A left ventricular ejection fraction (LVEF) of 35% or less is the major determinant of ICD implantation for SCD primary prevention in patients with ischemic or nonischemic LV dysfunction[3]. Even in the recent European Society of Cardiology (ESC) guidelines[3], the only suggested markers of arrhythmic risk to guide ICD implant are LVEF and NYHA functional class (Table 1). However, it is now well-known that ejection fraction alone has limited sensitivity and specificity as a risk marker for SCD, because it is not able to distinguish the risk of sudden death from death caused by heart failure or other non-cardiac diseases. Subsequently, many patients implanted for primary prevention according to current guidelines will have little benefit from their ICD, with a low rate of appropriate ICD therapy (2%-4%/year)[9], while they can suffer from side effects (even > 10%/year overall), in particular inappropriate shocks, lead failure and infections[10,11]. On the other side, many patients who are at risk of SCD are missed when using only LVEF, because the largest part of sudden arrhythmic death patients have only mildly depressed ejection fraction[9,12,13]. Anyway, the substrates of SCD are particularly complex, so it is unlikely for a single test to achieve significantly better predictive accuracy than LVEF. To overcome this limitation, a combination of markers has been proposed[9], for example, combining ejection fraction with different tests that investigate different arrhythmic mechanisms (LGE-CMR, T-wave alternans, programmed ventricular stimulation, evaluation of autonomic tone, etc.).

The pathophysiology of VA in structural heart diseases is due - in most cases - to re-entrant circuits. Electrophysiological studies and anatomic mapping have highlighted, in these cases, the presence of extensive areas where the electrical potentials are absent (indicating the absence of viable myocardium, scar and fibrosis) and areas with low-amplitude, fragmented, late potentials compared to healthy myocardium (conduction with high anisotropy and low speed)[6-8]. The classic arrhythmogenic substrate of re-entry arrhythmias is represented by a mix of these areas, with inflammation often acting as a trigger. Thanks to its ability to identify both areas of myocardial scar/fibrosis and inflammation, CMR can provide essential information in this context[8,9].

Myocardial fibrosis can be evaluated with the LGE imaging technique. Gadolinium-based contrast agents are washed out by viable myocytes and accumulate in extracellular spaces, such as areas of fibrotic tissue, where cardiomyocytes have been replaced by collagen, or in areas of acutely damaged myocardium[7,8]. The LGE imaging techniques have been validated by histology in several studies with animal models[14]. To date, due to the high spatial resolution (approximately 2 mm), it is the most accurate method to detect myocardial fibrosis and to precisely identify its location and extension, distinguishing in particular endocardial, epicardial or transmural involvement. The pattern of LGE distribution is particularly useful in the differential diagnosis between ischemic and non-ischemic fibrosis[15-17]. Virtually all patients with ischemic cardiomyopathy have LGE, presenting with a subendocardial or transmural distribution in myocardial segments following a coronary artery territory[16]; the most common pattern consists of core dense fibrosis within a heterogeneous peri-infarct (gray) zone, indicating the presence of both viable and nonviable myocardium[9]. On the other side, in non-ischemic dilated cardiomyopathy fibrosis is present only in about 30%-40% of cases and it shows a “midwall” pattern, mostly located in the interventricular septum[17] (Figure 1). From an etiological and therapeutical point of view, this is a very important issue: non-ischemic cardiomyopathy on the basis of a traditional definition (clinical history, ECG, echocardiogram and coronary angiography) may be reclassified as ischemic cardiomyopathy thanks to CMR in about 20% of cases[15].

Open in New Tab   Full Size Figure   Download Figure

Figure 1 Late gadolinium-enhanced cardiac magnetic resonance images of ischemic (left panel) and non-ischemic (right panel) dilated cardiomyopathy. In ischemic cardiomyopathy LGE has a subendocardial or transmural distribution in myocardial segments following a coronary artery. In non-ischemic etiology, it shows a “midwall” pattern (panel B, C) or subepicardial distribution (panel D). LGE: Late gadolinium enhancement.

Numerous studies have demonstrated that LGE is a powerful predictor of VA events both in ischemic and non-ischemic cardiomyopathy patients, with moderately to severely depressed LVEF[18-21]. An overview of 19 studies, all with an arrhythmic endpoint, for a total of 2692 patients, indicated that the presence and extension of myocardial fibrosis, documented by LGE, predicted VA both in ischemic and non-ischemic diseases, even in patients with only mildly depressed LVEF[9,18,19]. Furthermore, CMR increased the negative predictive value for SCD prediction to 95%[9,20-22].

