Advantage and disadvantage of different techniques for myocardial deformation analysis.
Evaluation of myocardial mechanics, although complex, has now entered the clinical arena, thanks to the introduction of bedside imaging techniques, such as speckle-tracking echocardiography.
- cardiac mechanics
- strain rate
- speckle tracking
Speckle-tracking imaging (STI) is a non-invasive ultrasound technique that allows an objective and quantitative evaluation of global and regional myocardial function, independently from the angle of insonation and partly from cardiac translational movements [1–4].
Echocardiographic estimation of segmental left ventricular contractility is routinely accomplished through visual interpretation of endocardial motion and myocardial thickening. This method is subjective and requires a relatively experienced observer. Quantitative analysis based on tracing of the endocardial border may also be hampered by endocardial “dropout” and trabeculations.
Tissue Doppler imaging (TDI) has been previously used in deriving myocardial velocities and assessing fundamental parameters of myocardial deformation (strain and strain rate) . Myocardial tissue velocities represent the net effect of the contractile and elastic properties of the area under investigation and the motion caused by traction and tethering from other regions. In contrast, strain is a dimensionless index reflecting the total deformation of the ventricular myocardium during a cardiac cycle, as a percentage of its initial length. Strain rate is the rate of deformation or stretch. Strain techniques are, in principle, the optimal modalities for the assessment of regional myocardial function. The major limitation of TDI has been its angle dependency , requiring alignment of the ultrasound beam parallel to the direction of tissue movement. Thus, deformation study was substantially limited to the analysis of the tissue moving toward or away from the probe (Table 1).
STI is based on bi-dimensional (2D) echocardiographic technology, not limited by Doppler analysis [6–8]. Segments of myocardial tissue show a pattern of gray values in the ultrasound. This pattern, resulting from the spatial distribution of gray values, is commonly referred to as speckle pattern, characterizes the underlying myocardial tissue acoustically and is unique for each myocardial segment. Speckle tracking allows the measure of all in-plane components of the velocity vector, in all pixels . More recently, the addition of the third dimension (3D) has partly expanded the scope of this technology.
2. Human myocardium
Human myocardium is made up of multiple layers [10–12]. According to the helical model described by Torrent-Guasp, a single myocardial muscle band folding upon itself creates varying orientations of fibers throughout the myocardium, with two principal loops, basal (transverse) and apical (oblique). Myocardium finally consists of three separate layers: transversely oriented circular fibers that wrap (around both ventricles), the inner oblique layer (clockwise rotation), and the outer oblique layer (counterclockwise rotation). Briefly, the myofiber geometry of the LV changes gradually from a right-handed helix in the subendocardium to a left-handed helix in the subepicardium.
The mechanics of the LV is complex, but three principal components contribute to systole: inward motion, longitudinal motion (base moving toward the apex), and differential rotation of the apex and base (twisting). Diastolic mechanics is the opposite of systolic motion. LV torsion (or twist) has an important role in cardiac systo-diastolic mechanics. During the cardiac cycle, there is a systolic twist and an early diastolic untwist around left ventricular long axis, due to opposite apical and basal rotations. Systolic apical rotation is counterclockwise and basal rotation is clockwise . LV rotation is a sensitive indicator of changes in regional and global LV function.
2.1. Cardiac mechanics and limits of conventional indices
Cardiac function is the result of force development (inotropism: opening of the cardiac valves) and deformation (shortening of the myocytes: volume ejection) and can be evaluated globally (pump performance) or regionally . Correct evaluation of systolic function should focus on the intrinsic properties of the fibers and myocytes that represent the real actors behind heart function as well as a pathologic process. The deformation of a myocardial segment during the cardiac cycle is a complex phenomenon and consists of normal deformation (longitudinal shortening/lengthening, radial thickening/thinning, and circumferential shortening/lengthening) and shear deformation (base-apex twisting, epi-endo circumferential shear, and epi-endo longitudinal shear). There is a clear gradient from base to apex in both velocity and displacement (stationary of the apex within the thorax while the base is moving toward it). In contrast, deformation is more or less homogenous throughout the (normal) myocardial wall. It is important to understand the relation between intrinsic function (contractility) and the resulting motion/deformation. A myocardial segment develops a force but it is also subject to the forces developed by near segments. The intrinsic contractile force of the myocardium (inotropism) is the most important determinant of myocardial performance, but any segment of myocardium is part of a ventricle, with external forces acting up on it (mostly in the opposite direction of the contractile force) and resulting from local wall stress, caused by the intracavitary pressure (related to local geometry of the ventricle) and the interaction with neighboring contracting segments (“pulling” the segment itself). The relationship between acting forces and the resultant deformation is influenced by regional elasticity, which by itself is not a constant; due to the structure of the tissue, the more the myocardium is stretched, the more difficult it becomes to stretch it even further.
