Comparative Analysis of Cardiac Magnetic Resonance Viability Indexes to Predict Functional Recovery After Successful Percutaneous Coronary Intervention in Acute Myocardial Infarction

The aim of this study was to examine the relative value and the influence of the association of 4 cardiac magnetic resonance (CMR) viability indexes for predicting segmental functional recovery after optimal pharmacologic therapies and early percutaneous coronary intervention in acute myocardial infarction (AMI). CMR has been shown to predict functional recovery after AMI. The relative predictive value of CMR viability indexes remains disputed and has not been described in AMI reperfused within the first 12 hours. Sixty-nine patients with a first reperfused (<12 hours) Thrombolysis In Myocardial Infarction grade 3 AMI (61 men, 57.6 ± 12.6 years) were studied on day 5 ± 2. Low-dose (10 μg/kg/min) dobutamine response (DOB), microvascular obstruction (MVO), relative delayed enhancement extent (DE), and transmural DE pattern (TMDE) were assessed in each of the 17 left ventricular segments. Segmental functional outcome was assessed by CMR at 3 months. Logistic regression and Bayesian probabilities evaluated the association between viability indexes and functional segmental outcome. At rest, 27% of segments (314 of 1,173) were dysfunctional of which 53% (165 of 314) recovered at follow-up. Odd ratios for dobutamine response, MVO, DE, and TMDE were 15.8, 5.9, 2.6, and 2.5 respectively. The probability of segmental recovery was 0.84 when dobutamine response was positive and increased successively to 0.91 when adding MVO absence, 0.94 when adding TMDE absence, and 0.97 when adding DE absence. In conclusion, contractile response to low-dose dobutamine is the best predictive factor of segmental recovery after Thrombolysis In Myocardial Infarction grade 3 early reperfused AMI. Its value is further increased by other CMR viability indexes.

Acute myocardial infarction (AMI) is associated with myocyte death and dysfunction. Early reperfusion may salvage viable myocardium resulting in reversible dysfunction. The recovery of segmental function depends on various factors including time to reperfusion, collateralization, and microvascular integrity. Early knowledge of the amount of viable myocardium is crucial to evaluate patient’s prognosis and to guide medical therapy aiming at prevention of ventricular remodeling. Assessment of segmental viability after AMI is possible using cardiac magnetic resonance (CMR) imaging of contractile reserve, microvascular obstruction (MVO), and delayed enhancement extent (DE). Whether low-dose dobutamine CMR has a higher diagnostic accuracy for prediction of recovery than other viability indexes, mainly DE, remains debatable. Some studies have investigated low-dose dobutamine CMR compared to DE in early reperfused AMI. Only a few studies have evaluated the accuracy of MVO to predict recovery after reperfused AMI. However, a direct comparison of these viability indexes from a large population optimally treated within the first 12 hours is currently lacking. The purpose of the present study was to examine the performance of the segmental response to low-dose dobutamine infusion, MVO, and DE analysis for predicting segmental outcome in patients with optimally reperfused recent AMI in a comparative and a combined manner.


All patients gave their written informed consent before inclusion and this prospective study was approved by our institution ethics committee. Seventy-two consecutive Caucasian patients (61 men, mean ± SD 57.6 ± 12.6 years of age, range 31 to 86) were included from February 2007 to September 2008. Inclusion criteria were (1) the presence of a first ST-segment elevation MI, defined by prolonged chest pain with persisting ST-segment elevation (>0.1 mV) in ≥2 contiguous leads and (2) successful reperfusion with Thrombolysis In Myocardial Infarction grade 3 flow in the infarct-related artery. Exclusion criteria were unsuccessful reperfusion (Thrombolysis In Myocardial Infarction grade ≤2 flow or residual stenosis ≥30%), hemodynamic instability, severe arrhythmia, and contraindications to CMR or to dobutamine infusion.

During transport for percutaneous coronary intervention, all patients received intravenous low-molecular-weight heparin (0.5 mg/kg) followed by subcutaneous enoxaparin 1 mg/kg, intravenous aspirin (≥250 mg), and loading doses of clopidogrel (≥300 mg). Some received prehospital thrombolysis. At admission, patients were evaluated with history and physical examination, electrocardiography, serum, and troponin T. All patients underwent percutaneous coronary intervention with bare metal stent implantation in the culprit artery. In case of unsuccessful thrombolysis, defined by persisting symptoms or ST-segment elevation ≥60 minutes after administration of thrombolysis, patients underwent rescue percutaneous coronary intervention. Conversely, after successful thrombolysis, all patients underwent coronary angiography at admission to evaluate the coronary arteries. In patients with Thrombolysis In Myocardial Infarction grade 3 flow, percutaneous coronary intervention was performed within 6 to 24 hours of the administration of thrombolysis. In all but 5 patients, glycoprotein IIb/IIIa inhibitors (abciximab) were administered during the procedure, according to the decision of the cardiologist in charge. Coronary flow before and after revascularization was graded according to the Thrombolysis In Myocardial Infarction study group classification by 2 blinded observers. A successful procedure was defined when Thrombolysis In Myocardial Infarction grade 3 flow and residual stenosis <30% were obtained. All interobserver disagreements were resolved by consensus in a joint session. Patients with no contraindications were treated with aspirin, clopidogrel, statins, β-blocking agents, and angiotensin-converting enzyme inhibitors, according to the American College of Cardiology/American Heart Association practice guidelines.