Taking into account only non-ischemic dilated cardiomyopathy, the cut-off for risk definition was the presence or absence of fibrosis and its midwall location. These markers were successfully used to dichotomize patients at high vs low risk of ventricular arrhythmic events[23-32]. The largest prospective study in non-ischemic cardiomyopathy by Gulati et al[26] included 472 patients followed for > 5 years. In this paper midwall fibrosis was an independent risk factor for ventricular tachyarrhythmias [hazard ratio (HR) = 4.61], while combining ventricular fibrosis with LVEF significantly improved risk reclassification for the arrhythmic endpoint. A recent meta-analysis of 29 studies including 2948 patients with idiopathic dilated cardiomyopathy[32,33] confirmed that the presence of ventricular fibrosis, identified by LGE, was an important risk factor for arrhythmic endpoints (SCD, VT, VF and ICD therapies): Clinical events occurred in 21% of LGE positive vs 4.7% of LGE negative patients, with an annual event rate of 6.9% and 1.6%, respectively.

In ischemic dilated cardiomyopathy, the issue is more complex: The majority of studies evaluating total LGE or “gray zone” (peri-infarct area) reported a statistically significant dose-response effect for arrhythmic risk, with larger and more heterogeneous scar associated with the higher risk of VA during follow-up[34-39]. Currently, there is not a definite cut-off value of fibrosis/scar extent to adequately differentiate patients at high vs low risk of arrhythmic events, especially in ischemic etiology[18]. The presence of a large amount of ventricular fibrosis/scar has been generally used as a marker of higher risk. However, a great variety of analysis methods and diagnostic thresholds exists[34-39]: Standardization of LGE-CMR should be a target to reach before spreading practical use of this technique for arrhythmic risk stratification. Moreover, no randomized study has been concluded so far: The DETERMINE study[40] was planned to demonstrate the role of LGE-CMR in decision-making for ICD implantation in patients with ischemic cardiomyopathy, but it was prematurely terminated due to a low rate of patient enrollment.

Even with the above limitations, a polyparametric approach, using a combination of different risk markers (including LGE-CMR), could help to refine risk stratification in at least two subsets of patients who are not adequately assessed by current guidelines[9].

The first group is represented by patients with LVEF less than 35% and high risk of death due to heart failure or non-cardiac causes. In this setting, the absence of LGE-CMR (non-ischemic etiology) or a small extension of fibrosis/scar (ischemic etiology), expecially if coupled with negative T-wave alternans test, identifies patients with a relatively low risk of sudden arrhythmic death (about 1%/year) for whom ICD implantation should be critically considered because they will hardly have a benefit[9,21,22,32].

The second group includes patients with LVEF of 35%-50% and high risk of SCD defined by: (1) presence or high burden of fibrosis on LGE-CMR; (2) VT/VF inducibility by programmed ventricular stimulation in post-infarction etiology; and (3) lamin A/C pathological mutation associated with familial sudden death in idiopathic cardiomyopathy. For these patients, even if the current guidelines do not recommend the use of ICD, such a therapy could be critically evaluated, discussed and offered case by case[9,37,38].

Finally, a small portion of patients without LGE at CMR will suffer from sudden death, especially in non-ischemic disease. LGE imaging is not suited to detect diffuse fibrosis that may be present in idiopathic dilated cardiomyopathy. New imaging techniques, such as T1 mapping, are able to detect and quantify diffuse fibrosis by means of extracellular volume fraction and preliminary data show that this pattern is associated with worse outcome in non-ischemic patients[41].

Currently, neither American nor European guidelines support CMR as a first-line tool for risk stratification in dilated cardiomyopathies, so further studies are needed to define its role in this context.

Hypertrophic cardiomyopathy (HCM) is the most common genetic cardiomyopathy and cause of SCD in young people, including competitive athletes. It is caused by mutations in genes encoding cardiac sarcomere proteins and has a prevalence of 1:500 in the general population[42]. HCM is defined by the presence of unexplained LV hypertrophy (wall thickness ≥ 15 mm), associated with non-dilated ventricular chambers, in the absence of other cardiac or systemic diseases that might cause hypertrophy[42,43]. Hypertrophied myocytes are arranged in a chaotic architecture with increased extracellular matrix[42]. The myocardium may also present ischemic areas, caused by microvasculature obstruction, with replacement fibrosis and scar[42-44]. This modified cardiac structure predisposes to the risk of malignant VA such as VT and VF[42,43].

