Introduction

Advanced imaging techniques have allowed not only more precise assessment for the severity of cardiovascular diseases, but also earlier detection of changes in the cardiac structure and function. Cardiac magnetic resonance (CMR) imaging has progressively emerged as a diagnostic tool, although it remains to be completely resolved regarding a lack of availability, portability, and cost-effectiveness. On the other hand, because of its ability to provide real-time images combined with high spatial and temporal resolution, echocardiography is invaluable for assessing chamber size, wall thickness, and systolic and diastolic function, which allows diagnostic, therapeutic, and prognostic considerations in various clinical conditions. Among patients with known or suspected cardiomyopathy in which myocardial ischemia is hardly involved as pathogenesis, the advent of echocardiographic tissue imaging (ETI), including tissue Doppler, strain, and strain rate, has provided information on the aforementioned matters in addition to better understanding of global and regional myocardial mechanics.

In this review, we discuss usefulness of ETI and other ultrasound techniques (Table 1) for detecting clinical and subclinical characteristics of cardiomyopathies, focusing on non-ischemic dilated cardiomyopathy (DCM), hypertrophic cardiomyopathy (HCM), and cardiac amyloidosis (CA), all of which have the myocardium diffusely affected. We also refer to some about “postsystolic shortening (PSS),” which is a unique phenomenon of the myocardial mechanics occurring even in a non-ischemic etiology, and considered to be important in defining cardiac function.

Table 1 Echocardiographic techniques and indices for assessing myocardial characteristics and function appearing in this review
Full size table

Dilated cardiomyopathy

DCM is characterized by an increased chamber size and diffusely compromised contractile function of the ventricle. Impairment of diastolic function is also common. These abnormalities can be expressed by most of the findings on ETI. In DCM, tissue Doppler imaging (TDI)-derived indices of systolic (s′), early diastolic (e′), and late diastolic (a′) annular velocities, all reflective of longitudinal fiber function, are attenuated. E/e′ serves to substitute for left ventricular (LV) filling pressure and thus to guide treatment strategies especially in patients with reduced LV ejection fraction [1]. Among patients with ischemic and non-ischemic cardiomyopathy, depression of s′ (< 3.0 cm/s) and e′ (< 3.0 cm/s) was reported to be associated with worsening of outcome [2].

At the advanced stage of disease process, TDI components can be recorded chaotic in both amplitude and timing (Fig. 1). Such a finding seems to indicate a condition of end-stage DCM involved with extensive structural and electrical derangement of the myocardium, although more research is needed to determine its clinical significance.

Fig. 1
figure 1

The apical 4-chamber view overlaid with TDI (a) and the mitral annular velocity profile at the septal side (b) from a patient with DCM

Full size image

Strain and strain rate techniques, based on both TDI and speckle tracking echocardiography (STE), provide measures on myocardial deformation. Generally, compared with normal individuals, DCM patients have decreased global and regional strain with a large variability of strain values between myocardial segments [3, 4]. DCM is associated with high prevalence of myocardial segments with postsystolic shortening (PSS) [5], the details of which are explained later. Such functional damages of the myocardium were found to improve with appropriate heart failure treatment [5, 6].

Impaired myocardial deformation, indicated by a decreased global longitudinal strain (GLS), has been shown to be associated with incremental prognostic information to LV ejection fraction in DCM patients [7]. GLS plays an important role in detecting subclinical myocardial damage and preventing progression to overt symptomatic heart failure [8]. Particularly, in patients undergoing chemotherapy, the measurement of GLS is recommended in the standard echocardiography protocols [9].

Three-dimensional speckle tracking echocardiography

Two-dimensional STE (2D-STE) has several limitations such as beat-by-beat alteration of cardiac function particularly in case of unstable cardiac conditions and disappearance of some speckles within the region of interest during a cardiac cycle referring to as “out-of-plane phenomenon” [10]. Because of the full-volume data acquisition, 3D-STE is expected to cover aforementioned limitations on 2D-STE, and thus has the potential for more detailed assessment of global and regional myocardial function than 2D-STE. 3D-STE also allows easier data acquisition and analysis for complicated measures of LV function such as dyssynchrony and twisting, and even for those of the right ventricle [10,11,12].

