The distribution of LVH in most patients is asymmetric and often separated by areas of normal wall thickness, providing a non-contiguous pattern [14]. The anterior free wall and contiguous basal anterior ventricular septum are the most common location of LVH (Fig. 1) [13]. In a small subset of patients, the LVH is only confined to one or two LV segments, which can explain why nearly 20% of HCM patients have normal LV mass index values in CMR [13, 16]. Patients with a LV wall thickness $\ge$30 mm represent a subgroup at risk for progressive drug-refractory HF secondary to left ventricular outflow tract (LVOT) obstruction and arrhythmic sudden death [17, 18].

Fig. 1.

Asymmetric septal nonobstructive hypertrophic cardiomyopathy (HCM). Steady-state free precession (SSFP) magnetic resonance images of a patient with a maximal left ventricular wall thickness of 31 mm in the mid inferoseptal are shown (A) in horizontal long-axis, (B) in midventricular short-axis, and (C) along the left ventricular outflow tract (LVOT).

The second most common phenotype is concentric HCM, characterized by global thickening of the LV with no regional preferences and reduced LV cavity dimensions [19]. Differentiating concentric HCM from acquired causes of concentric LVH, like aortic stenosis, hypertension, or athlete’s heart, is frequently challenging. Aortic stenosis can be ruled out by echocardiography in most cases. In complex cases, to obtain a more precise measurement of the aortic valve area and the transvalvular pressure gradient, cine imaging and velocity-encoded imaging can be combined [20]. In hypertensive heart disease with LV wall hypertrophy $\ge$15 mm, the presence of an elevated indexed LV mass, the lack of both mid-wall LGE and systolic anterior motion (SAM) of the mitral valve are significant disease predictors [21]. Also, the native T1 value is an independent discriminator between HCM and hypertensive heart disease [22]. Regarding athlete’s heart, LVH rarely exceeds 16 mm, LV cavity size is normal or enlarged, and there is no evidence of diastolic dysfunction nor extensive LGE despite LV remodeling [11, 19]. Regression in LV wall thickness after a period of systemic deconditioning is expected in an athlete’s heart [15].

Other infiltrative diseases such as cardiac amyloidosis (CA), Anderson-Fabry disease (AFD), or iron overload cardiomyopathy (IOC) can also manifest as diffuse myocardial hypertrophy as discussed later in this review.

Further phenotypic variations occur, such as midventricular HCM (involvement of midseptum and free wall) and apical HCM (wall thickening confined to the apex and the distinctive “ace of spades” configuration at end-diastole in CMR) [15]. Occasionally, these patients develop a dyskinetic or akinetic apical aneurysm. These features are linked to a higher risk of arrhythmic sudden death and thromboembolic events [23, 24]. Intracardiac thrombus, if present, appears as a low-signal intensity mass distinguishable from surrounding high-intensity structures in contrast-enhanced CMR (Fig. 2) [25]. Since echocardiography can be suboptimal in evaluating the apex, CMR plays an important role depicting apical hypertrophy, apical akinesis, and small-to-moderately sized apical aneurysms [26].

Fig. 2.

Midventricular hypertrophic cardiomyopathy (HCM). Steady-state free precession (SSFP) magnetic resonance images are shown (A) in horizontal long-axis and (B) along the left ventricular outflow tract (LVOT), depicting apical aneurysm (asterisk) and millimetric intracardiac thrombus within the aneurysm (white arrows).

Even though HCM is classically a disease of the LV, several patients with HCM have RV hypertrophy and/or dysfunction; therefore, RV assessment should be included in the imaging workup of these patients [27].

Some morphological abnormalities in genetic carriers without LVH, like myocardial crypts (blood-filled fissures in the myocardium of LV), have been described as a subtle marker of HCM [28, 29]. However, recent reports have supported the hypothesis that crypts are benign anatomical variants, frequent in healthy subjects, and not connected to major adverse cardiovascular events [30].

