Phenotypes of hypertrophic cardiomyopathy. An illustrative review of MRI findings

Myocardial clefts or crypts are congenital abnormalities related to myocardial fibre or fascicle disarray and have been described in healthy volunteers, as well as those with classic HCM (< 5%) [8, 20, 21].

These crypts are V- or U-shaped blood-filled myocardial invaginations perpendicular to the endocardial LV edge that penetrate more than 50% of the compact myocardium at end-diastole and collapse at end-systole (Fig. 1) [10, 21, 22]. Usually, myocardial crypts are multiple and located in the basal inferoseptal and inferior LV wall at the sites of right ventricle (RV) insertion [9, 22].

These myocardial crypts should be differentiated from trabeculations characteristic of LV noncompaction, congenital LV diverticula and aneurysms.

Unlike crypts, trabeculations of LV noncompaction are parallel to the endocardial border and do not penetrate the normal compacted myocardium [22].

Congenital LV diverticula are characterised by an outpouching that contains endocardium, myocardium, and pericardium, and they are a narrow connection to the cavity and collapse at end-systole. The primary diagnostic feature to differentiate diverticula from crypts is that the congenital diverticula extend outside the epicardial border while crypts remain confined inside the myocardial margin [23].

Congenital LV aneurysm represents an outpouching containing endocardium, epicardium and thinned myocardium that shows akinetic or dyskinetic motion. A thinned myocardium shows LGE without pericardial enhancement [23, 24].

Several echocardiographic studies have reported that the mitral valve is elongated in HCM, particularly in the obstructive form of the disease [25, 26].

Cardiac MR studies have shown elongated anterior and posterior mitral valve leaflets in patients with non-genotyped HCM compared to control subjects [11] and elongated anterior mitral valve leaflets in subjects with HCM-causing mutations without LV hypertrophy (Fig. 2) [11, 12, 27]. These studies suggest that anterior mitral valve leaflet elongation in young preclinical HCM patients without other phenotypic expressions of HCM represents an independent and primary component of HCM disease expression independent of age, LV thickness or the presence of LV outflow tract obstruction.

Other recent studies suggest that mitral valve elongation does not constitute the primary phenotypic expression of HCM and that the elongation of mitral valve leaflets in HCM seems to be associated with body size and left ventricular remodelling [28].

Left ventricular non-compaction is a genetically heterogeneous disorder and shares genetic mutations with HCM. Both entities can coexist and have been described as being associated with mutations in different sarcomeric genes [29].

Abnormal LV apical myocardial trabeculations also can be detected in HCM sarcomere gene mutation carriers without evident hypertrophy (Fig. 3). Both diseases can have the same mutation of myocardial sarcomere, and they are indistinguishable [12]. The mechanism for these abnormal trabeculations remains unclear, similar to myocardial crypts, they could represent the persistence of the embryological form into adulthood [10, 12].

The development of fibrosis is an important step in the LV remodelling process of HCM. Recent studies have demonstrated that HCM genotype-positive and phenotype-negative patients have increased levels of biomarkers of collagen deposition. This finding indicates that increased levels of profibrotic markers are also found in patients without ventricular hypertrophy, suggesting that hypertrophy is preceded by the fibrosis process. In genotype-positive, phenotype-negative HCM subjects, several cardiac MR studies found that the myocardial mid-wall LGE consistent with fibrosis was attributed to episodes of myocardial ischaemia due to coronary microcirculation remodelling [18]. Other studies have shown that both native T1 values and extracellular volume (ECV) fraction obtained after gadolinium administration can detect diffuse interstitial fibrosis and are elevated in regions without LGE in subjects with a mutation known to cause HCM without sufficient phenotypic expression to diagnose the disease based on current imaging criteria [17].

Native T1 mapping and the ECV fraction could be used at the initial phase of HCM for the early detection of interstitial diffuse myocardial fibrosis in these patients and to monitor disease progression, as well as to evaluate novel disease-modifying therapy, targeting interstitial fibrosis [8, 17].

