1 - Strain imaging applications and techniques

1

1Strain Imaging Applications

and TechniquesThomas H. Marwick, MBBS, PhD, MPH, FACC,

and Wojciech Kosmala, MD, PhD

BACKGROUNDThe assessment of left ventricular (LV) function is a fun-damental requirement in many clinical situations in car-diology. Numerous techniques are used for this purpose, and a number of parameters can be employed, although ejection fraction (EF) is by far the most established. Echocardiography is the most widely used technique for the assessment of LV function because it is widely avail-able and relatively inexpensive and can be taken to the patient’s bedside. The problem is that echocardiographic assessment of EF is subject to a number of limitations, both technological and physiologic (Table 1.1), which particularly compromise the test-retest variation of this

method.1 The development of strain as a clinical tool over the past 30 years has provided a technique that is accurate and reliable. The year 2020 represented a culmination of this work to produce an additional assessment of LV function because a new Current Procedural Terminology (CPT) code (93356) was introduced for the application of strain in clinical practice. The purpose of this chapter is to provide the technical background regarding this modality, summarize its potential applications, and pre-pare the reader for subsequent chapters about the main clinical applications of myocardial strain.

The adoption of a new tool can be difficult. Funda-mental to finding the motivation for this effort is a recognition of the limitations of assessment of both global and regional function using current techniques (see Table 1.1).1 There are two aspects of a desirable mea-surement: validity and reliability (Fig. 1.1). Validity (or accuracy) pertains to the comparison of the measure-ment against some external standard. The accuracy of strain has been validated experimentally against in vivo measurement of tissue excursion with sonomicrometry and clinically against magnetic resonance tagging tech-niques.2-5 Although validity is important to clinicians, in most circumstances, we are less worried about consistent bias (when the measurement is within, for example, 5% of the reference standard) but more concerned about random variation of the measurement that would pre-clude our ability to compare individuals or compare measurements at different points of time. This is a reflec-tion of reliability or precision. In sequential testing, the tool has to be sufficiently sensitive to pick up small differ-ences, and reliability can be measured as the ratio of true variance and true1error variance. What is important to

1

Global (EF) Regional (WMSI)

Variability VariabilityInsensitive to mild

dysfunctionDistinction of infarcted and

ischemic myocardiumInexact in LVH Inadequate in LBBBLess precise in tachycardia Less accurate in severe LVDInability to assess RV and

atrial functionInability to assess

diastolic function

TABLE 1.1 Limitations of Current Echocardiographic Techniques for the Assessment of Global and Regional Left Ventricular Function

EF, Ejection fraction; LBBB, left bundle branch block; LVD, left ventricular dysfunction; LVH, left ventricular hypertrophy; RV, right ventricular; WMSI, wall motion score index.

2 CHAPTER 1 Strain Imaging Applications and Techniques

realize is that both validity and reliability of existing tech-niques are inadequate and especially of limited reproduc-ibility. For example, the 95% confidence intervals of two-dimensional (2D)–EF measurement exceed 10%,6 implying that an apparent reduction of EF from 54% (normal) to 45% (mildly impaired) may simply represent the variability of the measurement. Although three-dimensional (3D)–EF has less variability, it probably still exceeds 5%.7

THE PHYSICS OF STRAIN

DefinitionStrain represents a fundamental property of matter, namely its deformation in response to an applied force. This force might be extrinsic to the heart, which is funda-mental to the performance of elastography (Chapter 10). However, most commonly, it is intrinsic to the heart, based on myocardial contraction associated with every heartbeat. Because the myocardium is assumed to be in-compressible, shortening in one direction is associated with thickening in the orthogonal plane.

