Author + information
- Received October 28, 2015
- Accepted November 6, 2015
- Published online December 1, 2015.
- Piet Claus, PhD∗,
- Alaa Mabrouk Salem Omar, MD, PhD†,‡,
- Gianni Pedrizzetti, PhD§,
- Partho P. Sengupta, MD, DM† and
- Eike Nagel, MD, PhD‖∗ ()
- ∗Laboratory for Cardiovascular Imaging and Dynamics, Department of Cardiovascular Sciences, KU Leuven, Leuven, Belgium
- †Zena and Michael A. Wiener Cardiovascular Institute, Icahn School of Medicine at Mount Sinai, New York, New York
- ‡Department of Internal Medicine, Medical Division, National Research Centre, Dokki, Cairo, Egypt
- §Department of Engineering and Architecture, University of Trieste, Trieste, Italy
- ‖Institute of Cardiovascular Imaging, Goethe University Frankfurt and German Centre for Cardiovascular Research ([DZHK], partner site Rhine-Main), Frankfurt, Germany
- ↵∗Reprint requests and correspondence:
Prof. Eike Nagel, Institute of Experimental and Translational Cardiovascular Imaging, DZHK Centre for Cardiovascular Imaging Interdisciplinary Cardiovascular Imaging, Internal Medicine III and Institute for Diagnostic and Interventional Radiology, University Hospital Frankfurt-Main, Frankfurt, Germany.
Tissue tracking technologies such as speckle tracking echocardiography and feature tracking cardiac magnetic resonance have enhanced the noninvasive assessment of myocardial deformation in clinical research and clinical practice. The widespread enthusiasm for using tissue tracking techniques in research and clinical practice stems from the ready applicability of these technologies to routine echocardiographic or cardiac magnetic resonance images. The technology is common to both modalities, and derived parameters to describe myocardial mechanics are the similar, albeit with different accuracies. We provide an overview of the normal values and reproducibility of the clinically applicable parameters, together with their clinical validation. The use of these technologies in different clinical scenarios, and the additive value to current imaging diagnostics are discussed.
- cardiac mechanics
- feature-tracking cardiac magnetic resonance imaging
- myocardial function
- speckle tracking echocardiography
Optimal management of patients with cardiovascular disease is increasingly based on algorithms that use cutoff values or continuous variables, rather than simple binary (“yes” or “no”) decision trees. Accordingly, cardiac imaging techniques such as echocardiography and cardiac magnetic resonance (CMR) imaging have been developed for providing diagnostic information that are often expressed numerically. Both techniques can measure cardiac muscle motion and deformation; however, more recently, the clinical focus for noninvasive deformation imaging is moving from tailored acquisitions, such as tissue Doppler imaging or myocardial tagging, to post-processing of standard grayscale B-mode or cine imaging, resulting in easier access and wider availability. This review focuses on the current status of these post-processing methods, commonly known as tissue tracking, irrespective of the imaging modality. Speckle tracking echocardiography (2-dimensional [2D] and 3-dimensional [3D]) and feature tracking in CMR are compared and contrasted for elucidating relative strength and pitfalls and solutions that address modality-related differences.
Technology of Tissue Tracking
As shown in the Central Illustration, the technology of tissue tracking refers to methods of identifying a peculiar pattern along a curve on 1 image, such as the endocardial border, and recognizing the same pattern within a second image taken a few instants later. In this way, displacement of myocardial segments can be estimated. In echocardiography, images are characterized by the presence of speckles with a certain persistence (1,2); thus, the technique is commonly referred to as speckle tracking echocardiography (STE) both in 2 and 3 dimensions; in CMR, where tissue regions are identified by individual anatomical features, they are usually referred to as feature tracking (FT) (3).Thus, the intrinsic differences in both techniques are related to what region of myocardium is tracked, which may lead to differences in measurements.
All tracking techniques are more robust and reproducible for global rather than regional values. Dependent on image quality, 2D-STE is making it difficult in instances of poor echogenic windows, ultrasound dropouts, and reverberations. Through-plane motion may limit measurements in 2D-STE due to difficulties in tracking speckles that fall out of plane in subsequent tracking frames. In the distal part of the ultrasound sector, image quality deteriorates, resulting in a better tracking quality of speckles that are located proximally (4).
The problem of through-plane motion can be solved by using 3D echocardiographic techniques (i.e., 3D-STE) that recently became possible with advancements in matrix-array ultrasound transducers. In principle, the same tracking technology can be applied to 3D volumetric regions without conceptual differences, resulting in the availability of several 3D tissue tracking solutions. Unlike 2D-STE, speckles can be tracked simultaneously in all directions to derive all deformation parameters, alleviating the effect of through-plane motion and allowing for assessment of speckles independent from specific imaging planes. However, 3D images present a substantially (at least 3× to 4×) lower spatial as well as lower temporal resolution than their 2D counterpart does. This limitation in combination with the remaining dependency of the data on image quality so far makes 3D-STE assessments still controversial and their true value remains unknown (5). Further technical developments are required for improving 3D-STE accuracy.
