Statistics from Altmetric.com
Heart failure has become a major problem in clinical cardiology, with recent estimations showing that 4.9 million patients in the USA have chronic heart failure, with 550 000 new patients diagnosed annually, resulting in 970 000 hospitalisations.1 It appears that coronary artery disease is the underlying cause of heart failure in > 70% of patients.1 The available therapeutic options include optimised medical treatment, heart transplantation, revascularisation (with optional left ventricular (LV) restoration and/or mitral valve repair) and cardiac resynchronisation therapy. Currently, substantial effort is invested in the development of gene and cell therapy for treatment of heart failure. In daily clinical practice, the choice is frequently between medical treatment and revascularisation. From this perspective, assessment of viability is important to guide management of patients with ischaemic LV dysfunction; patients with viable myocardium may improve in LV function after revascularisation, whereas patients with only scar tissue will not improve. In view of the increased risk for (peri-) operative complications, a pre-operative evaluation for myocardial viability is thus warranted, to select the patients who may benefit from surgery. Currently, many techniques are available for identification of dysfunctional but viable myocardium.w1 The most frequently used techniques in the clinical setting include nuclear imaging with positron emission tomography (PET) and single photon emission computed tomography (SPECT) to assess myocardial metabolism, perfusion, cell membrane and mitochondrial intactness; and echocardiographic imaging using dobutamine stress to assess contractile reserve/ischaemia or contrast agents to assess myocardial perfusion. More recently cardiac magnetic resonance (CMR) has become popular for the assessment of myocardial viability. This technique has an excellent spatial resolution and is currently the only imaging modality that allows distinction between transmural and subendocardial processes. Various techniques with CMR provide information on myocardial viability; these will be discussed extensively in this article. Besides assessment of viability, CMR provides for the surgeon additional information needed to select the optimal surgical strategy, including information on LV function, LV volumes, the presence of LV aneurysms (which need LV aneurysmectomy), and ischaemic mitral regurgitation (which require mitral valve repair in terms of restrictive annuloplasty). This article discusses the value of CMR (in particular the different CMR techniques to assess viability) for the evaluation of patients with ischaemic cardiomyopathy. Before entering this discussion, some definitions of viability and the clinical relevance of viability assessment are addressed.
The observational work by Rahimtoolaw2 resulted in the awareness that LV dysfunction is not necessarily an irreversible process, but that improvement of LV function is possible after revascularisation. This improvement has been related to the presence of dysfunctional but viable myocardium, which has the potential to recover in function after adequate restoration of blood flow. Over the years, many studies have focused on the identification of viable myocardium and prediction of improvement of function post-revascularisation. Considerable confusion has been raised about the appropriate definition of viable myocardium. Viable myocardium theoretically includes an entire spectrum, ranging from normal myocardium to epicardial regions of viable myocardium in non-transmural infarction. Viable, normal myocardium is not associated with chronic dysfunction. In the clinical setting, viability is only important in regions with chronic contractile dysfunction, to assess whether revascularisation will result in improvement of function. Rahimtoola popularised the concept of hibernation, which refers to a condition of chronically reduced/absent contraction secondary to chronic hypoperfusion in patients with coronary artery disease, in whom revascularisation will result in recovery of function.2 Studies with PET, however, showed that chronically dysfunctional myocardium frequently had (near-) normal resting blood flow instead of reduced blood flow. Further studies subsequently revealed that not resting blood flow, but rather flow reserve was reduced in patients with chronically dysfunctional myocardium. These findings have led to the hypothesis that repeated ischaemic attacks may result in chronic contractile dysfunction, with flow remaining normal or mildly reduced—a situation referred to as “repetitive stunning”.3 From a clinical viewpoint, the differentiation between repetitive stunning and hibernation may not be that important, since revascularisation is required in both conditions in order to improve contractile function; from a practical viewpoint, both conditions can be grouped as “jeopardised myocardium”.
Many studies (using all different imaging techniques) aiming at the prediction of functional improvement post-revascularisation reported a lower specificity, indicating that many dysfunctional segments that were classified as viable did not improve in function post-revascularisation. This is (partially) caused by regions that contain subendocardial scar tissue. In these regions, the epicardial layers still contain viable tissue. However, the question in these regions is whether these epicardial layers contain viable, but normal, myocardium or jeopardised myocardium (fig 1). Only in the latter condition can improvement of function be anticipated. Differentiation between these two conditions is currently one of the most difficult and challenging issues in assessing myocardial viability. All available imaging techniques do not have the spatial resolution to differentiate epicardial and endocardial processes. With CMR however, this differentiation is possible.
