Cardiac Magnetic Resonance in the Evaluation and Management of Nonischemic Cardiomyopathies

Abstract

Cardiac magnetic resonance is an essential tool in the evaluation and management of patients with suspected or diagnosed nonischemic cardiomyopathies (NICMs). It is considered the gold standard for accurately quantifying cardiac chamber dimensions, function, and mass, aiding in the initial diagnosis of myocardial dysfunction and/or dilation. Its unique value lies in its ability to characterize the myocardium using techniques such as late gadolinium enhancement, T1, T2, and T2* mapping, and extracellular volume fraction estimation, all of which have been histologically validated in specific clinical scenarios. These features enable specific diagnoses of the etiology of NICM, assess prognosis, guide management decisions, and potentially evaluate treatment response. This has allowed the widespread and increasing use of cardiac magnetic resonance in NICMs.

CENTRAL ILLUSTRATION CMR in the Evaluation and Management of NICMs

This chart demonstrates a proposed pathway of when to use cardiac magnetic resonance (CMR) in the evaluation of nonischemic cardiomyopathies (NICMs). Although it is reasonable to obtain a CMR to help diagnose all NICMs, in certain scenarios as outlined this may be of higher diagnostic yield. CMR provides multiparametric information which, along with clinical decision-making, can help arrive at specific etiologic diagnoses for NICMs. Notably, some overlap may exist between diagnoses such as genetic and dilated cardiomyopathies. Knowledge of specific NICMs diagnoses and associated prognostic data obtained from CMR parameters may guide management decisions. Created with BioRender.com. LGE = late gadolinium enhancement.

Cardiac magnetic resonance (CMR) is being increasingly used in the diagnosis and management of individuals with or at risk of heart failure¹ (Central Illustration). CMR is considered a reference standard for quantification of chamber size and ejection fraction. CMR enables assessment of myocardial tissue characteristics using parametric–T1, T2, T2*–mapping and myocardial scar with late gadolinium enhancement (LGE). These techniques have been validated against histopathological specimens in various clinical conditions. Pathognomonic LGE patterns help diagnose specific cardiomyopathies, and quantitative estimation/progression of LGE can guide management decisions. Parametric mapping techniques are useful in assessing disease progression of various cardiomyopathies and their responses to therapy.

In this review, we aim to describe the role of CMR in the diagnosis, management, and prognostication of various nonischemic cardiomyopathies (NICMs). We acknowledge previously described methods of classifying cardiomyopathies, including the comprehensive MOGE(S) criteria that incorporate morphofunctional phenotype, organ involvement, genetic inheritance, etiologic annotation, and functional status. However, for simplicity and to harmonize this review with the CMR data that we reference, we have classified these as genetic, infiltrative, dilated, and inflammatory cardiomyopathies. Additionally, the cardiomyopathies we describe represent protypical examples of a considerably larger number of specific cardiomyopathies. Of note, other imaging modalities such as echocardiography, bone scintigraphy, positron emitted tomography have complementary roles in various specific cardiomyopathies.

GENETIC CARDIOMYOPATHIES

HYPERTROPHIC CARDIOMYOPATHY.

Hypertrophic cardiomyopathy (HCM) is common, with a reported prevalence of 1:200 to 1:500 in the United States.² About 40% of patients have an identifiable sarcomere sequence variation, and the rest have a polygenic background with environmental factors such as hypertension and/or obesity superimposed. Traditionally, a clinical diagnosis of HCM has been established with an end-diastolic maximal wall thickness by imaging of ≥15 mm anywhere in the left ventricular (LV) myocardium, or ≥13 mm if associated with a sarcomere sequence variation or family history. However, these maximal wall thickness criteria have been derived from small studies. Diagnostic thresholds may be lower in some populations, such as those with apical HCM, women, and patients of Asian ancestory.^3,4 It has been recently shown that, due to the influence of age, sex, and body size on normal maximal wall thickness, using fixed thresholds can bias left ventricular hypertrophy (LVH) ascertainment, and using a demographic-adjusted approach to determining LVH can improve the accuracy of diagnosing HCM.⁵ With superior spatial resolution, CMR enables accurate wall thickness measurements and can better identify focal areas of LVH that may not be well visualized by transthoracic echocardiography, such as at the LV apex.⁶ Accordingly, guidelines ascribe a class I indication to CMR for screening of suspected patients in whom transthoracic echocardiography is inconclusive.²

Additionally, CMR can assess abnormalities of the mitral valve and subvalvular apparatus, quantify the severity of mitral regurgitation, and measure the left ventricular outflow tract (LVOT) gradient using velocity-encoded imaging. Notably however, echocardiography is superior to CMR in assessing dynamic LVOT obstruction and can more readily use stress maneuvers such as Valsalva or exercise to provoke LVOT gradients that may not be present at rest.² Indexed LV mass, a sensitive marker for predicting adverse outcomes,⁶ can be accurately measured with CMR. CMR tissue characterization accurately identifies potential HCM mimics such as infiltrative and storage diseases or athletic remodeling. It was shown that CMR helped reclassify ~20% of patients with presumed HCM referred to a tertiary Center of Excellence.⁷