FT-CMR has been explored on stacks of 2D cine images, with a typical slice distance of 6 to 8 mm and strong contrast between blood pool and myocardium, but with a lower in-plane spatial (1 to 2 mm) and temporal resolution (commonly 30 phases per heart cycle) than with 2D-STE. FT-CMR on 2D stacks suffers also from through-plane motion effects. Temporal averaging may result in lower strain values for FT-CMR in comparison to STE. Due to the lack of intramyocardial features, FT-CMR algorithms focus on the endocardial and epicardial borders with a stronger weighing of endocardial deformation explaining some of the differences in results found in direct comparisons of FT-CMR and STE. From the displacement estimations provided by these tracking methodologies a series of deformation parameters relevant to assess the mechanics of the myocardium can be derived (Table 1).
Normal Values, Reproducibility, and Clinical Validation
Tissue tracking can be used in the assessment of the mechanics of all cardiac chambers. Whereas the assessment of left ventricular (LV) and right ventricular (RV) deformations are of established clinical benefit, tissue tracking has been also used in studies to assess atrial deformations, mostly of the left atrium (LA); however, clinical applications remain to be validated.
Speckle tracking echocardiography
Global longitudinal strain (GLS) averaged from the apical views is the most robust and reproducible of all LV deformation parameters and has been shown to be a powerful diagnostic and prognostic tool. The use of LV- and RV-GLS is recommended in most recent guidelines for the quantitative assessment of LV and RV function (6). Due to a wide range of normal cutoff values (−17.3% to −21.5%) when measured by different vendors (6), guidelines are not recommending a clear cutoff value for LV-GLS but suggest that peak systolic LV-GLS in a healthy individual should be approximately –20%. Lower values should be expected in male patients and with increasing age and heart rate, as well as loading conditions, will influence the results (6,7). Serial measurements in an individual patient should be carried out using the same vendor machine and software (6).
RV-GLS and strain rate can be derived from tracking the RV free wall and the LV septum or only from the RV free wall in a RV focused apical 4-chamber view; however, RV free wall measurements were found to be of more prognostic value (6). Pooled data from single-center studies suggest that normal RV-GLS derived from the free wall should be lower than –20% (6).
It is important to note that GLS has better reproducibility than other LV deformation indices. Normal values of global circumferential strain (GCS) range between −20.9% and −27.8%, and global radial strain (GRS) between 35.1% and 59.0%, with mean values of −23.3% and 47.3%, respectively (8). The assessment of LV rotational mechanics suffers additionally from the lack of a standardized method of LV apical short-axis acquisition. Only small studies defining age-specific normal ranges of left ventricular twist (LVT) and untwist are available. Therefore, GCS, GRS, and LVT remain deformation parameters with potential clinical value awaiting further validation.
CMR feature tracking
Comparative normal values can be found in Table 2. For this review we included data from studies that have reported more than 1 strain component.
Across normal value studies, the most consistent parameters were GCS and GLS. In contrast, variations in GRS between studies were large and radial strains measured from long-axis acquisitions were often lower than from short-axis acquisitions, possibly attributed to through-plane motion in the short-axis scans. Normal values do not depend on field strength (9) when acquired with similar acquisition parameters including similar spatial resolution. So far no systematic studies have been performed to assess sex or age differences for FT-CMR deformation parameters or dependency on imaging parameters.
RV circumferential and radial strain and strain-rate values showed overall lower values than their LV counterparts did (10). This is consistent with RV geometrical features, having thinner walls and larger radius of curvature of the free wall.
Normal values for LA-GLS can be derived during the reservoir (29 ± 5%), conduit (21 ± 6%), and atrial contraction phases (8 ± 3%) of the atrial deformation and show an increase in atrial contraction in elderly subjects consistent with physiology of normal aging (11,12).
Summary: Deriving normal values for different deformations studied by STE and FT-CMR is not easy because of variability among studies and different users. GLS is the most robust deformation parameter for STE and is recommended in current guidelines for the assessment of LV function. Normal values around –20% for the LV and lower than –20% for the RV have been recommended for clinical practice. For FT-CMR, GCS is more reproducible than GLS with similar normal values as STE. Radial strain shows large ranges between studies and variability of segmental strains remains too high to use them as single clinical measures.
Transmurality of Dysfunction Corresponds to Patterns of Myocardial Deformation
Myocardial dysfunction can be classified according to the involved myocardial layer into predominantly subendocardial myocardial dysfunction, transmural myocardial dysfunction, and subepicardial dysfunction (Central Illustration) (13). This classification scheme corresponds to the structural and functional subunits of the LV that govern systolic and diastolic performance (13), both of which depend on the functional integrity and the synergistic coupling of the subendocardial and subepicardial fibers’ layers. Using the transmurality of myocardial involvement allows identifying cohorts of patients with similar presentations, prognosis, and response to therapy.