VIABILITY: CLINICAL RELEVANCE
Viability assessment is clinically relevant for patient management. Improvement of function after revascularisation is still considered the “gold standard” for viability. Pooled data from 105 viability studies (using nuclear imaging or stress echocardiography) showed a mean sensitivity and specificity of 84% and 69% to predict recovery of regional function after revascularisation.4 Improvement in LV ejection fraction (LVEF) was evaluated in 28 studies; patients with viable myocardium exhibited an improvement in LVEF from on average 37% to 45%, whereas patients without viability did not improve in LVEF (36% before revascularisation versus 36% post-revascularisation).4 Finally, viability is related to prognosis. The available 17 prognostic studies (seven with FDG PET, four with thallium-201 imaging, and six with dobutamine echocardiography) showed that patients with viable myocardium who are treated medically had a high event rate (20%), as compared to a low event rate in viable patients who underwent revascularisation (7%).4 Thus, the available evidence suggests that patients with viable myocardium should undergo revascularisation, although it is important to realise that prospective, randomised trials on the prognostic value of viability are still lacking.
CMR TECHNIQUES TO ASSESS VIABILITY
Several CMR techniques have been proposed for the assessment of myocardial viability. These techniques include resting CMR (which provides information on end diastolic wall thickness (EDWT)), dobutamine CMR (which provides information on contractile reserve), and contrast enhanced CMR (which provides information on scar tissue). The available studies using these various techniques to predict recovery of function post-revascularisation are summarised below and in tables 1–4.
RESTING CMR TO ASSESS LV END DIASTOLIC WALL THICKNESS
Following acute myocardial infarction, structural changes occur within the infarct zone. In particular, in the presence of extensive, transmural infarction, wall thinning occurs in the centre of the infarct zone. Various studies have demonstrated that wall thinning is frequently associated with transmural scar tissue. Perrone-Filardi et alw3 performed a direct comparison between CMR and PET using metabolic imaging with F18-fluorodeoxyglucose (FDG) in patients with ischaemic LV dysfunction. The authors demonstrated that an EDWT < 8 mm yielded a sensitivity and specificity of 74% and 79% for prediction of absence of metabolic activity. Similarly, Baer et al5 performed a head to head comparison between resting CMR and FDG PET in 35 patients with chronic ischaemic LV dysfunction. It was shown that regions with an EDWT < 5.5 mm had significantly reduced FDG uptake, whereas regions with an EDWT ⩾ 5.5 mm had preserved FDG uptake. This cutoff value for EDWT is in good agreement with data derived from necropsy studiesw4; in these studies it was shown that regions with a chronic infarction had a wall thickness < 6 mm. An example of a patient with a previous transmural infarction and severe wall thinning is shown in fig 2.
In a subsequent study, Baer et al6 tested the value of EDWT for prediction of functional recovery post-revascularisation. The authors showed that segments with an EDWT < 5.5 mm virtually never showed recovery of function post-revascularisation. The alternative was not true: segments with an EDWT ⩾ 5.5 mm did not always improve in function post-revascularisation; this is related to the aforementioned issue of non-transmural infarction. Segments with an EDWT ⩾ 5.5 mm frequently contain subendocardial scar tissue, with residual viability in the epicardial layers. In the absence of jeopardised myocardium though, recovery of function will not occur after revascularisation. Three studies (with a total of 100 patients) have used EDWT to predict functional recovery; pooling of the data (table 1) confirmed an excellent sensitivity (95%, range 94–100%) to predict recovery, with a lower specificity (41%, range 19–53%). Thus, severe wall thinning appears to indicate scar tissue and has a high accuracy to predict no recovery after revascularisation. However, it was recently demonstrated that even in the presence of severe wall thinning, recovery of function may occur, but only when contrast enhanced CMR excludes scar tissue.7
DOBUTAMINE STRESS CMR TO ASSESS CONTRACTILE RESERVE
In addition to assessment of EDWT to identify viable myocardium, the presence of contractile reserve is frequently used to detect viable myocardium. The hallmark of viability is the improvement of contraction in dysfunctional myocardium that is elicited by the infusion of low dosages of dobutamine (5–15 μg/kg/min). Baer et al5,6w5 extensively explored this approach and demonstrated that dobutamine stress CMR can adequately predict improvement of regional LV function after revascularisation. The authors showed that an increased systolic wall thickening > 2 mm during dobutamine infusion was a reliable marker of predictor of functional recovery.6 A total of nine studies with 252 patients using dobutamine stress CMR to predict recovery of function have been published with a mean sensitivity of 73% (range 50–89%) and a mean specificity of 83% (range 70–95%). Thus, dobutamine stress CMR has a high specificity with a slightly lower sensitivity.
