https://orcid.org/0000-0001-9584-5505
Abstract
Heart failure is a clinical syndrome with an increasing prevalence and incidence worldwide that impacts patients’ quality of life, morbidity, and mortality. Implantable cardioverter-defibrillator and cardiac resynchronization therapy are pillars of managing patients with HF and reduced left ventricular ejection fraction. Despite the advances in cardiac imaging, the assessment of patients needing cardiac implantable electronic devices relies essentially on the measure of left ventricular ejection fraction. However, multi-modality imaging can provide important information concerning the aetiology of heart failure, the extent and localization of myocardial scar, and the pathophysiological mechanisms of left ventricular conduction delay. This paper aims to highlight the main novelties and progress in the field of multi-modality imaging to identify patients who will benefit from cardiac resynchronization therapy and/or implantable cardioverter-defibrillator. We also want to underscore the boundaries that prevent the application of imaging-derived parameters to patients who will benefit from cardiac implantable electronic devices and orient the choice of the device. Finally, we aim at providing some reflections for future research in this field.
CIED, cardiac implantable electronic device; CMR, cardiac magnetic resonance; CT, computed tomography; HFrEF, heart failure with reduced ejection fraction; ICD, implantable cardioverter-defibrillator; LV, left ventricular; MMI, multi-modality imaging; RV, right ventricle.
Introduction
Heart failure (HF) is a clinical syndrome with an increasing prevalence and incidence worldwide. Once developed, HF significantly impacts patients’ quality of life, morbidity, and mortality, with a 1-year mortality rate of 7% and a 1-year hospitalization rate of 31.9%.1
Cardiac implantable electronic devices (CIEDs) such as implantable cardioverter-defibrillator (ICD) and cardiac resynchronization therapy (CRT) are pillars of the management of patients with HF and severely reduced left ventricular ejection fraction (LVEF).2–4
However, several studies have shown that LVEF is a poor indicator of the overall left ventricular (LV) performance and several grey zones exist concerning the selection of the best candidates for ICD or CRT implantation.5,6 While the benefit of ICD for the primary prevention of ventricular arrhythmias in patients with ischaemic heart disease and severely reduced LVEF is fully established, the Danish trial has underscored that this benefit is modest in patients with non-ischaemic dilated cardiomyopathy.7 These results pave the way for improving the selection criteria for ICD implantation beyond the simple assessment of LVEF. In the field of CRT, substantial underuse of CRT devices in the HFrEF population has been underscored.8 This is substantially attributable to the disproportionate attention given to the concept of ‘non-response’ to CRT and supports the role of CRT as a disease modifier, irrespective of the presence of LV remodelling. Additionally, some non-randomized trials have shown that cardiac imaging might be useful to identify patients who could benefit from the implantation of CRT, despite a Class IIB or no indication for the device before advanced cardiac imaging.3,9
The present paper aims to (i) provide an overview of the main advances in cardiac imaging for the characterization of patients undergoing CIED; (ii) underscore the boundaries that prevent the application of imaging-derived parameters to patients who will benefit from CIED and orient the choice of the device; and (iii) provide some indications for future research in this field to show how a rational multi-modality imaging (MMI) approach can help in (i) characterizing patients before CIED implantation; (ii) stratifying the arrhythmic risk to better identify ICD candidates; (iii) refining the indications for CRT implantation; and (iv) paving the way for further studies on the role of cardiac imaging to support and guide CIED implantation.
Role of imaging techniques before implantation
Echocardiography
Echocardiography is the first step imaging modality for the assessment of patients with HF. An LVEF threshold of 35% currently remains the only imaging criterion to guide the selection of patients for CIED implantation, both ICD and CRT.3
Although LVEF carries a significant prognostic value in HF with reduced ejection fraction (HFrEF), malignant arrhythmias might occur also in patients with LVEF >35%, whereas some patients with severely reduced LVEF receiving an ICD will never experience a threatening arrhythmic event.10,11 Because of the poor accuracy of LVEF for the selection of ICD candidates, the challenge is to find a risk stratification tool, including clinical, imaging, and electrophysiological parameters, that adequately selects patients at high risk of ventricular arrhythmias (VAs) and sudden cardiac death (SCD). Speckle tracking echocardiography allows the detection of subtle changes in LV function12 and has been shown to improve the prognostic stratification of patients with HF independently of the LVEF.13 Global longitudinal strain (GLS) measurements show excellent reproducibility compared with other echo-derived variables, The issue of inter-vendor variability is higher for segmental strain measure and modest for GLS.14,15 This should not prevent the application of GLS in clinical practice and the follow-up of patients but should be considered when performing clinical trials.
In patients with ischaemic cardiomyopathy, both regional16,17 and GLS17 values have shown to be independent predictors of VA, and similar results have been observed in patients with dilated cardiomyopathy18 These data have been recently confirmed in a consecutive ‘real-world’ population of patients with structural heart disease undergoing ICD implantation and underscore the continuous relationship between progressive reduction in GLS values and arrhythmic events.19
LV mechanical dispersion (LVMD) assessed using speckle tracking strain corresponds to the standard deviation of strain peaks and is a marker of ventricular electromechanical incoordination,18 which is associated with myocardial fibrosis. In patients with ischaemic cardiomyopathy and moderately abnormal LVEF, a cut-off of 61 ms for LVMD has 85% sensitivity and 73% specificity for the prediction of VA,20 whereas in patients with non-ischaemic dilated cardiomyopathy the performance of LVMD seems weaker because a cut-off of 72 ms has shown to predict VA with sensitivity and specificity of 67% and 89%, respectively.20 In a retrospective study on 203 patients with structural heart disease receiving ICD according to recommendations, Guerra et al. have shown that LVMD is not a predictor of VA.19 However, a large meta-analysis of 12 studies including 3198 patients with different cardiac diseases and a weighted mean LVEF of 46% suggested that 60 msec might be a reasonable cut-off for the risk stratification of VA in patients with cardiac disease21 (Figure 1A).
