OUP user menu

CMR imaging for the evaluation of myocardial stunning after acute myocardial infarction: a meta-analysis of prospective trials

Jorge Romero, Jonathan Kahan, Iosif Kelesidis, Harikrishna Makani, Omar Wever-Pinzon, Hector Medina, Mario J. Garcia
DOI: http://dx.doi.org/10.1093/ehjci/jet040 1080-1091 First published online: 10 April 2013


Background Myocardial stunning is an important sequela of acute coronary syndromes and its determination might affect decisions on defibrillator implantation and assist devices after myocardial infarction (AMI). The aim of the study was to evaluate and compare the sensitivity, specificity, negative predictive value (NPV), and positive predictive value (PPV) of cardiac magnetic resonance imaging (CMR) assessing myocardial stunning after acute myocardial infarction using low-dose dobutamine (LDD), end-diastolic wall thickness, and contrast delayed enhancement (DE).

Methods and results A systematic review of Medline, Embase, and Cochrane for all prospective trials assessing myocardial stunning by CMR following AMI was performed using a standard approach for meta-analysis for diagnostic test and a bivariate analysis. Search results revealed 9384 studies, out of which 17 met criteria. A total of 634 patients (mean age 59 years, 85% male, mean left ventricular ejection fraction: 52%) were included. DE-CMR had a weighted sensitivity of 87% and specificity of 68% to detect myocardial stunning using 50% transmurality as a cut-off, with a PPV and NPV of 83 and 72%, respectively. With an overall diagnostic accuracy of 82%, LDD-CMR had a sensitivity of 67% and a specificity of 81%, with a PPV and NPV of 82 and 63%, respectively. LDD showed an overall accuracy of 74%.

Conclusion DE-CMR has a higher sensitivity, whereas LDD-CMR has a higher specificity for the detection of viable stunned myocardium following myocardial infarction. Whether the combination of DE and LDD may improve the prediction of myocardial recovery remains to be determined.

  • Myocardial stunning
  • Cardiac magnetic resonance imaging
  • Revascularization
  • Myocardial viability


Myocardial stunning refers to reversible myocyte dysfunction seen after blood flow has normalized post-reperfusion following an ischaemic event. Numerous scenarios can cause myocardial stunning, including coronary artery disease, global ischaemia during coronary artery bypass grafting (CABG), and plaque disruption with distal embolization during percutaneous coronary intervention (PCI).1 Stunned myocardium is considered a potentially reversible acute injury that can lead to left ventricular (LV) dysfunction and heart failure if unnoticed and/or untreated making the detection of dysfunctional myocardium post-reperfusion therapy clinically relevant.2 Myocardial stunning when coupled with haemodynamic instability prompts the utilization of several treatments. These include, among others inotropic agents, intra-aortic balloon pump (IABP) and temporary ventricular assist device support. It is important to mention that the utility of IABP support remains controversial based on the IABP-SHOCK II trial results.3 Furthermore, that the feasibility of cardiac magnetic resonance imaging (CMR) in haemodynamically unstable patients for the detection of myocardial stunning is limited.3 Furthermore, the detection of myocardial stunning with CMR while a patient is haemodynamically unstable is currently unfeasible.3 Similarly, the potential for recovery of a severely reduced left ventricular ejection fraction (LVEF) in this setting should be taken into account and the need for a cardiac defibrillator following standard therapeutic strategies could be utilized earlier based on the fact that the highest incidence of sudden cardiac death has clearly been demonstrated to be in the first 30 days after myocardial infarction (AMI).4 Nevertheless, this potential therapeutic modification has yet to be investigated as a consequence of detected myocardial stunning.57 Moreover, repeated bouts of ischaemia leading to episodes of stunning can lead to chronic LV dysfunction due to hibernating myocardium.8

There are few non-invasive modalities available to determine viability and stunned myocardium in the ischaemic heart, including echocardiography, nuclear perfusion imaging, and CMR. Dobutamine stress echocardiography detects contractile reserve, and positron emission tomography with fluorine-18-deoxyglucose shows viability via intact metabolism.9 Single photon emission computed tomography (SPECT) uses the radiotracers thallium-201, technetium-sestamibi-99 m, and technetium tetrofosmin-99 m to assess viability via myocardial metabolism and perfusion.10