CMR LGE allows the identification and quantification of myocardial replacement fibrosis. Characteristic patterns of LGE are patchy, midwall involvement, particularly at the right ventricular (RV) septal insertion sites and in segments with extensive hypertrophy.⁶ LGE is commonly present in HCM populations, with a majority of studies reporting a >50% prevalence. It is well-established that the LGE burden, rather than presence, is associated with an increased risk of sudden cardiac death (SCD) and cardiovascular and all-cause mortality.^6,8 Although HCM is regarded a common cause of SCD in young adults, the overall risk of SCD in HCM is <1%.⁹ Identifying discriminatory thresholds of LGE associated with the greatest event risk can guide the need for primary prevention implantable cardioverter-defibrillators. LGE ≥15% has been associated with a doubling of SCD risk and is a Class 2B (Level of Evidence: B, Nonrandomized [LOE B-NR]) indication for implantable cardioverter-defibrillator placement, in addition to other imaging features such as wall thickness >30 mm, left ventricular ejection fraction (LVEF) <50%, and/or the midcavity obstruction with apical aneurysm subtype, which are 2A (LOE B-NR) indications.² This area is evolving; a meta-analysis of 11 studies (5,550 patients), which defined LGE as having a signal intensity of >6 SDs from normal myocardium, found that a 10% LGE threshold most accurately predicted SCD.¹⁰

In recent years, parametric mapping in HCM has been increasingly studied. Native T1, signifying interstitial fibrosis, performs well at discriminating healthy from pathologic myocardium in HCM and HCM from its phenocopies.⁶ Myocardial extracellular volume (ECV) can be increased in genotype-positive patients with HCM who have not developed LVH, and thus it may serve as a means of tracking disease progression or response to future targeted treatments.⁶ Elevated ECV has been associated with diastolic dysfunction in HCM.¹¹ High T2 signal intensity, signifying myocardial edema, has been associated with life-threatening arrhythmias¹² and may be associated with a worse prognosis than those with LGE and normal T2.¹³

Coronary microvascular dysfunction (CMD) is implicated in the pathogenesis of replacement fibrosis in HCM. Emerging techniques of CMR quantitative stress myocardial perfusion have shown significant correlation between hypertrophied, fibrotic segments and CMD.¹⁴ Abnormal myocardial perfusion has also been seen in sequence variation carriers in the absence of LVH, which may be useful in early phenotype detection for therapeutic prevention.¹⁵

The National Institutes of Health–funded, multicenter Hypertrophic Cardiomyopathy Registry, which incorporates the CMR, genetic, and biomarker data of 2,750 patients, has identified 6 distinct morphologic subtypes in descending order of prevalence, including localized basal septal hypertrophy, reverse curvature septal hypertrophy, apical HCM, concentric HCM, midcavitary obstruction with apical aneurysm, and other (Figure 1). Of these, 50% were found to have LGE. Combined genetic and morphologic data identified 2 broad populations: genotype-positive patients who are more likely to have reverse septal curvature morphology, LGE, and no significant LVOT obstruction; and genotype-negative patients who are were more likely to have isolated basal septal hypertrophy, less LGE, and more LVOT obstruction.¹⁶ Forthcoming outcome data will provide insight into specific risks associated with these subgroups, which should improve risk prediction. Characteristic CMR features of HCM are summarized in Table 1.

FIGURE 1. Subtypes of Hypertrophic Cardiomyopathy.

Row 1: (left) Four-chamber cine-CMR still-frame of reverse septal curvature type HCM, with (middle) patchy midwall LGE (yellow arrows) in the hypertrophied segment and (right) associated stress-induced perfusion defect. Row 2: Four-chamber cine-CMR still frame of localized basal septal hypertrophy (left) subtype of HCM with associated (right) midwall LGE at RV insertion sites (yellow arrows). Row 3: (left) Four-chamber cine-CMR still frame depicting apical subtype of HCM with (right) associated LGE in the hypertrophied region. Row 4: (left) Four-chamber cine-CMR still frame of midcavitary obstruction with apical aneurysm (blue arrow) and associated LGE in the midventricular segments (yellow arrows). CMR = cardiac magnetic resonance; HCM = hypertrophic cardiomyopathy; LGE = late gadolinium enhancement; RV= right ventricular.

TABLE 1.

Pathognomonic Features of Specific NICMs on CMR

LGE schematics created with BioRender.com.

ALVC = arrhythmogenic left ventricular cardiomyopathy; ARVC = arrhythmogenic right ventricular cardiomyopathy; BSA = body surface area; ECV = extracellular volume; LGE = late gadolinium enhancement; LV = left ventricle; RV = right ventricle.

ARRHYTHMOGENIC CARDIOMYOPATHIES.

Arrhythmogenic cardiomyopathy (ACM) is caused by sequence variations in genes encoding cardiac desmosomal proteins, resulting in fibrofatty myocardial replacement of the RV, LV, or both and clinically predisposing individuals to malignant ventricular arrhythmias.¹⁷ Though initially described as an RV predominant process (arrhythmogenic right ventricular cardiomyopathy [ARVC]), subsequent genotype–phenotype correlations and increasing CMR use have shown that the disease also affects the LV (arrhythmogenic left ventricular cardiomyopathy) or both ventricles. The diagnostic features of RV involvement include systolic dysfunction, dilation, regional akinetic, dyskinetic, or bulging segments. The so-called accordion sign, representing focal dyssynchrony or “crinkling” of the myocardium in the subtricuspid region of the RV, is a well-recognized qualitative pattern seen on CMR in individuals with genetically confirmed ARVC.