CMR was performed within a few days after AMI and 3 months later using a 1.5-T MR system (Sonata, Siemens Medical Solutions, Forchheim, Germany) with a phased-array coil and electrocardiographic triggering. Beta blockers were ceased 48 hours before CMR in all patients. Regional function of the left ventricle was assessed using breath-hold steady-state free precession sequences. Single-slice images were acquired in 2 orthogonal long-axis views (echo time 1.5 ms, repetition time 20 to 45 ms, flip angle 75°, slice thickness 6 mm, matrix 146 × 256). To encompass the entire left ventricle during 1 apnea, short-axis views were acquired using radial k-space segmented steady-state free precession sequences (echo time 1.5 ms, repetition time 20 to 30 ms, flip angle 60°, slice thickness 6 mm, matrix 64 × 128). Acquisitions were performed at rest and during the last minute after dobutamine infusion at 5 and then 10 μg/kg/min for 5 minutes at each dose. For first-pass perfusion analysis, images were obtained in 3 short-axis and 1 long-axis planes. In each plane, 5 to 8 10-mm-thick images were acquired every 2 heartbeats, resulting in 50 successive images. A T1-weighted multishot inversion recovery sequence (repetition time 6.6 ms, echo time 1.3 ms, inversion time 240 ms, flip angle 25°, matrix 128 × 128) was performed after an intravenous bolus of gadolinium 0.1 mmol/kg (Dotarem, Laboratoire Guerbet, Villepinte, France) followed by an 20-ml flush of saline solution was injected at a 5-ml/s rate using an automated device (Accutron MR, Medtron AG, Saarbrücken, Germany). A dose of gadolinium 0.1 mmol/kg was immediately reinjected for DE evaluation 15 minutes after contrast administration using a 3-dimensional turbo fast low-angle shot inversion recovery sequence (echo time 1.4 ms, repetition time 700 ms, flip angle 10°, matrix 150 × 256) resulting in 5-mm contiguous sections encompassing the left ventricle in short- and long-axis views. At follow-up, patients were evaluated clinically and using thallium myocardial scintigraphy to detect restenosis. The same CMR protocol was performed at rest only.

Names and study dates were deleted before analysis on a workstation (Leonardo, Siemens Medical Solutions). Left ventricular ejection fraction was calculated from end-diastolic and end-systolic volumes after the endocardial border were semiautomatically outlined on short-axis views. Regional analysis of each segment as defined by the American Heart Association was performed on short-axis (segments 1 to 16) and long-axis (segment 17) views. Before evaluation, segment location and slice to be analyzed were determined on the first CMR image by consensus of 2 observers. The same projections were recalled for analysis at follow-up. To perform the analysis within a practical clinical routine, 3 viability indexes were evaluated as categorical variables and 1 as a continuous variable. The continuous variable was the DE in each segment. Using the previously described segmentation of left ventricular myocardium, the area of DE was manually outlined within each segment and expressed as the percentage of the total segment area. Categorical indexes were the presence or absence of MVO, any segmental transmural DE (TMDE), and improvement of left ventricular thickening during dobutamine infusion. MVO was defined as a persisting subendothelial defect at first-pass perfusion and/or any region of hypoenhancement within the hyperenhancement area on late-enhancement studies. TMDE was defined as any DE involving the entire wall thickness, no matter what proportion of the segment involved. Left ventricular wall thickening was assessed at rest and during inotropic stimulation using the following score: 0 = normal, 1 = mild hypokinesia, 2 = severe hypokinesia, 3 = akinesia, and 4 = dyskinesia. Responding segments to dobutamine infusion were defined as dysfunctional segments with ameliorated wall thickening during inotropic stimulation (i.e., dobutamine score minus score at rest >0). Conversely, a difference ≤0 defined nonresponding segments. At follow-up, functional recovery was defined by a difference of scores between follow-up CMR and first examination at rest >0, and absence of recovery by a difference ≤0. The ratio number of responding segments/number of nonresponding segments was calculated in each patient.