SCD represents the most feared complication, occurring in about 5% of patients[43,44]. In patients with HCM, at high risk for SCD, ICD reduces mortality rate to 0.5% per year[44]. A primary prevention risk model has been proposed to identify high risk patients and guide ICD implant[43,44], based on: (1) family history of premature HCM-related SCD, in close or multiple relatives; (2) unexplained non-reflex syncope, particularly if recent and in young patients; (3) nonsustained VTs on ambulatory ECG, particularly if multiple, repetitive or prolonged; (4) hypotensive or attenuated blood pressure in response to exercise; and (5) extreme hypertrophy (wall thickness ≥ 30 mm). Although current risk factor model is effective, not all high-risk patients are identified and the absence of conventional risk factors does not eliminate the risk of SCD.

In this context, CMR is increasingly considered an important tool, in particular for the evaluation of fibrotic areas (LGE-CMR) and wall thickness[44]. Moreover, it allows more precise characterization of the phenotype, which helps to differentiate HCM from other causes of LV hypertrophy.

Approximately 50%-60% of HCM patients demonstrate LGE-CMR which, when present, occupies on average 10% of the LV myocardial mass. LGE can be observed in any location or distribution, although most frequently in the ventricular septum and free wall (> 30% of patients), with mid-myocardial distribution, and less often involving the apex and the right ventricular insertion into ventricular septum[42] (Figure 2). Moreover, patients with LGE have greater maximal LV wall thickness and LV mass index than patients without[43]. A large number of studies demonstrated that the presence of LGE-CMR identifies areas of myocardial fibrosis where life-threatening VA can originate and is associated with a significant higher risk of SCD, even in patients without conventional risk factors[45-50]. LGE extension, expressed as a percentage of myocardial mass, correlates with the risk of developing life-threatening VA, in particular if LGE exceeds 15% of LV mass[50]. On the contrary, patients without LGE have a low arrhythmic risk and can be reassured.

Open in New Tab   Full Size Figure   Download Figure

Figure 2 Late gadolinium enhancement patterns in hypertrophic cardiomyopathy patients. A: LGE in the lateral wall (small arrows) and in the interventricular ventricular septum; B: LGE in the LV apex and inferior wall; C: LGE localized to the insertion area of the RV wall into the anterior and posterior ventricular septum; D: Transmural LGE involving the ventricular septum. LGE: Late gadolinium enhancement.

CMR also enables the identification of other high-risk subsets of patients such as those with massive LV hypertrophy and apical aneurysms (the latter being a subgroup at increased risk for VA and thromboembolic stroke)[51,52]. Notably apical HCM may be underlooked by echocardiography, while CMR can precisely visualize apical segments and detect hypertrophy and aneurisms.

Current schemes for SCD risk determination, such as American algorithm[51] and ESC risk calculator[53], are not completely effective and precise in risk evaluation. CMR, instead, has shown to improve stratification, providing additional information in patients for whom the current markers underestimate the risk (for example, young asymptomatic patients without conventional risk factors but with LGE) and in patients for whom decision-making about ICD implantation is difficult and ambiguous (for example, patients with a single risk factor and at intermediated risk), and potentially acting as an “arbitrator”. Anyway, at the moment, neither American nor European guidelines support CMR as a first-line tool for risk stratification in HCM[51,53].

Arrhythmogenic cardiomyopathy (AC) is a group of heart muscle disease clinically characterized by life-threatening VA and pathologically by a progressive dystrophy of the ventricular myocardium with fibro-fatty replacement[54,55]. AC affects mostly young males < 40 years old. Its estimated prevalence ranges between 1:2000 and 1:5000, therefore it is considered among rare diseases[54,55]. It is mostly caused by autosomal dominant genetic mutations (with incomplete penetrance and variable expressivity) of desmosomal proteins like desmoplakin and plakoglobin[56,57]. The desmosomal complex, situated in the cardiac intercalated disk, is responsible for tissue strength and stability, binding cells to one another. Consequently, a defective desmosomal complex can cause cell loss with fibro-fatty tissue replacement[58-60]. In its most common form the disease affects the right ventricle, but in a minority of patients it may affect both ventricles or only the left ventricle, thus supporting the use of the more general term AC[54,55]. Clinical manifestations differ in the different phases of the disease, from asymptomatic patients to patients with heart failure and VA or SCD[56-60]. AC is a major cause of sudden death in young and athletes, with VT and VF occurring at any stage[54-61].