Determination of RV dilation and dysfunction on CMR were previously made in compliance with specific indexed RV end-diastolic volume and ejection fraction cutoffs defined by the 2010 Task Force,¹⁸ which have been postulated to be more specific than sensitive. Recognizing that these cutoffs were derived from older cine-CMR techniques with lower spatial resolution, the 2020 Padua criteria recommend determination of RV dilation or dysfunction on CMR according to test-specific nomograms rather than specific numeric cutoffs.¹⁷ LGE of the RV free wall is a specific marker of ARVC¹⁹ and is now considered a major diagnostic criterion, but it can be challenging to identify due to the thin RV wall. Fatty infiltration of the RV free wall is nonspecific and not part of these criteria. CMR feature-tracking derived RV strain, readily obtained by postprocessing images available from standard protocols, is a sensitive method of detecting subclinical functional changes. In a study of 74 patients with definite and borderline ARVC by 2010 Task Force criteria, incorporating subthreshold RV global longitudinal strain values with the conventional diagnostic criteria improved the diagnostic accuracy for those with a borderline diagnosis.²⁰

Biventricular or left-dominant dominant variants are recognized by the presence and pattern of LGE, often preceding LV dilation or dysfunction. Histologically, LGE in ACM represents fibrofatty myocardial replacement and progresses from the epicardium to endocardium.¹⁹ Subepicardial and midmyocardial LGE involving the inferolateral regions of the LV, with or without septal involvement, are typically described with arrhythmogenic LV cardiomyopathy. Specific genetic variants have also been associated with more circumferential, ringlike striae patterns of subepicardial or midmyocardial LGE^21,22 (Figures 2A and 2B).

FIGURE 2. Late Gadolinium Enhancement Patterns in Various NICMs.

(A, B) Short-axis CMR LGE images of ringlike subepicardial LGE (arrows) in patients with arrhythmogenic cardiomyopathy with suspected desmosomal protein encoding sequence variations. (C, D) Left: 4-chamber CMR LGE image of a 55-year-old man who presented with acute complete heart block with multifocal patchy LGE (yellow arrows), diagnosed with cardiac sarcoidosis. Right: Short-axis CMR LGE image of characteristic “hook and triangle” pattern LGE in the basal anterior and inferoseptal segments extending into adjacent RV in cardiac sarcoidosis. (E) Midwall LGE in the basal lateral segments in Duchenne muscular dystrophy. (F) Four-chamber CMR LGE image of a septal midwall stripe (arrows) in dilated cardiomyopathy. NICM = nonischemic cardiomyopathy; other abbreviations as in Figure 1.

ACM can be associated with an inflammatory component, particularly during so-called hot phases of the disease, wherein both the clinical presentation and CMR pattern with T2 elevation can be suggestive of myocarditis. This can lead to diagnostic difficulty, particularly in such acute cases. Though cardiotropic viruses have been detected on pathologic specimens from sporadic ACM cases, their potential causal/contributory role in ACM is unclear.²³

ACM with LV involvement has been associated with an increased risk of SCD.²⁴ Further, the presence of ringlike LGE in those with left or biventricular involvement was shown to be an independent predictor of sustained ventricular arrhythmias.²⁵ Characteristic CMR features of ACM are summarized in Table 1.

MUSCULAR DYSTROPHIES.

Muscular dystrophies are a heterogeneous group of inherited disorders characterized by progressive skeletal muscle degeneration associated with cardiac dysfunction. The X-linked recessive dystrophinopathies, such as Duchenne (DMD, 1:4,000 to 1:6,000 live male births) and Becker muscular dystrophy (BMD, 1:18,000 live male births) have the greatest incidence and the most associated CMR data.²⁶ Cardiac involvement in these diseases is characterized by a progressive dilated cardiomyopathy (DCM), with subepicardial or midmyocardial LGE of the lateral free wall (Figure 2E). CMR techniques can help to identify early LV dysfunction, allowing the early institution of therapies that can delay the progression to overt cardiomyopathy and potentially reverse remodeling.⁶ Presence of LGE precedes declines in ejection fraction and increases with both age and a decreasing LVEF.⁶ In a case-control study of DMD patients (N = 60) undergoing serial CMRs, global circumferential strain at baseline predicted a 3-year interval decline in LVEF, independent of LGE.²⁷

CMR-derived biomarkers are prognostic in muscular dystrophies. In a prospective study of 78 DMD patients, the extent of LGE, LVEF, and indexed LV volumes were predictive of all-cause mortality.²⁸ In another study (N = 148), including those with dystrophinopathies, limb-girdle, and type 1 muscular dystrophies who underwent CMR with 3-dimensional myocardial deformation analysis, peak 3-dimensional strain had incremental prognostic value over LVEF.²⁹

Female carriers of DMD/BMD, though musculoskeletally asymptomatic, often develop cardiomyopathy. Cardiac screening is thus recommended. Lateral wall LGE in a pattern similar to patients with DMD/BMD has been frequently seen in these carriers: in 1 study up to 45%³⁰ with associated incipient features of heart failure. Thus, CMR may serve as a screening tool for these women.

INFILTRATIVE CARDIOMYOPATHIES

CARDIAC AMYLOIDOSIS.