Applications of speckle tracking echocardiography in patients receiving cardiac implantable electronic devices. (A) Assessment of global longitudinal and mechanical dispersion in two patients with severe non-ischaemic dilated cardiomyopathy and very different values of global longitudinal strain and ventricular mechanical dispersion. (B) Assessment of strain patterns in two patients with ‘typical’ and ‘atypical’ (right panel) left bundle branch block. The systolic stretch index is also defined. (C) Assessment of myocardial work in patients with left bundle branch block without scar (left panel); anteroseptal scar (central panel); and lateral scar (right lateral). The bullseyes and the pressure–strain loops show a significant difference in myocardial work between the septal and the lateral wall according to heart failure aetiology. GLS, global longitudinal strain; LBBB, left bundle branch block; LVMD, left ventricular mechanical dispersion.
Despite the large application of CRT in patients with HF with reduced ejection fraction and widened QRS, there is no univocal definition of CRT response.22 From the point of view of imaging, CRT response has often been referred to as the presence of a significant LV reverse remodelling, more often defined as a > 15% reduction in LV end-systolic volume.23 This binary interpretation of the effects of CRT led to a substantial underestimation of its benefit, which can be evident -despite to a lesser extent- also in patients without significant reverse LV remodelling.24
The identification of imaging-derived parameters associated with CRT response can provide some interesting insight into the pathophysiology of LV conduction delay.
The assessment of opposite wall delay, which has been initially proposed as an echo-derived method to evaluate LV mechanical dyssynchrony, has been abandoned because of the deceiving results of the PROSPECT trial.23 This is because the assessment of opposite wall delay cannot identify the specific mechanical substrates, amenable to be electrically corrected, and associated with LV reverse remodelling after CRT.25 The visual assessment of specific contraction patterns such as septal flash and apical rocking is associated with significant LV reverse remodelling (sensitivity: 79% and 84%, specificity: 74% and 79%, respectively) and better outcome (HR 0.45, 95% CI: 0.34–0.61; HR 0.40, 95% CI: 0.30–0.53, respectively) in CRT candidates. The value of septal flash and apical rocking for the prediction of CRT response is independent of LVEF and QRS width, which are traditionally used to guide CRT implantation.26 Interestingly, the value of the visual assessment of LV dyssynchrony for the prediction of LV reverse remodelling and prognosis after CRT delivery is evident also in patients with a QRS between 120 and 150 msec,24 which only have a class IIa indication for CRT3,4 and has been shown to improve the prognostic stratification of guideline-based patient selection.9
Moreover, these motion patterns have proven to predict CRT response in patients with chronic right ventricular pacing needing a pacemaker upgrade to CRT.27 The correct identification of septal flash and apical rocking in CRT candidates is strictly limited by the experience of the echo-reader. However, on top of the visualization of 2D motion, the qualitative analysis of strain curves looking for specific patterns of septal flash28 or the assessment of the systolic stretch index,29 have also proven to be predictors of CRT response and prognosis (Figure 1B). The localization and extension of the myocardial scar has a pivotal role in the evaluation of CRT candidates. In the case of lateral scar the assessment of residual LV contractility by the analysis of strain curves, and the identification of the latest activated region of the myocardium can be used to optimize the site of CRT delivery.30,31 This approach has been associated with a greater proportion of volumetric response to CRT at 6 months (70% vs. 55%, P < 0.031),31 and has shown to have an incremental prognostic value as compared to the simple assessment of clinical parameters and LV dyssynchrony (Harrel’s C = 0.75).30 Nevertheless, this method neglects the fundamental role of the LV septum in the pathophysiology of LBBB and therefore, should be considered complementary and not an alternative to a comprehensive analysis of global and regional LV function. Aalen et al. have shown that the presence of a large septal-to-lateral myocardial work difference and the presence of viability at CMR can predict LV remodelling after CRT, with a reduction of at least 15% in LV end-systolic volume, with an AUC of 0.88 (0.81–0.95) and are associated with better survival (HR 0.21, 95% CI: 0.072–0.61)32 after CRT implantation (Figure 1C). Finally, a comprehensive approach including the evaluation of both the left atrium and the right ventricular function seems reasonable for the risk stratification of CRT candidates, allowing the identification of subjects with more advanced stages of HF, less contractile reserve, and diastolic dysfunction.33,34
Cardiovascular magnetic resonance
Cardiovascular magnetic resonance (CMR) allows the assessment of multiple morpho-functional features in a one-stop shop technique and is therefore increasingly used in multi-parametric arrhythmic risk stratification.