In a small number of children with DCM, Balasubramanian et al. showed that 3D-STE was superior to 2D-STE in terms of reproducibility measurement [13]. Similar to 2D-STE, 3D-STE provides DCM patients with lower strain values for three directions of the strain (longitudinal, circumferential, and radial) compared with healthy individuals [14]. With anaysis of torsional mechanics, 3D-STE was shown to discriminate DCM patients with wide QRS complex from those with narrow QRS complex [15].

Figure 2 shows a 3D-STE image from a patient with DCM, where global and regional strain values are decreased overall, with some of the myocardial segments associated with PSS.

Fig. 2
figure 2

Parametric images of multi-planar reconstruction by 3D-STE from a patient with DCM

Full size image

Postsystolic shortening

PSS is a phenomenon of myocardial contraction that occurs or sustains after the end of the ejection phase [16]. It can be assessed by TDI, strain, and strain rate not only in patients with ischemic myocardium but also in those with non-ischemic etiologies and even in normal individuals [5, 17,18,19]. With strain analysis, PSS is defined when the strain after end-systole (after aortic valve closure) is greater than that at end-systole (Fig. 3). Both clinical and experimental studies found that PSS was correlated with LV diastolic function and diminished with appropriate interventions [5, 18, 20]. Figure 4 shows bull’s eye displays of postsystolic index mapping before and after effective medical treatment in a DCM patient.

Fig. 3
figure 3

Schematic presentation of longitudinal strain profile with PSS (gray curve) and normal strain (black curve). AVC indicates aortic valve closure

Full size image
Fig. 4
figure 4

Bull’s eye displays of postsystolic index mapping before (a) and after (b) symptomatic and echocardiographic improvement with medical treatment in a patient with DCM

Full size image

The underlying mechanism of PSS remains to be elucidated. Claus et al. and later Bijnens et al. proposed that occurrence of myocardial deformation after aortic valve closure is caused by an imbalance in thickness between the segment that would develop PSS and the surrounding segments [21, 22]. PSS is initiated by a sudden decrease in LV pressure, which changes the balance between the forces acting locally on a myocardial segment and leads to thickening while the surrounding segments thin. It is also suggested that under conditions of preserved myocardial elasticity (not scarred), the amount of PSS is related to the amount of imbalance between the myocardial contractility of a segment and that of the surrounding segments [21].

“Pre-ejection stretch,” a pattern of myocardial deformation occurring even in a normal heart, may contribute to development of PSS [23] (Fig. 5). This phenomenon is considered to represent a precondition to optimize initial tension of the myocardium for the subsequent contraction and to play a pivotal role in determining synchronous myocardial motion [24]. In more severe condition of impaired LV systolic function, however, appearance and behavior of pre-systolic stretch in relation to PSS does not seem to be coherent (Fig. 1). Further work is needed to clarify significance of pre-ejection stretch and its relevance to PSS in various clinical settings.

Fig. 5
figure 5

The mitral annular velocity profile by TDI (a) and the strain curves for 17 myocardial segments with the bull’s eye display (b) from a patient with DCM. Arrows indicate strain curves with PSS and arrow-heads those with pre-systolic stretch

Full size image

Hypertrophic cardiomyopathy

During systole, LV contraction is initiated by shortening of longitudinal and circumferential myocardial fibers and by thickening of radial myocardial fibers [25, 26]. In HCM, besides reduced longitudinal shortening, circumferential fiber shortening is preserved or even exaggerated, reportedly shown by both TDI and STE [27, 28], perhaps in order to counterbalance the reduction of longitudinal fiber shortening. This, in HCM patients, indicates that the measurement of LV ejection fraction is largely dependent on the quantification of circumferential fiber shortening, and thus attention should be paid not to underestimate the degree of systolic impairment. Longitudinal shortening may play an important role in determining early myopathic process based on the finding that longitudinal fiber myocardial deformation was linked to myocardial disarray [29, 30].

