imaging or conventional spin echo pulse sequence.15–17 The DHE technique that was used in this study has several advantages over these other techniques including improved spatial resolution and superior contrast- and signal-to-noise ratios. In addition, this technique allows for simultaneous evaluation of infarct size, transmurality, and MVO (Figure 5). As such, evaluation of MVO is currently routinely performed with the delayed enhancement technique.

Our study is in agreement with that of Gerber et al5 who found that degree of DHE is a more powerful predictor of segmental LV recovery than MVO. Not surprisingly, most (80%) segments with DHE transmurality of 1% to 50% had recovery of regional LV function. There was no difference in improvement according to baseline function between groups with 1% to 25% and 26% to 50% transmurality. Interestingly, compared with cohorts with chronic myocardial infarction,4 most segments (68%) with DHE transmurality of 51% to 75% demonstrated improvement in function on follow-up CMR, presumably secondary to myocardial edema associated with acute infarction. However, only 2% of segments with DHE transmurality ⬎75% demonstrated

Coronary Artery Disease/CMR to Predict LV Recovery After MI

215

Table 3 Quantitative analysis of segmental left ventricular function relative to degree of infarct transmurality and microvascular obstruction at baseline and follow-up CMR Degree of transmurality Baseline CMR Baseline systolic wall thickening Follow-up systolic wall thickening Change in thickening Degree of MVO Second CMR Baseline systolic wall thickening Follow-up systolic wall thickening Change in thickening

⬍50% 0.39 ⫾ 0.21 0.56 ⫾ 0.26 0.18 ⫾ 0.25

51–75% 0.24 ⫾ 0.15 0.39 ⫾ 0.27 0.15 ⫾ 0.23

76–100% 0.17 ⫾ 0.10 0.19 ⫾ 0.14 0.01 ⫾ 0.14

⬍0.0001 ⬍0.0001 0.0002

No MVO 0.37 ⫾ 0.20 0.54 ⫾ 0.20 0.18 ⫾ 0.26

MVO ⬍50th percentile 0.26 ⫾ 0.16 0.37 ⫾ 0.20 0.11 ⫾ 0.13

MVO ⱖ50th percentile 0.17 ⫾ 0.10 0.26 ⫾ 0.23 0.09 ⫾ 0.22

⬍0.0001 ⬍0.0001 0.04

Values are means ⫾ SD or percentages.

Figure 5. (A to E) Patient 1 demonstrates akinesis of the anteroseptal, anterior, and anterolateral walls on (A, B) cine imaging on baseline CMR. (C) DHE imaging shows a near transmural myocardial infarction with a high degree of MVO in the same territory as the segmental wall motion abnormalities. (D, E) Follow-up examinations show no improvement in segmental LV function. (F to J) Patient 2 demonstrates hypokinesis of the anteroseptal and anterior walls on baseline examination. (H) DHE imaging displays an infarct of ⬍50% transmurality and no MVO. (I, J) Follow-up examinations show significant improvement in segmental LV function.

recovery. We also show that degree of MVO is a function of degree of DHE. This association is not surprising because degree of DHE and MVO are related to the total size of the myocardial infarction. We found a moderate correlation between perfusion defect size on first-pass imaging and MVO as determined by the DHE technique (R ⫽ 0.78). This finding is in agreement with that of Lund et al18 who found a correlation of 0.71. Given the close correlation between MVO as assessed by these 2 techniques, it is not surprising that our results are so similar to those of Gerber et al. With regard to first-pass perfusion images in this cohort, it is important to note that (1) the size of the hypoenhanced area observed on the first-pass perfusion images was larger than the hypoenhanced area (surrounded by hyperenhancement) on the DHE images (10.6 ⫾ 6.3 vs 4.2 ⫾ 2.8 g) and (2) there was a smaller but persistent area of hypoenhancement on first-pass perfusion imaging but no area of hypoenhancement on DHE imaging on follow-up scans. These 2 observations suggest that the perfusion defect seen on first-pass perfusion may represent more than just MVO. First-pass perfusion imaging may overestimate degree of MVO because it is acquired very soon after administration of gadolinium. A study that validated the detection of microvas-

cular obstruction against histology found a good correlation at ⬃3 minutes after infusion of gadolinium.16 There are several limitations to this study. The relatively small sample makes it potentially difficult to generalize our findings. However, the results of our study are concordant to those of Gerber et al. Although we did employ strategies to semiquantitatively and quantitatively assess segmental LV function, more sophisticated methods, such as measuring circumferential shortening strain at baseline and follow-up to detect segmental LV recovery, may provide more reproducible data. Because we evaluated LV function, DHE, and MVO in the short axis only, the true apex was not included in the analysis. Observers were not blinded to whether they were analyzing the first or second CMR examination and this might have introduced bias. 1. Judd RM, Lugo-Olivieri CH, Arai M, Kondo T, Croisille P, Lima JA, Mohan V, Becker LC, Zerhouni EA. Physiological basis of myocardial contrast enhancement in fast magnetic resonance images of 2-day-old reperfused canine infarcts. Circulation 1995;92:1902–1910. 2. Kim RJ, Fieno DS, Parrish TB, Harris K, Chen EL, Simonetti O, Bundy J, Finn JP, Klocke FJ, Judd RM. Relationship of MRI delayed contrast enhancement to irreversible injury, infarct age, and contractile function. Circulation 1999;100:1992–2002.