Regional myocardial deformation is thus the result of:
Intrinsic contractility (influenced by tissue perfusion and electrical activation).
Intracavitary pressure (afterload/preload; ventricular geometry);
Collagen amount (fibrosis).
Despite this pathophysiological insight, echocardiography is still mainly based on global and indirect indices such as ejection fraction, a volume-based parameter that does not reflect contractility, being based on geometric assumptions. For example, “supranormal” values of ejection fraction (EF) are often found in hypertrophied or volume-overloaded ventricles, thus not reflecting real changes in contractility. EF is also load dependent and, per se, is only a global index, without regional implications and not taking into account segmental interactions, which do not contribute to pump function. Moreover, indices for global functional assessment reflect mainly radial function, ignoring longitudinal function, which is usually altered long before changes occur in radial indices.
2.2. The concept of myocardial deformation
When two neighboring points of the myocardium move at different velocities, myocardium changes its shape (deforming). Otherwise, myocardium is moving but not deforming. When the velocity of the tissue is known, several other parameters can be derived.
Displacement is the integral of the velocity over time (Eq. (1)).
Strain and strain rate are measures of changes in shape, that is, deformation.
For mono-dimensional deformations, that is, shortening or lengthening, the simplest measurement is conventional or Lagrangian strain (Eq. (2)).
The Greek letter epsilon (ε) is commonly used as a symbol for conventional strain. The strain value is dimensionless and can be presented as a fractional number or as a percentage. For Lagrangian strain, a single reference length (L0) is defined, against which all subsequent deformation (L(t)) will be measured. Strain is positive if L is major than L0 (an object has lengthened) and negative if L is smaller than L0 (shortening). If L equals L0, the strain is thus zero.
Natural strain (ε’) is defined as :
Natural strain employs a reference length that changes as the object deforms. It therefore describes the instantaneous length change and it is independent of reference times. Compared to that of conventional strain, the natural strain amplitude is smaller for positive strains and larger for negative strains. This concept applies in principle to all three one-dimensional (longitudinal, circumferential, and radial) displacement and strain components.
In two or three dimensions, we should also consider shear strain, i.e., measurement of deformation in angle. It is also mandatory to specify directions and magnitudes of maximal and minimal strain.
The strain rate is the temporal derivative of the strain (Eq. (4)).
Whereas strain indicates the amount of deformation, strain rate indicates the rate of the deformation. The spatial gradient in myocardial velocities represents the rate of myocardial deformation, that is, the strain rate. The unit of the strain rate is normally 1/s or s−1. Strain rate is more uniformly distributed along the different regions of the LV, whereas myocardial velocity decreases from base toward apical parts of the LV. Strain can subsequently be derived by temporal integration of the strain rate curve. Indeed, if the rate of deformation is known at each time instance during the cardiac cycle, the total amount of deformation can easily be calculated. A positive strain rate means that the length of the object is increasing, whereas a negative strain rate means that the length is decreasing. If the length is constant, the strain rate is zero. Therefore, whereas strain is a measurement of deformation relative to a reference state, strain rate is an instantaneous measurement. When the strain rate has been calculated for each time point during the deformation, the strain can be found as the temporal integral of the strain rate (Eq. (5)).
2D strain comprises four measurements: two natural strains and two shear strains .
A 3D model allows the evaluation of three natural strain and six shear strain measurements along x, y, and z or azimuthal axes.
2.3. Fundamentals of speckle tracking
“Speckles” are small groups of myocardial pixels created by the interaction of ultrasonic beams and the myocardium, with specific gray scale characteristics. A speckle is commonly defined as the spatial distribution of gray values in the ultrasound image. The result of a speckle-tracking procedure (followed by regularization process) is an estimate of the in-plane velocity vector in all pixels in each of the frames of the ultrasound data set (dynamic velocity vector field). The spatial distribution of the gray values within the ultrasound image is due to constructive and destructive interference of reflections from the individual scatterers within the myocardium. Reflections occur at transitions between different types of tissues or at specific sites, and are much smaller than the wavelength. Constructive interference generates a high-amplitude signal, destructive a low-amplitude one. The exact scatter positions determine the speckle characteristics. Speckle-tracking technology offers the ability to identify and track the same speckle throughout the cardiac cycle .