All categorical indexes were randomly and blindly analyzed by 2 observers. The first observer was a radiologist with a 2-month experience in CMR. The second observer consisted of an examination in consensus by a radiologist with a 10-year experience in CMR and a cardiologist with a 6-year experience in CMR.

Agreement between observers was evaluated using AC-1 statistics. Comparisons between patients’ global functional characteristics at baseline and follow-up were achieved using paired t tests. In dysfunctional segments at rest, the influence of each index for predicting recovery was evaluated using odds ratios computed by logistic regression. A receiver operating characteristic curve was built to determine the threshold value for DE and allowed separation of segments as follows: segments were DE0 when no DE was found, DE1 when DE was lower than the determined threshold value, and DE2 when it was higher. Sensitivity and specificity of indexes were compared using Fisher’s exact test. In a per-patient analysis, the influence of the number of segments presenting with each index for predicting left ventricular remodeling was evaluated using logistic regression. Remodeling was defined as an increase in left ventricular end-diastolic volume ≥20% at follow-up. Bayes statistics were used to determine the prognostic value of each index in all segments (i.e., normal and dysfunctional segments at rest) and viability indexes association. All statistics were computed using NCSS (2001, NCSS Statistical Software, Kaysville, Utah) except AC-1 and Bayesian probabilities (2003, Excel, Microsoft Corporation, Richmond, Virginia). A p value <0.05 was considered statistically significant.


During the study, 2,105 patients were treated for AMI according to the universal definition of MI. Of these, 610 presented with a first ST-segment elevation MI. One hundred thirty-two patients were excluded based on exclusion criteria. Three hundred forty-six patients were not enrolled due to a lack of availability of an MR scanner in the few days after percutaneous coronary intervention. Sixty others declined to participate to this study, arguing the absence of individual benefit.

Three patients were excluded from the final analysis. During the time between the 2 CMR examinations, 1 patient presented with a new AMI, another presented with an intrastent stenosis, and the third underwent implantation with a pacemaker. The final analysis included 69 patients whose characteristics are listed in Table 1 . Twenty-eight patients (40.5%) presented with significant stenosis involving 1 coronary vessel, 28 (40.5%) had stenoses of 2 vessels, and 13 (19%) 3-vessel disease.

Table 1

Patient characteristics and angiographic findings (n = 69)

Age (years) 58 ± 12
Men/women 58/11
Body mass index (kg/m 2 ) 26.4 ± 4.2 (19.5–43)
Smoker 55 (80%)
Hyperlipidemia 46 (67%)
Hypertension 27 (39%)
Diabetes mellitus 7 (10%)
Peak troponin level (ng/ml) 5.2 ± 3.5 (0.6–13.9)
Initial thrombolysis
Number 50 (73%)
Successful 12 (17%)
Unsuccessful 7 (10%)
Time to reperfusion (hours) § 4.8 ± 3 (1–12)
Infarcted-related coronary artery
Left anterior descending 28 (40%)
Left circumflex 15 (22%)
Right coronary 26 (38%)
Thrombolysis In Myocardial Infarction flow grade before percutaneous coronary intervention
0 44 (64%)
1 7 (10%)
2 6 (9%)
3 12 (17%)
Platelet glycoprotein IIb/IIIa inhibitors 64 (93%)
Number of stents implanted 1.2 ± 0.6 (1–4)
Medication at discharge/follow-up
Aspirin 69 (100%)/69 (100%)
Clopidogrel 69 (100%)/69 (100%)
β blockers 68 (98%)/68 (98%)
Angiotensin-converting enzyme inhibitors 68 (98%)/68 (98%)
Statins 68 (98%)/69 (100%)

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

Hyperlipidemia was defined as total cholesterol >6.5 mmol/L, low-density lipoprotein cholesterol >4 mmol/L, or high-density lipoprotein cholesterol <1.2 mmol/L.

Hypertension was defined as systolic blood pressure ≥140 mm Hg and/or diastolic blood pressure ≥90 mm Hg.

Peak troponin was determined from successive blood samples obtained on admission in the intensive care unit, then after 6 hours, 1 time each day during 3 days, and the seventh day after onset of acute myocardial infarction.

§ Time to reperfusion was defined as time to (first) intravenous injection of thrombolytics or time to arterial puncture in patients treated with percutaneous intervention only.

First examination was obtained 5.2 ± 1.7 days (range 3 to 10) after AMI. Average examination time was 42 ± 14 minutes. All patients completed all steps of the dobutamine infusion. Five patients (7%) developed asymptomatic premature ventricular beats. Baseline heart rate was 67.3 ± 7.5 beats/min and increased significantly to 77.1 ± 8.2 beats/min during dobutamine infusion at 10 μg/kg/min (p = 0.01).