A minority of HCM patients develop the so-called end-stage phase of the disease, with regression of hypertrophied and hypercontractile non-dilated LV to severe systolic dysfunction (ejection fraction 50%), enlarged ventricular chambers, increased end-diastolic and end-systolic volumes, defined as LV remodeling [31].

Apart from the presence of LVH, other structural manifestations of HCM include left atrial remodeling, myocardial bridging of coronary arteries, microvascular dysfunction, myocardial fibrosis, areas of LV noncompaction, LVOT obstruction, and abnormalities of papillary muscles and mitral apparatus [7].

Around 70% of patients with HCM present with LVOT obstruction (gradient $\ge$30 mmHg) at rest or with provocative maneuvers [32]. Mitral valve and subvalvular apparatus morphologic and functional abnormalities contribute to LVOT obstruction, and the diagnosis of these findings may dictate a specific management strategy [33, 34, 35]. Abnormal acceleration and misdirection of the flow toward the LVOT during systole create a drag effect pushing the mitral leaflets anteriorly, also known as SAM. It causes septal contact, flow resistance, and incompletely coaptation of anterior and posterior leaflets leading to mitral regurgitation [36]. CMR imaging using perpendicular and longitudinal cine sequences through the LVOT allows excellent depiction of the SAM and the signal void caused by high-velocity jet across the LVOT as well as mitral regurgitation (Fig. 3) [7, 15].

Fig. 3.

Asymmetric septal obstructive hypertrophic cardiomyopathy (HCM). Longitudinal cine steady-state free precession (SSFP) magnetic resonance images through left ventricular outflow tract (LVOT) at (A) end-diastole and (B) early systole depicting systolic anterior motion (SAM) of anterior mitral valve leaflet (white arrow) and thickened ventricular septum (asterisk).

Regarding structural abnormalities of the mitral valve, the most frequent is the marked elongation of both mitral leaflets [33]. Subvalvular abnormalities include hypertrophy and variation of the papillary muscle (PM) morphology (double bifid PM and anteroapical displacement of the anterolateral PM), which may also contribute to the severity of LVOT obstruction by anteriorly displacing the mitral valve plane toward the interventricular septum [37, 38].

Late gadolinium enhancement depends on varying uptake and washout patterns between normal and abnormal myocardium [39]. Following intravenous administration, gadolinium-based contrast agents accumulate in areas with expanded extracellular space due to either necrotic myocardium or scar tissue, resulting in increased signal intensity on T1-weighted gradient echo sequences [40]. This is depicted by applying an inversion recovery preparation pulse to achieve myocardial signal nulling while the scar tissue and blood remain bright [41].

The prevalence of LGE in patients with HCM is 60% [42]. It is well established that LGE is a noninvasive marker of increased risk for ventricular arrhythmias and HF progression with systolic dysfunction [42, 43, 44, 45]. The most common pattern or distribution of LGE in HCM is a punctuate and/or patchy mid-wall hyperenhancement at a noncoronary vascular distribution in the hypertrophic segments (Fig. 4), although HCM may present practically with any pattern [19]. The extent of LGE appears to have more discriminatory value than its presence, in particular when LGE is $\ge$15% of the LV mass, which demonstrated a significant increase in SCD risk [46]. However, LGE may also be found in hearts with no histopathological evidence of microvascular ischemia, mainly when circumscribed to areas of RV insertion to the ventricular septum. It is believed that these areas represent an expanded extracellular matrix created by the confluence of intersecting myofibrils [47].

Fig. 4.

Late gadolinium enhancement (LGE) in hypertrophic cardiomyopathy (HCM). (A) Short axis view with hyperenhancement (white arrows) at the right ventricular insertion to the ventricular septum. (B) Horizontal long-axis view and (C) vertical long-axis view show a patchy mid-wall hyperenhancement (white arrows) in the hypertrophic segments.