In the ultrasound image, we see the speckle pattern occurring at a position further away from the transducer. To correctly detect speckles, the motion of the tissue should be slower than the motion of the ultrasound beam (image lines). Sound waves propagate through tissues at an average velocity of 1530 m/s, while myocardial tissue moves at velocities in the order of centimeters per second: the basic condition is thus clearly met .
There are different algorithms used by different vendors in tracking these speckles. Some speckle-tracking methods are based on so-called block matching, where a region in the image is selected (the kernel) and is followed in the next image frame by subsequently trying out different positions and by determining the similarity between the kernel and the pattern observed in that position. The position where the similarity between the kernel (“fingerprint”) and the observed pattern is maximal is assumed to be the new position of the speckle pattern . Another common approach is based on conservation of gray value, that is, it is assumed that gray values do not change over time. Radio frequency (RF) speckle—used in block-matching method—is a high-frequency signal, so that small between-frame motion can be detected, whereas its corresponding gray-scale speckle—used in gray-scale tracking—is derived from lower-frequency signals, being less sensitive to small displacements. Importantly, speckle tracking of gray-scale images does not necessarily perform well on high frame-rate data . Then, RF-based methods allow to obtain a higher spatial, temporal, and velocity resolution because they use a signal with a higher-frequency content; at the same time, these methods are more sensitive to decorrelation and noise, requiring more severe regularization, which in turn might limit their resolution. Because both RF and gray-scale approaches offer advantages, a hybrid method was recently proposed.
So far, it is possible to evaluate the direction of movement, the speed of movement, and the distance of such movement at any point in the myocardium, independently from the transducer, relative to adjacent segments. The semi-automated nature of speckle-tracking echocardiography guarantees good intra-observer and inter-observer reproducibility .
Given that the velocity vector field is known for all pixels within the image, the axes are known with minimal user interaction. The radial, longitudinal, or circumferential velocity profiles throughout the cardiac cycle can be reconstructed, independent of the angle between the ultrasound image line and the direction of motion as in the conventional Doppler imaging . The process of correcting the initial velocity vector estimates by applying additional boundary conditions based on a priori knowledge about the characteristics of the velocity field is called regularization. Regularization can consist of median filtering, weighted smoothing, elastic model, and myocardial boundaries definition.
Velocity vector imaging is partly analogous to 2D STI as it too tracks the speckles using 2D echocardiography, but utilizes additional physiological information to more robustly track the speckle kernels . Each vector is an expression of direction and the magnitude of the velocity. The qualitative evaluation of the velocity is determined by comparing vectors along the tracked contour. Longitudinal strain is the percentage decrease in the length of the myocardium during systole (movement of the base toward the apex). It is expressed as a percent negative value (decrease in length in systole) . Longitudinal strain may be calculated as an endocardial strain, midline strain, epicardial strain, or averaged over the entire cardiac wall. There is currently insufficient evidence to favor one way over another. Radial strain refers to the thickening of the myocardial wall during inward motion of the ventricle, measured in the short-axis views. The value is traditionally defined as percent positive (thickening in systole). Circumferential strain represents the change in the length along the circular perimeter, by definition percent negative in systole. Strain parameters can be individualized for each myocardial segment or can be expressed as global strain (averaging of all segments). Strain rate (evaluated globally or for each segment) represents rate of longitudinal, radial, or circumferential deformation in time. It has a marked systolic negative peak (S) with two positive peak in early (E) and late diastole (A).
Relevant strain values along strain curves are, but are not limited to:
End-systolic strain: the value at end-systole;
Peak systolic strain: the peak value during systole;
Positive peak systolic strain;
Peak strain: the peak value during the entire heart cycle. The peak strain may coincide with the systolic or end-systolic peak, or may appear after aortic valve closure (AVC) (it may be described as “post-systolic strain”) .