LVEF has long been considered the strongest predictor of adverse outcomes and is still the key indicator for ICD implantation.3 Although primary prevention ICD implantation relies on the evidence of reduced LVEF, upcoming studies are increasingly focusing on the mildly reduced and preserved LVEF cohorts, as discussed later. Despite current cut-offs for LVEF that have been derived from studies using transthoracic echocardiography (TTE), it is now well established that CMR is the gold standard for the assessment of biventricular function, as it does not rely on any geometric assumption and is highly reproducible,2 suggesting that CMR can be used in difficult cases to better quantify LVEF.35
LVEF, however, is not an optimal predictor of SCD because up to 80% of patients experiencing VA have an LVEF > 35%.36
In these patients, CMR-based myocardial tissue characterization is an accurate diagnostic tool, able to disclose the presence of myocardial scar, or diffuse fibrosis, and has been shown to carry additional prognostic data over LVEF in different cardiomyopathies. T2-weighted and T2 mapping sequences, detecting myocardial oedema, identify patients with acute and reversible processes for whom implantation of a permanent device may be postponed or even not indicated. Moreover, tissue characterization-guided device implantation is particularly useful in diagnosing specific disease aetiology, such as sarcoidosis, amyloidosis and myocarditis that show typical features on CMR.3,37
Despite fewer existing studies on the prognostic role of new parametric mapping techniques, a recent meta-analysis of 1524 patients with ischaemic and non-ischaemic cardiomyopathy has shown a higher incidence of cardiovascular death and combined cardiac events in patients with higher extracellular volume values.38 LGE analysis finally allows for the identification and quantification of myocardial tissue heterogeneity, which has been associated with VT inducibility and increased ICD discharge rate.39
A meta-analysis of 2850 patients with LVEF ≤30% of ischaemic and non-ischaemic aetiology has confirmed the prognostic value of late gadolinium enhancement (LGE) on CMR to predict arrhythmic events (SCD, ventricular tachycardia (VT)/ventricular fibrillation and ICD discharge), with a pooled OR of 5.62, which increased up to 9.56 in patients.40 A model based on CMR-LVEF ≤35% or CMR-LVEF ≤35% and LGE detection, as compared to TTE-LVEF, determined net reclassification of 47% and 41% of patients evaluated for primary prevention ICD implantation, respectively.35 A study on 252 patients with dilated cardiomyopathy has reported a survival benefit of CRT-D over CRT-P, in terms of lower mortality (HR 0.23), total mortality or HF hospitalisation (HR 0.32), and total mortality or hospitalisation for major adverse cardiovascular events (HR 0.30) only in patients with positive LGE41 (Figure 2). In these patients, the specific location of LGE also provides prognostic information, as left ventricular septal LGE has been associated with increased arrhythmic risk, irrespective of LGE extent; on the other hand, postero-lateral LGE has been shown not only to reduce CRT response but also to increase arrhythmic risk if the pacing is provided over the scarred area.42 Moreover, a multicentre, randomized, blinded, prospective trial, the CRT-reality study (NCT04139460), will aim at assessing the benefit of CRT-D or CRT-P in HF patients with non-ischaemic dilated cardiomyopathy and no evidence of LGE on CMR.43
Applications of cardiovascular magnetic resonance in a patient with non-ischaemic dilated cardiomyopathy before cardiac implantable electronic device. Four-chamber (A) and three-chamber (B) cine steady state-free precession images showing dilated left ventricle. Four-chamber (C) and three-chamber (D) post-contrast images show extensive subepicardial late gadolinium enhancement (LGE) of the anterolateral wall (C, arrowheads) with the almost transmural extent in the inferolateral wall (D, arrowheads). Short-axis mid-cavity post-contrast image showing subepicardial LGE of the inferior and inferolateral walls (E, thin arrows) and mid-wall LGE stria of the inferoseptum (E, thick arrow).
The role of CMR in the assessment of LV dyssynchrony in CRT candidates is limited by the low temporal resolution of this imaging modality. However, as underscored in the previous paragraph, the localisation of myocardial scar in the septum combined with the evaluation of the septal-to-lateral wall difference in myocardial work at transthoracic echocardiography has shown to be a significant predictor of CRT response and prognosis.32,44
Cardiac computed tomography
Cardiac computed tomography (CT) is an attractive tool in the pre-procedural evaluation for CIED implantation, given its superior spatial resolution (0.5–0.75 mm slices) and rapid acquisition amongst non-invasive imaging modalities.45,46
One of the main applications of cardiac CT before CRT implantation is the analysis of the coronary sinus and its tributaries to look for venous branches that course to the posterior and lateral left ventricular wall for optimal lead positioning47,48 (Figure 3). Attention is paid to assessing the location, size, angle, Thebesian valve and branching pattern of the coronary sinus system from the right atrium.49 Additionally, CT comprehensively evaluates transvenous access to the heart from the periphery, identifying congenital and/or acquired abnormalities such as left-sided superior vena cava and subclavian venous stenoses.
Example of the application of cardiac computed tomography for the assessment of the pulmonary vein anatomy in the lateral (left panel) and posterior (right panel) left ventricular wall. CX, circumflex coronary artery; CS, coronary sinus; LatV, lateral veins; PIV, posterior interventricular vein; RCA, right coronary artery.
Assessment of myocardial sympathetic innervation with 123I-metaiodobenzylguanidine (MIBG) in a patient with previous inferior myocardial infarction and episodes of ventricular tachycardia. The MIBG planar images (A) showed reduced cardiac uptake of the tracer both in early and delayed images (right panel), with a Normal wash-out (left panel). The single-photon emission computed tomography (SPECT) images (B) showed large areas of innervation/perfusion mismatch in the lateral wall and the apex in the bullseye. The electroanatomical map (C) shows that the location of the radiofrequency application to block the channel originating ventricular arrhythmia (circles) is located in the area of innervation mismatch.
Cardiac CT might have several other roles in these patients. Retrospective ECG-gated (‘4-dimensional CT’) image acquisition can capture the entire heart cycle to assess cardiac chamber size, wall thickness, function, and regional wall motion abnormalities with a high correlation to CMR.50 CT measures of dyssynchrony have been developed to further assess indications or responses to CRT.51 A recent randomized trial has shown that these measures were associated with adverse cardiovascular events, while lead placement in regions of maximal wall thickness was associated with improved outcome.51 Cardiac CT is also an alternative to CMR to evaluate myocardial perfusion and scar, especially when CMR is contraindicated.52,53 Hypoperfusion and ischaemic scar appear as subendocardial or transmural hypoattenuation relative to normal myocardium, in early and late imaging after contrast administration respectively. As described before with LGE CMR, the scar burden and location have important implications for lead placement and the indications for ICD and CRT, respectively. Furthermore, assessments of coronary artery disease, left atrial appendage thrombus, pulmonary venous anatomy and extracardiac pathologies are possible. Finally, the fusion imaging technique enables overlapping CT with fluoroscopy, echocardiography and ablation mapping images for intraprocedural guidance and improved navigation.54
Despite the appraisal of the application of CT and fusion imaging to ease the implantation of CIED, we need to underscore the importance of reducing the number of unnecessary and inadequate exposure of the patients by privileging -when possible- the imaging modality with the lower radiation dose.55
Nuclear imaging
Nuclear cardiac imaging techniques can as well contribute to the stratification of patients’ arrhythmic and death risk56–58 and could be of help for risk stratification before ICD implantation. The zone between areas of fibrosis and viable myocardium is suspected of being the site where re-entry wavefronts originate in patients with ischaemic heart disease. Accordingly, and beyond perfusion and viability considerations, the evaluation of myocardial autonomic nervous system innervation with scintigraphic nuclear-based techniques may unmask the role of myocardial denervation in the triggering of arrhythmias.