Among asymptomatic or mildly symptomatic patients with HCM, Kitaoka et al. found that those who experienced cardiovascular events including death and heart failure admission had low e′ and a′ compared with those who did not [31]. TDI-derived s′ is reduced despite normal or supernormal LV ejection fraction, even for the myocardial segments not demonstrating overt hypertrophy [32]. These findings may relate to the presence of interstitial fibrosis, cardiomyocyte hypertrophy, and impaired coronary microcirculation [33, 34].

There has been an intriguing observation that TDI-derived s′ and e′ enable to identify subclinical disease involved with adolescent subjects from HCM families who have been yet to demonstrate phenotypic hypertrophy [35]. This observation implies that repeated imaging, hopefully at yearly interval, is essential to find out later development of LV hypertrophy. In contrast, a study from Netherland found that both s′ and e′ were not so sufficiently sensitive for screening family members in the HCM population [36], although a large-scale study would be warranted to solve such discrepancies between the studies.

The left atrial (LA) longitudinal strain represents an excellent parameter for analyzing LA function in several conditions [37]. Aly et al. demonstrated that along with information on LA volumes, decreased peak atrial longitudinal strain (PALS) was useful to detect early myocardial changes in mutation carriers of HCM [38]. Given that no changes in global LV function were observed and that LA dysfunction is the first manifestation of LV diastolic dysfunction [38], 3D-STE would become crucial for differentiating HCM carriers from non-carrier individuals.

As HCM is a common cause of premature sudden cardiac death leading to increased mortality, identification of the underlying etiology of LV hypertrophy is of clinical importance. TDI helps differentiate various conditions resulting in LV hypertrophy. A cut-off value of s′ < 9.0 cm/s, averaged for the septal and lateral mitral annulus, was shown to differentiate pathological LV hypertrophy (due to HCM or hypertensive heart disease) from physiological LV hypertrophy [39]. However, diagnosing HCM seemed to be difficult only with TDI, especially in elderly patients with mild concentric LV hypertrophy [40]. Using strain rate, Kato et al. proposed combined use of reduced longitudinal shortening and presence of asymmetric hypertrophy to identify biopsy-proven HCM patients apart from hypertensive patients [40].

Diastolic dysfunction

Diastolic function comprises continuum of early relaxation and late compliance [41]. Impairment of ventricular relaxation is a hallmark of HCM. With TDI, e′ is considered as a measure of early relaxation whereas a′ as a measure of late compliance. To the best of our knowledge, there is no convincing evidence on late compliance in HCM such that a′ enables to detect subclinical cardiac damage, to estimate LV filling pressures, or to predict clinical outcome. This is presumably because HCM has a specific myocardial/chamber property during diastole [42].

Pak et al. observed in a hemodynamic study that pressure-volume relationship in HCM was quite different from that in hypertensive heart disease, the former demonstrating that with volume overloading, the pressure-volume curve shifts superiorly with the rightward shift only a little [42]. Such a complexity of diastolic property, in HCM, may render introduction of robust markers for late compliance challenging. In HCM patients, both Doppler LV filling and pulmonary venous flow profiles are not as reliable for the non-invasive estimation of LV filling pressures as in DCM patients [43, 44].

On the other hand, in HCM, TDI-derived e′ is established as an index of myocardial relaxation, functional status, in addition to the degree of myocardial fibrosis [1], and E/e′ represents a non-invasive estimation of LV filling pressures [45]. Recently, the global strain rate during early diastole of the left ventricle (e′sr) has been proposed to assess LV filling pressures [19, 46]. It was augured that the e′sr had an advantage over E/e′ in terms of the coverage of the entire endocardium being assessed [19, 46]. Further investigations are needed to clarify relative usefulness of such a new index to E/e′ for cardiomyopathies.

Color kinesis (CK) is a real-time ultrasound technique providing information on the magnitude and timing of LV wall motion by tracking and color-encoding endocardial motion during the cardiac cycle [47]. CK has great usefulness particularly in the evaluation of LV relaxation [48]. Another strength of CK may be providing at-a-glance recognition of which LV segment has slowed relaxation. A study of HCM showed that CK imaged the regional LV filling time prolonged even in the non-hypertrophied segments [49].