Cardiac amyloidosis (CA) is characterized by the extracellular deposition of pathologic fibrillar proteins in the myocardium, resulting in progressive cardiac dysfunction. Light chain amyloidosis (AL-CA), caused by a plasma cell dyscrasia, has a poor prognosis. Transthyretin-related amyloidosis (ATTR-CA), a progressive disease caused by misfolded transthyretin, is associated with high morbidity though recent treatment advances have positively impacted survival and quality of life. Though its true prevalence is unknown, multimodal imaging has led to greater and earlier detection of ATTR-CA.³¹

Morphofunctional features of CA on CMR include predominantly global LV thickening (Figure 3A) without dilation and a preserved ejection fraction. Biatrial enlargement, atrial septal thickening, and RV thickening are also seen. Similar to echocardiography speckle tracking, CMR feature-tracking strain shows a reduction in global longitudinal strain with a base-apex gradient and relative apical sparing.³² The key advantage of CMR in the diagnosis of CA is myocardial tissue characterization. Typical patterns of LGE are diffuse, circumferential subendocardial and transmural LGE and may include atrial LGE (Figure 3B). Noncircumferential and patchy, midwall LGE with apical sparing may also be seen, perhaps in earlier stages of the disease.³¹

FIGURE 3. Transthyretin Cardiac Amyloidosis.

(A) 4-chamber CMR-cine still frame showing concentric hypertrophy and severe biatrial enlargement in an 86-year-old man diagnosed with transthyretin cardiac amyloidosis. (B) 4-chamber CMR LGE image with diffuse subendocardial left ventricular and atrial LGE (arrows). (C, D) ECV fraction maps in the same patient showing an interval increase in basal anteroseptal ECV due to progressive deposition of transthyretin amyloid fibrils despite interval initiation of tafamidis 2 years before the interval CMR. For reference, mean ECV in healthy individuals is reported to range from 25% to 27 %, with upper and lower limits of 20% to 30%. Though no strict references exist, a change in ECV of 5% or greater would be considered significant. ECV = extracellular volume; other abbreviations as in Figure 1.

Abnormal gadolinium kinetics are another hallmark of CA. On look-locker inversion recovery imaging, normally the blood pool in the LV cavity has a significantly shorter inversion time and therefore “nulls” before normal myocardium, but in CA it often has similar inversion/nulling times. In advanced stages of amyloid deposition, occasionally myocardium may null before the blood pool.⁶

Given the extracellular fibril deposition, ECV can be very high in CA, as can native T1. Using 1 or more of these features, CMR has a sensitivity and specificity of 86% and 97%, respectively, for the diagnosis of CA as determined by a recent meta-analysis.³³ Notably, CMR and bone scintigraphy play complementary roles in the diagnosis of ATTR-CA. CMR should be viewed as a screening technique to suggest the diagnosis, whereas bone scintigraphy is used to make the specific diagnosis of ATTR-CA.³⁴ The sensitivity of bone scintigraphy for the diagnosis of ATTR-CA ranges from 91% to 97%.³⁵ However, earlier stages of the disease and additional morphofunctional features raising suspicion for it may be missed on bone scintigraphy alone. Thus, expert consensus recommends bone scintigraphy as an additional step, either initial or confirmatory in the diagnosis of ATTR-CA.^34,36

ATTR-CA has been associated with a greater LV mass, thicker septum, and greater ECV than AL-CA, and AL-CA has been associated with higher T1. One study (46 AL-CA and 51 ATTR-CA patients) found a greater overall LGE burden in ATTR-CA with more frequent transmural and RV LGE, whereas AL-CA tended to have subendocardial (nonglobal) LGE and nonspecific patterns of patchy, midwall LGE.³¹ Nevertheless, CMR alone is unreliable in differentiating AL-CA from ATTR-CA. However, the combination of typical LGE and inversion recovery pattern on CMR along with negative serum light chains may have a higher specificity (up to 98%) for diagnosing ATTR-CA.³⁷

Multiple studies have evaluated the prognostic performance of T1, ECV, and LGE in CA. A meta-analysis of 13 studies found ECV most prognostic with a significantly greater HR (4.27; 95% CI: 2.97–6.37) compared with LGE and native T1.³⁸

Serial T1 and ECV measurements have been used to monitor therapeutic response in CA (Figures 3C and 3D). In a study of AL-CA patients on chemotherapy (N = 176), 38% had a 5% or greater decline in ECV at 2 years, and the ECV response at 6 months predicted mortality.³⁹ Another AL-CA cohort study (N = 221) showed a reduction in native T1 on chemotherapy, which was associated with favorable hematologic response and changes in traditional cardiac biomarkers. An increase in native T1 was associated with increased mortality.⁴⁰ Tafamidis has also been found to delay myocardial amyloid progression as measured by ECV.⁴¹ In a study of 16 hereditary ATTR-CA patients, treatment with patisiran was associated with ECV reduction, accompanied by improvement in functional status and heart failure biomarkers.⁴² Characteristic CMR features of CA are summarized in Table 1.

MYOCARDIAL IRON OVERLOAD.