Cardiac imaging with 123I-metaiodobenzylguanidine (MIBG) allows the assessment of cardiac sympathetic innervation activity, which is characteristically altered in patients at higher risk of malignant arrhythmic events and SCD and has been reported to provide a non-invasive method to assess cardiac sympathetic function and risk-stratify patients with HFrEF,59,60 independently from LVEF.61,62
In randomized trials, patients with equivalent LVEF but more evidence of extensive myocardial denervation had a higher risk of cardiac arrhythmias and SCD (Figure 4). MIBG based on these trials received FDA approval for clinical use in the US, but cost barriers prevented it from wide adoption by the cardiology community. The incorporation of MIBG imaging in the risk assessment of patients with HFrEF is however yet to be adopted by the European and American guidelines.
Myocardial SPECT in a patient with non-ischaemic cardiomyopathy and LV dyssynchrony. The comparison of rest and stress images (A) shows the absence of myocardial ischaemia. The analysis of contractility histograms and curves (B) and the corresponding polar map (C) allows the identification of septal-to-lateral wall dyssynchrony both at rest and after stress.
Nuclear cardiology can also provide useful information for the characterization of patients before CRT delivery. As previously described, the key factor for increasing the response rate to CRT is the identification of the optimal LV lead position by detecting the optimal pacing site defined as the myocardial area with maximal dyssynchrony but with residual viability and contractility,63–67 which can both be assessed by SPECT. In patients with non-ischaemic cardiomyopathy and LBBB, SPECT can be plagued by the occurrence of septal and apical perfusion artefacts which can be mistaken as reduced viability.68 However, the quantification of absolute myocardial blood flow using positron emission tomography can avoid misinterpretation of septal perfusion defects in patients with LBBB.69 In patients with ischaemic cardiomyopathy, the quantification of scar burden by SPECT has shown to be associated with positive CRT response,70 and outcomes.70 Moreover, compared to other imaging techniques, nuclear imaging allows the evaluation of myocardial viability, mechanical dyssynchrony and LV global function in one single scan (Figure 5). This ‘one-stop shop’ technique has the advantage to be operator-independent and reproducible in a wide variety of settings.63,64,71–74 Nevertheless, these benefits are counterbalanced by a low spatial and temporal resolution, which can impact the refined assessment of LV dyssynchrony and myocardial function, thus favouring other imaging modalities for these specific purposes.
Relationship between cardiac tissue heterogeneity disclosed at CMR after late gadolinium enhancement and the electroanatomical map obtained by intracardiac interrogation. Figures (A–F) show late contrast-enhanced images of the left ventricle from 4-chamber (A, C, E) and short-axis views (B, D, F) depicting an extensive lateral scar (arrows). By segmenting the left ventricle and the different intensities of contrast enhancement (A and B) in the left ventricular myocardium, colour-coded transformation can be obtained (E and F) and a 3D representation view can be depicted (G), mimicking the same view orientation as the electroanatomical map obtained by intracardiac interrogation (H). The line in figure G represents the border zone areas inside scar tissue that correspond to the channel originating ventricular arrhythmia depicted in the electroanatomic mapping (H) and highlighted with a thick line (I).
MMI approach
Whilst the results from the growing number of image guidance studies on CIED in HF is encouraging, one of the major limitations is that the data output from the imaging modality in each study is reviewed separately from, and alongside X-ray, rather than integrated. Given the radiolucency of the cardiac silhouette and high variability in the rotation of the left and right-sided chambers relative to one another, it is not surprising that using fluoroscopy to determine regional anatomy can be highly inaccurate, particularly with regard to CRT and lead position.75 In 2017, the Guide CRT study demonstrated the feasibility of real-time analysis and fusion of CMR-derived scar and dyssynchrony data to guide LV lead implantation.76 Upon coronary venography, the 3D-derived model of the patient’s left ventricle is instantaneously fused enabling the implanting physician to identify, using the individual’s coronary venous anatomy (detected by cardiac CT), scar distribution (evaluated by CMR of nuclear imaging), electrophysiology and mechanical contraction patterns, the patient-specific target locations for LV lead placement, improving the effectiveness of CRT. Despite the interesting perspectives provided by MMI in CRT, two large randomized trials applying a combination of echocardiography, CMR and CT did not show any survival benefit of the MMI-derived approach vs. the standard approach for CRT implantation.77,78 Nevertheless, these studies showed that patients receiving the LV lead in the optimal site had better LV remodelling and survival after CRT, supporting the usefulness of an MMI approach in selected cases.