216

The American Journal of Cardiology (www.AJConline.org)

3. Wagner A, Mahrholdt H, Holly TA, Elliott MD, Regenfus M, Parker M, Klocke FJ, Bonow RO, Kim RJ, Judd RM. Contrast-enhanced MRI and routine single photon emission computed tomography (SPECT) perfusion imaging for detection of subendocardial myocardial infarcts: an imaging study. Lancet 2003;361:374 –379. 4. Kim RJ, Wu E, Rafael A, Chen EL, Parker MA, Simonetti O, Klocke FJ, Bonow RO, Judd RM. The use of contrast-enhanced magnetic resonance imaging to identify reversible myocardial dysfunction. N Engl J Med 2000;343:1488 –1490. 5. Gerber BL, Garot J, Bluemke DA, Wu KC, Lima JAC. Accuracy of contrast-enhanced magntic resonance imaging in predicting improvement of regional myocardial function in patients after acute myocardial infarction, Circulation 2002;106:1083–1089. 6. Beek AM, Kuhl HA, Bondarenko O, Twisk JW, Hofman MB, vanDockum WG, Visser CA, van Rossum AC. Delayed contrast-enhanced magnetic resonance imaging for the prediction of regional functional improvement after acute myocardial infarction. J Am Coll Cardiol 2003;42:895–901. 7. Choi KM, Kim RJ, Gubernikoff G, Vargas JD, Parker M, Judd RM. Transmural extent of acute myocardial infarction predicts longterm improvement in contractile function. Circulation 2001;104: 1101–1107. 8. Baks T, van Geuns RJ, Biagini E, Wielopolski P, Mollet NR, Cademartiri F, van der Giessen WJ, Krestin GP, Serruys PW, Duncker DJ, de Feyter PJ. Effects of primary angioplasty for acute myocardial infarction on early and late infarct size and left ventricular wall characteristics. J Am Coll Cardiol 2006;47:40 – 44. 9. Petersen SE, Voigtlander T, Kreitner KF, Horstick G, Ziegler S, Wittlinger T, Abequewardene N, Schmitt M, Schreiber WG, Kalden P, et al. Late improvement of regional wall motion after the subacute phase of myocardial infarction treated by acute PTCA in a 6-month follow-up. J Cardiovasc Magn Reson 2003;5:487– 495. 10. Ingkanisorn WP, Rhoads KL, Aletras AH, Kellman P, Arai AE. Gadolinium delayed enhancement cardiovascular magnetic resonance correlates with clinical measures of myocardial infarction. J Am Coll Cardiol 2004;43:2253–2259.

11. Baks T, van Geuns RJ, Biagini E, Wielopolski P, Mollet NR, Cademartiri F, Boersma E, van der Giessen WJ, Krestin GP, Duncker DJ, Serruys PW, de Feyter PJ. Recovery of left ventricular function after primary angioplasty for acute myocardial infarction. Eur Heart J 2005;26:1070 –1077. 12. Taylor AJ, Al-Saadi N, Abdel-Aty H, Schulz-Menger J, Messroghli DR, Friedrich MG. Detection of acutely impaired microvascular reperfusion after infarct angioplasty with magnetic resonance imaging. Circulation 2004;109:2080 –2085. 13. Alpert JS, Thygesen K, Antman E, Bassand JP. Myocardial infarction redefined—a consensus document of the Joint European Society of Cardiology/American College of Cardiology committee for the redefinition of myocardial infarction. J Am Coll Cardiol 2000;36:959 –969. 14. Gerber BL, Rochitte CE, Melin JA, McVeigh ER, Bluemke DA, Wu KC, Lima JA. Microvascular obstruction and left ventricular remodeling early after acute myocardial infarction. Circulation 2000;101: 2734 –2741. 15. Wu KC, Zerhouni EA, Judd RM, Lugo-Olivieri CH, Barouch LA, Schilman SP, Blumenthal RS, Lima JA. Prognostic significance of microvascular obstruction by magnetic resonance imaging in patients with acute myocardial infarction. Circulation 1998;97:765–772. 16. Rogers WJ Jr, Kramer CM, Geskin G, Hu YL, Theobald TM, Vido DA, Petrulo S, Reichek N. Early contrast-enhanced MRI predicts late functional recovery after reperfused myocardial infarction. Circulation 1999;99:744 –750. 17. Rochitte CE, Lima JAC, Bluemke DA, Reeder SB, McVeigh ER, Furuta T, Becker LC, Melin JA. Magnitude and time course of microvascular obstruction and tissue injury after acute myocardial infarction. Circulation 1998;98:1006 –1011. 18. Lund GK, Stork A, Saeed M, Bansmann MP, Gerken JH, Muller V, Mester J, Higgins CB, Adam G, Meinertz T. Acute myocardial infarction: evaluation with first-pass enhancement and delayed enhancement MR imaging compared with 201 Tl SPECT imaging. Radiology 2004; 232:49 –57.