CONTRAST ENHANCED CMR TO ASSESS SCAR TISSUE
Contrast hyperenhancement on delayed resting CMR was first described more than 20 years agow5 w7 and is defined as regions of increased image intensity on T1 weighted images, acquired more than five minutes after the intravenous administration of a contrast agent. In a study in 2001, Kim et alw8 investigated an improved pulse sequence and measured image intensities in “delayed hyperenhanced regions” to be 485% higher than in normal regions. The mechanism underlying the hyperenhancement appears related to the interstitial space between collagen fibres, which is larger in scar tissue as compared to the densely packed myocytes in normal myocardium, and the contrast agent will be trapped in these areas in infarcted tissue.
Kim et al8 elegantly validated the value of contrast enhanced CMR to detect scar tissue in animal experiments. In chronically instrumented dogs with previous infarction, the authors showed a perfect agreement between the extent of scar tissue on contrast enhanced CMR and the histological extent of necrosis using TTC staining of the explanted hearts. The major advantage of contrast enhanced CMR over other imaging techniques is that because of the superior spatial resolution, differentiation between transmural and subendocardial infarction is possible. Examples of contrast enhanced CMR are shown in fig 3.
Klein and colleagues9 evaluated 31 patients with depressed LVEF (mean (SD) 28 (9)%) with FDG PET and contrast enhanced CMR. The agreement between both techniques for assessing scar tissue was 91%. Importantly, 11% of segments defined as viable on FDG PET had some extent of scar tissue on contrast enhanced CMR. This reflects the superior resolution of CMR allowing discrimination of small subendocardial infarcts.
To evaluate further the value of contrast enhanced CMR for predicting functional recovery, Kim and co-workers studied 50 patients with chronic infarction and LV dysfunction undergoing revascularisation.10 Contrast enhanced CMR and resting LV function were assessed before revascularisation, and recovery of function was assessed 11 weeks post-revascularisation. The likelihood of segmental recovery of function post-revascularisation paralleled the transmurality of infarction of the segments: improvement of function decreased progressively as the transmurality of scar tissue increased. In particular, 78% of dysfunctional segments without contrast enhancement improved in function, as compared to 2% of segments with scar tissue extending > 75% of the LV wall (fig 4). Using a cutoff value of 25% transmurality of scar tissue, the sensitivity and specificity were 86% and 61% to predict improvement of function. Changing the cutoff value to 75% transmurality, sensitivity and specificity would be 100% and 15%, respectively. With nuclear imaging, 50% tracer uptake is frequently used to assess viability; when a cutoff value of 50% transmurality was applied, the sensitivity and specificity were 97% and 44%, respectively. Pooling of the four available studies in patients undergoing revascularisation (total 132 patients) confirmed these findings and revealed a sensitivity of 95% with a specificity of 45% (table 3).
The suboptimal specificity is related to the presence of segments with subendocardial necrosis (and epicardial viability) that do not improve in function. The low specificity indicates that information on the constitution of the epicardial regions is needed: do they contain normal, viable tissue or jeopardised myocardium?
COMPARISON OF CMR TECHNIQUES TO PREDICT FUNCTIONAL RECOVERY
There are currently not many direct comparisons in large groups of patients undergoing revascularisation. Baer et al6 performed a head to head comparison between EDWT and dobutamine stress CMR in 43 patients undergoing revascularisation. The sensitivities of both techniques were comparable (92% v 89%), whereas the specificity was higher for dobutamine CMR as compared to EDWT (94% v 56%).
Two studies directly compared dobutamine stress CMR with contrast enhanced CMR. Lauerma et alw9 reported a superior sensitivity and specificity for dobutamine stress CMR in 10 patients undergoing revascularisation. Wellnhofer and colleagues11 demonstrated in 29 patients undergoing revascularisation a higher sensitivity for contrast enhanced CMR, with a higher specificity for dobutamine stress CMR to predict functional recovery.
When pooled data are compared, the differences in accuracy of the various CMR techniques to predict functional recovery post-revascularisation become more clear (table 4). The sensitivity of EDWT and contrast enhanced CMR are significantly higher than that of dobutamine stress CMR, as evidenced by the absence in overlap of 95% confidence intervals. Conversely, the specificity of dobutamine stress CMR is significantly higher than that of EDWT and contrast enhanced CMR.