Myocardial iron overload (MIO) or cardiac siderosis can occur either due to primary iron metabolism dysregulation, such as in hereditary hemochromatosis syndromes, or, more commonly, due to transfusional iron overload in hemoglobinopathies. This can manifest as acute or chronic heart failure syndromes. Because heart failure from MIO is a major cause of death, particularly in patients with transfusional iron overload, and because the cardiomyopathy is reversible if iron chelation is instituted in time, early detection is critical.⁴³

Local myocardial magnetic field inhomogeneities from iron deposition result in shortening of the tissue relaxation parameter T2* as measured by CMR. Myocardial iron is reproducibly quantified by T2* and correlates with progressive decline in ventricular dysfunction.³⁷ T2* <20 ms is considered diagnostic of MIO and predictive of the need for chelation therapy (Figure 4A). T2* <10 ms has been associated with LV dysfunction, significantly greater risk of developing heart failure and arrhythmic events within 1 year in thalassemia major patients, and is considered severe MIO.⁴⁴ A marked improvement in survival in thalassemia major patients was witnessed in the United Kingdom after the introduction of T2* imaging in 1999, with a 71% reduction in cardiac deaths over the ensuing 5 years.⁴⁵

FIGURE 4. Myocardial Iron Overload.

(A) Basal short-axis T2* map depicting severe MIO with a markedly reduced T2* value of 11 ms in the basal septum. T2* >20 ms indicates no MIO. (B) Corresponding basal short-axis T1 map showing significantly reduced myocardial T1 consistent with MIO. Normal T1 values range from 900–1,100 ms at 1.5-T and 1,100–1,300 ms at 3.0-T. MIO = myocardial iron overload.

The T2* technique has some technical drawbacks, including being sensitive to susceptibility artifacts. It has reduced sensitivity for detecting changes from mild or early MIO. Myocardial T1 is low in MIO, and this correlates to T2* values. In a study of 146 thalassemia patients, low T1 (Figure 4B) was detected in a substantial number of myocardial segments, suggesting that combined use of segmental T1 and T2* could improve the sensitivity for early detection of MIO. T1 may detect iron stored within ferritin, which occurs early in MIO, whereas T2* is predominantly determined by hemosiderin. T1 lacks specificity, however, particularly when myocardial fat or fibrosis are present.⁴⁶ Characteristic CMR features of MIO are summarized in Table 1.

ANDERSON-FABRY DISEASE.

Anderson-Fabry disease (AFD) is a rare X-linked lysosomal storage disorder wherein a GLA gene sequence variation causes α-galactosidase A deficiency, resulting in intracellular glycosphingolipid accumulation in multiple organs. Cardiac involvement is common, resulting in LV hypertrophy, myocardial inflammation, and fibrosis, leading to restrictive cardiomyopathy, arrhythmias, and predominantly cardiovascular mortality.⁴⁷

Based on CMR features and blood biomarkers, 3 distinct phenotypic stages of AFD have been proposed. The subclinical phase of accumulation often occurs during childhood and is characterized by low T1 due to intracellular sphingolipid accumulation, without any myocyte hypertrophy and normal LV mass. Low myocardial T1 is a pathognomonic feature of AFD. The second phase of myocyte hypertrophy and inflammation is characterized by the presence of midmyocardial LGE characteristically in the basal inferolateral segment with local T2 elevation signifying myocardial edema,⁴⁸ often with chronic troponin elevation. RV hypertrophy is also characteristic of AFD.⁴⁷ The third phase of fibrosis is characterized by progressive LGE in segments beyond the basal inferolateral segment with myocardial thinning and onset of clinical heart failure.⁴⁹

CMR provides prognostic information in AFD. The presence of LVH and LGE is associated with an increased risk of adverse cardiac events, including ventricular arrhythmias, heart failure, and cardiac death.⁵⁰ A prospective study (N = 200) found that a multivariable model including native myocardial T1 and indexed LV mass was most predictive for adverse cardiac outcomes in AFD.⁵¹

Because overt LVH and fibrosis negatively affect treatment outcomes in AFD, CMR techniques sensitive to detecting very early cardiac involvement are of interest. Global longitudinal strain by CMR feature-tracking in AFD inversely correlates to native T1, suggesting mechanical dysfunction occurs before sphingolipid deposition.⁵² Quantitative stress perfusion CMR in AFD has shown that even prehypertrophic patients have reduced myocardial perfusion compared with controls, supporting its use as a potential early disease marker.⁵³ Characteristic CMR features of AFD are summarized in Table 1.

DANON DISEASE.

Danon disease is a rare X-linked dominant lysosomal glycogen storage disease caused by a sequence variation in lysosomal-associated membrane protein-2 (LAMP-2). Like AFD, cardiac involvement with early-onset heart failure and ventricular arrhythmia is common. It is also a phenocopy of HCM. Men are more severely affected with skeletal myopathy and neurobehavioral problems in addition to cardiomyopathy. Danon disease may be under-recognized in women presenting later in adulthood, often with isolated cardiac involvement.⁵⁴ Concentric LVH is the most common phenotype, followed by asymmetric septal LVH.⁵⁵ Dilated cardiomyopathy is rare and when present occurs mostly in women. RV hypertrophy is common (up to 81% in 1 series). Extensive LGE is often present: the largest case series to date reported LGE involving a median 35% of LV mass.⁵⁶ Subendocardial, midwall, or transmural patterns of LGE predominantly involving the apex, lateral wall, and RV insertion sites have been described. Characteristic sparing of the basal-midseptum despite septal hypertrophy has been seen in most cases of Danon disease, distinguishing it from HCM. Elevated T2 indicative of myocardial edema/inflammation and impaired myocardial perfusion has also been seen.