CMR is one of the most versatile techniques to assess viability, with contrast delayed enhancement (DE)-CMR and low-dose dobutamine (LDD)-CMR being the main modalities used. DE-CMR uses gadolinium to determine the extent of fibrosis post-AMI in the LV wall, while LDD-CMR uses dobutamine in doses of 5–10 μg/kg/min range to assess contractile reserve by measuring wall thickness during contraction.11 A third method based on the determination of end-diastolic wall thickness (EDWT) has also been shown to be useful in patients with chronic ischaemic heart disease.12 Advantages of CMR are non-reliance on body habitus or acoustic windows, avoidance of ionizing radiation or angiography, and rapid data acquisition.11 Techniques in CMR, such as respiratory motion suppression, segment inversion recovery, and improved electrocardiographic gating, have allowed for increased quality in cross-sectional images of the heart, as well as more accuracy in determining the percentage of viable myocardial tissue.11,1315

To the best of our knowledge, no meta-analysis has been conducted to evaluate the accuracy of the different CMR techniques to detect myocardial stunning.


Search strategy

The objective of this study was to evaluate prospective studies where cine-CMR was used to assess LV regional and global function in patients admitted with ST-elevation myocardial infarction who underwent percutaneous coronary angioplasty revascularization or CABG. A search was performed using PubMed, Cochrane Central Register of Clinical Trials, and Embase using the term: (acute myocardial infarction OR myocardial infarction OR stunned myocardium OR stunning myocardium OR revascularization) AND (MRI OR magnetic resonance imaging OR magnetic resonance OR cardiac magnetic resonance OR cardiovascular magnetic resonance OR contrast-enhanced MRI). The only set limit of the search was to adult humans in peer-reviewed journals from 1966 to June 2012. Trials in abstract form without a peer-reviewed manuscript were excluded. References were also reviewed and original articles were included.

Selection criteria

Trials included in the analysis had to meet following criteria: (i) prospective study involving patients with acute myocardial infarction undergoing revascularization (i.e. PCI, pharmacologic thrombolysis or CABG); (ii) use of CMR in evaluating LV function recovery in the short (i.e. <4 weeks) and long terms (>2 months) following revascularization; (iii) allowed for sensitivity, specificity, PPV, and NPV calculations based on viability segment analysis; and (iv) use of standardized cut-offs to predict accuracy.

Data extraction

J.R, J.K., and O.W. extracted the data independently using standardized protocol and reporting form. Characteristics of each trial, initial and follow-up CMR scan, interval between revascularization, baseline demographics and number of viable segments found in each scan were extracted. In cases where data were not readily available, the main investigator of that trial was approached to supply the relevant information.

Quality assessment

The assessment of quality of each study was done by evaluating 14 items considered relevant to the review topic, based on the Quality Assessment of Diagnostic Accuracy Studies instrument (QUADAS).16

Statistical analysis

A bivariate random effects model was used to calculate weighted sensitivities, and specificities, PPV, and NPV were calculated.17 This eliminates the problem with standard summary receiver-operating curve (ROC) model, which are able to generate diagnostic odds ratios but cannot differentiate between a test that can detect viable segments (sensitivity) from a test that is better at detecting non-viable segments (specificity).18,19

We assessed between-study heterogeneity by first plotting inverse logistic sensitivity and specificity into a summary ROC curve. By doing this, we used a bivariate approach, allowing us to estimate heterogeneity, obtain PPV and NPV, compare sensitivities and specificities between different methods, and increase the selection of summary ROC curves available.2022 Figures and final analyses were created by using R (R 2.12.2 2001. The R foundation for Statistical Computing) and SAS (SAS 9.2, 2002–2008 SAS Institute, Cary, NC, USA), respectively. QUADAS and data entry was achieved by using Review Manager Revman, version 5.1 (Copenhagen, Nordic Cochrane Center, The Cochrane Collaboration, 2011).

Sensitivity analysis

Both CMR imaging methods were further evaluated depending on techniques and patient characteristics. For DE-CMR, standard deviation cut-offs, time between initial and follow-up CMR, age and per cent male were examined. For LDD-CMR, percent male, time between initial and follow-up CMR, and average age were examined. For both techniques, no factor appeared to affect the sensitivity of either DE or LDD-CMR.