The diagnosis of mild HCM in young competitive athletes may be challenging when the LV wall thickness is between 13 and 15 mm (or 12 and 13 mm in women), which identifies the “grey-zone” of overlap between the physiological adaptations to training and mild phenotypic expression of the disease. In such instances, the diagnosis can often be resolved by applying noninvasive markers. HCM is favoured with the identification of a pathogenic sarcomere mutation or a family history of HCM, an abnormal LV filling/relaxation, small/normal size LV, myocardial crypts or LGE on contrast cardiac MR and response to detraining [30, 31].

Recent studies have shown that native T1 values and myocardial ECV by T1 mapping can be used in the differential diagnosis between HCM and an athlete’s heart. While the ECV fraction increases with increasing LV hypertrophy in HCM (due to extracellular matrix expansion and myocardial disarray), the ECV fraction reduces in athletes with an increasing wall thickness (due to an increase in the healthy myocardium by cellular hypertrophy) [32].

The distribution of myocardial hypertrophy is extremely variable. Cardiac MRI is particularly useful for characterising the location and extent of LV hypertrophy (Fig. 4), offering a superior visualisation and a higher diagnostic accuracy compared with 2D echocardiography, particularly if the involved segments are the basal anterolateral free wall or the apex [34, 35].

In HCM, reduced ventricular volumes are frequent and the hyperkinetic appearance of systolic contraction translates into a normal or supernormal ejection fraction (EF) (> 65%) until the end-stage of the disease [8]. EF is, therefore, considered inadequate to evaluate the indication for medical treatment and cardiac transplantation [2].

Diastolic dysfunction is one of the early hallmark features of HCM, and it is related to myocardial disarray and fibrosis, even in the absence of LV hypertrophy [8, 38].

Doppler parameters are accurate in evaluating diastolic function in HCM. Accurate measurements of transmitral and transpulmonary vein flow parameters using phase-contrast cardiac MRI are lower than those echocardiography, which is probably due to the lower temporal resolution of the breath-hold phase-contrast sequence. A new semi-automated, three-dimensional model-based analysis of the LV fillings curve showed a similar agreement with Doppler parameters [39].

In addition, myocardial tagging [39, 40] enables the quantification of cardiac strain evolution during late diastole, which can be used to assess diastolic myocardial dysfunction [40]. However, despite recent advances of myocardial tagging methods, the fading of taglines due to T1 relaxation is a common inherent limitation of all existing methods, which also requires prolonged imaging acquisition and post-processing times [39].

Abnormalities in the morphology and function of the mitral valve apparatus are now considered a phenotypic expression of HCM that is independent of the hypertrophy distribution [11].

Mitral valve leaflets are increased in length in many HCM patients, and substantially elongated mitral valve leaflets are an important determinant of LVOT obstruction in some patients (Fig. 5) [8, 41].

Other abnormalities include hypertrophy of the papillary muscle heads, an increased number of papillary muscles, anterior and apical displacement of the papillary muscles, direct insertion of the anterolateral papillary muscle to the ventricular aspect of anterior mitral valve leaflet and LV accessory muscle bundles extending from the apex to the mid-ventricular or basal levels of the LV anterior or septal walls (Fig. 6) [42].

LVOT obstruction is present in 70% of classic HCM phenotype patients and is associated with an increased risk for heart failure-related death [2, 8]. The mechanisms responsible for LVOT obstruction are related to complex anatomical relationships between the basal septum, LVOT, mitral valve, and papillary muscles [8, 41].