Some limitations to this technique have been reported, especially when diffuse myocardial fibrosis is present and the myocardial signal intensity is nearly isointense and in patients with subtle diffuse enhancement [48, 49]. T1 mapping is a novel and robust CMR technique, which offers quantitative measures of the myocardial signal. It creates a pixel-wise parametric map, in which each pixel reflects the absolute value of T1, coded in color (Fig. 5) [50]. Moreover, it directly measures ECV fraction from T1 values before (native T1) and after administration of gadolinium [14]. Increased native myocardial T1 values and an elevated ECV fraction were found in HCM, even in non-hypertrophic segments with preserved contraction function or in patients without LGE, suggesting that myocardial tissue remodeling may precede morphological and functional changes [49, 51].

Fig. 5.

Native T1 mapping and late gadolinium enhancement (LGE) images in hypertrophic cardiomyopathy (HCM). Native T1 mapping with high signal at (A) mid-septal wall and (B) mid inferoseptal segment (white arrow). (C,D) Corresponding phase-sensitive inversion recovery (PSIR) imaging shows late gadolinium enhancement in the same regions.

Besides diagnosis, phenotypic characterization, and planning for septal reduction procedures, CMR is also increasingly important for risk stratification and prognosis in HCM. According to the 2020 American Heart Association (AHA)/American College of Cardiology (ACC) guideline for the diagnosis and treatment of patients with HCM [8], CMR plays a crucial role in the risk stratification and subsequent decision to ICD placement. On the contrary, the ESC guideline does not endorse LGE for risk stratification purposes and continues to recommend the HCM Risk-SCD score, despite debate on the value of the latter approach [28, 52, 53]. CMR allows assessment of maximum LV wall thickness, LV ejection fraction, LV apical aneurysm, and extent of myocardial fibrosis with LGE, all of which are used in ICD decision-making [8]. Massive LVH and systolic dysfunction (ejection fraction 50%) are common indications for ICD placement, being CMR the gold standard to measure ventricular wall thickness and function [8]. Apical aneurysms, which echocardiography may fail to detect, are also better assessed with CMR [24].

As stated above, the extent of scar, imaged by LGE is associated with increased risk for potentially life-threatening ventricular arrhythmias and cardiac death [46, 54, 55]. Based on this evidence, the AHA/ACC guideline recommends ICD placement in the presence of extensive LGE ($\ge$15% of LV mass) (2b recommendation) [8]. Further randomized trials are needed to confirm and assess the value of CMR for SCD risk stratification in HCM.

Systemic amyloidosis comprises a group of diseases caused by the deposition of insoluble amyloid fibrils in the extracellular space of tissues and organs. This deposition ultimately leads to progressive organ failure [57]. CA is an infiltrative cardiomyopathy characterized by increased biventricular wall thickening, restrictive LV filling, and, frequently, a non-dilated LV cavity with preserved or mildly depressed LV systolic function [58]. Previously thought to be rare, CA is now understood to be underdiagnosed [59].

Most cases of cardiac involvement occur in two types of amyloidosis: light chain immunoglobulin (AL) and transthyretin amyloidosis (ATTR). ATTR is further classified into the wild-type ($>$90% of cases) and hereditary (10% of cases) [4]. The main treatment options in AL-CA are chemotherapy or autologous stem-cell transplant, aimed at the underlying hematological condition. Treatment of ATTR is mainly based on stabilization and reduction of transthyretin production. Liver and/or cardiac transplantation can be considered only in end-stage disease of familial ATTR [4]. Tafamidis showed reduction of all-cause mortality and cardiovascular hospitalizations in hereditary and wild-type ATTR, mostly in those patients with New York Heart Association (NYHA) class I and II at baseline [60].

Echocardiography is often the first imaging modality performed in patients presenting with HF. CA’s echocardiographic features are frequently present in patients with advanced disease but absent earlier [61]. In recent years, the application of CMR in CA has increased, since it provides information regarding the presence, distribution, and location of hypertrophy, visualization of cardiac amyloid infiltration with LGE imaging, and measurement of cardiac amyloid burden with T1 mapping and ECV [62, 63, 64]. Additionally, CMR findings appear to have prognostic significance [65]. Technical development of new CMR sequences has contributed to CA awareness and recognition [1]. Furthermore, the emergence of new treatment strategies for CA, such as tafamidis, entails a need to establish the diagnosis during the earlier stages of the disease to prolong survival and improve outcomes.