DILATED CARDIOMYOPATHIES

DCM, characterized by the presence of LV dilation and systolic dysfunction, is the product of 1 or more nonischemic causes, which may be genetic or not, with superimposed environmental/epigenetic influences. Several nongenetic causes, such as viral myocarditis, endocrine disorders, nutritional deficiencies, substance use, and antineoplastic therapies have been identified. Genes responsible for DCM typically encode cytoskeletal or sarcomeric proteins such as titin and lamin A/C. DCM is quite prevalent, with modern estimates in the range of 1:250.⁵⁷ In addition to confirming the diagnosis of DCM by accurate estimation of ventricular volumes, mass, and function, CMR’s ability to characterize both interstitial (by native T1 values and ECV) and replacement fibrosis (by LGE) aids in prognostication and management.⁶

Longitudinal or patchy striae of midwall LGE along the septum has been seen in ~30% patients with DCM and distinguishes it from ischemic cardiomyopathy⁶ (Figure 2F). The presence of characteristic midwall fibrosis in DCM is independently prognostic of all-cause mortality and SCD.⁵⁸ A greater LGE burden in DCM is associated with a lack of response to guideline-directed heart failure therapies and a lower probability of LV reverse remodeling.⁵⁹ LGE is a strong predictor of ventricular arrhythmia and SCD in DCM.

CMR may guide which DCM patients benefit most from a primary prevention implantable cardioverterdefibrillator. DCM patients with midwall LGE, even with LVEF ≥40%, are at an increased risk of SCD.⁶⁰ Furthermore, in a large study of ~1,200 patients with DCM, a combination of LVEF and LGE was superior at predicting arrhythmic risk compared with the LVEF ≤35% cutoff alone.⁶¹ Mechanical dyssynchrony indices assessed by CMR strain techniques and lack of scar at the LV lead location have been shown to predict both functional improvement and LV remodeling after cardiac resynchronization therapy.^62,63 These techniques have potential to optimize the use of device therapy in DCM but have not been assessed in randomized trials and are currently not widely used in clinical practice.

The myocardium in DCM may have subtle, diffuse interstitial fibrosis, even in the absence of visually apparent replacement fibrosis/LGE. T1 mapping and ECV, which can quantify such fibrosis, have prognostic significance in DCM. Mean native T1 and ECV are associated with cardiac-related death or transplantation in DCM patients without LGE, and combining T1, ECV, and LGE may improve risk stratification.⁶⁴ Characteristic CMR features of DCM are summarized in Table 1.

CMR has become a cornerstone in the surveillance of cancer therapy–related cardiac dysfunction because it is more sensitive for diagnosing LV dysfunction. Echocardiography may overestimate LVEF in adult survivors of pediatric cancers, resulting in an underdiagnosis of cardiomyopathy subsequently diagnosed by CMR.⁶⁵ CMR also accurately measures LV mass, a decline in which is associated with greater adverse events in those with anthracycline exposure.⁶⁶ CMR can aid in the subclinical detection of myocardial dysfunction in cancer therapy–related cardiac dysfunction by identifying abnormal myocardial strain. In a study of cancer patients receiving chemotherapy and radiation undergoing serial CMR, preserved baseline circumferential and longitudinal strain measured by fast strain-encoded CMR was more predictive than LVEF, myocardial T1 or T2 values, or biomarkers for the development of overt or subclinical cardiotoxicity. In the same study, more preserved strain after the development of toxicity was most predictive of recovery when treated with cardioprotective medications.⁶⁷

It is suggested that excessive trabeculations of the LV, often referred to as LV noncompaction, are no longer considered a distinct cardiomyopathy. “Non-compaction” is considered a misnomer because there is no embryologic evidence of formation of compact from noncompact myocardium.⁶⁸ By CMR, excessive trabeculation is often defined as a trabecular to compact myocardial ratio of >2.3 in end-diastole, assessed in a 4-chamber view. Other measures of total trabecular volume are increasingly being investigated as better approaches for defining excessive trabeculations.

Although excessive trabeculations may be associated with dilated or hypertrophic cardiomyopathies, they may also occur as a physiologic response to increases in myocardial preload, such as athletic training or pregnancy. A wide range of genes are associated with an excessive trabeculation phenotype, fewer than half of which are also associated with cardiomyopathies such as TTN, MYH7, and MYBPC3.⁶⁹ In most cases, excessive trabeculations and cardiomyopathy were concurrent. Hence, no causal relationship between excessive trabeculations and cardiomyopathy has been established. No independent prognostic value of excessive trabeculation alone in patients without LGE or ventricular is known.⁷⁰ Thus, it is advised that when excessive myocardial trabeculations are present, management should be in accordance with guideline recommendations for the underlying cardiomyopathy (if present) without regard to the trabecular pattern.

INFLAMMATORY CARDIOMYOPATHIES

CARDIAC SARCOIDOSIS.

Sarcoidosis is a multi-system disease of unknown etiology histologically characterized by non-necrotic inflammatory granulomas. The prevalence of cardiac sarcoidosis (CS) ranges from 20% to 30% of patients with extracardiac sarcoidosis, although this varies by region, race, and sex.⁷¹ Manifestations of CS include LV dysfunction, high-grade bradyarrhythmias, and malignant ventricular arrhythmias.