Noninvasive imaging techniques currently do not assess myocardial fiber mechanics. However, layer-specific strain measurements are being developed and awaiting clinical validation. The differences in direction of the subendocardial and subepicardial fibers have been related to the components of myocardial deformation for defining the transmural extent of myocardial dysfunction (Central Illustration). Contraction of the subendocardial fibers contributes to longitudinal shortening, whereas contraction of the subepicardial fibers contributes to circumferential shortening. Both aspects contribute to radial thickening. Myocardial rotation due to shearing forces created by sliding of myocardial fibers in the longitudinal-circumferential direction is dominated by the subepicardial layer because of the larger radius of rotation (14).
Most progressive myocardial diseases predominantly cause subendocardial dysfunction in their early stages, leading to reduction in longitudinal LV mechanics (15). Because epicardial fibers remain spared, circumferential strain and twist mechanics of the LV show normal or even increased values, compensating for the longitudinal mechanical dysfunction and thus preserve stroke volume and ejection fraction (EF). Given the coronary anatomy, ischemia causes mainly subendocardial fiber dysfunction, for example, in the presence of increased LV afterload. The development of subepicardial hypertrophy attempts to compensate for the loss of longitudinal function and reduce subendocardial wall stress, which in turn leads to the increased circumferential and rotational mechanics (16). Blood pressure control in hypertensive patients has been shown to reduce circumferential strain and LVT values with concomitant regression of subepicardial hypertrophy (17).
Transmural involvement results in concomitant subendocardial and subepicardial dysfunction resulting in attenuation of myocardial mechanics in all directions (13), with impairment of LV ejection performance. An acute large transmural infarction is a classic example where simultaneous impairment of LV longitudinal and circumferential mechanics is seen, resulting in LV systolic dysfunction.
Summary: Myocardial dysfunction can be classified according to the extent of endocardial or transmural involvement. Endocardial involvement results in worsening of longitudinal involvement, whereas other components may remain preserved, thereby preserving EF. Whereas transmural dysfunction leads to loss of myocardial mechanics in all directions, resulting into LV dilation and reduction of LVEF.
Specific Clinical Scenarios
Subclinical myocardial dysfunction
The longstanding presence of cardiovascular risk factors such as hypertension, diabetes mellitus, and obesity may cause disruption in the myocardial interstitial matrix as a result of microvascular ischemia, intramyocardial fibrosis, or collagen degradation product accumulation, leading to microscopic structural changes of the myofiber (18–20). These changes usually result in subclinical myocardial dysfunction at the endocardial level that can be detected using strain imaging as described earlier (reduction in GLS with compensatory increase in GCS and LVT with deterioration of diastolic function) (13). It is to be noted that similar changes also occur during the early subclinical stages and mild cases of infiltrative cardiac diseases. Because GLS is more reproducible and less variable than other parameters are and has proven clinical value, it is probably the single most important variable that can be used to identify and follow-up these patients.
Summary: Long-standing risk factors cause microscopic changes at the cellular level that affect the endocardial layer leading to depressed GLS and GRS, whereas GCS and LVT are relatively preserved due to the spared epicardial layer. GLS is of particular value in identifying and monitoring serial changes in these patients.
Coronary artery disease
In patients with coronary artery disease (CAD), STE imaging has been applied extensively to reduce interobserver variability, to provide absolute cutoff values, and to assess “functional” transmurality of an infarction. Whereas visual assessments of regional systolic function, such as wall motion score index, are accepted methods of assessing myocardial ischemia and coronary heart disease, both at rest and during stress testing, these methods are greatly variable between readers and are expertise-demanding. On the other hand, 2D-STE strain imaging offers a relatively more robust and simpler method of assessing global and regional functions in patients with coronary heart disease at rest and during stress testing, allowing for less variability in the assessment of hypokinetic and dyskinetic segments (21).
Given the predominantly subendocardial involvement in ischemia, reduction of global and segmental GLS is an early abnormality seen in patients with CAD (Figures 1A and 1C). Reduced GLS can be used to diagnose CAD and the location of myocardial infarction (22), whereas augmentation of GLS during dobutamine stress echocardiography indicates global myocardial viability (Figures 1B and 1D). Segmental assessment of longitudinal strain (LS) is superior to visual assessment in identifying viable myocardial segments during low-dose dobutamine echocardiography (23), can be used to identify ischemic territories (24) and segments with ischemia-related post-systolic shortening (25,26), and to estimate infarct size.
Numerous studies have suggested clinical utility of regional strain in CAD (Table 3); however, improvement in the precision of regional strain will be required for making this technique clinically useful.
