Cardiovascular magnetic resonance (CMR) imaging, with its versatility and ability for soft tissue characterisation in conjunction with the lack of ionising radiation, has over the past decade evolved as a first-line imaging tool for several diagnostic problems.1 Developments in this field, especially the introduction of the ‘late enhancement’ (LE) imaging technique by the working group of Kim and Judd2,3 in the late 1990s, have led to a widespread clinical acceptance of CMR. For this technique, gadolinium-based contrast agents are used that act via a shortening of T1 relaxation and cannot enter normal myocytes with intact, selectively permeable cell membranes; hence, they are restricted to the extravascular interstitial space. The loss of cell membrane integrity, for example due to irreversible ischaemic injury, enables the contrast agent to enter the intracellular space and consecutively increases its volume of distribution. In addition, contrast wash-out kinetics of damaged myocardium are also delayed.4–7 By the use of a dedicated pulse sequence (the so-called inversion recovery gradient echo technique) 10–30 minutes after contrast administration, it is possible to null the signal of normal myocardium and exaggerate the contrast between viable tissue and the gadolinium-enhanced scar region. While initial applications of this technique focused exclusively on ischaemic heart disease, other forms of cardiomyopathies and systemic disease have recently been investigated. This review aims to cover both established and still evolving applications of LE imaging.
Late Enhancement in Coronary Artery Disease
Global left ventricular (LV) systolic function is a prognostic factor in patients with coronary artery disease.8–11 Revascularisation of dysfunctional but viable myocardium has shown to improve global function,12–13 clinical symptoms14 and patient outcome.15 In contrast, in the absence of viable myocardium revascularisation, procedures carry a risk of a higher rate of death and non-fatal events.15 Thus, the discrimination of myocardial dysfunction due to infarcted myocardium with fibrosis and scar tissue – due to chronically hypoperfused but viable myocardium (so-called hibernating myocardium) is of pivotal clinical importance. Various non-invasive and invasive techniques have been evaluated for their usefulness to distinguish reversible from irreversible damaged myocardium, with nuclear medicine techniques such as single photon emission computed tomography (SPECT) having gained the widest clinical acceptance.
Positron emission tomography (PET) until recently has been regarded as the gold standard in non-invasive viability assessment.16–18 Nuclear techniques, however, carry important limitations. They have limited spatial resolution, expose the patient to substantial ionising radiation and, with regard to PET imaging, are not widely available and are highly expensive. Apart from the lack of radiation, CMR is becoming increasingly attractive because of its three- to five-fold higher spatial resolution19 and its ability to allow for simultaneous evaluation of regional wall motion, myocardial perfusion and associated cardiac pathology such as valve disease, presence of pericardial effusion, etc. Several studies have compared the LE technique with SPECT20–23 or PET.24–26 Concordantly they showed a close agreement to nuclear imaging with a superior performance of CMR in the detection of small and very small subendocardial infarcts. For example, in a study by Wagner et al.20 SPECT was unable to detect a fixed perfusion defect in 47% of segments with less than 50% transmural extent of LE.
The close correlation between the extent of hyperenhancement and infarct size in histopathology has been extensively validated.27–30 Furthermore Rehwald et al.31 could demonstrate that reversible injured myocardium does not enhance on LE images. The LE technique has shown an excellent reproducibility32 and several studies could demonstrate its potential for predicting myocardial contractile reserve after revascularisation.19,33,34 Kim et al. studied 50 patients with ischaemic dysfunction before and after revascularisation with cine and LE. They applied a segmental approach with grading of the transmural extent of LE and wall thickening on a five-point scale. While a single cut-off point for prediction of functional recovery could not be defined, an increasing LE transmurality gradually reduced the likelihood of functional recovery after revascularisation. Notably, none of the segments with at least severe hypokinaesia and a transmural extent of hyperenhancement of 76–100% showed improved contractility at follow-up.
Assessment of Infarct Tissue Heterogeneity
In addition to the sole detection of myocardial scarring, CMR LE imaging can take infarct imaging a step further. Frequently hypoenhanced regions surrounded by hyperenhanced tissue can be seen within the infarct zone, which resemble vital myocardium (see Figure 2). These areas have been identified as regions of microvascular obstruction (MVO) that at the time of image acquisition have not yet been reached by gadolinium.35,36
Initial studies have associated the presence of MVO with adverse outcome.35,37 Wu et al.35 followed 44 patients with myocardial infarction and observed that the 11 patients with MVO had more cardiovascular events (death, reinfarction, congestive heart failure or stroke) than those without (45% versus 9%; p=0.016). Furthermore, microvascular status remained a strong prognostic marker even after control for infarct size. These results have been confirmed by Hombach et al.37 in a CMR study of 110 patients early after myocardial infarct. Multivariable analysis revealed LV end-diastolic volume, LV ejection fraction and MVO as significant predictors for the occurrence of major adverse cardiac events.