Modern software allows display of results in bull’s eye (polar map) similar to single-photon emission computed tomography (SPECT). This is more familiar to cardiologists as it depicts single myocardial segments with relative values of strain, strain rate, and time to peak strain/strain rate (synchronicity). A more unfamiliar method to display results in a monoplane view is the so-called curved anatomic M-mode (CAMM) which depicts timely variation of single parameters evaluated for a specific segment of interests from base to the apex and from septal to lateral wall. This offers a unique opportunity for timing and recognizing precise phases of a cardiac cycle (relaxation) and for the evaluation of AVC. End-systole coincides with AVC and can be visualized in the parasternal or apical long-axis view or by detecting the closure click on the spectral tracing of the pulsed-wave Doppler of aortic valve flow .
Rotation is the measure of the rotational movement of the myocardium in relation to an imaginary long-axis line from apex to base drawn through the middle of LV cavity 4]. Clockwise rotation is defined as negative, while counterclockwise rotation has a positive value. Twist is the algebraic difference in rotation between the apex and the base. Torsion is the twist normalized for the length of the LV cavity (degrees per centimeter). LV rotation or twisting motion has an important role in LV systolic and diastolic function. Normal values for LV rotation and twist angle have shown high variability (technique used, location of the region of interest, age, and loading hemodynamics of the ventricle). The increase in LV twist angle with age observed in literature can be explained by less opposed apical rotation, resulting from a gradual decrease in subendocardial function with aging. Worsening of diastolic relaxation and reduced diastolic suction is, however, associated with an early reduced and delayed diastolic untwisting.
Myocardial strain and Strain Rate (SR) are sensitive parameters for the quantification of diastolic function. Diastolic SR signals can be recorded during isovolumic relaxation, during early filling, and in late diastole. The hemodynamic determinants of protodiastolic strain rate include LV relaxation, regional diastolic stiffness, systolic function, end-systolic wall stress, and filling pressures. In addition, protodiastolic strain rate can assess interstitial fibrosis and can be used to identify viable myocardium after stunning and infarction. Measurement of diastolic strain and strain rate may be useful for research applications but is presently not recommended for routine clinical use.
The detection of myocardial fibrosis and viability depends on the evaluation of myocardial characteristics and shape during the cardiac cycle. Fibrotic tissue may be focal (as occurs in patients with myocardial infarction [MI]) or diffuse (systemic or metabolic disturbances). Fibrosis is actually accurately identified using myocardial late enhancement or T-weighted mapping with cardiac magnetic resonance imaging (MRI), but speckle tracking (especially systolic and protodiastolic strain rate) has a good correlation with tissue fibrosis, evaluated via cardiac magnetic resonance or biopsy.
All these parameters can be measured not only for the LV but also for the right ventricle (RV) and left and right atria (LA and RA, respectively), but have not been fully validated and, still together, commercial applications to process these chambers do not exist.
Timing peak strain is pivotal in defining dyssynchrony as well as for the evaluation of ischemia (post-systolic thickening or shortening).
2.4. Image acquisition
Gated images are obtained during end-expiratory breath holding with stable electrocardiographic traces, avoiding foreshortening of the ventricle and proper visualization of endocardial border. Images acquired should be of high quality. Optimal frame rate should be 60–110 frames per second (FPS). The operator should keep the sector width and depth minimal to focus on the structure of interest. Usually, three consecutive cardiac cycles are obtained and the values averaged for the final processing. Low FPS limits tracking efficacy, while higher FPS “smooths” speckle pattern and the final quality of the analysis. Apical four-chamber, two-chamber, and three-chamber views are necessary for estimation of LV strain and strain rates by 2D STI. This finally offers global longitudinal strain (GLS) value, that is, the average of longitudinal strain for all segments in all views. Parasternal short-axis views (basal, papillary muscles, and apex) are necessary for radial and circumferential strains (finally averaged in global radial and circumferential strain) and strain rates as well as for rotation, twist, and torsion analysis. The ways myocardial segments are divided widely vary among vendors, but in general, a 16–18-segment LV model is used. Myocardium is divided into six segments: basal septal, mid septal, apical septal, apical lateral, mid lateral, and basal lateral. For the timing determination of cardiac events, mitral inflow and LV outflow velocities are recorded using pulsed-Doppler echocardiography and the aortic and mitral valve closure/opening (AVC/O and MVC/O, respectively) times are obtained, as well as visually (AVC in apical long-axis view) or semi-automatically (evaluation of CAMM). The recordings are analyzed offline using semi-automated computer software for estimation of strain and strain rate by 2D STI. A region of interest (ROI) has to be outlined manually, tracing the endocardium. The epicardium is automatically traced by the system, but the wall thickness can be manually adjusted.