Summary: GLS can be effectively used in CAD and myocardial infarction for identifying ischemic territories and infarct size, respectively. Improvement of GLS and segmental LS values with dobutamine infusion is suggestive of global and segmental viability. Scarred and ischemic segments are associated with a reduction in other deformation parameters. Circumferential strain and LVT is preserved in subendocardial ischemia and is reduced in patients with transmural ischemia and infarction.
Dilated cardiomyopathy (DCM) can be idiopathic or secondary to a variety of causes including ischemic heart disease, alcohol consumption, peripartum, chemotherapy, ventricular noncompaction, sarcoidosis, hemochromatosis, and myocarditis. The significant transmural myocardial remodeling that occurs in these patients is usually accompanied by loss of myocardial mechanics in all directions in parallel to the severity of LV systolic dysfunction. Both STE and FT-CMR studies have shown that the depressed GLS in these patients is of particular prognostic value and can be used to assess response to therapy and predict the occurrence of major cardiovascular events (27).
In addition to the abnormal strain values in patients with DCM, the increased sphericity of the LV apex affects the overall rotational mechanics and is usually accompanied by significant decrease in apical rotation and overall LVT. LV rotational mechanics, in addition to GLS, are other potential predictors of the response to cardiac resynchronization therapy (CRT) and rejection after transplantation (28,29).
Using FT-CMR, it was found that GLS is an independent predictor of survival in DCM and has incremental information for risk stratification beyond clinical parameters, biomarker, and standard CMR (30). STE-based studies found that strain can be a sensitive tool in early detection of cardiotoxicity especially in cancer patients being treated with chemotherapy. In female breast cancer patients treated with anthracyclines, GLS in combination with ultrasensitive troponin I was found to be predictive of the occurrence of chemotherapy-induced cardiotoxicity after 15 months of the completion of anthracyclines (31). GLS was also an independent predictor of LVEF in female breast cancer patients treated with trastuzumab (32). A recent systematic review found that early reduction of GLS by 10% to 15% is the most useful parameter in prediction of cardiotoxicity in patients receiving cancer chemotherapy (33).
In patients requiring CRT, an increase in GLS after implantation, in addition to the standard echocardiographic measurements, was found to predict responders and all-cause mortality at 1-year follow-up (34). A combined index of the magnitude and time of segmental radial strain was also found to be predictive of responders and survival 6 months post-implantation (35). Because of its ability to map temporal changes of strain distribution over the entire LV chamber, 3D-STE is considered a better modality for the analysis of LV dyssynchrony than 2D-STE is (36). Therefore, 3D-STE strain mapping may be promising in prediction and assessment of response after CRT, but more studies and developments are needed.
Summary: DCM is associated with a significant decrease of myocardial deformation in all directions. GLS is of particular clinical value being able to monitor response to therapy, predict future events, and response to CRT.
Both fibrosis and hypertrophy contribute to abnormal myocardial mechanics in patients with hypertrophic cardiomyopathy (HCM). The increased wall stress and the relative endocardial ischemia together with the fibrotic changes lead to endocardial dysfunction and thus a prominent decrease in GLS (Central Illustration). Therefore, GLS can be effectively used to differentiate pathological hypertrophy from physiological hypertrophy that occurs in conditioned hearts of trained athletes, who show normal GLS values. Longitudinal septal strain, in particular, was found to be inversely related to the fibrotic changes in histopathology and to be a more powerful predictor of arrhythmia than late gadolinium enhancement is (37). GLS was also shown to predict adverse events in adults with HCM (38). The endocardial dysfunction noticed in HCM leads to depression of GRS; however, because of the epicardial layer thickening, these patients usually have increased LVT with delayed LV untwist (39). In 1 report (40), surgical myomectomy was found to normalize GCS and LVT in these patients.
Using FT-CMR, it was reported that patients with HCM have reduced longitudinal, radial, and circumferential strains compared with control subjects with the ability of radial and longitudinal strain to predict clinical outcome (41). Similar to STE studies, FT-CMR–derived LVT was found to be increased in patients with HCM, whereas LV untwist was delayed. In addition, peak LVT was significantly and independently related to the percentage of LV late gadolinium enhancement as well as myocardial thickness (42).
Summary: In patients with HCM, endocardial dysfunction causes decreased GLS and GRS, whereas the epicardial thickening leads to preserved GCS and LVT. GLS can be used to predict future response and, more importantly, to differentiate LV hypertrophy in HCM from that in athletes’ hearts.
Restrictive cardiomyopathy (RCM) is caused by several infiltrative and storage diseases such as amyloidosis, sarcoidosis, systemic sclerosis, fibroelastosis, hemochromatosis, endomyocardial fibrosis, and eosinophilic endomyocardial disease (43). The microscopic deposition of these materials results in increased myocardial stiffness, which causes impaired ventricular filling with normal or decreased diastolic volume of either or both. Wall thickness may be normal or increased depending on the underlying cause.
