Chronic contractile dysfunction can not only be related to scar tissue, but can also be related to hibernation or chronic stunning. In the presence of scar tissue, recovery of function will not occur, but in hibernation or chronic stunning, recovery of function may occur after revascularization. Viable myocardium is generally referred to as “alive myocardium,” independent of the contractile status of the myocardium. This however may not be preferred in the clinical setting, since the aim is to eventually predict improvement of function post-revascularization. Accordingly, the starting point in the discussion on viability should be the identification of regional dysfunction (usually assessed by echocardiography, but can also be detected by MRI or MDCT).

The distinction between hibernation and chronic stunning can be based on the assessment of myocardial blood flow; in hibernation resting blood flow is reduced, whereas in chronic stunning resting flow may still be preserved but flow reserve will be reduced.15 Clinically this differentiation may not always be feasible, but may also not be needed, since both entities require revascularization to improve in function and can be referred as viable, ischemically jeopardized myocardium.

In contrast, the mixture of normal (non-jeopardized) myocardium and scar tissue (as may be encountered with subendocardial scar formation) will not improve in function and can be considered as viable, non-jeopardized myocardium (Figure 2).16

A variety of non-invasive imaging techniques have been developed to detect viable myocardium in patients with ischemic heart failure. The different imaging techniques target different characteristics of viable myocardium (Table 2) and thus provide different information, which translate in different accuracies to predict improvement of function after revascularization.17

Most experience has been obtained with nuclear imaging using SPECT and PET; both techniques permit assessment of cardiac metabolism and perfusion. With SPECT, cell membrane or mitochondrial integrity can also be assessed.

Resting echocardiography or MRI can be used to assess LV end-diastolic wall thickness; it has been demonstrated that thinned myocardium (<6 mm) does not often improve in function after revascularization.18 Both echocardiography and MRI can be used in combination with low-dose dobutamine infusion to detect contractile reserve in dysfunctional myocardium. All these techniques aim at the detection of viability. It has been demonstrated that 40-50% of dysfunctional segments without contractile reserve may still have preserved perfusion and/or metabolism; some of these segments may still recover function after revascularization. It has been shown that loss of contractile reserve is associated with more severe ultrastructural damage and fibrosis formation and the severity of damage may eventually determine whether functional recovery is still possible.

In contrast to all the techniques discussed above which detect viable myocardium, contrast-enhanced MRI is the technique of choice, with the highest resolution, to detect scar tissue.

Most protocols and techniques have been described extensively over the last decades and have become widely implemented in the clinical setting.17

A recent, very extensive report summarized the results of the viability studies focusing on prediction of functional recovery, and the results are summarized in Figure 5, and Tables 3 and 4.17

The highest sensitivity was obtained by FDG PET (92%), followed by SPECT with thallium-201 (87%) and technetium-99m-labeled agents (83%). The most specific approach was dobutamine stress echocardiography (specificity 78%). Improvement of global LV function (LVEF) is clinically more meaningful than improvement of segmental function. It is unclear how much viable myocardium is needed to result in the improvement of global LV function. Various studies have shown that recovery of global LV function may occur when at least 25% of the dysfunctional segments are viable.21-27 It should be emphasized however that various other factors also influence recovery of function (these are discussed below). Not only improvement of function is important, but also improvement in exercise capacity and heart failure symptoms are relevant, particularly from a patient’s perspective. This has been addressed in 8 studies,22,23,28-33 and the pooled analysis indicated that patients with significant viable tissue revealed an improvement in New York Heart Association (NYHA) functional class from 2.9 to 1.6 after revascularization, whereas the NYHA class did not change significantly in patients without viable tissue. In addition, 3 studies evaluated exercise capacity (expressed in METS) in relation to viability on FDG PET;21,34,35 in the viable patients, the exercise capacity improved from 4.4 to 5.7 after revascularization, whereas the change was minimal in non-viable patients (from 5.1 before to 5.9 after revascularization).