The number of dysfunctional segments per patient was 4.6 ± 2.9 (range 0 to 12). The number of segments with DE was 5.2 ± 3.0 (range 1 to 12). The number of segments presenting with a dysfunctional pattern or DE was 6.0 ± 3.0 (range 0 to 12). The second CMR image was obtained 94.4 ± 8.1 days after AMI (range 78 to 110).

There was no interobserver disagreement for MVO and TMDE assessment (AC-1 = 1). Interobserver agreement when assessing segmental response to stress and recovery at follow-up was very good (AC-1 = 0.79, 95% confidence interval 0.77 to 0.81). Interobserver disagreements were resolved by consensus.

Receiver operating characteristic analysis using DE indicated a cut-off value of 51% that allowed correct classification of 62% os segments (727 of 1,173, area under the curve 0.66). This cut-off value was used for further analysis and allowed separation of segments with a higher DE (DE2) from segments with a lower DE (DE0 and DE1). Results for each index are presented in Table 2 . At rest, 27% of segments (314 of 1,173) were dysfunctional; among them, 174 did not respond after dobutamine injection and 149 did not recover at follow-up. Five segments were normal at rest and hypokinetic under inotropic stimulation and at follow-up. They were considered nonresponding segments that did not recover. Sixteen other segments were normal at rest and during dobutamine infusion but hypokinetic at follow-up. These were considered normal segments whose function deteriorated.

Table 2

Magnetic resonance imaging results

Initial MR Imaging Follow-up MR Imaging
Normal Recovery No Recovery
(n = 838) (n = 165) (n = 170)
Response to dobutamine infusion
No dysfunction at rest 838 (100%) 0 (0%) 16 (9.5%)
Nonresponding segment 0 (0%) 46 (28%) 133 (78%)
Responding segment 0 (0%) 119 (72%) 21 (12.5%)
Microvascular obstruction
Absent 820 (98%) 128 (78%) 73 (43%)
Present 18 (2%) 37 (22%) 97 (57%)
Delayed enhancement
No delayed enhancement 742 (89%) 42 (25.5%) 32 (19%)
<51% delayed enhancement 77 (9%) 47 (28.5%) 34 (20%)
≥51% delayed enhancement 19 (2%) 76 (46%) 104 (61%)
Any transmural delayed enhancement
Absent 815 (97%) 106 (64%) 80 (47%)
Present 23 (3%) 59 (36%) 90 (53%)

Data are cross-tabulations showing segment outcome according to the presence or absence of each viability index.

Results for each index are presented in Table 3 . Taking into account sensitivity, specificity, and prevalence, their classification from the best to the poorest was dobutamine response > MVO > DE and TMDE. The probabilities of no recovery and recovery for dysfunctional segments at rest and all segments according to viability indexes are presented in Table 4 . The probability for a normal segment to deteriorate at follow-up was 0.09. The highest probability of having a normal function at follow-up was associated with a normal function at rest (p = 0.91) and then with lack of DE (p = 0.67).

Table 3

Outcome of dysfunctional segments at rest

Initial MR Imaging Follow-up MR Imaging Odds Ratio (95% confidence interval) p Value
Recovery No Recovery Recovery No Recovery
Response to dobutamine infusion
Responding segment 119 (38%) 21 (7%) 0.84 0.16
Nonresponding segment 46 (14%) 128 (41%) 15.8 (8.9–28.0) 0.25 0.75
Correctly classified segments (%) 72 86
Microvascular obstruction
Absent 128 (41%) 55 (17%) 0.68 0.32
Present 37 (12%) 94 (30%) 5.9 (3.6–9.7) 0.26 0.74
Correctly classified segments (%) 78 § 63 §
Relative surface of delayed enhancement
No to <51% delayed enhancement 70 (22%) 35 (11%) 0.64 0.36
≥51% delayed enhancement 95 (30%) 114 (37%) 2.6 (1.6–4.1) 0.43 0.57
Correctly classified segments (%) 42 § 77
Any transmural delayed enhancement
Absent 106 (34%) 62 (20%) 0.61 0.39
Present 59 (19%) 87 (27%) 2.5 (1.6–4.0) 0.38 0.62
Correctly classified segments (%) 64 § 58 §

Only gold members can continue reading. Log In or Register to continue

Stay updated, free articles. Join our Telegram channel

Dec 23, 2016 | Posted by in CARDIOLOGY | Comments Off on Comparative Analysis of Cardiac Magnetic Resonance Viability Indexes to Predict Functional Recovery After Successful Percutaneous Coronary Intervention in Acute Myocardial Infarction

Full access? Get Clinical Tree

Get Clinical Tree app for offline access