Study selection

We identified 9384 articles, out of which 6301 abstracts were retrieved and reviewed for inclusion (Figure 1). Seventeen studies (Tables 1 and 2) with 634 patients (mean age 59 ± 3 years, 85% male, mean LVEF: 52% ± 3) with a total of 7076 LV segments were included in the study, using the standard American Heart Association (AHA) 17-segment model.23 Nineteen studies were excluded from final analysis because they did not meet inclusion criteria (Figure 1). Resting EDWT was not analysed because only one study met inclusion criteria.24

View this table:
Table 1

Baseline characteristics of DE-CMR studies

First author (ref. no.)Study designnMale (%)Age (years)LVEF (%)RevascularizationInitial MRI (days)Follow MRI (weeks)Technique to assess LVEFTime after Gadolinium Admin (min)Hyper-enhancement (SD above normal intensity)Cut-off for viability (%)
Beek et al.30Prospective30905951PCI7 ± 313MRI10–15>6<50
Bodi et al.25Prospective40925749PCI4 ± 124MRI10>2<50
Choi et al.33Prospective249054N/APCI<710MRI10>2<50
Engblom et al.26Prospective22996247PCI7 ± 628MRI20N/A<50
Gerbaud et al.29Prospective728058N/APCI5 ± 212MRI15N/A<50
Gerber et al.31Prospective207061N/APCI428MRI10>2<50
Janssens et al.28Prospective67825756PCI4 ± 112MRI10–20N/A<50
Kitagawa et al.34Prospective186060N/APCI5 ± 139MRI15N/A<50
Motoyasu et al.32Prospective23906351PCI25 ± 625MRI10>2<50
Natale et al.35Prospective46N/A65N/APCI5 ± 220MRI10–15N/A<50
Shapiro et al.27Prospective17806055PCI6 ± 426MRIN/AN/A<50
  • PCI, percutaneous coronary intervention; MRI, magnetic resonance imaging; N/A, none available; n, number of participants in each trial.

View this table:
Table 2

Baseline characteristics of LDD-CMR studies.

First author (ref. no.)Study designnMale (%)Age (years)LVEF (%)RevascularizationInitial MRI (days)Follow MRI (weeks)Technique to assess LVEFDobutamine dose (μg/kg/min)Cut-off viability (mm)
Barmeyer et al.36Prospective50845653PCI3.5 ± 1432MRI52
Barmeyer et al.36Prospective50845653PCI3.5 ± 1432MRI10 2
Bodi et al.25Prospective40925749PCI4 ± 124MRI102
Dendale et al.37Prospective209057N/APCI<716MRI52
Gerbaud et al.29Prospective728058N/APCI5 ± 212MRI102
Motoyasu et al.32Prospective23906451PCI25 ± 628MRI102
  • All study designs were prospective, n = number of participants in each trial. PCI, percutaneous coronary intervention; N/A, none available; MRI, magnetic resonance imaging.

Figure 1

Selection of studies. Nineteen studies were also excluded from the final analysis because they did not meet inclusion criteria. AMI, acute myocardial infarction; MRI, magnetic resonance imaging; DE, contrast delayed enhancement; LDD: low-dose dobutamine.

Baseline characteristics

Of the 16 studies, 11 studies2535 enrolling 379 patients (mean age 60 years ± 3; 83% male) and analysing 5699 LV segments evaluated stunning myocardium by DE-CMR. All 11 studies used cine-CMR for follow-up. Correspondingly, for the LDD method, six studies25,29,32,36,37 with 255 patients (mean age 58 ± 3 years; 87% male) and 1377 LV segments were evaluated for stunned myocardium, again with all using cine-MRI for follow-up (Tables 1 and 2).

Quality assessment

Using the recommended 14-item list for evaluating imaging studies using QUADAS items, 11–14 were scored poorly or were considered unclear: item 11: (‘Reference standard results blinded?’), item 12: (‘Relevant clinical information?’), item 13: (‘Uninterpretable results reported?’), and item 14: (‘Withdrawals explained?’). Item 11 refers to blinding and may affect diagnostic accuracy, otherwise known as review bias. Item 12 refers to the baseline characteristics of the patients, and may lead to selection bias. Item 13 refers to any segments that may have been uninterpretable, and can, therefore, lead to false elevations of test accuracy and item 14 may lead to test performance bias, as patients unfit are removed to improve accuracy. One article did not explain the baseline characteristics of the patients beyond per cent male and age which may have led to selection bias.29 Otherwise, all studies showed high-quality scoring on the remaining 10 items (Figures 2 and 3).