Considering the most frequent variant, arryhtmogenic right ventricular cardiomyopathy (ARVC), the diagnosis is based on a score obtained from the assessment of several parameters combined into major and minor criteria[56-59], as there is no single gold standard diagnostic test. CMR has an important role for a comprehensive and precise assessment of right ventricular volumes, function and kinesis[56-59]. Typically, in ARVC myocardial disarray involves the entire ventricular wall, in particular the subtricuspid region and the right ventricle outflow tract (“triangle of dysplasia”), leading to aneurysm formation. In these regions wall motion abnormalities (akinesia or dyskinesia) and aneurysms can be detected by CMR, representing one of the criteria for diagnosis (Figure 3). The usefulness of CMR to detect fatty replacement or fibrosis is limited because the right ventricle has a thin wall and the differentiation between normal epicardial and intramyocardial infiltration is challenging. Therefore, to date, tissue characterization by CMR is not considered in the diagnostic work-up for ARVC[54-59]. Arrhythmic risk stratification in ARVC is based on multiparametric evaluation mainly based on clinical variables; patients at higher risk indicated for ICD implantation are those resuscitated after cardiac arrest, those with sustained and unstable monomorphic VT or exercise-induced unexplained syncope[60,61]. The role of CMR for risk stratification in ARVC is marginal, although significant: the extension of the disease to the left ventricle, identified by LGE-CMR, seems to to associated with a worse arrhythmic outcome and must be looked for[60-62].

Open in New Tab   Full Size Figure   Download Figure

Figure 3 Subtricuspid involvement in arrhythmogenic right ventricular cardiomyopathy. Dilated right ventricle with bulging of the subtricuspid region (arrow). The right ventricular apex is relatively spared.

On the other side, and even more rare, left-dominant arrhythmogenic cardiomyopathy (LDAC) is characterized by epicardial or midmyocardial fibrotic or fibro-fatty replacement in postero-infero-lateral LV wall (“isolated nonischemic scar”) associated with life-threatening VA exceeding the degree of LV dysfunction (LVEF is often normal)[63-65]. LDAC is increasingly recognized as a cause of SCD in young athletes[66-68]. ECG often shows T-wave inversion in infero-lateral leads and low-voltage QRS complexes; VTs have right bundle branch block configuration and are often exercise-induced. In genetic familiar forms, usually autosomal-dominant, gene mutation mostly concerns components of cardiac desmosomes. In non-familiar forms LDAC phenotype can be the result of myocarditis leading to disruption of desmosomal architecture[63-65]. LGE-CMR plays a major diagnostic role because subepicardial/midmyocardial scar location is usually missed by echocardiography (Figure 4). Risk stratification is not well defined: By extrapolation from ARVC, ICD is indicated in patients who survived VF, with poorly tolerated sustained VT, or exercise-induced syncope. LGE-CMR also helps in risk stratification because a “stria” pattern in postero-lateral LV wall has been recently associated with a higher arrhythmic risk compared to the “benign” junctional “spotty” pattern, in a population of young athletes[68].

Open in New Tab   Full Size Figure   Download Figure

Figure 4 Left dominant arrhythmogenic cardiomyopathy. Long-axis (A and C) and short-axis (B and D) postcontrast CMR views of two 34-year-old identical twin brothers showing a subepicardial/midmyocardial stria of LGE involving the lateral and inferolateral left ventricular wall (white arrows). CMR: Cardiac magnetic resonance; LGE: Late gadolinium enhancement.

Sarcoidosis is an idiopathic non-caseating granulomatous disease that affects several organs, mostly the lungs, but also the heart, skin, liver, spleen, eye, and lymph nodes. Sarcoidosis occurs worldwide, being more frequent in African-American and Northern Europeans, especially women. Disease prevalence ranges between 4.7 and 64 in 100000[69]. Cardiac involvement is clinically evident in approximately 5% of patients, in form of: (1) conduction abnormalities; (2) VA including unexpected SCD; and (3) heart failure with reduced LVEF. Moreover, about 25% of patients with systemic sarcoidosis have asymptomatic cardiac involvement. At CMR cardiac sarcoidosis can appear as LGE in a patchy pattern or in longitudinal striae in the midwall or subepicardium, usually located in basal septum or LV wall. Delayed enhancement represents focal scarring, while inflammation areas can be detected with T2-weighted and STIR sequences[70,71]. CMR is also useful for differential diagnosis with ARVC that sometimes can resemble cardiac sarcoidosis. A recent consensus document[69] provided guidance for diagnosis and management of this disease, with a particular focus on arrhythmias. There are few data to help with SCD risk stratification[71]; in general, evidence from major randomized ICD trials of dilated cardiomyopathy is applicable, both in primary and secondary prevention. The presence of inflammation in both ventricles may increase ventricular arrhythmic risk; indeed, patients with implanted ICD have more frequent therapies from their devices, compared to other non-ischemic cardiomyopathies[69,70]. Current consensus recommendations[69] consider the use of CMR and the presence/absence of LGE (combined with electrophysiological study) to guide decision-making about ICD implant.