In addition, both these studies focused on the localisation of the myocardial scar in the lateral wall and neglected the value of septal scar and of the septal-to-lateral wall interplay that has shown to be a key element to understand the pathophysiology of LBBB and predict CRT response and prognosis.26,27,32 The recent development of a wireless intracardiac LV endocardial electrode for CRT delivery (WiSE-CRT, EBR systems) represents a unique opportunity to use integrated MMI for optimal LV site selection.79 The use of MMI-based guidance strategies may be particularly suitable for this approach with the ability to truly target any region on the LV wall without the constraint of the coronary venous anatomy. Finally, the combination of data derived from different MMI techniques provides a comprehensive evaluation of the potential response to CRT. Specifically, echocardiography can assess mechanical dyssynchrony (due to its higher temporal resolution) and the quantification of septal-to-lateral differences in myocardial work, whereas CMR can quantify the extent and assess the localization of myocardial scar. If CMR is not available or cannot be performed, SPECT might be useful for the localization of the scar. Data derived from these imaging modalities are complementary to identify a potential mechanical abnormality amenable to be electrically corrected with CRT as well as to predict the extent of the response according to the localization and extent of scar, which is inversely related to contractile reserve and potentially recruitable myocardium.
The absence of real integration between the different techniques can be also observed in the planning of ICD treatment. From the histological and anatomical background, we know very well that the structural LV changes that underlie HF with reduced LVEF include expansion of the extracellular matrix with an increased proportion of collagen and fibroblasts. Scar or fibrotic tissues (bundles of collagen) are not per se pro-arrhythmic since they are electrically inert tissue. However, the areas where the fibrous/scar tissue intermingles with viable myocytes create a substrate where re-entry circuits may form and lead to VAs.80 As previously described, strain echocardiography, LGE CMR and nuclear imaging techniques have individually provided parameters with incremental value over LVEF to identify the patients who may benefit from an ICD. These imaging modalities permit visualization of both triggers (e.g. myocardial ischaemia) and substrates (e.g. scar) for SCD in ischaemic and non-ischaemic cardiomyopathy. The use of hybrid imaging and the combination of data derived from these MMI techniques might enhance the possibility to detect risk areas and select the best treatment choice for each patient (Figure 6).
Table 1 summarizes the main advantages and disadvantages of different imaging modalities for CIED implantation. Table 2 list the main studies underscoring the role of imaging for the prediction of VAs.
Advantages and disadvantages of imaging modalities for CIED implantation
| Pro | Contra | |
|---|---|---|
| Echocardiography | Costs Availability High temporal and spatial resolution | Feasibility, low image quality, and poor acoustic window in some patients (obesity, pulmonary disease, and so on) |
| CMR | No radiation exposure High spatial resolution Low contrast-related risk Accuracy (gold standard for the assessment of LV size and function) Tissue characterization Assessment of scar localization and extent | Costs Availability Long acquisition time Low temporal resolution Contraindicated or difficult to realize in case of previous devices Beam-hardening artefacts associated with previous devices |
| CT | High spatial resolution Rapid acquisition Assessment of scar localization and extent | Costs Availability Radiation exposure Contrast-related risk Beam-hardening artefacts associated with previous devices |
| SPECT/PET | Evaluation of myocardial innervation Evaluation of myocardial viability, ischaemia, and scar localization | Costs Availability Radiation exposure Low temporal and spatial resolution |
| Pro | Contra | |
|---|---|---|
| Echocardiography | Costs Availability High temporal and spatial resolution | Feasibility, low image quality, and poor acoustic window in some patients (obesity, pulmonary disease, and so on) |
| CMR | No radiation exposure High spatial resolution Low contrast-related risk Accuracy (gold standard for the assessment of LV size and function) Tissue characterization Assessment of scar localization and extent | Costs Availability Long acquisition time Low temporal resolution Contraindicated or difficult to realize in case of previous devices Beam-hardening artefacts associated with previous devices |
| CT | High spatial resolution Rapid acquisition Assessment of scar localization and extent | Costs Availability Radiation exposure Contrast-related risk Beam-hardening artefacts associated with previous devices |
| SPECT/PET | Evaluation of myocardial innervation Evaluation of myocardial viability, ischaemia, and scar localization | Costs Availability Radiation exposure Low temporal and spatial resolution |
CMR, cardiac magnetic resonance; CT, computed tomography; PET, positron emission tomography; SPECT, single-photon emission computed tomography.
Advantages and disadvantages of imaging modalities for CIED implantation
| Pro | Contra | |
|---|---|---|
| Echocardiography | Costs Availability High temporal and spatial resolution | Feasibility, low image quality, and poor acoustic window in some patients (obesity, pulmonary disease, and so on) |
| CMR | No radiation exposure High spatial resolution Low contrast-related risk Accuracy (gold standard for the assessment of LV size and function) Tissue characterization Assessment of scar localization and extent | Costs Availability Long acquisition time Low temporal resolution Contraindicated or difficult to realize in case of previous devices Beam-hardening artefacts associated with previous devices |
| CT | High spatial resolution Rapid acquisition Assessment of scar localization and extent | Costs Availability Radiation exposure Contrast-related risk Beam-hardening artefacts associated with previous devices |
| SPECT/PET | Evaluation of myocardial innervation Evaluation of myocardial viability, ischaemia, and scar localization | Costs Availability Radiation exposure Low temporal and spatial resolution |
| Pro | Contra | |
|---|---|---|
| Echocardiography | Costs Availability High temporal and spatial resolution | Feasibility, low image quality, and poor acoustic window in some patients (obesity, pulmonary disease, and so on) |
| CMR | No radiation exposure High spatial resolution Low contrast-related risk Accuracy (gold standard for the assessment of LV size and function) Tissue characterization Assessment of scar localization and extent | Costs Availability Long acquisition time Low temporal resolution Contraindicated or difficult to realize in case of previous devices Beam-hardening artefacts associated with previous devices |
| CT | High spatial resolution Rapid acquisition Assessment of scar localization and extent | Costs Availability Radiation exposure Contrast-related risk Beam-hardening artefacts associated with previous devices |
| SPECT/PET | Evaluation of myocardial innervation Evaluation of myocardial viability, ischaemia, and scar localization | Costs Availability Radiation exposure Low temporal and spatial resolution |
CMR, cardiac magnetic resonance; CT, computed tomography; PET, positron emission tomography; SPECT, single-photon emission computed tomography.