The most important issue is however the impact of viability on prognosis. A total of 28 retrospective viability studies have been reported including 3848 patients.17 The annualized mortality was lowest in the viable patients who underwent revascularization (3.6%), whereas annualized mortality was comparable in the other groups (ranging from 9.1% to 11.6%) (Figure 6). These retrospective studies suggested a direct relation between viability, improvement of function post-revascularization, translating in improved outcome. This topic was subsequently addressed in 1 prospective, randomized controlled trial: the STICH (Surgical Treatment for Ischemic Heart Failure) trial.36 In the larger part of the trial, 1212 patients with LVEF ≤ 35% and chronic coronary artery disease amenable to surgical revascularization were randomized to medical therapy (n = 602) or medical therapy and coronary artery bypass grafting surgery (CABG) (n = 610). In the medically treated group, death or hospitalization for cardiovascular events occurred in 411 patients (68%), as compared to 351 (58%) in the CABG group (hazard ratio with CABG, 0.74; 95% confidence interval, 0.64 to 0.85; P < 0.001). In the substudy, including 601 patients (298 medical therapy and CABG, 303 medical therapy alone), the relation between viability, therapy (surgery or medical therapy), and outcome was evaluated.37 The mortality was less in the viable patients (37%) as compared to the non-viable patients (51%, hazard ratio for death among patients with viable myocardium, 0.64; 95% confidence interval, 0.48 to 0.86; P = 0.003). Importantly, following correction for other baseline variables, this association with mortality was not significant (P = 0.21): these variables included LVEF, LV volumes, severity of symptoms, all indicators of more severe disease. In addition, there was no significant interaction between viability status and treatment regarding mortality (P = 0.53). Various issues could have affected the results of the STICH trial. First, as was pointed out in the beginning of this chapter, the focus should be on dysfunctional segments and in the STICH analysis all segments were included, and a large percentage of segments may be viable but have normal contractile function (and these cannot improve in function, and may thus affect the current results). This is also reflected in the distribution between viable and non-viable patients: 487 vs 114, while in the daily practice, the majority of patients with previous infarction does not contain viability in the infarcted area. In the current trial, the small cohort of non-viable patients is then randomized to 60 patients receiving medical therapy and 54 undergoing CABG. In addition, the viability techniques used were dobutamine stress echocardiography and ⁹⁹technetium SPECT imaging, whereas FDG PET (the most sensitive viability assessment technique) and contrast-enhanced-MRI (the most accurate scar assessment technique) may have been preferred. Probably the most important issue is the beginning of this article: viability is only one part of the diagnostic and prognostic work-up of the patients with ischemic heart failure, and all other factors indicated in Table 1 will affect outcome. In the 2014 European Society of Cardiology guidelines on myocardial revascularization, it is recommended that myocardial revascularization should be considered in patients with chronic ischemic heart failure (LVEF ≤ 35%) in the presence of viable myocardium (class IIA, level of evidence B).38

Left ventricular volumes and LVEF are important parameters for the risk stratification and management (particularly device therapy) of heart failure patients. The wide availability, accuracy, and safety make 2-dimensional transthoracic echocardiography the imaging technique of first choice (video 1).9 Besides LV dimensions, geometry, and LVEF, echocardiography provides information on diastolic function, valvular heart disease, pericardial disease, right ventricular function, and pulmonary systolic pressure. Internal LV linear dimensions can be measured from M-mode recordings from the parasternal long-axis view which is a reproducible and high temporal resolution method. However, in abnormal LV geometry, these measurements are not accurate, and assessment of LV volumes using the biplane disk summation method from the apical 2- and 4-chamber views is recommended (Figure 7).39 The addition of intravenous echocardiographic contrast is recommended when ≥2 contiguous LV endocardial segments are poorly visualized (video 1). These contrast agents enhance endocardial border definition and provide more reproducible measurements of volumes and LVEF, and compare well with MRI.40 Further improvement can be obtained by 3-dimensional echocardiography (Figure 7).41 Besides echocardiography, other 3-dimensional techniques such as MRI42 or ECG-gated MDCT provide high-quality measurements of LV volumes and LVEF (Figure 7, video 1).43 Finally, with ECG-gated myocardial perfusion SPECT accurate 3-dimensional measurements of LV volumes and EF can be obtained from the time volume curves (Figure 7).