Amyloidosis is a disease characterized by protein misfold, aggregating into fibrils, and depositing extracellularly with disruption of organ architecture and function. There are two main types which affect the heart: Light chain (AL) amyloidosis and transthyretin cardiac amyloidosis (ATTR), both associated with the risk of VA and SCD[72,73]. Systemic amyloidosis occurs in more than 10 per million person-years in the United States population, with about 2000 new cases of AL amyloidosis occurring each year, approximately half of whom with significant cardiac involvement. The median age at presentation is 55-60 years, especially affecting women. The gold standard for diagnosis is endomyocardial biopsy, but CMR is increasingly used because it provides an accurate tissue characterization without the invasiveness of biopsy. At CMR the most frequent finding is a global subendocardial and circumferential LGE that is specific for cardiac amyloidosis. CMR findings (in particular LGE) have also been associated with prognosis and arrhythmic risk stratification, with the potential for guiding decision about ICD implant[73].

Left ventricular non-compaction (LVNC) is a relatively rare congenital disease, caused by an embryogenesis arrest, in which LV seems to be spongy. Ventricular wall anatomy is characterized by prominent LV trabeculae, a thin compacted layer, and deep intertrabecular recesses[74]. Clinical symptoms are related to neuromuscular disorders, heart failure with reduced LVEF, ventricular arrhythmic events and systemic thromboembolism. CMR can accurately identify this pathology, delineating hypertrabeculations of the apex and the LV lateral wall with subendocardial, midwall or transmural LGE[75]. CMR also helps to differentiate true LVNC from normal variants of increased trabeculations that can be found expecially in young athletes. There are also recent data about the role of CMR for risk stratification. In a recent prospective multicenter study[75], 113 patients underwent CMR, looking for diagnostic criterion of noncompacted/compacted ratio > 2.3 in end-diastole and LGE assessment. At a mean follow-up of 48 ± 24 mo the degree of LV trabeculation had no prognostic impact on the primary outcome (a composite of thromboembolic events, heart failure hospitalizations, VA and cardiac death) above LV dilation and dysfunction. LGE-CMR, instead, showed a significant correlation with life-threatening VA events and SCD.

Myocarditis is a group of heart-specific immune diseases classified by clinical and histopathological manifestations. Myocarditis may resolve spontaneously, recur or become chronic, leading about 30%-40% of biopsy-proven cases to dilated cardiomyopathy (DCM), death or heart transplantation. In the 2013 ESC myocarditis Task Force report[76], the disease was defined histologically as an inflammatory disease of the myocardium diagnosed on endomyocardial biopsy (EMB). Although EMB remains the diagnostic gold standard for diagnosis, it is not widely used. Traditionally, when the diagnosis is only based upon the histological Dallas criteria, myocarditis results to be a relatively rare disease. However, the use of highly sensitive immunohistochemical and molecular tools applied to EMB and of CMR suggests that there is a substantial clinical underestimation of its frequency and of its role in DCM[77,78]. CMR sequences have important diagnostic and prognostic value. T₂-weighted CMR sequences detect edema or water, and T₁-weighted sequences detect inflammation or fibrosis. LGE imaging can help in distinguishing nonischemic patterns of myocyte damage and fibrosis from ischemic injury, and T₂-weighted and early gadolinium enhancement imaging detect other inflammatory features of edema, capillary leakage and hyperemia[78,79]. LGE has been associated with a higher (3.7%/year) risk of a composite of cardiovascular adverse events and its extent also predicted a composite endpoint of cardiac death, heart failure hospitalization, VT, and sudden death[80].

Anderson-Fabry disease is a X-linked disorder due to a deficiency of the alpha-galactosidase enzyme that causes an inability to catabolize glycosphingolipids, leading to their accumulation in several organs, including the heart[81]. The storage of lipids causes an increase of the ventricular wall thickness that simulates HCM and leads to heart failure[81,82]. Diagnosis can be made with CMR showing LGE within the basal infero-lateral wall but tipically sparing the endocardium, related to myocardial collagen scarring that represents the substrate for re-entry mechanism and SCD. Patients who have significant fibrosis on MRI and those with nonsustained VT are at higher risk for arrhythmic complications and may be considered for ICD[82,83].