1. Introduction
Cardiomyopathies are myocardial disorders in which the heart is structurally and functionally abnormal. They are currently sub-classified on the basis of cardiac morphology as hypertrophic cardiomyopathy (HCM), dilated cardiomyopathy (DCM), arrhythmogenic cardiomyopathy (ACM), and left ventricular noncompaction cardiomyopathy (LVNC). The European Society of Cardiology (ESC) divides dilated cardiomyopathy into two groups, familial and nonfamilial
[1]. Conventionally, idiopathic DCM with at least one first- or second-degree relative with confirmed idiopathic DCM is defined as familial DCM
[1,2][1][2]. These non-ischaemic cardiomyopathies are less common than ischaemic cardiomyopathy
[3]. Of non-ischaemic aetiologies, idiopathic DCM is the second most common aetiology, accounting for 31% of cases. Non-ischaemic DCM is more common in female, non-white, and younger individuals
[3].
Dilated cardiomyopathy accounts for up to one-third of heart failure cases and is one of the leading causes for cardiac transplantation. The reported prevalence of DCM from epidemiological data is 36.5/100,000 individuals in Western populations
[3]. This is likely to be an underestimate, however, since its diagnosis has relied on modalities with low sensitivity, such as echocardiography or angiography. Hershberger and colleagues, in their review, were able to report a higher prevalence of DCM of > 1 per 250 individuals on the basis of recent data
[1,4][1][4].
Estimations of the prevalence of
familial DCM range widely, between 2% to 65%, averaging 23% in a meta-analysis of 23 studies
[5]. This is due partly to heterogeneity of the diagnostic criteria as well as increasing diagnosis over time related to more systematic clinical screening
[5]. Familial DCM has the following subtypes with already mapped genetic loci (>40): autosomal dominant, autosomal recessive, X-linked, and mitochondrial forms. These might comprise either a pure cardiomyopathy or may have associated myopathy
[6,7][6][7]. The penetration of familial DCMs is incomplete, variable, and age-dependent
[6,7][6][7].
Amongst familial DCM, monogenic causes account only for approximately 30–40% of cases
[8,9][8][9]. The implication of this is that that traditional Mendelian considerations will leave more than half of cases without a defined monogenic cause. It is, therefore, likely that complex mechanisms underlie familiar DCM rather than monogenic transmission alone. It has accordingly been proposed that common variants predispose to DCM in the appropriate environmental exposure, while rarer variants may underlie monogenic forms. Including rare variants in the genetic panel increases the yield of genetic testing for DCM, with a genetic diagnosis achieved in approximately 40% of apparently familial cases
[4]. Pathogenic genetic variants can be identified in 15–25% of sporadic DCM
[8,9][8][9]. Titin (
TTN) mutations are the most common aetiology of familial DCM, occurring in ~25% of familial cases of DCM and in 18% of sporadic cases
[1].
Rare variants in more than 30 genes can produce a DCM phenotype, some of which also underlie other cardiomyopathies, inherited muscle diseases, or myopathic syndromes. These genes encode both contractile and non-contractile proteins, such as cytoskeletal proteins, abnormalities of which result in the DCM phenotype. This results in reduced resistance to mechanical stress as well as abnormalities of intracellular calcium handling, myocellular energetics, and sarcolemmal ion channel function
[1]. De novo mutations are rare and defined when none of the biological parents carry the offspring’s mutation and confirm the pathogenic status of genetic variants
[9]. A multicentre study tested the hypothesis that both familial and non-familial DCM have a rare variant genetic basis and concluded that most idiopathic DCM have a genetic basis
[10].
2. Diagnosis
The diagnostic process consists of a detailed clinical history alongside at least a three-generation family pedigree, comprehensive cardiac imaging, biochemical profile, and genetic testing, where clinically indicated. Systematic screening with electrocardiography (ECG) and echocardiography of first-degree relatives of patients with idiopathic DCM has been proposed to identify subclinical forms. European Society of Cardiology Guidelines recommend screening with an ECG and echocardiogram in all first-degree relatives of an index patient with DCM, irrespective of family history
[11].
Mestroni et al.
[6] define
familial DCM by either the presence of two or more affected relatives within a single family or in the presence of a first-degree relatives of a DCM patient with SCD below age 35 years
[6]. Criteria proposed to diagnose the index case includes imaging criteria, family history, ECG, imaging criteria, and exclusion of competing causes, such as significant coronary artery disease, chronic alcohol excess, uncontrolled hypertension, persistent arrhythmia, pericardial disease, congenital heart disease, and cor pulmonale. There are some inherent difficulties in identifying the index case. Firstly, the varying clinical presentation and course of the disease lends diagnostic challenges. Secondly, acquired disorders, such as hypertension, excess alcohol intake, and systemic inflammatory diseases may produce phenocopies of idiopathic DCM or act as environmental factors, unmasking rare variants
[6].