Figure 2

Methodological quality summary. QUADAS, quality assessment of diagnostic accuracy studies.

Figure 3

Methodological quality graph.

Publication bias

There was no publication bias using Egger's test (P = 0.421). Similarly, no indication of publication bias was found using Peter's test (P = 0.111).

Delayed-enhancement CMR

A total of 11 studies evaluated myocardial stunning using DE with initial CMR performed during the first week after PCI, or thrombolysis and the second CMR being performed between weeks 5 and 39 with a mean of 22 ± 9 weeks. The standard deviation used to define hyper-enhancement ranged from >2 to >6 SD and did not appear to influence the test performance characteristics. Gadolinium was the contrast agent used in all studies and images were obtained 5–20 min after administration.

All studies used a 50% wall thickness cut-off to define the presence (<50%) or absence (>50%) of viability. Seven studies also reported the results when using 0%, <25 and <75% cut-offs, allowing us to compare the performance characteristics using variable definitions. One study included the 50% cut-off in a wider range (1–75%) in the follow-up CMR and, therefore, could not be included in final analysis.38 Also, one study involved bone marrow stem cell transplantation after revascularization, therefore, only control cases were included.28 The weighted mean sensitivity and specificity for the <50% cut-off was 87% (95% CI: 86–88) and 68% (95% CI: 66–70), respectively. The PPV was found to be 83 (95% CI: 82–84) and NPV was 72 (95% CI: 70–74; Tables 3 and 4). This technique had an overall weighted accuracy of 82% (95% CI: 81–83; Table 3).

View this table:
Table 3

Sensitivities/specificities and predictive values of DE-CMR

Beek et al.3081 (221/273)63 (143/227)73 (221/305)73 (143/195)70 (364/500)
Bodi et al.2568 (40/59)93 (87/94)85 (40/47)82 (87/106)80 (127/153)
Choi et al.3392 (292/316)48 (100/208)73 (292/400)81 (100/124)70 (392/524)
Engblom et al.2679 (168/214)39 (75/195)58 (168/288)62 (75/121)60 (243/409)
Gerbaud et al.2991 (908/1003)61 (104/170)93 (908/974)52 (104/199)90 (1012/1173)
Gerber et al.3182 (179/219)64 (109/170)75 (179/240)73 (109/149)70 (288/389)
Janssens et al.2869 (42/61)71 (93/132)52 (42/81)83 (93/112)70 (135/193)
Kitagawa et al.3489 (141/158)66 (38/58)88 (141/161)69 (38/55)80 (179/216)
Motoyasu et al.3283 (146/175)72 (74/103)83 (146/175)72 (74/103)80 (220/278)
Natale et al.3592 (1065/1152)75 (310/412)91 (1065/1167)78 (310/397)90 (1375/1564)
Shapiro et al.2753 (80/150)87 (130/150)80 (80/100)65 (130/200)70 (210/300)
Weighted mean8768837282
View this table:
Table 4

Sensitivities/specificities and predictive values of dobutamine stress CMR

AuthorSensitivity (%) segmentsSpecificity (%) segmentsPPV (%) segmentsNPV (%) segmentsAccuracy
Barmeyer et al.3649 (78/158)71 (64/90)75 (78/104)44 (64/144)60 (142/233)
Barmeyer et al.3656 (87/155)71 (55/78)79 (87/110)48 (55/123)60 (142/248)
Bodi et al.2542 (25/59)93 (87/94)78 (25/32)72 (87/121)70 (112/153)
Dendale et al.3757 (43/76)88 (66/75)82 (43/52)67 (66/99)70 (109/151)
Gerbaud et al.2972 (119/165)86 (128/149)85 (119/140)74 (128/174)80 (247/314)
Motoyasu et al.3288 (156/177)81 (82/101)89 (156/175)80 (82/103)90 (238/278)
Weighted mean6781826374

Delayed-enhancement CMR using quartiles

The seven studies that reported results using additional cut-offs included a total number of 201 patients with 5558 segments analysed.

  • Cut-off 0%:30,33,34 The weighted sensitivity and specificity were 62% (95% CI: 58–65) and 81% (95% CI: 78–85, while the PPV and NPV were 81% (95% CI: 77–84) and 55% (95% CI: 52–59), respectively.

  • Cut-off <25%:26,28,30,3234 The weighted sensitivity and specificity were 68% (95% CI: 65–70) and 72% (95% CI: 69–75), while the PPV and NPV were 74 (95% CI: 72–77) and 62% (95% CI: 59–65), respectively.