Sudden cardiac death (SCD) represents a major cause of mortality after myocardial infarct38 with the presence of infarct tissue forming the substrate for malignant re-entry arrhythmias.39,40 Histopathological studies have shown that especially the infarct border zone can exhibit marked spatial heterogeneity with areas of necrosis interspersed with bundles of viable myocytes.41–43 Tissue heterogeneity in the peri-infarct zone can originate areas of slow electrical conduction leading to life-threatening re-entrant tachycardias.44,45 Hence, recent work also aimed to look closer at the heterogeneity of the infarct periphery in LE images.46–48 While Bello et al.49 had already demonstrated a significant correlation between infarct surface area/total infarct size and the inducibility of tachycardias, a study by Yan et al.48 reported an association between the extent of the peri-infarct zone by LE and all-cause mortality in patients with ischaemic heart disease. Schmidt et al.46 studied a high-risk group of 47 patients who underwent implantation of an implantable cardioverter defibrillator (ICD) for primary prevention of sudden cardiac death. Quantification of tissue heterogeneity at the infarct periphery was strongly associated with inducibility for monomorphic ventricular tachycardia and remained the single significant factor in a stepwise logistic regression. Although these preliminary results hold the potential for risk stratification in post-infarct patients, further studies are required to explore the reproducibility and the prognostic capability in a large post-infarct population.
Detection of Intra-cardiac Thrombi
LV thrombi present a frequent complication after myocardial infarct with a substantial risk of systemic embolisation occurring in approximately 13% of patients.50 Transthoracic echocardiography (TTE) is generally used as the main diagnostic technique.51 However, due to insufficient image quality and problems assessing the LV apex (near-field probe) thrombi can be difficult to image and therefore can be missed.
On the other hand, false-positive findings are not infrequent on TTE.52 On LE images the LV cavity shows a homogeneous, strong enhancement after gadolinium administration, with abnormal intraventricular structures having a dark appearance (see Figure 2).53 LE imaging allows for the visualisation of small thrombi (<1cm3), which are missed on cine CMR and TTE, especially when trapped within trabeculations. Mollet et al.54 could demonstrate this instance in a study of 57 patients with acute myocardial infarction, chronic myocardial infarction, or ischaemic cardiomyopathy. LE CMR detected mural thrombi (size ranging from 0.5–8.6cm3) in 12 of 57 patients whereas only six and five of them were visible on cine CMR and TTE, respectively. With LE considered as the gold standard, TTE falsely suggested an apical thrombus in three patients. Although differentiation of mural thrombus and zones of MVO at times can be challenging, LE imaging yields a better identification of LV thrombi than presently used clinical imaging modalities.
Late Enhancement in Non-ischaemic Myocardial Disease
LE techniques have, over time, also gained increasing interest for the evaluation of non-ischaemic forms of myocardial disease. The following section will cover the latest advances.
Discrimination of Ischaemic and Dilated Cardiomyopathy
Initial observations by Wu et al.55 had shown that in contrast to ischaemic cardiomyopathies, none of the control patients with non-ischaemic cardiomyopathy nor any of the enrolled healthy volunteers demonstrated LE areas. The ability of differentiating ischaemic from non-ischaemic cardiomyopathies has subsequently been tested in a prospective manner in several trials.56–57 McCrohon et al. found areas of LE in 41% of patients, which were located in the mid-wall in the majority of cases (28%). In 13% LE resembled the pattern of prior myocardial infarct with subendocardial enhancement. In contrast, all patients with angiographically proven significant coronary artery disease showed LE with the subendocardium involved. Another study could demonstrate LE areas in 81% of patients with coronary artery disease and in only 9% with angiographically classified non-ischaemic cardiomyopathy. Although not diagnostic as a sole investigational tool, the absence or the pattern of LE can, in selected cases, point to a non-ischaemic cardiomyopathy.
Hypertrophic Cardiomyopathy
Patients with hypertrophic cardiomyopathy (HCM) are at risk of arrhythmic SCD. However, due to diverse phenotypic expression, the definition of high-risk subgroups is challenging. Concordantly, studies by Choudhury et al.58 and Moon et al.59 demonstrated areas of LE in 81% and 79%, respectively. Based on conventional clinical parameters Moon et al. could further show that the extent of LE was greater in patients with risk factors for SCD and in patients with progressive disease. In this excellent study he also tried to correlate different LE patterns to future cardiovascular risk. These findings are encouraging and it is conceivable that LE, in the future, could add important information for risk stratification in HCM.