Postsystolic shortening in hypertrophic cardiomyopathy

As mentioned, HCM patients tend to be accompanied by non-uniformity of regional systolic function; however, information on PSS in these patients is limited. A group of ours found that HCM patients had PSS in the considerable extent of LV wall with its magnitude being increased compared with normal controls [18]. This is in turn supportive of HCM patients being involved with myocardial disarray and fibrosis with various distributions of myocyte hypertrophy. There was an interesting observation that in patients with hypertrophic obstructive cardiomyopathy (HOCM), PSS attenuated after alcohol septal ablation, the degree of which correlated with E/e′ [50].

Apical asynergy with cavity obliteration of the left ventricle is a unique associate in HCM. The dyskinetic apical region may account for thrombus formation [51]. In one third of the cases, a paradoxical flow toward the left ventricular outflow tract (LVOT) during early diastole resulting from intraventricular obstruction is detected [51]. With TDI-based strain and strain rate analysis, Støylen et al. observed that the myocardium of the apical region associated with delayed relaxation and subsequent contraction during early diastole (PSS in a broad sense) was accounting for the paradoxical flow [52].

Figure 6 presents cardiac images from a case of HCM with apical cavity obliteration we experienced before. There is a paradoxical flow toward the LVOT with duration extending into early diastole, during which PSS is recorded both with TDI and strain rate methods.

Fig. 6
figure 6

Angiographic left ventriculogram (a) and images of continuous Doppler (b), TDI (c), and strain rate with curved M-mode (d) from a patient with HCM demonstrating apical asynergy with cavity obliteration. The black arrow indicates the site of intraventricular obstruction, arrowheads indicate paradoxical jets toward LV outflow with duration extending into early diastole, the yellow arrow indicates a strain curve of the apical site (arising from blue ROI) with PSS, and asterisks indicate a myocardium with its contraction sustaining into diastole (PSS). AVC and MVO indicate aortic valve closure and mitral valve opening, respectively

Full size image

Left ventricular outflow tract obstruction

About one third of HCM patients have systolic anterior motion of the mitral valve, a phenomenon involving most commonly the anterior mitral valve leaflet [53]. Mechanistically, it occurs as result of drag forces and the Venturi effect, which pushes and drags, respectively, the leaflets into the LVOT, thereby resulting in septal-mitral contact and subsequent LVOT obstruction [53]. However, detailed myocardial mechanics accompanied by this phenomenon have merely been evaluated.

Recently, with STE-based strain utilized in HOCM patients, Giraldeau et al. found that LV dyssynchrony induced by biventricular pacing was associated with reduced LVOT pressure gradient at 12 months follow-up without deterioration of LV ejection fraction or GLS [54]. The mechanism of which was implied as earlier recruitment of the LV free wall into systole and concomitantly released LVOT obstruction by the pacing, resulting in the preserved LV systolic function [54].

Other investigators proposed utilization of right ventricular pacing to attenuate LVOT obstruction by inducing “longitudinal” LV dyssynchrony, as determined by TDI, especially for the ventricular septum [55]. During a pacing study combined with TDI-based strain and strain rate analysis, in addition, it was shown that the myocardium of the basal septum “shortened” during OFF-mode pacing and “lengthened” during ON-mode pacing, with the OFF-mode setting being associated with greater LVOT pressure gradient [56].

Studies observed that HOCM patients who underwent alcohol septal ablation for intractable LVOT obstruction had improvement of ETI-assessed cardiac function [57, 58], supporting the notion that alcohol septal ablation gives more functional and prognostic benefit to HOCM patients than dual-chamber pacing, in which both systolic and diastolic function can be deteriorated after the procedure [59, 60]. In addition, LA function was shown to improve after septal ablation therapy in HOCM patients [61].