CMR LGE can identify and quantify myocardial inflammation/fibrosis/scar in sarcoidosis with high spatial resolution. Endomyocardial biopsy has limited sensitivity for the diagnosis of CS due to the patchy nature of myocardial involvement. However, CMR is able to identify small areas of fibrosis/scar. A large meta-analysis of the diagnostic performance of various imaging studies found that CMR had a sensitivity of 95% and specificity of 85% for the diagnosis of CS with clinical diagnostic criteria used as the reference.⁷² CMR is now a widely used modality for diagnosing CS and is incorporated in contemporary diagnostic algorithms.⁷⁰

Though patterns of LGE can vary in CS, a meta-analysis of patients with histologically proven CS at autopsy or cardiac transplant found that the patterns of LGE found in >90% of these patients–termed pathology-frequent LGE–were LV subepicardial, LV multifocal, septal and RV free wall involvement.⁷² Another study (N = 149) noted that basal inferoseptal triangular LGE with wedge-shaped extension into the inferior RV or LV with or without endocardial RV septal LGE–the so-called hook and triangle sign–had a 98% specificity for the diagnosis of CS with clinical diagnostic criteria as reference⁷³ (Figures 2C and 2D).

CMR may also be able to accurately identify CS patients with active myocardial inflammation who would warrant anti-inflammatory therapy by the presence of elevated native myocardial T2. A decline in T2 has also been noted in CS patients treated with anti-inflammatory therapy.⁷⁴ Fluorodeoxyglucose positron emission tomography (FDG-PET) has a higher sensitivity for detecting active inflammation in cardiac sarcoidosis and remans the standard modality for diagnosing active disease.⁷¹ It can also serve as a confirmatory test when cardiac sarcoidosis is suspected in the absence of known extracardiac sarcoidosis.

CMR biomarkers are prognostic of adverse outcomes in CS. LV dysfunction with EF ≤40%−50% is independently predictive of all-cause mortality, heart failure hospitalizations, and ventricular arrhythmias.⁷⁵ Multiple large meta-analyses have found that LV LGE is significantly associated with greater odds of all-cause mortality and arrhythmic events,^71,76 especially when the LGE burden is >5%, regardless of LVEF.⁷⁷ RV involvement portends a worse prognosis in CS. Reduced RV function and presence of any RV LGE has been associated with worse outcomes in CS.^70,78 Characteristic CMR features of CS are summarized in Table 1.

CARDIOMYOPATHIES ASSOCIATED WITH AUTOIMMUNE RHEUMATIC DISEASES.

Autoimmune rheumatic diseases are a heterogeneous group of disorders where inappropriate activation of the immune system results in multiorgan tissue damage. Women are more commonly affected than men. Cardiovascular involvement is characterized by inflammation, fibrosis, and microvascular and macrovascular ischemia, and CMR is well suited to detect and monitor these pathologic changes.⁷⁹ Autoimmune rheumatic diseases with potential cardiovascular involvement include systemic lupus erythematosus (SLE), rheumatoid arthritis, systemic sclerosis, idiopathic inflammatory myopathies including dermatomyositis and polymyositis, and small, medium, and large vessel vasculitis.⁸⁰

SLE is associated with cardiac involvement in >50% cases. Prevalence of myocarditis is estimated at 5% to 10% in symptomatic patients, but this may be higher subclinically. SLE patients with clinically suspected myocarditis have been found to have significantly elevated T1 and T2 values suggestive of myocardial edema and fibrosis, in addition to greater LV mass and reduced global longitudinal strain.⁸¹ LGE patterns described in these patients have included patchy epi/endomyocardial LGE and small amounts of midmyocardial LGE in the interventricular septum.⁶ CMR can distinguish subclinical myocardial involvement, with elevated T1, T2, ECV, impaired strain and LGE detected in SLE patients without suspected cardiac involvement, highlighting its use as a potential screening tool.⁶ CMR may also be useful for therapeutic monitoring in SLE because a reduction in T1 and T2 values after anti-inflammatory treatment in these patients has been observed.⁸¹

Stress perfusion CMR has demonstrated that CMD is prevalent in SLE. In 1 study among 20 female SLE patients with anginal symptoms who underwent stress CMR, none had obstructive coronary artery disease, and 44% had evidence of CMD with reduced myocardial perfusion reserve.⁶ At 5-year follow-up evaluation, a majority had persistent symptoms, and nearly 50% had similar or worse myocardial perfusion reserve, suggesting an alternative to accelerated obstructive atherosclerosis as an explanation for persistent symptoms in this population.⁸²

Cardiovascular involvement in rheumatoid arthritis predominantly affects young women and accounts for 40% to 80% of the associated premature mortality. It has been shown that rheumatoid arthritis patients have diffuse myocardial fibrosis as evidenced by greater native T1 and ECV values and impaired peak circumferential strain. In this study, T1 and ECV correlated with increased disease activity, and 46% had LGE, predominantly midmyocardial in the basal inferior and lateral wall.⁸³