The ROI is defined at end-diastole by :
Each of these contours can be user defined or generated automatically.
Topographic definitions of the myocardial ROI in apical views are :
Vendors have incorporated tools to help users identify tracking reliability. Various methods are utilized. Some vendors have introduced protocols that identify segments where tracking is suboptimal and is excluded from the final results. In addition, some vendors provide accuracy indices to guide the user in tracking performance estimates.
Longitudinal strain is more robust and reproducible than other parameters. The values tend to be partly different for different walls and segments. There is a gradient of longitudinal strain values from base to apex (higher values for apical segments) as well as from endo to epicardium (higher values of strain in the subendocardial region). Table 2 depicts the recently published data on normal values for different strains of LV, while Table 3 depicts the principal advantages and pitfalls of different strain imaging techniques .
21 ± 4
24 ± 6
39 ± 16
19 ± 4
22 ± 7
37 ± 18
17 ± 4
21 ± 7
37 ± 17
21 ± 7
22 ± 6
37 ± 17
19 ± 6
24 ± 6
37 ± 19
19 ± 6
26 ± 11
39 ± 15
A recent meta-analysis identified normal values for strain as (GLS) -19.7% (95% CI, −20.4 to −18.9%), global circumferential strain (GCS) −23.3% (95% CI, −24.6 to −22.1%), global radial strain (GRS) 47.3% (95% CI, 43.6–51.0%). Age, gender, body mass index, systolic blood pressure, frame rate, and equipment vendor were considered the variables most likely to influence GLS. In a general linear model, only mean blood pressure was independently associated with higher values of strain. The differences in each strain component are probably actually linked to technical motives: the superiority of longitudinal strain is linked to the reliability of measurements in the axial plane respective to azimuthal one; the variability of radial strain may reflect the limited amount of tissue to track in the short-axis view of the non-hypertrophied heart; the ROI, which is user defined, may affect the strain amplitude . 2D strain parameters have been validated against tagged MRI studies and sonomicrometry studies [22–24]. 2D strain data correlate well with TDI-derived ones, although with higher strength and reproducibility .
|TDI||2D STI||3D STI|
|Temporal resolution (strain curves)||+++||+||–|
|Spatial resolution (strain curves)||++||+||+|
|Validation: in vitro||+(+)||+(+)||+|
|Validation: in vivo||++||++||++|
|Validation: other techniques in clinical scenario||++||+++||+|
|Defined normal values||+++||+++||–|
2.5. 3D strain
With developments in ultrasound transducer technology and both hardware and software computing, systems capable of acquiring real-time volumetric LV data are now widely available. Reasonable spatial and temporal resolution of 3D data sets can now be achieved. The ability to estimate true 3D myocardial motion and deformation using various STI approaches may provide cardiologists with a better view of regional myocardial mechanics, which may be important for diagnosis, prognosis, and therapy. These 3D approaches can measure all strain components in all LV segments from a single acquisition [25–27]. Furthermore, they are angle independent, do not suffer from strain estimation errors associated with out-of-plane motion, and may in theory allow more precise calculations of LV twist and assessment of shear strain components [28–31]. This tool is promising for the evaluation of deformation parameters, although only preliminary data are available. A single apical full-volume acquisition is performed, according to standard modalities, with an FPS between 18 and 25 (3D temporal resolution is lower than that obtained with 2D images) . This avoids multiple acquisitions making readily and instantaneously available evaluation of strain parameters and torsion. The operator is able to limit foreshortening and properly identify walls and segments [33, 34]. 3D strain offers a combined assessment of longitudinal and circumferential strain . To evaluate transmural (radial) deformation, due to image quality and tracking limitation, a derivative parameter, area strain, has been introduced. However, it is important to note that full volume is the result of a stitching process, which can limit tracking of speckles. Frame rate and lateral resolution can also limit good tracking.
2.6. Recent advances and consensus: need for standardization
Recognizing the critical need for standardization in strain imaging, in particular in order to derive a common standard for GLS, the most affordable parameter, in 2010, the European Association of Echocardiography and the American Society of Echocardiography (ASE) invited technical representatives from all interested vendors to participate in a concerted effort to reduce intervendor variability of strain measurement [36, 37]. In order to obtain a perfectly defined strain, synthetic ultrasound images simulated from mathematically modeled ventricles (phantoms) were developed. Jan D’hooge and colleagues from the University of Leuven generated cine loops mimicking normal, hypertrophied, and dysfunctional ventricles (Figures 1–3), and provided them to the vendors: after several attempts, results were similar for the principal vendors.