Tissue characterization by ultrasound integrated backscatter

About two decades ago, when TDI still was not clinically available, myocardial tissue characterization using echocardiographic integrated backscatter (IB) had attracted researchers in the field of cardiac ultrasound. A finding of IB enabling to detect rejection against a cardiac allograft used to be astonishing [62]. There were two parameters analyzed with this technique: IB intensity and cyclic variation of IB, the former reflecting the degree of myocyte hypertrophy or myocardial fibrosis, and the latter regional myocardial function.

A group of ours demonstrated in DCM patients that IB intensity, but not cyclic variation, discriminated patients who responded to beta-blocker therapy from those who did not [63], and that elevated IB intensity was associated with a higher rate of cardiac death or worsening of heart failure [64]. For HCM patients, it was found that IB intensity was increased and the cyclic variation was decreased, a finding of which was clearly different when compared with hypertensive patients [65].

Figure 7 shows the result of IB analysis from a patient with HCM in contrast to a normal subject.

Fig. 7
figure 7

IB images (left) and sequential plots of IB intensity (right) from a normal subject (a) and from a patient with HCM (b). Note that the HCM patient has relatively high IB intensity compared with the normal subject throughout cardiac cycles

Full size image

T1-mapping by cardiac magnetic resonance

The degree of myocardial fibrosis as an arrhythmogenic substrate is an important clinical consideration in the management of HCM patients. The late gadolinium enhancement (LGE) on CMR is sensitive to the finding of fibrotic lesions [66]. T1-mapping, an emerging technique of CMR, is also useful to detect interstitial fibrosis [67]. Notably, T1-mapping provides us with data that are not currently available with ETI. For example, a decrease in native T1 value is highly specific to detect disease severity related to myocardial iron overload or glycosphingolipid accumulation in Anderson-Fabry disease, whereas an elevated native T1 value is observed in HCM, amyloidosis, and other conditions associated with intercellular edema [67].

Several reports have evaluated usefulness of T1-mapping for diagnosing HCM and assessing its severity. It was reported that the native T1 value was increased in known HCM patients than normal controls, and that among HCM patients, LV segments with LGE were associated with higher native T1 value than those without LGE [68]. In a multicenter study, T1-mapping was found to discriminate HCM from hypertensive heart disease and to detect early changes in family members with sarcomere-gene mutations [69]. Unfortunately though, there have been no studies on HCM that compare early detection of the disease by CMR versus ETI.

Figure 8 shows images of native T1-mapping and the bull’s eye display of longitudinal strain from a patient with HCM, in contrast to those from a patient with light-chain CA.

Fig. 8
figure 8

Color-coded images of native T1-mapping in the 4-chamber view (left) and bull’s eye displays of longitudinal strain mapping (right) from a patient with HCM (a) and from a patient with light-chain CA (b). The asterisk indicates pericardial effusion

Full size image

Cardiac amyloidosis

CA is involved with the aggregation of insoluble deposits of misfolded proteins. The echocardiographic findings of CA include increased wall thickness with normal or mildly decreased LV ejection fraction, restrictive LV filling pattern, and dilated atria [70]. The amyloid deposits occasionally show a granular sparkling with increased granular myocardial echogenicity [70]. Only with conventional methodologies, however, the differentiation of CA from other types of cardiomyopathies such as HCM and hypertensive heart disease is not always straightforward. ETI, in that point, may come into play [71,72,73].

Given that regional systolic and diastolic dysfunction is the main feature of functional abnormality in CA, TDI, strain, and strain rate represent marked decreases in parameters derived even in the absence of restrictive physiology [72, 73]. STE has proven to be useful in distinguishing CA patients. In contrast to HCM and hypertensive heart disease, CA tends to be characterized by a distinctive pattern of regional variations in myocardial deformation, with relative sparing of function in the apical segments [71]. This finding is referred to as “apical sparing” (Figs. 8 and 9). Moreover, CA has been reported to be associated with decreased circumferential strain compared with HCM patients (Fig. 10), although GLS is impaired in both conditions [71]. One would see how easily recognizable on apical sparing and circumferential strain reduction in CA patients, seemingly an accurate means for segregating CA from HCM patients.