The most investigated cause of RCM is amyloidosis. Overall, GLS is decreased in amyloid heart disease as expected in a case of RCM and was found to be a negative predictor of survival (47). However, GLS in cardiac amyloidosis has a characteristic pattern with reduced strain at the LV base and progressively increased strain near the LV apex. LV twisting and untwisting motions are initially increased in patients with light-chain amyloidosis or systemic amyloidosis with no cardiac involvement (48,49), but they may normalize or become reduced with ongoing cardiac involvement (48). This may be helpful in differentiating amyloid-associated cardiac hypertrophy from HCM (46).
In patients with endomyocardial fibrosis (Fabry disease), in addition to the lower GLS (50), diastolic strain rates are also reduced (51,52), which correlated with the late gadolinium enhanced CMR (50). In patients with sarcoidosis and systemic sclerosis (53), GLS is also significantly reduced, and LVT might be slightly increased (54), and radial strain variations can differentiate sarcoidosis patients from those with DCM (55).
Summary: In patients with RCM, GLS is depressed significantly, whereas GRS and GCS might be preserved. In amyloid heart disease, beside the depressed GLS, LV twist and untwist might be preserved early and progressively decrease with the progression of cardiac involvement.
Constrictive pericarditis (CP) is a pericardial disease with a particular clinical importance, because its clinical picture can be mistaken for RCM, as both involve diastolic dysfunction with different mechanisms. Thus differentiation between both conditions is challenging using the conventional methods. STE-derived deformation imaging can usually give a clue to differentiate both conditions, as different myocardial layers are affected and thus each case exhibits different mechanical behaviors. In RCM, endocardial dysfunction causes depression in GLS and GRS. The relatively spared epicardial fibers cause preservation of GCS and LVT (Figures 2A to 2C). In contrast, in CP, mainly the subepicardial fibers are affected with relative sparing of subendocardial fibers, which leads to preserved GLS and GRS, whereas GCS and LVT are usually depressed (Figures 2D to 2F) (44).
Summary: Deformation imaging can be used to assess patients with pericardial diseases and to differentiate CP from RCM. In CP, the epicardial dysfunction leads to depressed GCS and LVT, whereas GLS and GRS are preserved, compared with RCM where the endocardial dysfunction causes depression of GLS and GRS with preserved GCS and LVT.
Valvular heart disease
In patients with aortic stenosis (AS) and preserved EF, the subendocardial ischemia resulting from the pressure-overload causes attenuation of GLS with relative preservation or increase in LVT (56). Treatment of AS by surgical or transcatheter aortic valve replacement cause normalization of all LV mechanics (56,57).
FT-CMR was used to compare the impact of a transfemoral to a transapical approach of transcatheter aortic valve replacement on myocardial mechanics (58). No differences in peak radial or longitudinal strain were found for the basal and mid-segments. Not surprisingly, in the transapical approach peak radial and longitudinal strain were reduced with respect to the transfemoral approach in the apical segments and the apical cap.
In patients with mitral regurgitation (MR) and normal EF, the development of insidious myocardial dysfunction is usually associated with depression in GLS, GCS, and GRS (59), whereas LVT might be preserved until the development of evident reduction of EF (60). In MR patients undergoing mitral valve repair, pre-procedural GLS was found to predict post-procedural EF reduction (61). In patients with secondary significant MR undergoing mitral clip, 3D-STE showed overall improvement in both LV and RV strain after clip implantation (62).
In patients with mitral stenosis, GLS and GCS were found to be lower than in control subjects, with continuous normalization within 72 h after balloon mitral valvuloplasty (63). In addition, LVT was found to be lower in patients with mitral stenosis (64). It is still controversial whether the basal decrease in LV mechanics in patients with mitral stenosis is due to LV dysfunction or as a result of decreased pre-load caused by the stenosed valve.
Summary: In AS, endocardial dysfunction leads to depressed GLS, whereas LVT is usually preserved. All deformation parameters tend to normalize after valve replacement. In MR, depression of strain parameters is a sign of insidious systolic dysfunction. GLS can predict post-mitral valve repair EF reduction.
Myocardial mechanics in heart failure
Systolic heart failure
Symptoms of heart failure can develop despite preserved EF (i.e., heart failure with preserved ejection fraction [HFpEF]). It was previously believed that the main pathophysiological changes only involve abnormal diastolic function; however, LV strain studies have shown that concomitant systolic abnormalities do occur (15,65). Ongoing progression of the underlying pathology may result in abnormalities in strain in all directions resulting into heart failure with reduced ejection fraction (HFrEF).