Characteristics of the main studies underscoring the role of imaging in the prediction of ventricular arrhythmias in patients with ischaemic and non-ischaemic dilated cardiomyopathy
| Article, reference | n | HF aetiology | Outcomes | Imaging modality | Parameter | Predictive value |
|---|---|---|---|---|---|---|
| Ng et al.17 | 424 | IDCM | Mortality, ICD therapy | Echo | Peri-infarct strain | Prediction of ICD therapy: HR = 1.22, 95% CI 1.09 to 1.36 |
| Haugaa et al.18 | 96 | NIDCM | VA | Echo | MD (cut-off = 72 msec) | Se 67%, Sp 85%, AUC 0.80 |
| Guerra et al.19 | 203 | IDCM and NIDCM | ICD therapy | Echo | GLS | Prediction of ICD therapy: HR = 1.94; 95% CI: 1.30–2.91 |
| Nguyen et al.20 | 476 | IDCM | VA | Echo | MD (cut-off = 62 msec) | Se 85%, Sp 73%, AUC 0.84 |
| Pontone et al.35 | 409 | IDCM and NIDCM | MACES (including VA) | CMR | LVEF ≤ 35% and LGE+ | Se 63%, Sp 71%, AUC 0.67, NPV 84.7, PPV 43.6 |
| Disertori et al.40 | 2850a | IDCM and NIDCM | VA | CMR | LGE+ | Se 92.4%, Sp 36.7%, PLR 1.46, NLR 0.21 |
| Leyva et al.41 | 252 | NIDCM with CRT-D or CRT-P | Mortality, HF, MACES | CMR | LGE+ | CRT-D associated with reduced event rates Events only in LGE+ |
| Fallavollita et al.58 | 204 | IDCM | VA | 11C-HED PET | Sympatetic denervation | Prediction of VA: HR = 1.069, 95% CI: 1.023–1.117 |
| Jacobson et al.60 | 961 | IDCM and NIDCM | HF, VA, cardiac mortality | MIBG | H/M ratio > 1.60 | HF: HR = 0.49, 95% CI: 0.32–0.77; VA: HR = 0.37, 95%CI: 0.16–0.85; Cardiac mortality: HR = 0.14, 95% CI = 0.03–0.58 |
| Article, reference | n | HF aetiology | Outcomes | Imaging modality | Parameter | Predictive value |
|---|---|---|---|---|---|---|
| Ng et al.17 | 424 | IDCM | Mortality, ICD therapy | Echo | Peri-infarct strain | Prediction of ICD therapy: HR = 1.22, 95% CI 1.09 to 1.36 |
| Haugaa et al.18 | 96 | NIDCM | VA | Echo | MD (cut-off = 72 msec) | Se 67%, Sp 85%, AUC 0.80 |
| Guerra et al.19 | 203 | IDCM and NIDCM | ICD therapy | Echo | GLS | Prediction of ICD therapy: HR = 1.94; 95% CI: 1.30–2.91 |
| Nguyen et al.20 | 476 | IDCM | VA | Echo | MD (cut-off = 62 msec) | Se 85%, Sp 73%, AUC 0.84 |
| Pontone et al.35 | 409 | IDCM and NIDCM | MACES (including VA) | CMR | LVEF ≤ 35% and LGE+ | Se 63%, Sp 71%, AUC 0.67, NPV 84.7, PPV 43.6 |
| Disertori et al.40 | 2850a | IDCM and NIDCM | VA | CMR | LGE+ | Se 92.4%, Sp 36.7%, PLR 1.46, NLR 0.21 |
| Leyva et al.41 | 252 | NIDCM with CRT-D or CRT-P | Mortality, HF, MACES | CMR | LGE+ | CRT-D associated with reduced event rates Events only in LGE+ |
| Fallavollita et al.58 | 204 | IDCM | VA | 11C-HED PET | Sympatetic denervation | Prediction of VA: HR = 1.069, 95% CI: 1.023–1.117 |
| Jacobson et al.60 | 961 | IDCM and NIDCM | HF, VA, cardiac mortality | MIBG | H/M ratio > 1.60 | HF: HR = 0.49, 95% CI: 0.32–0.77; VA: HR = 0.37, 95%CI: 0.16–0.85; Cardiac mortality: HR = 0.14, 95% CI = 0.03–0.58 |
CI, confidence interval;11C-HED, 11C-meta-hydroxyephedrine; CMR, cardiac magnetic resonance; CRT, cardiac resynchronization therapy; CRT-D, cardiac resynchronization therapy-defibrillation; CRT-P, cardiac resynchronization therapy-pacing; 3D, three-dimensional; Echo, transthoracic echocardiography; GLS, global longitudinal strain; HF, heart failure; H/M, heart on mediastinal ratio; HR, hazard ratio; IDCM, ischemic dilated cardiomyopathy; LGE, late gadolinium enhancement; MACEs, major adverse cardiac events; MD, mechanical dispersion; MIBG, Myocardial Iodine-123 Meta-iodobenzylguanidine; NIDCM, non-ischemic dilated cardiomyopathy; NLR, negative likelihood ratio; PET, positron emission tomography; PLR, positive likelihood ration; Se, sensitivity; Sp, specificity; VA, ventricular arrhythmias.
Meta-analysis.