In addition, LV shape is important, and particularly the formation of large LV aneurysms after infarction. Thrombus formation frequently occurs at these aneurysmatic sites, and surgical resection of dysfunctional myocardium and/or scar tissue can reduce LV size and restore normal geometry.38 Information on location and extent of LV aneurysms can be provided by echocardiography, MRI, MDCT, and gated SPECT (Figure 8, video 1).

Secondary mitral regurgitation occurs frequently in patients with ischemic heart failure, and is characterized by a combination of reduced LV closing forces (due to LV dysfunction or dyssynchrony) and global and regional LV remodeling which leads to distortion of the subvalvular apparatus of the mitral valve, displacement of the papillary muscles, tethering of the mitral leaflets, and failure of mitral valve coaptation. Secondary mitral regurgitation results in LV volume overload, which further worsens LV remodeling and mitral valve incompetence. The presence of significant secondary mitral regurgitation provides incremental prognostic information over LVEF.13 In 1256 heart failure patients (60% ischemic etiology), significant secondary mitral regurgitation was associated with increased risk of heart failure hospitalization and all-cause mortality.13 Therefore, accurate quantification of secondary mitral regurgitation severity is crucial for clinical decision making in heart failure patients. Two-dimensional echocardiography remains the mainstay imaging technique to assess the severity and mechanism of secondary mitral regurgitation.44 Severe secondary mitral regurgitation is defined by an effective regurgitant orifice area ≥0.2cm² and a regurgitant volume ≥30 mL/beat quantified using the proximal isovelocity surface area (PISA) method.44 While in central regurgitant jets, the PISA method is relatively accurate to quantify the severity of mitral regurgitation, in eccentric or multiple regurgitant jets or non-hemispheric regurgitant orifices, this method may be less accurate (Figure 9, video 2). In those cases, 3-dimensional imaging techniques may provide more accurate quantification of secondary mitral regurgitation severity (Figure 9, video 2).45 Furthermore, secondary mitral regurgitation is highly dynamic and strongly influenced by pressure and volume conditions. Therefore, in symptomatic heart failure patients with mild to moderate secondary mitral regurgitation, exercise echocardiography may help unmask severe mitral regurgitation. An increase in effective regurgitant orifice area ≥0.13 cm² at peak stress has been associated with poor survival.46

The suitability for surgical restrictive mitral valve annulopasty is influenced by several geometrical aspects of the mitral valve and left ventricle, and also by the presence of extensive myocardial scar, which has been associated with increased mortality rates after surgical repair.47,48

Patients with ischemic heart failure and depressed LVEF (≤35%) have an increased risk of arrhythmic death.49 Various trials have demonstrated the prophylactic benefit of an ICD in these patients.50-53 However, the percentage of patients that require ICD therapy (appropriate shocks) to prevent ventricular tachycardia/fibrillation (VT/VF) at follow-up is relatively low, suggesting that a substantial percentage of patients may not benefit from ICD.54 Currently, the LVEF (<30-35%) is used as the main selection criterium for ICD therapy.55 The precise substrate for VT/VF however is unknown, but may be related to scar tissue (infarct zone) and the heterogenous border zone around the infarct core (mixture of viable myocardium interspersed with fibrous and scar tissue). Non-invasive imaging may eventually help in patient selection for ICD implantation. The presence, extent, and characteristics of myocardial scar assessed with contrast-enhanced MRI have incremental prognostic value over LVEF to predict the occurrence of VT/VF.56 Subsequent studies quantifying the extent of contrast-enhanced tissue (scar mass) revealed that the arrhythmic risk increased significantly when the scar mass exceeded >1.4-5% of the LV volume (Figure 10).57,58 However, this arrhythmic risk reached a plateau at larger levels of scar mass suggesting that other factors contribute to the arrhythmogenic substrate, and the extent of the border zone has also been associated with increased risk of VT. Roes et al59 demonstrated that an extent of the border zone ≥16.7 g identified a group of patients with high incidence of appropriate ICD therapies. Each 10 g increase of border zone increased the risk of appropriate ICD therapies by 1.49 (95% confidence interval 1.01-2.2, P = 0.04) independently of LVEF and extent of infarct core.59