3.1.2 Late Gadolinium Enhancement

In CA, there is an expansion of the extracellular space due to increased amyloid deposition, leading to high gadolinium concentrations in the myocardium and prolonged impermeability, resulting in LGE [70, 71]. LGE imaging provides highly characteristic findings with high diagnostic accuracy [72]. CA presents a distinctive pattern of LGE with global subendocardial or transmural distribution, associated with dark blood-pool due to abnormal gadolinium kinetics [73]. Three LGE patterns were identified in CA: none, subendocardial and transmural (Fig. 6). These patterns are a continuum and correlate with the extent of myocardial infiltration [74]. Transmural LGE is associated with the highest amyloid infiltration in advanced CA and has shown to carry the most adverse prognosis [74]. As opposed to myocardial infarction, LGE in CA is not limited to characteristic vascular territories nor is sharply defined [75]. RV LGE is also frequently present [67]. LGE is present in most cases of CA (100% LGE in LV and 96% in RV), being more common in ATTR (vs. AL-CA), but does not differentiate between CA subtypes [67].

Fig. 6.

Cardiac amyloidosis. Steady-state free precession (SSFP) magnetic resonance images at (A) end-diastolic horizontal long-axis and (B) midventricular short-axis showing concentric thickening of left ventricular walls (septal wall measuring 26 mm), right ventricular thickening, biatrial dilatation, interatrial septal thickening (9 mm), and pericardial effusion. The corresponding late gadolinium enhancement (LGE) images (C and D) show extensive amyloid infiltration with generalized gadolinium uptake.

As previously stated, traditional LGE imaging is a comparison technique, based on differences between normal and abnormal myocardium. The null inversion time is chosen based on what the operator considers normal myocardium. By convention, the areas with the most contrast concentration should have a brighter display, and normal myocardium should be presented as black or “nulled” on LGE imaging. Nulling the normal myocardium in diffuse infiltrative diseases such as CA can be challenging given the lack of normal myocardium areas for comparison. The myocardial signal intensity may be globally “nulled,” thus looking like normal tissue [1, 67]. It is a clear indication of the presence of amyloidosis when achieving the usual contrasts in LGE imaging is a challenging task [75]. More recent techniques, such as phase-sensitive inversion recovery (PSIR), are less operator-dependent and more precise in defining the degree of cardiac involvement. A combination of LGE with T1 mapping is another alternative to address this issue [1, 72, 74].

Other limitations of LGE include the use of gadolinium which is relatively contraindicated in patients with severe renal dysfunction, common in patients with amyloid disease, and lack of quantitative results, limiting the ability to track changes over time. However, T1 mapping can overcome these limitations and potentially detect amyloid infiltration at an earlier stage of the disease than LGE [61, 76].

Iron overload is a systemic process characterized by an increased level of plasma iron and iron accumulation in parenchymal cells [82]. Its detection is imperative owing to multisystem manifestation such as myocardial disease, type 2 diabetes mellitus, or liver cirrhosis with increased risk of hepatocellular carcinoma [82]. IOC describes the different forms of cardiac dysfunction in the setting of increased gastrointestinal iron absorption (e.g., hereditary hemochromatosis) or as a result of transfusion-dependent anemias (e.g., thalassemia major and sickle cell disease) [83]. Early stages of IOC present as restrictive cardiomyopathy with diastolic LV dysfunction. If iron overload persists and no proper therapy is started, dilated cardiomyopathy with chamber dilatation and impaired systolic function may develop [84].