Characteristics of the main studies underscoring the role of imaging in the prediction of ventricular arrhythmias in patients with ischaemic and non-ischaemic dilated cardiomyopathy
| Article, reference | n | HF aetiology | Outcomes | Imaging modality | Parameter | Predictive value |
|---|---|---|---|---|---|---|
| Ng et al.17 | 424 | IDCM | Mortality, ICD therapy | Echo | Peri-infarct strain | Prediction of ICD therapy: HR = 1.22, 95% CI 1.09 to 1.36 |
| Haugaa et al.18 | 96 | NIDCM | VA | Echo | MD (cut-off = 72 msec) | Se 67%, Sp 85%, AUC 0.80 |
| Guerra et al.19 | 203 | IDCM and NIDCM | ICD therapy | Echo | GLS | Prediction of ICD therapy: HR = 1.94; 95% CI: 1.30–2.91 |
| Nguyen et al.20 | 476 | IDCM | VA | Echo | MD (cut-off = 62 msec) | Se 85%, Sp 73%, AUC 0.84 |
| Pontone et al.35 | 409 | IDCM and NIDCM | MACES (including VA) | CMR | LVEF ≤ 35% and LGE+ | Se 63%, Sp 71%, AUC 0.67, NPV 84.7, PPV 43.6 |
| Disertori et al.40 | 2850a | IDCM and NIDCM | VA | CMR | LGE+ | Se 92.4%, Sp 36.7%, PLR 1.46, NLR 0.21 |
| Leyva et al.41 | 252 | NIDCM with CRT-D or CRT-P | Mortality, HF, MACES | CMR | LGE+ | CRT-D associated with reduced event rates Events only in LGE+ |
| Fallavollita et al.58 | 204 | IDCM | VA | 11C-HED PET | Sympatetic denervation | Prediction of VA: HR = 1.069, 95% CI: 1.023–1.117 |
| Jacobson et al.60 | 961 | IDCM and NIDCM | HF, VA, cardiac mortality | MIBG | H/M ratio > 1.60 | HF: HR = 0.49, 95% CI: 0.32–0.77; VA: HR = 0.37, 95%CI: 0.16–0.85; Cardiac mortality: HR = 0.14, 95% CI = 0.03–0.58 |
| Article, reference | n | HF aetiology | Outcomes | Imaging modality | Parameter | Predictive value |
|---|---|---|---|---|---|---|
| Ng et al.17 | 424 | IDCM | Mortality, ICD therapy | Echo | Peri-infarct strain | Prediction of ICD therapy: HR = 1.22, 95% CI 1.09 to 1.36 |
| Haugaa et al.18 | 96 | NIDCM | VA | Echo | MD (cut-off = 72 msec) | Se 67%, Sp 85%, AUC 0.80 |
| Guerra et al.19 | 203 | IDCM and NIDCM | ICD therapy | Echo | GLS | Prediction of ICD therapy: HR = 1.94; 95% CI: 1.30–2.91 |
| Nguyen et al.20 | 476 | IDCM | VA | Echo | MD (cut-off = 62 msec) | Se 85%, Sp 73%, AUC 0.84 |
| Pontone et al.35 | 409 | IDCM and NIDCM | MACES (including VA) | CMR | LVEF ≤ 35% and LGE+ | Se 63%, Sp 71%, AUC 0.67, NPV 84.7, PPV 43.6 |
| Disertori et al.40 | 2850a | IDCM and NIDCM | VA | CMR | LGE+ | Se 92.4%, Sp 36.7%, PLR 1.46, NLR 0.21 |
| Leyva et al.41 | 252 | NIDCM with CRT-D or CRT-P | Mortality, HF, MACES | CMR | LGE+ | CRT-D associated with reduced event rates Events only in LGE+ |
| Fallavollita et al.58 | 204 | IDCM | VA | 11C-HED PET | Sympatetic denervation | Prediction of VA: HR = 1.069, 95% CI: 1.023–1.117 |
| Jacobson et al.60 | 961 | IDCM and NIDCM | HF, VA, cardiac mortality | MIBG | H/M ratio > 1.60 | HF: HR = 0.49, 95% CI: 0.32–0.77; VA: HR = 0.37, 95%CI: 0.16–0.85; Cardiac mortality: HR = 0.14, 95% CI = 0.03–0.58 |
CI, confidence interval;11C-HED, 11C-meta-hydroxyephedrine; CMR, cardiac magnetic resonance; CRT, cardiac resynchronization therapy; CRT-D, cardiac resynchronization therapy-defibrillation; CRT-P, cardiac resynchronization therapy-pacing; 3D, three-dimensional; Echo, transthoracic echocardiography; GLS, global longitudinal strain; HF, heart failure; H/M, heart on mediastinal ratio; HR, hazard ratio; IDCM, ischemic dilated cardiomyopathy; LGE, late gadolinium enhancement; MACEs, major adverse cardiac events; MD, mechanical dispersion; MIBG, Myocardial Iodine-123 Meta-iodobenzylguanidine; NIDCM, non-ischemic dilated cardiomyopathy; NLR, negative likelihood ratio; PET, positron emission tomography; PLR, positive likelihood ration; Se, sensitivity; Sp, specificity; VA, ventricular arrhythmias.
Meta-analysis.
Imaging in patients’ selection and clinical benefit
The ESC guidelines on cardiac pacing and cardiac resynchronization therapy were most recently updated in 2021,3 while the most recent version of the US guidelines was released in 2018.4 In a similar manner to the US guidelines, the ESC guidelines provide a Class I recommendation for CRT in patients with symptomatic HF, LVEF ≤35%, and a QRS duration ≥ 150 msec with an LBBB QRS morphology. While previous versions of the ESC guidelines did not recommend CRT for non-LBBB patients with LVEF ≤35%, symptomatic HF, and a QRS duration of 130–149 ms, the current iteration offers a Class IIb recommendation, which is roughly in line with the US guidelines.3,4 Although the ESC guidelines make no mention of patients with Class NYHA I HF, LVEF ≤35%, and QRS duration ≥ 150 msec, such patients receive a Class IIb indication for CRT implant in the US guidelines.4
The application of cardiac imaging, particularly the assessment of LV dyssynchrony and myocardial scar, can be useful to strengthen the benefit of CRT in patients having a class IIa indication and to identify off-label indications (Class IIb and III) for CRT implantation.9,32 The accumulated evidence from multiple, large non-randomized trials for the use of mechanical dyssynchrony and scar localization to select CRT candidates is conform with the ESC requirements for a Class I, level B indication for CRT delivery. This means that in some patients, the disdain for imaging can favour the underuse of CRT by denying therapy to subjects who might benefit. The potential application of MMI to improve and ease the selection of CRT candidates is proposed in Figure 7A.