Moreover, a great effort has been made to standardize, speedup, and automatize (less subjective approach) GLS calculation, in order to offer immediate results to clinicians (bedside) and avoid errors in calculations (heart rate variability).
2.7. Clinical applications
|Field of Application||Explanation|
|Myocardial ischemia||Reduction in strain by 2D STI more objective than WMSI.|
Longitudinal, radial, and circumferential strain reduction in ischemia.
|Myocardial infarction||Differentiation of transmural from subendocardial infarction.|
|Myocardial viability||Objective evaluation during stress echo.|
Differentiation of active contraction from passive tethering.
|Heart failure with normal LVEF (HFnEF)||Twisting/untwisting.|
|Cardiac resynchronization therapy (CRT)||Longitudinal strain from TDI velocity with 2D STI radial strain.|
Longitudinal strain delay index.
Radial strain and survival.
|Takotsubo cardiomyopathy||Impaired longitudinal strain.|
|Restrictive cardiomyopathy||Impaired longitudinal deformation and twist.|
|Constrictive pericarditis||Impaired LV circumferential deformation and torsion.|
|Detection of subclinical myocardial disease||Systemic hypertension, diabetes mellitus, systemic sclerosis, amyloidosis, and Duchenne’s muscular dystrophy.|
|Valvular heart disease||Decreased radial, circumferential, and longitudinal strain in patients with severe aortic stenosis despite normal EF. Septal strain and mitral regurgitation.|
|Congenital heart disease||Right ventricular longitudinal strain and strain rate.|
Moreover, a recent meta-analysis presented the incremental value respective to EF retained by GLS . It is essential to understand that technology development has today made available a fast, objective (automatized), and standardized definition of GLS, with final representation of bull’s-eye plot of longitudinal strain value making it appealing, easily recognizable, and aligned with standardized segmentation of LV wall (Figures 4 and 5).
2.7.1. Ischemic cardiomyopathy
In coronary artery disease, an assessment of myocardial ischemia other than simple visual wall motion score estimation (high inter-observer variability) has been invoked from a long time. Regional velocities (TDI) as well as peak systolic strain rate and systolic strain (TDI and STI) reduce linearly with a reduction in regional perfusion, time-delay prolongation to treatment, and the presence of fibrosis [14, 50, 51]. Subendocardial longitudinal fibers are the most vulnerable ones, resulting in an early deterioration of longitudinal strain, followed by radial and circumferential strains. Anyway, it’s rather improper to consider a “radial function,” because there are no myocardial fibers going in the radial direction. Actually, wall thickening measured by radial strain is a function of wall shortening, as the heart muscle is incompressible. At the same time, circumferential strain does not reflect circumferential fiber contraction because circumferential shortening is mainly due to the inward movement of mid-wall circumference as the wall thickens (inwards—as described by the eggshell model) [52–54]. Briefly, there would have been circumferential shortening even without circumferential fibers. These aspects, together with less standardized values than GLS, make GRS and GCS considerably reliable.
Dobutamine stress echocardiography is an important area of interest because in normal tissues an increased deformation occurs (continuously increasing strain/strain rate) as long as filling is not reduced by increased heart rate. On the contrary, acutely ischemic tissue during stress test shows less deformation and post-systolic deformation (PSD, thickening/shortening with radial/longitudinal strain, respectively), that is, the continued contraction of the myocardium after AVC. PSD is a common finding in myocardial ischemia. All these alterations are proportional to the severity of ischemia and persist in the experimental setting for up to 2 h after the ischemic insult resolution, with a peculiar time decay [55–57]. A noteworthy fact is that the stunned myocardium is characterized by decreased systolic deformation and PSD at rest, but almost normal systolic deformation and disappearance of PSD with dobutamine . This behavior could be secondary to the heterogeneous contractile properties of the myocardium, probably linked to myofibrillar edema reducing the effective force myocardium can develop . Furthermore, interstitial myocardial edema results in a sudden and temporary increase in end-diastolic wall thickness (this behavior is observed also in infarcted segments at the moment of reperfusion) [59, 60]. In chronic infarction, dobutamine is associated with low or no deformation increase, depending on the fibrosis extension (from subendocardial to transmurality involvement) .