Possibly, another characteristic of the border zone is the presence of viable but (partially) denervated myocardium; the functional integrity of the sympathetic nerve terminals in this area can be evaluated with using radiolabeled norepinephrine analogs.60 This concept was recently addressed in the Prediction of ARrhythmic Events with Positron Emission Tomography trial.61 In 204 heart failure patients receiving an ICD, cardiac sympathetic innervation was quantified with ¹¹C-meta-hydroxyephedrine (¹¹C-HED) PET, perfusion was assessed with ¹³nitrogen-ammonia and viability was quantified with FDG. Patients with vs without ICD shocks had significantly larger areas of denervated myocardium (33 ± 10% vs 26 ± 11%, P < 0.001) and larger areas of denervated, viable myocardium (10 ± 6% vs 7 ± 5%, P = 0.02) despite having comparable LVEF. In addition, the extent of denervated myocardium was independently associated with ICD shocks (hazard ratio 1.069 per 1% increment, 95% confidence interval 1.023-1.117; P = 0.003).61 Similarly, ¹²³-iodine (¹²³I)-metaiodobenzylguanidine (MIBG) SPECT has been used to evaluate cardiac innervation. From the planar images, the heart-to-mediastinum (H/M) ratio and the washout rate quantify the global myocardial uptake of ¹²³I-MIBG (Figure 11). The AdreView Myocardial Imaging for Risk Evaluation in Heart Failure trial, enrolling 961 heart failure patients (68% ischemic heart disease), showed that an H/M ratio <1.6 identified a cohort of patients with increased incidence of ventricular arrhythmias.62 The H/M ratio was an independent predictor of cardiac events (hazard ratio 0.36, 95% confidence interval 0.17-0.75; P = 0.006). In addition, from SPECT images, regional abnormalities in ¹²³I-MIBG uptake have been associated with ventricular arrhythmias.63 Among 116 ICD recipients (74% ischemic heart failure), a late ¹²³I-MIBG SPECT summed defect score >26 identified a subgroup of patients with high cumulative rates of ventricular arrhythmias at 3 years follow-up (52% vs 5%, P < 0.01).63 These studies suggest that imaging may potentially help in patient selection for ICD therapy.

Cardiac resynchronization therapy can also improve clinical outcomes of heart failure patients in NYHA functional class II-IV despite optimal medical therapy, with LVEF <35% and QRS duration >120 ms.64 The individual response varies significantly and imaging may help improve patient selection. Nuclear imaging techniques provide information on 2 important pathophysiological determinants of response to CRT: LV mechanical dyssynchrony65-67 and the extent and location of LV scar tissue.66,68,69

Phase analysis measures the onset of mechanical contraction by approximating the variation of regional maximal counts over the cardiac cycle (which represents myocardial wall thickening) in SPECT myocardial perfusion imaging with Fourier harmonic functions (Figure 12).70 From the onset mechanical contraction phase distribution, the phase standard deviation and the phase histogram bandwidth are derived as quantitative indices of LV mechanical dyssynchrony. The relevance of these LV mechanical dyssynchrony indices to predict response to CRT was demonstrated in 40 patients with heart failure (70% ischemic cardiomyopathy).65 After 6 months of follow-up, 60% of patients revealed a symptomatic improvement (responders). Compared with non-responders, these patients had significantly larger histogram bandwidth (94 ± 23° vs 68 ± 21°, P < 0.01) and phase standard deviation (26 ± 6° vs 18 ± 5°, P < 0.01). Furthermore, accumulating evidence has shown that an LV lead placed at the latest activated region of the left ventricle is associated with good response to CRT.71 The phase polar maps of the left ventricle can be subdivided into 17 segments and the mean phase of each segment, which represents the timing of onset of mechanical contraction, can be displayed. The segment with the largest mean phase is identified as the latest activated segment. Furthermore, the presence of transmural scar at the segments targeted by the LV lead and the presence of significant scar burden have been associated with lack of response to CRT.68,69

Eventually, the information on LV dyssynchrony, the site of latest activation, and scar tissue can be fused with fluoroscopic venograms into a 3-dimensional LV epicardial surface extracted from ECG-gated SPECT myocardial perfusion imaging (Figure 13).72 The clinical role of this tool needs to be demonstrated in prospective studies, but it highlights the potential use of nuclear imaging to select patients for CRT.