(A) Proposed multi-modality imaging approach for patients undergoing CRT implantation. (B)Proposed multi-modality imaging approach for patients undergoing ICD implantation. ApR, apical rocking; CT, computed tomography; CMR, cardiac magnetic resonance; HED, hydroxyephedrine; H/M ratio, heart-to-mediastinal ratio; LGE, late gadolinium enhancement; LBBB, left bundle branch block; LVEF, left ventricular ejection fraction; LVMD, left ventricular mechanical dispersion; MIBG, metaiodobenzylguanidine; MMI, multimodality imaging; NYHA, New York Heart Association functional class; OMT, optimized medical therapy; PET, positron emission tomography; SF, septal flash; SPECT; single-photon emission computed tomography.
The clinical benefit of ICD implantations for the primary prevention of VA relies on the likelihood of SCD in patients having a low risk of death from comorbidities or worsening HF. As we have seen above, the benefit of ICD is established in patients with ischaemic cardiomyopathy and LVEF < 35% (Class I-A indication in both European and US Guidelines), but it is still controversial in the case of dilated cardiomyopathy of non-ischaemic aetiology (Classe IIa-B in both European and US Guidelines7).3,4Table 2 displays a list of the main publications on the role of cardiovascular imaging to guide ICD implantation.
The potential application of different imaging modalities to improve and facilitate the selection of ICD candidates is depicted in Figure 7B.
Gaps in knowledge
The negative prognostic role of LVEF ≤35% is well established and is reflected in all international guidelines as a risk factor, key to patients’ management.
Over the last decade, the increasing use of MMI has led to the identification of other morpho-functional parameters that have shown to overcome or be complementary to reduced LVEF as a marker of the arhythmic risk, LV remodelling, and prognosis after CRT. Some of these parameters have shown their usefulness in multiple, large non-randomized trials, which conforms to the B level of evidence of ESC recommendations.
Nevertheless, in the absence of randomized-controlled trials (RCTs) proving the impact of imaging-derived parameters on CIED implantation and prognosis, it seems that the role of cardiac imaging will remain marginal and will not impact future recommendations. Future studies should be designed to overcome this gap and provide evidence on the use of specific imaging-derived tools for the management of patients with HF needing CIED. In the field of VA, increasing interest is presently given to the prevention of SCD in patients with mildly reduced and preserved LVEF, where the arrhythmic risk stratification is still challenging. Similarly, the identification of patients needing CRT-P or CRT-D, particularly in the case of non-ischaemic cardiomyopathy, is a matter of concern. In these patients, the localization and quantification of fibrosis at CMR has shown its usefulness for the stratification of the arrhythmic risk in non-randomized trials35,41 and metanalysis.40 Nevertheless, current recommendations on VA and SCD do not take into account the role of CMR besides the assessment of LVEF and LV dysfunction aetiology. The ongoing randomized CMR GUIDE trial (NCT 01918215) aims at understanding the role of myocardial fibrosis and the risk of developing ventricular arrhythmias in patients with mild-moderate LV dysfunction (LVEF 36–50%), on optimal HF treatment.81 Similar studies are advocated to confirm whether myocardial fibrosis might guide ICD implantation in patients with non-ischaemic cardiomyopathy and severely reduced LVEF and in CRT candidates.
In this same field, we advocate the realization of large prospective trials aiming at confirming the role of MD in the selection of ICD candidates.
In the field of CRT, MMI can provide useful information on the pathophysiological mechanisms of LBBB and ease the identification of the electromechanical substrates that are amenable to being corrected by resynchronization. The information provided by cardiac imaging seems especially valuable for patients with ischaemic cardiomyopathy because the localization and extent of myocardial scar together with the assessment of septal-to-lateral work asymmetry can impact LV remodelling and survival, also in patients with a Class II indication for CRT.32 In the upcoming years, the prospective randomized AMEND-CRT trial (NCT NCT04225520) will indicate if CRT delivery based on the visual assessment of mechanical dyssynchrony can improve the selection and outcomes of CRT candidates. A similar RCT should be proposed to evaluate the role of the combined evaluation of scar localization and myocardial work in the selection of CRT candidates.
Finally, the combination of cardiac activation signals with the upper body geometry to obtain body surface potential mapping might provide interesting insight into various pathophysiological processes such as the identification of re-entry and of the site of abnormal activation or conduction, which might turn useful for the selection of candidates for CRT and/or ICD implantation.82,83
Conclusions
ICDs and CRTs have proven prognostic benefits in patients with systolic HF. The role of various cardiac imaging modalities is evolving from simple assessment of LVEF to identifying patients who would benefit most from ICD or CRT implantation. It is clear from this review that a single imaging modality is insufficient in providing a complete picture to assess chamber function, venous anatomy, dyssynchrony, myocardial scarring, and denervation. An upfront comprehensive assessment using a swiss knife kit of echocardiography, CT, CMR, and nuclear techniques may be best for selecting appropriate patients who would benefit most from advanced therapies with ICD/CRT.
Further research is needed to potentially generate a risk/benefit calculator model from all these techniques to personalize patients’ management and inform decisions for device implantation.
Funding
None declared.
Data availability
No new data were generated or analysed in support of this research.
References
Author notes
These authors have contributed equally to this work and share the second authorship
Conflict of interest: None declared.