2.7.2. Volume overload
Deformation is also closely related to ventricular geometry. Dilation is the end stage of most of the cardiomyopathies and heart valve diseases because for a given volume, the object with the smallest surface area is the sphere. This means that the same deformation (determined by the contractile force) can generate a larger stroke volume in a dilated heart. Similarly, in a dilated heart the same amount of stroke volume can be generated with less contractility and less deformation, that we can directly evaluate with strain(-rate) . Remodeling in the long term will lead to irreversible damage to the heart muscle and finally ventricular dysfunction, but there is no specific diagnostic method to detect subclinical changes in systolic function. STI and TDI might potentially be useful in detecting subclinical changes in cardiac function .
Mitral regurgitation (MR) is a typical volume overload condition. Primary MR leads to cardiac remodeling, increased left ventricular filling pressure, pulmonary arterial hypertension, and myocardial dysfunction. Conventional 2D, M-mode, and Doppler examination play a critical role in the initial and longitudinal assessment; anyway, most variables are load dependent and both afterload and preload are altered during the disease course. TDI and STI provide new parameters to assess regional and global myocardial performance and may help in identifying asymptomatic patients and choosing the optimal time for surgical correction. It is worthy to note that patients with severe primary degenerative MR may have near-normal left ventricle ejection fractions (LVEFs) because of disproportionately higher compensation in GLS. Moreover, the higher the GLS, the higher the risk of substantial reduction in LVEF (>10%) during the immediate postoperative period . On the contrary, in patients tested at 6 months after surgery, when LV reverse remodeling has already settled, LVEF reductions >10% were associated with lower baseline strains . Chronic ischemic MR instead is not a valvular disease per se but is rather a “ventricular disease.” In particular, inferior MI has been recognized as the most frequent cause of ischemic MR because of the geometric distortion in the papillary muscle-bearing segments . Therefore, the site of MI might be a more important determinant of MR degree in LV dysfunction than the extent of post-MI LV remodeling. Under normal conditions, the basal rotation, which is determined mainly by inferior and posterior myocardial segments , shortens the distance between the mitral valve and the head of the papillary muscles, contributing to MVC and counterbalancing the tethering forces. When local remodeling occurs, as in patients with inferior-posterior MIs, the basal strain and basal rotation are significantly lower and fail to shorten the distance between the papillary muscles and the mitral valve. Interestingly, Zito et al. demonstrated that basal rotation but not basal strain was an independent predictor of the severity of MR .
2.7.3. Pressure overload
Considering pressure overload pathology, aortic stenosis (AS) is the most common valve lesion causing chronic pressure overload on the LV. The development of symptoms in AS heralds a malignant phase of the condition and prompt aortic valve replacement results in a clear reduction in mortality . In contrast, the management of patients with severe AS in the absence of symptoms remains one of the most controversial and debated areas in modern cardiology [67, 68].
The increased afterload leads to left ventricular hypertrophy and the basal septal is the first to show changes, due to increased wall stress according to the Laplace law . It first shows a decrease in strain(-rate) and the development of PSD is observed, as well as the development of localized hypertrophy (septal bulge) . Recently, GLS has been shown to be an independent predictor of outcomes in patients with severe asymptomatic AS, incremental to other echocardiographic markers . Not to forget, the role of exercise testing in asymptomatic AS is well established, and recommended by guidelines in equivocal cases .
2.7.4. Mechanical dyssynchrony
Searching for the presence of mechanical dyssynchrony to identify potentially recruitable function, rather than looking only for electrocardiogram (ECG) manifestations of ventricular conduction delay, could increase the rate of cardiac resynchronization therapy (CRT) responders . Mitigation of intraventricular dyssynchrony is currently thought to be the primary mechanism of improved myocardial performance with CRT. Anyway, many patients eligible for CRT have dilated ventricles with complex motion, especially if infarcted areas are present. Moreover, in dilated hearts local motion is importantly influenced by other myocardial segments and even by right ventricular motion . That’s why to this day, M-mode, 2D e TDI analyses have expressed modest sensitivity and specificity in measuring dyssynchrony and improving patient selection for CRT .
STI-based methods, mostly assessing the time difference between maximal values form different myocardial segments (e.g. septal to posterior wall motion delay , septal rebound stretch , wasted work ratio) [77, 78], are promising, but they must be validated in multicenter randomized trials.