Among autoimmune rheumatic diseases, systemic sclerosis is reported to have the highest overall risk of cardiovascular disease.⁸⁴ Systemic sclerosis patients have higher native myocardial T1 and ECV, which likely represent low-grade inflammation in diffuse myocardial fibrosis. In 1 study, >50% had LGE predominantly in a patchy, midmyocardial septal, inferior, and inferolateral distribution.⁸⁵ In a prospective study of 130 systemic sclerosis patients, reduced LVEF or the presence of LGE suggestive of primary myocardial involvement (nonischemic and non–RV-insertion site) was associated with a 2-fold increase in death or major adverse cardiac event.⁸⁵ Systemic sclerosis has been associated with stress-induced myocardial perfusion defects in a non-segmental distribution, which highly correlated to Raynaud phenomenon and digital ulceration, postulating a common mechanism of microvascular dysfunction.⁸⁶

CMR plays a similar role in medium and small-vessel vasculitides that can involve the myocardium. Antineutrophil cytoplasmic antibody (ANCA) associated small vessel vasculitides including granulomatosis with polyangiitis and eosinophilic granulomatosis with polyangiitis are associated with LGE, with a 43% prevalence in 1 study (N = 37). This was predominantly in a nonischemic epicardial, midmyocardial, or noncoronary subendocardial pattern. Moreover, elevated T1, T2, and ECV were observed independent of LGE, suggesting their potential role in subclinical detection or differentiating stages of the disease.⁸⁷

PRACTICAL CONSIDERATIONS FOR CMR IN NICMs

Patients with NICMs often undergo therapeutic or preventive cardiac implantable electronic device (CIED) implantation. In patients where this is anticipated, obtaining a CMR beforehand would be prudent, both to obtain prognostic information that may guide the decision for implanting a particular type of CIED and to avoid some practical challenges associated with CMR in patients with CIED. Although most contemporary CIEDs are magnetic resonance imaging conditional, and scans can be safely performed 6 weeks after device implantation,⁸⁸ the need for additional monitoring during CMR may pose logistic challenges and delays in scheduling. Multiple CMR techniques can be used to mitigate CIED-related impairments in image quality due to artifacts and improve image quality. With the advent of wideband LGE pulse sequences, morphology, function, and LGE can be reliably obtained in a patient with a CIED.⁸⁹

All the CMR techniques used for diagnostic purposes in NICMs and described earlier are widely available and should ideally be performed in expert and nonexpert centers alike. Patients with NICMs often have some degree of renal dysfunction. Concerns regarding gadolinium contrast use in patients with chronic kidney disease due to the potential risk of nephrogenic systemic fibrosis have historically limited its use in these patients. With contemporary group II macrocyclic agents, the risk of nephrogenic systemic fibrosis is exceedingly low, and the American College of Radiology recommends against witholding it when a CMR is indicated.⁹⁰ Gadolinium contrast is also to be administered judiciously and can often be foregone when obtaining interval CMR to assess changes in function or tissue characteristics.

FUTURE DIRECTIONS

Multiple aspects of CMR are constantly evolving with developments aimed at more efficient, accessible, and patient-friendly scans. Virtual native enhancement is a deep learning–derived technique that generates images resembling conventional LGE by enhancing existing image signals in T1 maps and cine images without the use of gadolinium contrast. In a study of Hypertrophic Cardiomyopathy Registry patients (N = 124 for validation), virtual native enhancement had a high agreement with conventional LGE both in terms of distribution and quantitative burden, with superior image quality.⁹¹ This has the potential to reduce scan time and cost and to allow a contrast-free protocol for myocardial tissue characterization. All-in-one CMR aims to abbreviate scan times by using techniques or sequences that simultaneously image multiple processes such as cardiac motion, tissue characteristics from relaxation properties, and contrast agent dynamics rather than the current practice of serial individual pulse sequences.⁹² Real-time CMR techniques aim to accelerate the scan and avoid the need for breath-holding and/or electrocardiogram-gating. The latter technique is currently used clinically in a limited manner, typically as a last resort when such patient-related challenges arise, and it may have limited spatial and temporal resolution. With further technical advances and extensive clinical validation, these techniques are likely to significantly shorten scan times, improve access, and allow CMR’s use in broader, sicker cardiomyopathy populations.⁹³

CONCLUSIONS

The ability of CMR to characterize myocardium using techniques such as T1 mapping with ECV measurements, T2, T2* mapping, and LGE makes it an invaluable tool in the diagnosis of various NICMs as highlighted in this review. Additionally, these data can help risk stratify patients with NICMs, optimize their management plan, and potentially evaluate therapeutic response. The routine use of CMR in this population should be strongly considered.

HIGHLIGHTS.

• Cardiac magnetic resonance is a valuable tool in the evaluation and management of NICMs.

• Myocardial tissue characterization by different CMR parameters provides insight into the etiology of the cardiomyopathy.

• Late gadolinium enhancement, parametric mapping, extracellular volume fraction, and myocardial strain assessed by CMR are also of prognostic value in cardiomyopathies.

ABBREVIATIONS AND ACRONYMS

ACM

arrhythmogenic cardiomyopathy

AFD

Anderson-Fabry disease

AL-CA

light chain cardiac amyloidosis

ARVC

arrhythmogenic right ventricular cardiomyopathy

ATTR-CA

transthyretin cardiac amyloidosis

DCM

dilated cardiomyopathy

HCM

hypertrophic cardiomyopathy

LGE

late gadolinium enhancement

MIO

myocardial iron overload

NICM

nonischemic cardiomyopathy

SCD

sudden cardiac death

SLE

systemic lupus erythematosus