3. Genetics in Familial DCMs
Studying the genetic basis of DCM requires either multigeneration DCM pedigrees or genome-wide sequencing. Most studies use the latter approach, and once DCM-associated variants are identified, the numbers of DCM patients with these variants are compared with the number of individuals in the reference datasets carrying the same variants. A probability of 90% or more is required to be labelled pathogenetic, although this does not necessarily reflect causation
[1]. Sequencing of large numbers of genes is required due to low prevalence, heterogeneity of mutations, private mutations, modifier genes, and different mutations producing the same phenotype
[4]. Fortunately, there are large, easily available genetic datasets which enable evaluation of already identified pathogenic mutations. Another approach, used by most clinical centres, is to perform targeted next generation sequencing of high-risk variants.
Several genome-wide sequencing association studies (GWAS) have identified genetic variants associated with DCM. However, the yield has been limited by modest sample sizes (< 5000 cases). The prevalence of these pathogenic genetic variants is greater than the estimated disease prevalence. Hershberger suggested this mismatch is due either to lower penetrance of the mutations, non-pathogenicity of some, or that the actual DCM prevalence is higher than estimated
[4]. Villard et al.
[12] were the first to perform genome-wide study, while Esslinger et al.
[13] and Meder et al.
[14] carried out similar studies involving 3000 and 4000 DCM cases with matched controls, respectively.
4. Cardiac Imaging in Familial DCMs
DCM has been defined by echocardiography by the degree of systolic impairment, that is, fractional shortening (FS) less than 25% (> 2SD) and/or ejection fraction less than 45% (> 2SD) and the degree of LV enlargement; or LV end diastolic diameter (LVEDD) greater than 117% (2SD (112%) plus 5%) or end diastolic volume (LVEDV) greater than 2SD of the predicted value, as corrected for age and body surface area, excluding any known cause resulting in the myocardial abnormality observed
[6].
Echocardiography is often the first imaging test for assessing LV remodelling and also provides associated data, such as the presence and severity of functional mitral regurgitation. Speckle-tracking echocardiography uses the distinct speckle pattern in the myocardium to assess myocardial deformation. Abnormalities of strain and strain rate can be detected by echocardiography in first-degree relatives of patients with DCM, indicating a subclinical phenotype
[15].
Cardiovascular magnetic resonance imaging (CMR) is the reference standard for measurement of ventricular volume, ejection fraction, and myocardial mass. In addition, CMR detects myocardial oedema, which, when present, may suggest an inflammatory basis for the observed phenotype. Long native myocardial T1 time and high extracellular volume (ECV) fraction may be helpful in differentiating DCM from athletic heart adaptation or iron overload cardiomyopathy. The presence, pattern, and burden of late gadolinium enhancement (LGE) may be helpful in determining the risk of malignant ventricular arrhythmias. Echocardiography may be suboptimal in certain individuals when CMR is recommended as an alternative. Progressive increases in chamber dimensions, strain abnormalities, and LGE are features of early DCM
[16]. Longitudinal studies over many years with imaging are required to characterise DCM progression in genetically predisposed individuals. CMR offers higher repeatability in volumes and ejection fraction and may allow detection of subtle changes in surveillance of mutation carriers. LGE detects replacement fibrosis but not diffuse fibrosis, so it may be deceptively reassuring even in those with established diffuse fibrosis, where native T1 is long and ECV is high.
Amin et al.
[2] demonstrated that the combining CMR with genetic information allows better DCM stratification and results in a change in management, as per an ESC Position Paper
[2]. In DCM, the degree of fibrosis shown by LGE, is a predictor of mortality and hospitalisation, particularly ventricular arrhythmias
[17]. Hyper-trabeculation is also detected commonly in DCM (36%), although its presence was not associated with adverse outcomes, and this can be detected in normal hearts
[18,19][18][19].
Imaging traits of DCM have been used to investigate genetic variants involved in a DCM phenotype
[18,20][18][20]. valuation of approximately six thousand subjects from cardiac imaging from the Candidate Gene Association Resource (CARe) Study and more than a thousand individuals from the Multi-Ethnic Study of Atherosclerosis (MESA) with CMR yielded four associated genes
[21]. Pirrucello et al.
[7] analysed CMR-derived left ventricular measurements in 36,000 UK Biobank and 2000 Multi-Ethnic Study of Atherosclerosis participants, identified 45 new loci, and developed a polygenic score for prediction of DCM on the basis of variants most strongly associated with DCM phenotypic variables on CMR.