  • Cut-off <75%:2628,30,33,34 The weighted sensitivity and specificity were 93% (95% CI: 91–94) and 46% (95% CI: 43–50), while the PPV and NPV were 65 (95% CI: 63–68) and 80% (95% CI: 77–84), respectively.

Dobutamine stress CMR

Stunned myocardium was assessed in six studies. All studies defined viability as a 2 mm increase in LV wall systolic wall thickening during a variable infusion rate of 5–10 μg/kg/min. One study separated results for infusion rates of 5, 10, and 20 μg/kg/min, and these results were reported separately while the 20 μg/kg/min dose was excluded.36 The average age of patients was 58 ± 3 years, with 87% male, and average LVEF was 52 ± 2% (Table 2). Initial CMR was performed in the first week except for one study which performed an initial CMR at week 4 following PCI.32 Follow-up CMR was performed between weeks 8 and 39, the average being week 24 ± 8, which was not statistically significant (Table 2). Using this technique, the mean weighted sensitivity and specificity were 67% (95% CI: 64–70) and 81% (95% CI: 78–85). The PPV and NPV were 82% (95% CI: 80–86) and 63% (95% CI: 67–71), respectively. The weighted overall accuracy was 74% (95% CI: 72–77) (Table 4).

The bivariate model showing summary diagnostic accuracies and comparing both methods are shown in Table 5. Forest plots and ROC curves are displayed in Figures 4A–E and 5A and B, respectively.

View this table:
Table 5

Summary estimates for weighted mean sensitivity, specificity, positive predictive value, negative predictive value, and overall accuracy using bivariate model

Overall (%)Delayed-enhancement CMR (%)LDD-CMR (%)P-value
Mean weighted sensitivity84.65 (83.70–85.61)87.41 (86.40–88.42)66.96 (63.99–69.92)0.02
Mean weighted specificity72. 9 (70.71–74.07)68.46 (66.48–70.43)81.45 (78.26–84.65)0.05
Mean weighted PPV83.20 (82.14–84.25)83.25 (82.12–84.38)82.83 (79.92–85.75)0.54
Mean weighted NPV69.11 (67.39–70.83)71.65 (69.62–73.68)63.26 (60.01–66.51)0.22
Mean weighted acccuracy80.95 (80.08–81.83)82.17 (81.22–83.12)74.49 (72.30–76.68)0.46
Figure 4

Forest plots of diagnostic accuracies of DE-CMR and LDD-CMR. The size of the square-plotting symbol is proportional to the same size for each study. Horizontal lines are the 95% confidence intervals, and the summary sensitivity, specificity, and predictive values are calculated based on a random effects model.

Figure 5

Hierarchical summary receiver operating characteristic (HSROC) curves. HSROC plots of (A) DE-CMR and (B) LDD-CMR. Based on combined sensitivity and specificity weighted for sample size of each data set reflected by the size of the circles, showing average sensitivity and specificity estimate of the study results (solid square) and 95% confidence region around it.


Our meta-analysis showed that DE-CMR provides the highest sensitivity and NPV for detecting stunned myocardium, while LDD-CMR provides the highest specificity and PPV. As expected, increasing the enhancement cut-off value for DE-CMR led to a higher sensitivity but a lower specificity. Alternatively, decreasing the enhancement cut-off value led to a higher specificity but a lower sensitivity. Differences in performance between the same CMR techniques for detecting myocardial stunning versus hibernating may be secondary to the differences in timing of follow-up MRI to detect stunning, a lack of a standard definition on the number of weeks needed to classify stunned myocardium in LV segments with late gadolinium enhancement and a decrease in scar volume over time which may be expected weeks to month AMI.39

Nevertheless, the distinction between reversible and irreversible myocardial damage and what exactly defines stunned myocardium is difficult to make with alternative imaging techniques. The European Society of Cardiology in their guidelines for acute heart failure defines stunned myocardium as myocardial dysfunction that persists in the short-term following restoration of blood flow and recommends the use of IABP if myocardial stunning is detected, as this is a reversible cause of severe heart failure after an AMI.40 Of note, there is no strict definition of imaging technique or time course to define stunning. These recommendations were based on consensus as well as the study by Waksman et al.41 which examined a subset of 16 patients post-revascularization complicated by cardiogenic shock, showing that IABPs were effective in treating cardiogenic shock, although this recommendation may change after the results of IABP-SHOCK II trial.3 However, whether selective use of the IABP support may be useful in those patients with stunned myocardium remains to be determined. The AHA and the American College of Cardiology have no specific recommendations in their guidelines for the detection and management of myocardial stunning.42,43