Fig. 9
figure 9

Parametric images of multi-planar reconstruction by 3D-STE from a patient with transthyretin CA. Arrows indicate apical regions showing “apical sparing”

Full size image
Fig. 10
figure 10

Analyses of circumferential strain using STE at the mid-ventricular level from a normal subject (a), from a patient with HCM (b), and from a patient with transthyretin CA (c). Marked depression of averaged strain can be seen in the CA patient

Full size image

Assessment of myocardial mechanics utilizing 3D-STE on 12 CA patients showed that the more pronounced LGE detected on CMR, the lower myocardial strains to the corresponding segments [74]. A pattern of marked LGE from the base to apex may render such a finding more likely to result from CA rather than HCM, although further information is warranted to understand and differentiate these two conditions with a large number of patients.

Atrial involvement

Atrial involvement is also common in CA patients and may cause “restrictive” LV filling in terms of the A-wave velocity reduction. Analysis of LA strain and/or strain rate to differentiate CA from HCM patients has been performed by several investigators. It was found that PALS was attenuated to a more extent in transthyretin-related CA compared with HCM despite comparable alterations of LA volume and LV ejection fraction [75]. In a study of various types of CA (light chain, transthyretin with mutant, and transthyretin with wild type), Nochioka et al. demonstrated that transthyretin with wild type had worse LA reservoir and pump function than other types of CA [76]. The authors also found the severity of LA function to correlate with parameters of LV function such as E/e′ and longitudinal strain and suggested that in CA, LA function did not deteriorate independently, but did to a greater or lesser extent under influence of altered LV function [76].

There are few studies on CA patients that examined LA function with 3D-STE [77, 78]. Using 3D-STE, Mohty et al. demonstrated that PALS and total LA emptying fraction were correlated with the Mayo Clinic staging, which was based on the finding of NT-proBNP and cardiac troponin T levels, and that decreases in these LA indices were associated with 2-year mortality [78].

Prolongation of intra- and interatrial conduction time with heterogeneous propagation of the sinus impulses is a pathological feature of the atria leading to atrial arrhythmia. Recent studies have shown that the time interval between the beginning of electrocardiogram P wave and the beginning of TDI-derived a′ (atrial electromechanical delay: AEMD) (Fig. 11) relates to AF development, suggesting that AEMD reflects AF vulnerability [79, 80].

Fig. 11
figure 11

Measurement of AEMD at the lateral side of the mitral annulus from a healthy person

Full size image

In CA patients, there should be a considerable degree of AEMD prolongation, because the myocardium with infiltrated amyloid proteins and the alteration of loading conditions lead to functional and structural changes not only on ventricles but also on atria; however, very scarce information exists on AEMD in CA [81]. Iio et al. calculated AEMD by STE, but not TDI, in patients with CA and in those with HCM, and found that the lateral wall EMD was recorded longer in CA patients being associated with a higher mortality rate [81].

Figure 12 shows ETI images from a CA patient. There is a marked prolongation of AEMD, with TDI-derived a′ existing beyond the QRS complex such that the LV filling is looking like a fusion wave.

Fig. 12
figure 12

Echocardiographic images from a patient with light-chain CA. a Two-dimensional apical four-chamber view. b Pulsed-Doppler LV filling profile. c AEMD measurement at the lateral side of the mitral annulus. d Measurement of PALS by STE

Full size image

Conclusions

In this review, we highlight echocardiographic findings in relatively common types of cardiomyopathies (DCM, HCM, and CA) and discuss which ETI methodology or index can be used in accordance with clinical, morphological, and functional features inherent to each type of cardiomyopathy. For example, global systolic dysfunction for DCM, impairment of regional systolic and diastolic function for HCM, and significant atrial involvement for CA. Contemporary progression of ultrasound technologies over ETI seems to allow even better understanding of cardiac characteristics and function. Utilization of STE is rapidly glowing and particularly, implementation of 3D-STE has potential to provide an optimal and homogenous application for strain measurements. We hope that further development of technologies regarding ETI will enable to quantitate deformation measurements reflecting the degree of structural alterations, myocardial fibrosis, and even chamber/myocardial stiffness.