Integrated use of CMR techniques can be considered for optimal prediction of functional recovery. A very sensitive technique, such as contrast enhanced CMR, may serve as a first step. In the presence of minimal scar tissue (transmurality < 25%), recovery of function is likely to occur, whereas segments with extensive scar tissue (transmurality > 50–75%) will not recover (as shown in the study by Kim et al10). Segments with an intermediate extent of scar tissue (transmurality 25–50%) have an intermediate likelihood of recovery, and in these segments additional testing may be needed. Dobutamine stress CMR may serve as a second step in these segments in order to further differentiate between segments with low (without contractile reserve) and high likelihood (with contractile reserve) to improve in function post-revascularisation. Kaandorp et al12 recently demonstrated the feasibility of integrated assessment of contrast enhanced CMR and dobutamine stress CMR. The accuracy of this approach, however, needs further evaluation in patients undergoing revascularisation.
Integration of EDWT with another technique can also be considered, but the problem is as follows. It has been demonstrated that EDWT has a high sensitivity, but low specificity. Thus, as a first step, assessment of EDWT is possible, and segments with EDWT < 5.5 mm should be considered scar tissue. Then, in the remaining segments (with EDWT ⩾ 5.5 mm) a second test is needed. However, with the recent evidence that even in segments with EDWT < 5.5 mm recovery of function is possible,7 all segments need to undergo a second test.
ADDITIONAL INFORMATION PROVIDED BY CMR
With more aggressive surgical approaches, more information is needed preoperatively to determine the optimal surgical procedure. This includes information about the LV function—that is, the LVEF to assess the risk of surgery; patients with lower LVEF are at higher risk for (peri-)operative morbidity/mortality. CMR should currently be considered as the “gold standard” for assessment of LVEF.w10
In addition, LV volumes are important to determine the likelihood of recovery of function post-revascularisation. It was recently demonstrated that patients with severely dilated left ventricles have a low likelihood to improve in LVEF despite the presence of viable tissue.w11 In 61 patients with substantial viability (⩾ 4 viable segments), the likelihood of improvement in LVEF > 5% was minimal when the LV end systolic volume exceeded 153 ml.
Another issue of importance is the presence of ischaemic mitral regurgitation. This phenomenon is frequently observed in ischaemic cardiomyopathy, as a consequence of mitral annular dilatation,w12 and mitral valve repair should be performed in addition to revascularisation. As illustrated in fig 5, ischaemic mitral regurgitation can be visualised adequately with CMR. Moreover, the feasibility of precise quantification of regurgitant volume with CMR was reported recently.13
Finally, with the increasing use of LV aneurysmectomy or surgical LV restoration, information on the presence of LV aneurysms is needed.w13 In fig 6, an example of a large LV aneurysm (with thrombus formation) is demonstrated using contrast enhanced CMR.
With the increasing number of patients with ischaemic heart failure, information on myocardial viability is needed to guide patient treatment. Accurate viability assessment is possible with CMR using different techniques including EDWT assessment, dobutamine stress CMR, and contrast enhanced CMR. While dobutamine stress CMR has the highest specificity to predict functional recovery post-revascularisation, EDWT and contrast enhanced CMR have a higher sensitivity. Integrated use of particularly dobutamine stress CMR and contrast enhanced CMR may be preferred for optimal prediction of functional recovery.
CMR viability assessment in chronic ischaemic LV dysfunction: key points
Dobutamine stress cardiovascular magnetic resonance (CMR) has the highest specificity to predict functional recovery post-revascularisation
End diastolic wall thickness (EDWT) and delayed contrast enhanced CMR have a higher sensitivity to predict functional recovery
Integrated use of dobutamine stress CMR and contrast enhanced CMR may be preferred for optimal prediction of functional recovery
Additional information can be provided by CMR on left ventricular ejection fraction (LVEF), LV volumes, ischaemic mitral regurgitation, and LV shape
Finally, CMR can provide additional information on LVEF, LV volumes, ischaemic mitral regurgitation and LV shape (aneurysms), which can be used to plan the surgical strategy.
Additional references appear on the Heart website—http://www.heartjnl.com/supplemental
In compliance with EBAC/EACCME guidelines, all authors participating in Education in Heart have disclosed potential conflicts of interest that might cause a bias in the article