STE-based studies have shown that a decreased GLS can be used as a marker of insidious systolic function in HFpEF patients and is associated with higher levels of pro–B-type natriuretic peptide levels and worse diastolic function (Figure 3A) (66). Exercise is usually associated with further depression of GLS and provocation of symptoms in these patients. The development of abnormal endocardial function early in the course of heart failure is responsible for the abnormal longitudinal function. The reason why EF is preserved in these patients is that GCS and LVT remain unchanged or even increased (Figures 3B and 3C), compensating for the decreased GLS. With the development of HFrEF, GLS continues to decrease together with failure of the compensating mechanisms of LVT and GCS, thus myocardial pump failure and dilation usually ensue (Figure 4).
Summary: In HFpEF, endocardial dysfunction leads to depressed GLS, which correlates with levels of pro–B-type natriuretic peptide and diastolic functions. The preserved GCS and LVT usually compensates for the depressed GLS. Failure of these compensatory mechanisms later in the course of the disease lead to pump failure (HFrEF).
Diastolic dysfunction and assessment of filling pressures
Deformation imaging provides unique information during the diastolic period that can potentially be useful in the assessment of diastolic function. This includes the quantification of segmental and global diastolic strain rate, as well as the differences in the timing of transition from myocardial contraction to relaxation with strain rate imaging (67). Whereas most of these parameters were used in the assessment of regional fibrosis and ischemia, a few studies have shown a significant relation between segmental and global early diastolic strain rate and the time constant of LV relaxation (tau) (68,69).
LV diastolic untwist contributes to early LV filling through the generation of negative suction pressure (21). Untwisting parameters were shown to correlate with invasive indices of LV relaxation and suction (first derivative of pressure measured over time [dP/dt] and tau) but not with LV stiffness, suggesting that untwisting is a key mechanical event that aids LV early diastolic filling because of serving as a link between systolic compression and early diastolic recoil (21,67). Importantly, in cases with abnormal LV relaxation, untwist duration and time to peak untwist rate are directly related to tau (21).
However, untwist rate, in addition to LV relaxation, is also affected by systolic contraction and thus, changes in untwist rate might not be an accurate representation of the diastolic dysfunction in patients with HFpEF.
Although there are plenty of studies concerned with the assessment of diastolic functions using STE, there is relatively less data available for the use of CMR in the assessment of LV diastolic function. Nevertheless, studies have employed myocardial tagging to demonstrate abnormalities in LV untwisting in diastole in conditions such as asymptomatic individuals with LV hypertrophy and AS (70). In the MESA (Multi-Ethnic Study of Atherosclerosis), CMR tagging was found to be useful in the assessment of regional diastolic function noninvasively. In those patients, regional impairment of early diastolic strain rate was found to occur despite the preserved systolic function (70,71). Moreover, in a subset of patients in the CARDIA (Coronary Artery Risk Development in Young Adults) study, diastolic strain rate calculated from tagged magnetic resonance images was found to predict incident heart failure and atrial fibrillation (72).
Summary: Tissue tracking parameters can be used to assess LV relaxation and diastolic function. LV untwist and strain rate might be promising in this regard. In patients with diastolic dysfunction, prolonged and delayed LV untwist is particularly related to prolonged LV relaxation and tau and worse LV diastolic suction.
Clinical Use of Deformation Imaging in Other Cardiac Chambers
Right ventricular function
The use of RV-LS derived from the RV free wall is currently recommended for the assessment of RV functions in patients with suspected RV dysfunctions, such as congenital abnormalities, arrhythmogenic RV cardiomyopathy, pulmonary, primary arterial hypertension (PAH), and pulmonary thromboembolism. In those patients, RV-LS values are usually decreased and delayed (73). Interventions that cause improvement of hemodynamics in these patients are usually accompanied by improvement in RV-LS.
In general, RV function derived from STE was shown to have prognostic value in a variety of congenital abnormalities (74–76). In patients with tetralogy of Fallot, RV-LS correlated with exercise capacity (76).
However, LV affection in patients with RV dysfunction, as a manifestation of ventricular interdependence, is not uncommon. For instance, in patients with PAH, LV septal circumferential strain was found to be affected more than the RV free wall LS was (77). Moreover, RV strain was found to correlate with LV filling pressure, and combination of both parameters as a measure of interdependence might serve as a predictor for future outcome in patients with PAH (78).
In patients with arrhythmogenic right ventricular cardiomyopathy (ARVC), compared with control subjects, RV-LS derived from either STE or FT-CMR was found to be significantly depressed (10,79) and to be superior to conventional echocardiographic parameters when used for the diagnosis of these patients.
It was also found that in patients with ARVC, compared with control subjects, RV-LS fails to increase with exercise (62). Furthermore, 2D-STE have suggested a high incidence of LV involvement in phenotypic patients with ARVC as well as genetically predisposed relatives. LV involvement in addition to an enlarged RV outflow tract were found to be independent prognostic markers of event-free survival (80).