2.7.5. Diabetic cardiomyopathy
Another field of interest is represented by diabetes mellitus because studies showed that myocardial damage occurred in at least 30% diabetic patients when LV diameter and LVEF were normal . Diabetic myocardial diastolic dysfunction seems to precede systolic dysfunction, but this might be explained by the insensitivity of techniques for detecting LV systolic function. In many but small studies, longitudinal dysfunction (segmental GLS) occurred in early stages of diabetes, while LV torsion increased compensatively [80–84].
2.7.6. Cancer therapeutics-related cardiac dysfunction
Cardiotoxicity from cancer therapy has become a leading cause of morbidity and mortality in survivors [85, 86]. A careful consideration of potential cardiotoxicity during therapy and a focus on early detection and intervention are developing. Echocardiography is the cornerstone in the cardiac imaging evaluation of patients in preparation for, during, and after cancer therapy because of its wide availability, repeatability, versatility, lack of radiation exposure, and safety. The most commonly used parameter for monitoring LV function with echocardiography is LVEF. In addition, the calculation of LVEF should be combined with assessment of the wall motion score index . Anyway, 2D echocardiography appears to be reliable in the detection of differences close to 10% in LVEF . Because this is the same magnitude of change used to adjudicate cancer therapeutics-related cardiac dysfunction, the sensitivity of LVEF has been questioned. Moreover, detecting a decreased LVEF after anthracyclines may be too late for treatment , suggesting that more sensitive parameters of LV dysfunction could be helpful. The prognostic value of early measurement of systolic deformation indices (above all ΔGLS) measured in the prediction of subsequent LV systolic function has been evaluated in several studies, both in animals  and in humans [91–94].
2.7.7. Left atrium
Speckle tracking was recently applied to study the myocardial mechanics of a thin-walled structure such as the LA [95–98]. For the analysis, apical views are obtained using conventional 2D gray-scale echocardiography, during a breath-hold, with a stable electrocardiographic recording. The frame rate is set between 60 and 80 frames/s, and recordings are processed using acoustic-tracking software. The LA mechanical indices are calculated by averaging values observed in all LA segments (global strain) with a 15-segment or a 12-segment model. The software generates longitudinal strain and strain rate curves for each atrial segment. The radial deformation cannot be calculated because the LA wall is thin and the spatial resolution is limited .
2.7.8. Right ventricle
A recent methodological study has reported the feasibility, the reference values, and the reproducibility of right ventricular longitudinal strain measured by STI in normal patients and in patients with RV dysfunction . The technique is similar as for LV: global strain is the average of six single segments (ROI) traced semi-automatically and processed by software packages (today, a dedicated software for RV does not exist). The evaluation of right ventricular function with STI, considering the important limitations of other parameters and methods, could offer more detailed information about regional and global RV mechanics with important clinical implications for non-invasive evaluation of RV systolic function (subclinical RV dysfunction). Further prospective studies are necessary to define its role in the management of patients.
Rotation, deformation, and out-of-plane motion can cause speckle patterns to change between acquisitions (decorrelation). Loss of tracking can be limited acquiring at a proper frame rate. Image artifacts should be avoided. In general, high-quality acquisition is a prerequisite for optimal speckle-tracking results .
Among limitations of the method, we should include:
Lack of reproducibility: even if much less compared with TDI, every STI study should include an intra- and interobserver variability testing ;
Intervendor variability: lack of standardization results in changes in reference values ;
Oversimplification due to software processing algorithms: automatic tracking of epicardial border (assumption of uniform thickness of myocardium); averaging of parameters within a segment; drifting;
2D imaging limitations: image quality; artifacts; image dropout; frame rate (low frame: no tracking, as at high heart frequencies; high frame: limited lateral resolution); lateral resolution and depth; depth dependence of transverse tracking; out-of-plane motion in short axis limiting tracking of the same ROI; noise.
Evaluation of radial strain poses important technical challenges compared to longitudinal one.
This is due to:
Measurements in the axial plane are more reliable than those that depend on lateral and elevation (or azimuthal) resolution;
Limited amount of tissue to track in the short-axis view of the non-hypertrophied heart;
Placement of the ROI is user defined;
Speckle tracking is actually a reality in echocardiography. Simple protocols of acquisition and novel processing packages have made available deformation analysis in daily clinical arena.
Overcoming many of the previous limitations, thanks to technological development, including 3D introduction and STI, offers to the cardiologists the potential benefit of a solid, fast, easy, and reproducible quantization of myocardial mechanics .