Cardiac magnetic resonance is the only non-invasive technique for quantifying myocardial iron overload (MIO). It is considered the standard of care in diagnosing and monitoring iron overload diseases [85, 86]. The presence of myocardial iron deposits causes magnetic field inhomogeneity and accelerates the relaxation times [87, 88]. As the myocardial iron concentration increases, the signal intensity in T1-, T2- and T2*- weighted images decreases [89, 90]. CMR gradient-echo T2* relaxation time is currently the mainstay of IOC diagnosis, and the measure in a single slice of a full-thickness region of interest in the mid-ventricular septum is highly representative of global myocardial iron distribution [83, 91]. A multislice segmental approach is also validated and allows the identification of different patterns of iron distribution (homogenous MIO, heterogeneous MIO, and no MIO) that correlate with cardiac complications [92, 93]. A cardiac T2* relaxation time of 20 ms has been proposed as a cutoff value for diagnosing MIO and substantially better predictor for HF (T2* 10 ms) and arrhythmias (T2* 20 ms) than liver T2* or serum ferritin [94]. This non-invasive modality is also helpful in identifying the proper moment to initiate therapy and monitor therapy response by serial imaging, which has a significant positive impact on the survival of these patients [91, 95].

Anderson-Fabry disease is an X-linked congenital error of the glycosphingolipid metabolic pathway due to the absent or deficient activity of $\alpha$-galactosidase A ($\alpha$-Gal A) enzyme and is the most prevalent lysosomal storage disorder [96]. It results in a progressive glycosphingolipid accumulation within lysosomes in multiple cell types, including capillary endothelial, renal, nerve, and cardiac cells [97]. In adulthood, progressive cardiovascular involvement and renal failure account for most deaths associated with the disease [98]. Over 60% of patients present with cardiovascular signs such as arrhythmias and conduction abnormalities, aortic and mitral valves deformities, LV mild-to-moderate concentric hypertrophy, and HF [99, 100]. Since treatment with enzyme replacement therapy has been shown to reverse or slow disease progression, early diagnosis is critical, and CMR plays an important role in this field [100].

Non-invasive detection and monitoring of cardiovascular manifestations of AFD have focused on LVH, the classical morphological abnormality of the disease, which can mimic the HCM. In fact, the prevalence of AFD gene mutations in patients with unexplained LVH is 0.5% [101]. LV concentric thickening is the most common morphological manifestation of AFD, but the spectrum also includes apical and asymmetric septal hypertrophy [102]. CMR imaging allows improved visualization of LV geometry and identification of hypertrophy, particularly in segments not well characterized by echocardiography, and quantification of LV mass index [103, 104]. Due to its higher precision, CMR has also been proposed for LVH follow-up and response monitoring of patients treated with enzyme replacement therapy [100, 104].

LGE is identified in up to 50% of AFD patients in the basal and/or mid inferolateral wall, with a mid-myocardial pattern [99, 105, 106]. The presence and extent of LGE are associated with a greater risk of adverse cardiac events, such as ventricular arrhythmias and SCD [102, 107]. Unfortunately, it is known that LGE imaging is less sensitive in the setting of a diffuse process, and the advanced renal dysfunction of some of these patients often precludes the use of contrast [40].

Tissue characterization using T1 mapping is a powerful tool in diagnosing cardiac involvement in AFD since the sphingolipid deposition in the myocardium provides a low native T1 value allowing distinction from other causes of LVH, which generally have an elevated T1 [108, 109]. It has been reported that in AFD without LVH or fibrosis, native T1 values may still be lower than normal, suggesting that it could be used as an early marker of cardiac involvement and a prompt the initiation of ERT to achieve long-term improvements [108, 110]. Furthermore, compared with healthy controls, these patients present normal myocardial ECV, suggesting no diffuse fibrosis [106, 109].

In AC, CMR may reveal global and regional ventricular dilation, global ventricular dysfunction, and regional wall motion abnormalities (Fig. 7). Regional wall motion abnormalities occur preferentially in the subtricuspid region. The “accordion sign”, that represents a focal “crinkling” of the myocardium, is an example of these abnormalities, which is caused by a small region with dyssynchronous contraction [112].

Fig. 7.

Arrhythmogenic cardiomyopathy with isolated right ventricular involvement. Steady-state free precession (SSFP) magnetic resonance images at (A) end-diastolic and (B) end-systolic horizontal long-axis shows dilated right ventricle with mild-moderately depressed systolic function, akinesia/mild dyskinesia of the basal segment of the right ventricle free wall; and mild right atrial dilatation. Left ventricle morphology and systolic function are within the normal range. 