Myocarditis
Owing to the lack of currently available adequate diagnostic techniques, the diagnosis of myocarditis has, in the past, mainly been made on the basis of clinical examination, electrocardiogram and inflammatory laboratory markers. The considered gold standard of endomyocardial biopsy carries a certain peri-procedural risk with additional limited diagnostic sensitivity and specificity and is therefore clinically not widely used. A first report by Friedrich et al.60 in 1998 using the spin echo technique already indicated the diagnostic potential of contrast-enhanced CMR. With the development of the LE technique, CMR imaging of myocarditis has become a promising field. Mahrholdt et al.61 could demonstrate LE in 88% of 32 patients with clinically diagnosed myocarditis and guided endomyocardial biopsy in 21 of these patients, and revealed histological evidence of myocarditis in 91%. In concordance with findings at autopsy,62,63 LE was predominantly seen in the epicardial layer of the lateral free wall (see Figure 3). Moreover, a recent report of 128 patients with suspected myocarditis from the same author group could demonstrate differences in the clinical course and the LE pattern between patients with myocarditis caused by parvovirus B19 and human herpesvirus 6.64
Sarcoidosis
Sarcoidosis is a systemic inflammatory disease of unknown aetiology in which heart involvement presents as a major cause of death.65 While accurate assessment of cardiac involvement has so far proved difficult, its treatment can improve prognosis.66 In a recent study in 70 patients with biopsy-proven sarcoidosis, Patel et al.67 detected cardiac involvement by LE more frequently (24%) than had been detected by use of the Japanese Ministry of Health Criteria (14%).
The LE pattern was inconsistent with that of typical myocardial infarction in 88% (mostly mid-myocardial and/or epicardial enhancement) and, according to the results of Smedema et al.,68 LE is frequently localised in the basal and lateral segments. Since a majority of patients with cardiac sarcoidosis die of arrhythmic SCD,65 like HCM, LE possibly could have a future role in risk stratification for implantation of an ICD.
Cardiac Amyloidosis
Cardiac amyloidosis occurs in up to 50% of patients with light chain (AL) amyloidosis and is associated with a median survival of usually less than one year.69 Identification of cardiac involvement is critical, since therapy might improve cardiac function and prognosis.70,71 In a series of 30 patients with proven cardiac amyloidosis Maceira et al.72 detected global subendocardial LE in 69% coupled with abnormal myocardial and blood-pool gadolinium kinetics. The findings of LE distribution correlated to the transmural histological distribution of amyloid protein and the cardiac amyloid load in an autopsy study of one patient. The authors hence concluded that LE imaging “may prove to have value in diagnosis and treatment follow-up”.
Anderson-Fabry Disease
Anderson-Fabry disease (AFD), an X-linked disorder of sphingolipid metabolism, is characterised by the deposition of glycosphingolipid (GB3) within myocytes73,74 and can present a possible cause of left ventricular hypertrophy, especially in middle-aged men.75 Enzyme replacement therapy has recently become available and can lead to regression of hypertrophy and improvement of regional myocardial function in a subset of patients. Thus, identification of patients with AFD is desirable. In a study of 26 patients with confirmed AFD, LE was present in 13 patients, occurred in the basal inferolateral wall and was not sub-endocardial.76 The question arises of how an intracellular storage disease can cause focal LE, and could be answered in a subsequent autopsy study in one of these patients. Moon et al.77 demonstrated that LE is caused by focal myocardial collagen scarring, which also might be the arrhythmic substrate for SCD occurring in some patients.73,78
Chagas Disease
Chagas disease is caused by infection with Trypanosoma cruzi and predominantly occurs in Latin America, with an estimated 200,000 new cases annually.79 In approximately one-third of infected individuals Chagas disease shows cardiac involvement with development of myocardial fibrosis leading to progressive heart failure and cardiac arrhythmias including SCD. LE imaging in a study by Rochitte et al.80 showed areas of hyperenhancement in 69% of 51 patients with various stages of Chagas disease. The degree of LE increased progressively from the mildest to the most severe disease stages, thus there is a chance that LE can guide future development of new therapeutic interventions designed to halt myocardial fibrosis early in the subclinical phases of the disease process.
Arrhythmogenic Right Ventricular Cardiomyopathy
Arrhythmogenic right ventricular cardiomyopathy (ARVC) is characterised by fibro-fatty infiltration of the right ventricle81,82 leading to regional or global right ventricular (RV) dysfunction, ventricular tachyarrhythmias and SCD. Although CMR was recognised early on as a valuable tool in identifying patients having ARVC,83 diagnostic difficulties remain, especially in the early stages of the disease.84 LE may in future contribute to diagnostic accuracy. Tandri et al.85 studied 30 patients with suspected ARVC, of whom 12 met the Task Force criteria for ARVC.86 Eight of these 12 patients (67%) showed LE compared with none of the 18 remaining patients classified as not having ARVC. Furthermore, LE results showed an excellent correlation with histopathology and predicted inducible ventricular tachycardia on programmed electrical stimulation. However, further studies are needed to assign LE a possible role in the evaluation of patients with suspected ARVC.
Summary
While initially used for the diagnosis and sizing of myocardial infarction, LE imaging has the potential of contributing to other aspects of ischaemic heart disease, namely SCD risk stratification and detection of intra-cardiac thrombi. Additionally, it has already proven its usefulness in the evaluation of non-ischaemic forms of myocardial disease such as viral myocarditis, HCM and sarcoidosis. Possible future applications comprise imaging of eosinophilic myocarditis,87,88 Churg-Strauss Syndrome,89,90 rejection of heart transplants,91 cardiac involvement in muscular dystrophies92,93 and assessment of cardiotoxicity in chemotherapy.94,95 The relative simplicity and robustness of the LE imaging technique and its novel diagnostic potential will result in a further expansion of CMR in cardiac imaging.