5. CMR Sequences for DCM Assessment
The cardiac MRI protocol for DCM involves cine imaging and tissue characterisation, including mapping and gadolinium-enhanced images; flow mapping, oedema, and stress perfusion imaging may be considered. Cine images use a balanced steady-state free processing (SSFP) sequence, which provides high signal-to-noise ratio and myocardial blood pool contrast or flash (spoiled gradient echo) sequence in case of cardiac device or at 3T. These are usually retrospectively ECG gated along with breath-held-in expiration, with a minimum of 25 reconstructed phases, with a slice thickness and gap of 8 mm and 2 mm to give temporal resolution less than or equal to 45 ms and spatial resolution 2 mm.
Gadolinium-enhanced late images are inversion recovery or phase-sensitive inversion recovery gradient echo sequences, segmented as a routine, while single shot for patients with difficulty holding their breath or irregular rhythm. In plane resolution, it should be ~1.4–1.8 mm. These are generally acquired 10 min after the contrast injection; however, the delay should be tailored to the dose and protocol; images are acquired every other beat but can vary according to heart rate. A TI scout can be used as a guide for choosing inversion time to null myocardium; however, inversion time can be different due to different readout parameters. The late gadolinium-enhanced images are acquired in the same planes and slices as cine images.
T1 mapping sequences most frequently involves shortened MOLLI (ShMOLLI) during breath-holding, with the trigger delay set to end-diastole, slice thickness 6–8 mm, with in-plane resolution 1.6–2 mm. The native and 15 min post contrast maps are acquired as short axis slice, but number and plane of slices can be modified. Other sequences, such as SASHA, which involve saturation instead of inversion, are also available.
T2 mapping uses either a T2-prepared SSFP sequence or fast spin echo sequences, usually with motion correction.
Black blood T2-weighted short tau inversion recovery (STIR) sequence is used for oedema imaging. Images are usually acquired in three short axis slices as well as long axis views.
6. CMR Characteristics of Familial DCMs
Septal linear mid-wall fibrosis is characteristic of, although not specific to, familial DCM. Furthermore, its presence carries adverse prognosis
[22]. In a meta-analysis by Becker et al., LGE was present in approximately 45% of patients with DCM and was associated with adverse outcomes, including ventricular arrhythmia, hospitalisation, and death, while its absence predicted LV reverse remodelling
[20].
Table 1 summarises the features of genetic DCM subtypes, which are now detailed further by subtype.
Table 1.
Genetic dilated cardiomyopathies and their clinical, MRI, and arrhythmia features.
Mutation
|
Screening and Transmission
|
Age of Presentation/Prognosis
|
ECG and Clinical Features
|
Extra-Cardiac Features
|
Imaging Features/Volume and Function
|
Tissue Characterization
|
Arrhythmia/Conduction Disease Risk
|
Lamin
|
Recommended commencing 10–12 years by ECG and Echo
Autosomal dominant
|
Young age onset and progression
|
Conduction or depolarization abnormalities, such as IVCD/BBB/
Early abnormalities in strain or diastolic function
|
Arterial and venous thromboembolism
|
Isolated LV dilation or dysfunction; marked dilation unusual
Wall thinning not present
RWMAs (basal) in Lamin A/C
HNDCM to DCM progress
|
Prominent right ventricular epicardial fat
Early increase in ECV
LGE shows basal to mid septal, mid-myocardial fibrosis (extensive fibrosis only late in disease)
|
Increased arrhythmias and both AV and SA nodal conduction disease
|
Dystrophin
(DMD and BMD)
|
Cardiac screening recommended in mutation carrier females along with affected males
Consider CMR in screening alongside ECG and echo
X linked
|
Childhood onset, progression from HNDCM to typical DCM.