Lagrangian strain () is defined as the difference be-tween myocardial length at baseline (L0, usually end dias-tole) and myocardial length at the end of a time interval (L, usually at end systole) divided by L0. This corresponds to the strain measured by speckle strain (Fig. 1.2):

�Lagrange =L L

L� 0

0

In contrast, if strain is calculated by comparing instanta-neous length change during deformation over an infini-tesimal small time interval (dt), this strain is termed natural strain, and it corresponds to tissue velocity strain, in which initial segment length is not used as a reference2:

� � � � N t

tt L t dt L t

L t( ) ∫ ( ) ( )

( )0

In this case, the total amount of strain represents an aggregate of all infinitesimal strain contributions between t0 and t.

If the strains are small, Lagrangian strain approximately equates to natural strain, but deformations .10% (such as are observed in the cardiac cycle) using each approach are different. The majority of currently used imaging techniques (speckle-tracking echo cardiography, cardiac magnetic resonance, nuclear) measure Lagrangian strain.

Pro

babi

lity

dens

ity

A

BC

D E

Reference value

Accuracy

Precision

Low precision,low accuracy

Low precision,high accuracy

High precision,low accuracy

High precision,high accuracy

Value

+2% +2%

+2%

+1%

–1%

–2%–3%

+2%

+1%

–1%

–2%–3%

+1%

–1%

–2%–3%

+1%

–1%

–2%

+3% +3%

–3%

+3% +3%

X

XX

X

XXX X

X

X XXXX

XX

X

X

X

XX

X

X

XX X

X

Fig. 1.1 Cornerstones of imaging measurement—validity and reliability. These parameters pertain to the distribution of measurements from the reference “truth” and from each other (A). The schematics illustrate the deviation of measurements as an overestimation or underestimation of the “true” mea-surement in black at the center of the bull’s eye. (B) The sce-nario of low precision (discordance between measurements) and low accuracy (average displacement of 2% away from the reference). (C) Low precision (discordance between measure-ments), but high accuracy (average displacement overlies the reference). (D) High precision (measurements are concordant) but low accuracy; although imperfect, most clinicians can compen-sate for consistent overestimation or underestimation of function. (E) The ideal scenario of high precision and high accuracy.

3CHAPTER 1 Strain Imaging Applications and Techniques

Strain is a dimensionless parameter, expressed as a percentage. When applied to longitudinal or circumfer-ential shortening during systole, the resulting values are negative, since L is shorter than L0. Conversely, radial thickening is denoted by a positive value.

The measurement of strain takes no cognizance of the timeframe of myocardial motion, although this can be accounted for by assessing strain rate (SR), which repre-sents the average rate of deformation, having the unit s–1:

SR td t

dt( ) ( )

��

Individual myocardial segments may be measured, or strain can be assessed along the length of an entire wall or for the chamber as a whole, the latter being the basis of the most widely used parameter, global longitudinal strain (GLS). Even though this parameter is the most commonly used clinically, it is essential that strain is not used as a black box and that the underlying strain curves are assessed.

Although the expression of longitudinal strain (short-ening) as a negative number is completely correct from the standpoint of the underlying physics, it has led to significant confusion in the interpretation of strain by nonexperts. As strain moves from being the domain of researchers and cardiologists to other clinicians (e.g., oncologists in the setting of testing for cardiotoxicity), this has posed a significant problem. Recently, some authors have started to use an absolute (positive) value of GLS in reporting their data.8,9 A recent opinion piece from the editors of an imaging journal proposed that to increase the understanding of GLS, it should be quoted as an abso-lute, positive value,10 perhaps labeled as global longitudi-nal shortening. There is a precedent in the expression of GLS as a positive number because if EF were calculated in the same way, it also would be a negative parameter. This is a matter of ongoing discussion, but at present, the guidelines support its expression as a negative number.

THE PHYSIOLOGY OF STRAINThe three dimensions of LV and right ventricular (RV) strain calculation are longitudinal (obtained from the apical views), circumferential (obtained from short axis views), and a measure of myocardial thickening (radial in the short axis and transverse in the apical views). Of these three measurements, longitudinal strain is the most robust and most widely used. There seems to be little incremental information from using circumferen-tial strain, although it has been used to predict LV re-modeling.11 Radial and transverse strain are the least reliable, and one of the reasons for this might be that the amount of myocardium used for this calculation is sig-nificantly less than the others. Accordingly, the use of radial strain is not recommended. Moreover, although total deformation is quantified by measuring the three sheer strain components, sheer is currently neglected in clinical practice, with the possible exception of torsion.