An elongated anterior mitral leaflet (Fig. 5a) is considered an important determination of LVOT obstruction, particularly in patients in whom the mitral leaflet length exceeds the transverse dimension of the outflow tract at end-systole 2-fold [3, 8, 11]. Systolic anterior motion (SAM) of the mitral valve refers to the paradoxical movement of the anterior leaflet and/or chordae toward the interventricular septum during systole. The precise mechanism of SAM in HCM remains unclear. Many investigators hypothesise that anterior mitral leaflet elongation and thickening and hypertrophy with anterior displacement of the papillary muscles contribute to the development of SAM. The resultant flow forces during systolic anterior motion pulling the leaflets into the LVOT in mid- and late-systole also result in reduced leaflet coaptation and a posteriorly directed mitral regurgitant jet (Fig. 5b) [25, 26].

Echocardiography remains the technique of choice to evaluate LVOT obstruction. However, cardiac MR-derived planimetry at the rest of the LVOT area during systole can identify obstructive HCM. A cutoff value of 2.7 cm² has been reported as having an accuracy of 100% to differentiate obstructive from non-obstructive HCM [43]. Moreover, cine MRI in a long-axis cine view can provide a more accurate evaluation of the mechanism of outflow tract obstruction, demonstrating turbulent flow generated by systolic movement of the anterior mitral leaflet, chordae, and papillary muscle toward the interventricular septum [41] (Fig. 6).

The left atrium (LA) is often enlarged in HCM patients, and its size provides important prognostic information. An enlarged LA is related to an increased morbidity and mortality in cardiovascular patients, and it is considered a marker for risk for SCD and potentially lethal arrhythmic events among HCM patients [44‐46].

Volumetric MR quantification is the best approach to assess the LA size, but it may be too time-consuming. An area ≧ 15 cm²/m² and a transverse diameter ≧ 2.8 cm/m² in the four-chamber view are valuable alternatives to identify LA enlargement [47].

The cause of LA enlargement is multifactorial, but the most common mechanisms are SAM-related mitral regurgitation and elevated LV filling pressures [48]. Mild or moderate LA dilatation is common and is related to LV diastolic dysfunction [44]. Severe LA dilattion is common when mitral regurgitation is associated with LVOT obstruction [8, 41].

Atrial fibrillation is the most common arrhythmia in HCM (20%) and LA remodelling is the strongest predictor of atrial fibrillation independent of age and functional class [48].

The hallmarks of HCM include LV hypertrophy, fibrosis and microvascular ischaemia, and whether there is a causative link between these features has not been determined to date [49].

Microvascular ischaemia in HCM, which is attributed to increased oxygen requirements due to hypertrophy, impaired ventricular relaxation, microvascular dysfunction of intramyocardial arterioles, and LVOT obstruction, is associated with chest pain, clinical deterioration, diastolic dysfunction, and an adverse prognosis [50, 51]. It is most pronounced in the hypertrophied segments and in gene-positive HCM patients [52].

Although microvascular dysfunction may precede clinical deterioration for several years and is present in all stages of HCM, it is not associated with evidence of permanent ischaemic damage and replacement fibrosis [16, 19].

The presence and severity of myocardial ischaemia and reduced myocardial blood flow can be diagnosed using non-invasive imaging modalities, including positron emission tomography (PET) [51], single photon emission computed tomography (SPECT) [53] and first-pass perfusion cardiac MR sequences [8, 54]. Myocardial perfusion defects are usually found at subendocardium (most common) or at mid-wall, and it correlates with the areas of maximal wall thickness [54].

Myocardial fibrosis in HCM is histopathologically diffuse in both the replacement and interstitial forms, and it is assumed to be a substrate for tachyarrhythmias and SCD. LGE essentially shows the focal distribution of fibrosis on cardiac MR, is present in less than half of those with the classic HCM phenotype and occupies a small percentage of the LV mass (median value: 2%) [16, 55], suggesting that collagen deposition at this stage reflects an exaggerated activation of the matrix rather than a reparative process [56].

Myocardial LGE can present in a wide variety of patterns, although it never corresponds to a coronary vascular distribution (Figs. 4d and 7) [15, 57, 58]. It is usually localised in segments with the maximum LV wall thickness. Isolated or multiple patchy LGE at mid-wall (most common) (Fig. 7a, b) and along the RV insertion points on the septum (Fig. 7c) are the usual patterns of the classic HCM phenotype. LGE is unusual in non-hypertrophied segments [8, 59].