Nearly all patients with AC (96% of patients) present an abnormal RV on CMR [116]. The most frequent RV abnormalities are basal inferior wall dyskinesia (in 94% of cases) and basal anterior wall dyskinesia (87% of cases). Although clinical studies have focused predominantly on RV abnormalities, LV involvement is more common than previously thought. LV involvement occurs in 52% of patients, which has led to the reconsideration of the initial dysplasia triangle (posterior lateral wall of the LV, subtricuspid, and anterior wall of the RV) [116]. Clinically demonstrable RV involvement in biventricular variants of AC is an important additional criterion to distinguish AC from dilated cardiomyopathy [111].

Cardiac magnetic resonance provides a rigorous quantitative analysis of RV volumes, but significant interobserver variability in the RV free wall segmental contraction analysis and interpretation of qualitative findings have been reported [117]. RV’s standardized and regional biventricular analysis using cine steady-state free precession (SSFP) imaging that includes all RV segments may help overcome these limitations [118].

Directly assessing tissue composition of thin RV wall also remains a technical challenge on MRI. Some technical advances may be an encouraging approach to decrease variability in assessing RV function, including semi-automated myocardial deformation quantification, particularly feature tracking [119].

The pathologic hallmark of AC is the fibrofatty replacement of the myocardial tissue. CMR allows tissue characterization (fibrosis, fatty infiltration, and fibrofatty scar), but LGE is not included in the 2010 ITF criteria [113]. The visualization of fat and LGE by CMR as a diagnostic hallmark faces several problems which withheld their use: intramyocardial fat occurs in normal hearts and is not specific for AC without associated functional abnormalities [120]; detection of fibrosis in the RV is nonspecific and frequently difficulted by the thin RV wall, which in AC may be even more pronounced, making the LGE technique less dependable than for the LV [121]; distinguishing fat from fibrosis by LGE sequences is challenging; and finally, LV LGE is nonspecific and has a broad range of differential diagnosis (AC, sarcoidosis, myocarditis, CA, HCM, and dilated cardiomyopathy). The absence of RV myocardial fat in CMR does not exclude the diagnosis of AC, as seen in pediatric patients [117]. The presence of fat should not be interpreted alone [118]. Given the limitations of CMR at evaluation of fat and fibrosis, AC diagnosis is mainly based on establishing the consequence of intramyocardial fibrofatty replacement, such as static morphological abnormalities and regional RV wall motion abnormalities [118].

While these limitations exist, LGE may still be useful in AC evaluation. Diagnostic accuracy for AC was best (98%) when pre-/post-contrast signal abnormalities (including LV fat infiltration and LGE) were considered together with wall motion alterations [122]. LGE may also increase the sensitivity in the diagnosis of biventricular AC by demonstrating segmental subepicardial LGE in the LV wall, which in LV-dominant phenotypic variants of AC may be the only imaging feature [111].

In patients with suspected AC, a normal CMR had a negative predictive value of 99% for major clinical events in a 4.3-year follow-up, potentially becoming a “rule-out” test for AC [111, 123]. Furthermore, no events were observed among patients with a normal CMR during a 6-year follow-up [116]. On the other hand, an abnormal CMR was an independent predictor of events with a cumulative effect of several anomalies (including morphology, wall motion abnormalities, and fat/fibrosis) [123].

CMR has also been useful in the risk stratification for arrhythmic events. In a study with AC mutation carriers, sustained arrhythmias during follow-up were associated with structural abnormalities on CMR [116]. Another study showed that CMR was an independent predictor of ventricular arrhythmias [123]. In patients with AC and ventricular arrhythmias, regional wall-strain assessed on cine-CMR reliably predicts arrhythmogenic ventricular tachycardia-substrate. This technique permits safe and improved diagnostic accuracy in AC, preventing the need for invasive procedures, and facilitates the planning of ventricular tachycardia ablation procedures [124].

Table 1 summarizes the main CMR findings and markers of prognosis in non-ischemic cardiomyopathies.