BMD has an earlier and more severe disease course than DMD
|
Increased right precordial ECG R-to-S ratio
Deep Q waves in the lateral leads
Early abnormalities in strain
|
Typical skeletal muscle weakness and disability in affected
May be none in carriers
|
Wall thinning and RWMAs lateral wall, inferior, or septum
Early increased ECV, while still normal function
DCM or HCM phenotype
|
Mid-wall or subepicardial fibrosis in hypokinetic segments (inferolateral) in early disease (sometimes with preserved LV function)
Fibrosis becomes extensive or transmural in advanced stages
|
Conduction abnormalities and tachyarrhythmias (both SVT and VT)
Early onset in severe forms
|
Dystrophin
(EDMD and others)
|
Screening recommended
|
|
Early strain abnormalities
|
Typical skeletal muscle weakness and disability
|
DCM phenotype but no specific features
|
Fibrosis not early feature in EDMD
Septal mid-wall fibrosis early feature in limb–girdle type
|
Tachyarrhythmias and conduction abnormalities
|
Sodium channel
|
|
|
|
|
|
|
Arrhythmias
|
Filamin C
|
Cardiac screening recommended in mutation-positive or first-degree relatives of proband
Consider CMR
Autosomal
|
|
Low voltage and flat or inverted T waves inferolaterally
Early LGE with normal echo
ECG features of ACM absent in ACM phenotype
|
|
DCM or ACM (LV) phenotype
|
Subepicardial or mid-myocardial inferolateral fibrosis
“Ring-like” circumferential fibrosis, but not in all cases
|
Ventricular arrhythmias with a more marked and malignant course than average DCM cases
|
Sarcomere
|
Screening recommended early age
Exact gene determines inheritance mode and penetration
|
Severe and progressive disease if childhood onset
Mild and non-progressive if adult onset
|
Indistinguishable from others since no red flag features
|
|
DCM or HCM phenotype
|
Mid-wall, linear fibrosis may or may not be present.
|
|
Desmin
|
Screening recommended
Autosomal dominant
CMR may be considered alongside ECG and Echo
|
Variable onset and prognosis depending on exact mutation
|
|
May or may not have skeletal muscle involvement
|
RCM, ACM, or DCM phenotype
Focal wall hypertrophy early disease
Oedema in acute phase
Perfusion defect
|
Focal fibrosis by LGE (mid-wall or subepicardial) anywhere in LV (apex, anterior, septal, lateral, or inferior) or RV
Transmural scar in advanced stage disease
|
AV conduction abnormalities and ventricular arrhythmias
|
Danon
|
X-linked dominant
|
|
|
Syndromic facial features
|
HCM or DCM phenotype
|
LGE scar can be extensive and spares the mid-septum.
High ECV/T1
|
|
Titin
|
Cardiac screening recommended affected and carriers
Autosomal Dominant
|
|
|
Skeletal muscle, syndrome features
|
CHD or DCM
Thinner walls, lower LVEF
|
|
Arrhythmias
|
Mitochondrial
|
Autosomal recessive, maternally inherited or dominant
|
|
|
Neurological, endocrine, and ophthalmic
|
HCM or DCM phenotype depending on type of disease
|
Diffuse fibrosis (T1/ECV)
Extensive fibrosis only I end stage
|
Conduction abnormalities and Wolf Parkinson White syndrome
AVB prevalent in Kearns Sayre
|
7. Role of CMR in Guiding Device Therapy
In patients with DCM, the presence of LGE in CMR coupled with an LVEF ≤ 50% suggests consideration of an implantable cardiac defibrillator (ICD) for primary prevention
[72][23]. In a study of 452 patients with non-ischaemic cardiomyopathy and LVEF < 35% on optimal medical therapy who met the criteria for ICD insertion, ICD reduced all-cause mortality (HR, 0.45; 95% CI: 0.26–0.77) and cardiovascular death (HR, 0.51; 95% CI: 0.27–0.97) when LGE was present after a median follow-up period of 37.9 months
[73][24]. The DANISH-MRI study, a pre-specified sub-study of the DANISH trial (152), recruited 252 patients with non-ischemic DCM and an indication for primary-prevention ICD on optimal medical management
[74][25]. Overall prognosis in the patients with LV scarring was worse, but ICD implantation did not significantly reduce all-cause mortality in those with LV scarring compared with those without LV scarring (HR, 1.18; 95% CI: 0.59–2.38 vs. HR 1.00; 95% CI: 0.39–2.53,
p for interaction = 0.79)
[74][25].
CMR can also be used to predict and guide response to device therapy. Chalil et al.
[75][26] derived the tissue synchronization index (CMR-TSI) determined by segmental radial wall motion data, generating tissue synchronization polar maps of the LV, which is a global dyssynchrony measure. A CMR-TSI ≥ 110 ms was an independent predictor of death or unplanned hospitalisation in a cohort of 77 HF patients referred for cardiac resynchronisation therapy (CRT) after a median follow-up of 2 years
[75][26]. The presence of LGE may also guide lead placement in CRT. In 60 patients receiving CRT implantation, RV and LV leads over scarring was associated with non-response to CRT
[76][27]. LV lead positions at areas of LGE compared with viable myocardium were associated with an increased risk of CV death (HR 6.34;
p < 0.0001) or HF hospitalizations (HR 5.57;
p < 0.0001)
[77][28].