Myocardial fibers are arranged in a helix so that contraction leads to longitudinal shortening as well as radial thickening. This orientation provides a mechani-cal advantage, whereby approximately 10% shortening of each myofibril is able to translate to a 50% EF (Fig. 1.3). There are multiple myocardial layers, with predominantly longitudinal fibers in the subendocar-dium and epicardium, and predominantly radial fibers in the midmyocardium. These obliquely orientated myofibers generate an apical counterclockwise twist

Strain: dimensionless index ofchange in length

L

Strain (�) = L–L0/L0

L0

Fig. 1.2 Definition of strain. Strain is the ratio between the initial length and the final length of the measured structure.


Although LV-GLS and LVEF are both ejection phase markers, changes in each do not necessarily correlate. The reason is that EF is associated with GLS, global cir-cumferential strain (GCS), wall thickness, LV size, and diastolic volume.15 Although LVEF is proportional to the square of GCS, it is only linearly related to GLS. LVEF is inversely related to LV radius, so smaller LV cavities and LV hypertrophy (LVH) are associated with preserved EF, even if GLS is impaired. The implications of these rela-tionships are particularly important in valvular heart disease and are discussed in more detail in Chapter 7.

ECHOCARDIOGRAPHIC STRAIN METHODOLOGIES

Tissue Velocity ImagingThe fundamental work that underpins the assessment of strain is decades old and is based on observations from animal models using microcrystals. The initial application of this parameter in echocardiography was undertaken using tissue velocity imaging in the 1990s. Fundamentally, this technique examined the difference in velocity along

and a basal clockwise twist driving torsional ventricular contraction.12,13 Torsion correlates with untwist, which contributes to isovolumic pressure decay, generating an intraventricular pressure gradient responsible for dia-stolic suction and facilitation of LV filling (Fig. 1.4). Two observations are important in relation to this. First, the direction of measurement of strain pertains to the direction of the wall rather than the direction of the fibers themselves, so it is important to consider that these measurements are an approximation. Second, although there is much interest in the separate measure-ment of subendocardial and subepicardial function, this is dependent on the assumption that there could be differential movement of these components, which is unproven.

Strain is an ejection phase index and is therefore sus-ceptible to loading, particularly afterload, and to a lesser degree to preload and heart rate, both of which involve LV cavity size. Because of this contribution from loading, strain is not an optimal index of contractility. In fact, the closest deformation parameter to this is strain rate, which has been shown to correspond to dP/dt and thus better reflect the inotropic state of myocardium.14

�SL � 13%�epi � –7%

�mid � –15%

�endo � –26%�rad � �37%

EF � 60%

Diastole

Diastole

Systole

Systole

2.07 �

1.81 �

20 mL50 mL

1.0 cm1.4 cm

Fig. 1.3 Applied physiology of the transition from fiber shortening to ejection fraction (EF). A 10% change in fiber length translates to an epicardial strain (eepi) of 27%, midwall strain (emid) of 215%, and endocardial strain (eendo) of 226%, with a radial strain (erad) of 37% and EF of 60%.


the region of interest, providing strain rate, which was integrated to provide the assessment of strain (Fig. 1.5). The strength of this approach was very high temporal resolution, which allowed reliable assessment of strain rate, as well as recognition of transient events. However, the weakness was that Doppler is fundamentally directional, so deviation from parallel alignment of the ultrasound beam and the wall could lead to significant underestimation of deformation. A second problem was that the technique was very susceptible to signal noise.