A comprehensive meta-analysis comparing CMR to other imaging techniques was not performed secondary to the limited number of studies found. Regardless, comparing our results with isolated studies using different imaging modalities, the superior results are evident. Ward et al.44 looked at gated SPECT versus transthoracic echocardiography to evaluate post-ischaemic stunning and found a specificity of 41% and PPV of 8% for segments in the moderate-severe perfusion defect range. In a recent study using resting redistribution 99-mTc-sestamibi SPECT, Javadi et al.45 found similar specificities ranging from 49 to 52%, although sensitivities were higher, ranging from 80 to 93% depending on the follow-up time used. CMR has been shown to be superior in evaluating LV function to echocardiography in previous studies, as well as a studies included in this meta-analysis.10,4648 The improved accuracy of CMR in comparison with other imaging modalities evaluating stunning can be attributed to the fact that tomographic evaluation which is feasible with CMR, in which acoustic windowing optimization is negligible, with optimal assessment of the endocardial border definition and with the added value of late gadolinium enhancement. This difference in performance may be due to the fact that while echocardiography is adequate for evaluating myocardial contractility and nuclear scintigraphy for evaluating myocardial perfusion, CMR is the only method capable of evaluating the two criteria required to define myocardial stunning simultaneously.

It is worth pointing out that the DE-CMR studies included in this meta-analysis did not take into consideration microvascular obstruction (MVO) as an additional factor when assessing for myocardial stunning. MVO has clearly demonstrated to be an important factor for long-term prognosis after AMI regardless of infarct size.49

Data regarding potential treatments targeting stunned myocardium are rapidly accumulating. To date, ionotropic agents, such as dobutamine, dopamine, and milrinone and calcium sensitizers such as levosimendan are used to manage symptoms while cardiovascular shock recovers.7 Interestingly, Glucagon like peptide-1 inhibitors and ATP-potassium channel openers have been shown to improve myocardial metabolism and appear to be a promising lead for the treatment of myocardial stunning.50,51

Clinical implications

DE-CMR has a higher sensitivity than LDD-CMR for the detection of viable stunned myocardium following myocardial infarction, while LDD-CMR has a higher specificity. Whether the combination of DE and LDD may improve the prediction of myocardial recovery remains to be determined. This can be achieved by having patients undergo DE-CMR first to assess any areas of hyper-enhancement, followed by LDD-CMR to assess areas of wall motion abnormalities. The latter of course would be unnecessary if a patient has no hyper-enhancement in the LV wall indicating a higher probability of functional recovery. Likewise, if the area of DE is >50% of the LV wall, we could assume based of the high sensitivity and NPV of this technique that the LVEF will probably remain low over time. In this case, although unproven, it is feasible to consider the patients functional status, such as ejection fraction, set, and, therefore, the clinician may consider advanced therapies such as AICDs in an earlier phase where the incidence of sudden cardiac death is the highest.4 This approach might probably gain some acceptance if the results of VEST trial demonstrate that wearable defibrillator decrease mortality in this early period after AMI (Vest Prevention of Early Sudden Death Trial and VEST Registry, ClinicalTrials.gov Identifier: NCT01446965 VEST trial). It is crucial to recognize that myocardial stunning similarly to myocardial viability assessed by DE-CMR is not an all or none phenomenon given the fact that its diagnostic accuracies significantly depends on the amount of delayed enhancement as demonstrated in this meta-analysis.


One limitation of our study was that DE-CMR and LDD-CMR diagnostic accuracies could not be statistically combined in order to evaluate how adding these methods might improve sensitivity and specificity upon each technique separately, given the fact that only two studies implemented both DE and LDD techniques on the same patients and patient level data were not provided in these articles. Finally, EDTW was not included as no studies were found in order to complete a comprehensive meta-analysis.


DE-CMR has a higher sensitivity while LDD-CMR has a higher specificity for the detection of viable stunned myocardium following myocardial infarction. Whether the combination of DE and LDD may improve the prediction of myocardial recovery remains to be determined.

Conflict of interest: none declared.


View Abstract