Summary: RV-LS is a recommended measure to assess RV function in clinical scenarios with suspected RV dysfunction. RV dysfunction caused by PAH, congenital anomalies, and ARVC is usually associated by depressed RV-LS. RV-LS is also of prognostic value in these patients.
Left atrial function
STE and FT-CMR can be also used to assess LA longitudinal deformations. Because the LA is a thin-walled chamber, the measurement of the radial deformation is difficult. LA strain and strain rate can be used to study LA conduit, contractile, and reservoir functions (81) and are considered promising tools that can be used for the assessment of LV filling pressures (81,82). In patients with HCM and HFpEF, FT-CMR–derived total strain and positive strain rates were reduced and both LA conduit function strain during ventricular E-wave and first negative strain rate were impaired (12).
In a substudy of MESA, patients had lower atrial GLS and higher indexed minimal LA volume, both of which were independently associated with incident heart failure, even after adjusting for traditional risk factors, LV mass and N-terminal pro–B-type natriuretic peptide (70). FT-CMR–derived LA LS and strain rate were also found to reliably discriminate between patients with impaired LV relaxation and healthy control subjects (12).
LA strain measures, however, are not currently recommended for routine clinical use and remain mainly promising research tools, primarily because most studies concerned with their use are small.
Summary: LA-LS can be assessed by tissue tracking and is a promising tool that can be used for assessment of LV filling pressures and atrial function; however, it still remains as a research tool awaiting clinical validation.
Future Research Directions
Speckle tracking echocardiography
Variability in most of the STE-derived parameters remains a great concern and creates difficulties in standardizing their use and application in everyday clinical practice (5). The recently initiated European Society of Cardiovascular Imaging and the American Society of Echocardiography strain standardization task force is aiming at addressing the issue of variability by standardizing the use and applications of STE between different vendors’ software and users (4).
Importantly, because 3D-STE can simultaneously measure myocardial motion in all directions, the introduction of high-resolution 3D systems can be used for the assessment of myocardial principal strain, a newly developed concept of studying myocardial mechanics using 3D echocardiography, which simultaneously integrates myocardial deformations in all directions, alleviating the need for multidirectional strain assessments (83). In the future, limitations in the spatial-temporal resolution may be addressed with the incorporation of high-frequency ultrasound systems, which represent a new frontier for echocardiographic-based deformation imaging that promises high temporal and spatial resolutions (84).
Finally, appreciation of the LV intracavitary vortex motion and studying its interaction with myocardial deformations may open a new era of exploring the mechanical efficiency of cardiac contractions (85).
CMR feature tracking
As opposed to speckle tracking, FT-CMR has not been validated in the setting of phantoms or animal models to compare with ground truth assessment of myocardial deformation by, for example, crystals.
In many pathologies, early subclinical changes are detected by changes of LS. However, because GLS and especially global RV-LS are less reproducible than GCS is, optimization of the longitudinal tracking will be necessary to be applicable in the individual patient. The major advantage of FT-CMR is that it does not require special sequences and can be applied retrospectively. Therefore, it can be expected that this technique will see a large increase in published studies, as was the case after the introduction of STE replacing the tissue Doppler-based strain imaging techniques.
As opposed to STE (6), most FT-CMR studies have been using the same software for analysis, limiting this potential source of variability. It can be expected that when FT-CMR enters more into the clinical arena, new software will emerge and similar efforts of standardization like with STE will be needed (4).
There is a large transmural gradient (from endocardium to epicardium) of circumferential strain and a slightly different initial contour placement will influence circumferential strain values. The optimal choice here remains unknown and will need to be determined by further research or consensus. In this context, it is also important to remember that FT-CMR derives its information mainly from the endocardial border, as opposed to intramyocardial speckles as used for STE.
FT-CMR offers the possibility to use historical cine data. Although the effect of frame rate has not been studied, so far the standard frame rate used in CMR (usually approximately 30 frames/cardiac cycle) is below the recommended frame rates for STE, which may induce significant differences especially when considering strain rate values.
Both STE and FT-CMR currently offer reproducible measurements of global strain values that can be applied in different clinical scenarios to assess LV and RV function (Table 4). CMR has an established and continuously expanding role in tissue characterization and is the modality of choice for accurate evaluation of global function using volumetry. Whether the additional assessment of myocardial deformation yields information beyond direct tissue characterization remains to be evaluated. Similarly, the yield of STE versus CMR tissue characterization needs to be assessed. Most likely, a major clinical value of these techniques will only be achieved when a reliable segmental analysis becomes possible. For the time being, GLS in 2D-STE has a proven clinical value due to its sensitivity to early contractile changes in the presence of normal or preserved EF.
For the supplemental material, please see the online version of this article.