Histologically, fibrosis is often global, or diffuse, and often undetectable by standard LGE pulse sequences. T1 mapping is a novel and promising cardiac MR technique that provides an assessment of the total extent of the expanded extracellular space rather than the detection of regional areas of myocardial fibrosis identified by traditional LGE imaging. Patients with HCM have substantially higher-than-normal native T1 values and extracellular volume fractions, even in areas without LGE [60, 61] (Fig. 8), and reduced post-contrast myocardial T1 values consistent with the presence of diffuse interstitial fibrosis (Fig. 9). The imperfect concordance between native T1 mapping and LGE imaging for detecting focal replacement fibrosis may be explained by the distinct significance of both imaging methods because LGE (as well as postcontrast T1 mapping) reflects only the extracellular space, whereas native T1 mapping reflects a composite of both the intra- and extracellular compartments [62].

Myocardial hypertrophy can occur as a result of a variety of cardiac and systemic diseases, such as hypertensive heart disease, amyloidosis and Anderson-Fabry disease [8, 63‐65].

Systemic hypertension can lead to concentric LV hypertrophy, and discriminating between HCM and hypertensive heart disease is a frequent dilemma in clinical practice. Unlike HCM, the LV wall thickness in hypertensive heart disease rarely exceeds 16 mm, and usually, there is no LGE [63].

The differentiation of Anderson-Fabry disease from HCM remains a clinical challenge, given that there is overlap in their imaging findings. Recent studies have shown that native T1 values are significantly lower in patients with Anderson-Fabry disease compared with HCM patients (Fig. 10). This is probably the most sensitive and specific cardiac MR feature to differentiate between both diseases [64].

Hypertrophy in amyloidosis is usually moderate and relatively concentric, involving both ventricles, atria, inter-atrial septum and valves. Amyloidosis displays a specific pattern of global or segmental subendocardial LGE and frequently atrial enhancement as well [8]. In addition, the typical markedly elevated native T1 values contribute to differentiating amyloidosis from HCM [65].

Although SCD is infrequent in classic HCM (0.5–1%/year), the risk stratification for SCD should be performed in all patients with HCM due to the efficacy of the ICD in preventing SCD in this disease [1, 2, 44].

The major risk factors for SCD in HCM include a family history of SCD, age (< 40 years), unexplained syncope, an abnormal blood pressure response to exercise, nonsustained ventricular tachycardia, a maximal LV wall thickness > 30 mm, a LVOT gradient ≥ 30 mmHg at rest and the LA diameter [1, 2, 44]. Although current risk factor models, such as the American algorithm [1] and European Society of Cardiology HCM Risk-SCD calculator [46] are highly effective in identifying many patients with HCM who will benefit from ICD, some patients without conventional risk markers nevertheless remain at risk for SCD [66].

Currently, neither American guidelines nor the HCM Risk-SCD calculator incorporate LGE [1, 44]. However, multiple studies have demonstrated that LGE is correlated with areas of myocardial fibrosis as the arrhythmogenic substrate for ventricular tachyarrhythmias [27, 46, 67] and might be considered a potential risk modifier, particularly when it is seen in extensive areas of the myocardium (LGE ≥ 15% of the LV myocardium) [13, 14, 66, 68]. It has been postulated that T1 mapping may prove to be superior to LGE for risk stratification in HCM. However, to date, there has been no link between T1 mapping and cardiovascular outcomes within HCM [69].

Another imaging feature with a potential impact on the current risk stratification models could be the myocardial high T2 signal intensity areas. These areas have been related to markers of adverse disease progression, such as LGE, elevated troponin and non-sustained ventricular tachycardia [70] and they have been recently implicated as a potential tool for the prediction of SCD [71].