Dr. Claus receives financial support from KU Leuven under grant #PF/10/014. Dr. Omar is a consultant/advisor for Edwards Lifesciences. Dr. Pedrizzetti receives financial support from the Italian government (grant #PRIN N. 2012HMR7CF_002) and is a shareholder of AMID. Dr. Sengupta is a consultant/advisor for Edwards Lifesciences, Saffron Technology, Heart Test Labs, and TeleHealthRobotics; and receives research support from Forest Laboratories. Dr. Nagel receives support in kind (analysis software) from TomTec Imaging Systems, and from Circle Cardiovascular Imaging Inc. Drs. Claus and Omar contributed equally to this paper.
- Abbreviations and Acronyms
- arrhythmogenic right ventricular cardiomyopathy
- aortic stenosis
- coronary artery disease
- cardiac magnetic resonance
- constrictive pericarditis
- cardiac resynchronization therapy
- dilated cardiomyopathy
- ejection fraction
- feature tracking
- global circumferential strain
- global longitudinal strain
- global radial strain
- hypertrophic cardiomyopathy
- heart failure with preserved ejection fraction
- heart failure with reduced ejection fraction
- left atrium
- longitudinal strain
- left ventricle
- left ventricular twist
- mitral regurgitation
- primary arterial hypertension
- restrictive cardiomyopathy
- right ventricle
- speckle tracking echocardiography
- Received October 28, 2015.
- Accepted November 6, 2015.
- 2015 American College of Cardiology Foundation
- Hor K.N.,
- Gottliebson W.M.,
- Carson C.,
- et al.
- Mor-Avi V.,
- Lang R.M.,
- Badano L.P.,
- et al.
- Sengupta P.P.,
- Tajik A.J.,
- Chandrasekaran K.,
- Khandheria B.K.
- Jellis C.,
- Martin J.,
- Narula J.,
- Marwick T.H.
- Omar A.M.,
- Vallabhajosyula S.,
- Sengupta P.P.
- Asanuma T.,
- Uranishi A.,
- Masuda K.,
- Ishikura F.,
- Beppu S.,
- Nakatani S.
- Asanuma T.,
- Fukuta Y.,
- Masuda K.,
- Hioki A.,
- Iwasaki M.,
- Nakatani S.
- Nahum J.,
- Bensaid A.,
- Dussault C.,
- et al.
- Sawaya H.,
- Sebag I.A.,
- Plana J.C.,
- et al.
- Thavendiranathan P.,
- Poulin F.,
- Lim K.D.,
- Plana J.C.,
- Woo A.,
- Marwick T.H.
- Kydd A.C.,
- Khan F.Z.,
- O’Halloran D.,
- Pugh P.J.,
- Virdee M.S.,
- Dutka D.P.
- Almaas V.M.,
- Haugaa K.H.,
- Strøm E.H.,
- et al.
- Nucifora G.,
- Muser D.,
- Morocutti G.,
- et al.
- Hancock E.W.
- Sengupta P.P.,
- Krishnamoorthy V.K.,
- Abhayaratna W.P.,
- et al.
- Quarta C.C.,
- Solomon S.D.,
- Uraizee I.,
- et al.
- Yu C.M.,
- Lin H.,
- Yang H.,
- Kong S.L.,
- Zhang Q.,
- Lee S.W.
- Kraigher-Krainer E.,
- Shah A.M.,
- Gupta D.K.,
- et al.,
- for the PARAMOUNT Investigators
- Kato T.,
- Noda A.,
- Izawa H.,
- et al.
- Wang J.,
- Khoury D.S.,
- Thohan V.,
- Torre-Amione G.,
- Nagueh S.F.
- Habibi M.,
- Chahal H.,
- Opdahl A.,
- et al.
- Kempny A.,
- Fernández-Jiménez R.,
- Orwat S.,
- et al.
- Teske A.J.,
- Cox M.G.,
- De Boeck B.W.,
- Doevendans P.A.,
- Hauer R.N.,
- Cramer M.J.
- Bhan A.,
- Sirker A.,
- Zhang J.,
- et al.
- Sarvari S.I.,
- Haugaa K.H.,
- Zahid W.,
- et al.
- Ersbøll M.,
- Valeur N.,
- Mogensen U.M.,
- et al.
- Haugaa K.H.,
- Smedsrud M.K.,
- Steen T.,
- et al.
- Ersbøll M.,
- Valeur N.,
- Andersen M.J.,
- et al.
- Seo Y.,
- Ishizu T.,
- Enomoto Y.,
- et al.
- Technology of Tissue Tracking
- Normal Values, Reproducibility, and Clinical Validation
- Transmurality of Dysfunction Corresponds to Patterns of Myocardial Deformation
- Specific Clinical Scenarios
- Clinical Use of Deformation Imaging in Other Cardiac Chambers
- Future Research Directions