2D Speckle TrackingThe clinical application of strain became substantially more feasible with the development of speckle tracking in the early 2000s.16 The underlying principle behind this technique is similar to the optical mouse on a computer: The myocardium is characterized by a “texture,” created by speckles, which are persistent artifacts. As the myocar-dium moves, it is possible to track the speckles from frame to frame and therefore assess myocardial deforma-tion in relation to the original measurement. Inherent in

0

2.5

5

7.5

–15–10–50Peak LV untwisting, rad/s

IVP

G, m

mH

g

y = –0.21x – 0.11r = 0.78

y = –0.44x + 0.87r = 0.72

–15

–10

–5

00 20 40 60

Peak LV torsion, deg

Pe

ak

LVun

twis

ting,

ra

d/s

Torsion vs Untwist

Untwist vs IVPG

Counter-clockwise

Systole DiastoleClockwise

Time

Apical

Basal

Rotation as viewed from apex

Contraction

S LA

P Contract

Fig. 1.4 Measurement of left ventricular (LV) twist and untwist with circumferential strain at different myo-cardial levels. Torsion correlates with untwist, which is a marker of LV suction.

Velocity(cm/s)

Displacement(cm)

Strain(%)

Spa

tial

deriv

atio

n

Spa

tial

deriv

atio

n

Spa

tial

inte

grat

ion

Spa

tial

inte

grat

ion

Temporalderivation

Temporalderivation

Temporalintegration

Temporalintegration

Strain rate/s)

Fig. 1.5 Relation between displacement, velocity, strain, and strain rate. The mathematical relationship between these parameters is based on temporal and spatial integration and differentiation.


a frame rate of 50 to 90 frames per second. Although circumferential strain requires the acquisition of short-axis images, most clinical strain is performed with GLS, and this is derived from apical four, two, and long-axis views. Although there have been efforts to calculate strain from a single view, the resulting management risks greater levels of variability than from averaging three views19 and is therefore not recommended. Atten-tion to detail is important (Table 1.2). Detailed practical and technical guidance relating to strain measurement has been recently published.20,21

The dependence of strain measurement on the ability to accurately track the myocardium means that it is not feasible in every case, in a similar way that measurement of EF using the Simpson biplane technique is not always

this approach is the assumption that the speckles do not move out of plane. However, as the myocardium twists during systole, speckles do move out of plane, implying that late systolic and diastolic measurements are less reli-ably measured than measurements at the beginning of systole. Using standard 2D imaging, speckle imaging is performed at a frame rate of 50 to 90 frames per second. In many instances (e.g., during the tachycardia associated with stress echocardiography), this leads to undersam-pling and hence underestimation of strain and strain rate. New equipment is being developed that will improve temporal resolution by an order of magnitude, and this will permit interesting new applications, such as elastog-raphy, but at a cost to image quality.

3D Speckle TrackingThe 3D technique has the benefit of being able to track speckles out of plane and should therefore be able to pro-duce more reliable signals, particularly in late systole and diastole. However, the temporal resolution of 3D is lower than that of 2D imaging, with most 3D datasets recorded at 20 volumes per second. Generally there is a tradeoff between temporal and spatial resolution, with the latter suffering as temporal resolution increases. At present, al-though 2D speckle is imperfect, it is the most feasible and robust means of performing deformation imaging.

Multilayer StrainThere is a normal gradient of strain from the epicar-dium (lower) to the endocardium (greater), and multi-layer strain analysis may identify disturbance of this relationship, providing information about various dis-ease entities.17 However, although the myocardium is divided into specific layers, they are bound together by the interstitium, and it seems likely that deformation in one layer affects another. Moreover, as mentioned in relation to radial strain, small regions of interest pro-vide greater variance than larger ones, and strain pro-files seem to be influenced by not only measurement but also the expectations of the software. This point is emphasized by the fact that endocardial circumferential strain in infarcted segments does not become 0%, even in the presence of transmural infarction.18

ACQUISITION AND PROCESSING

Image AcquisitionThe measurement of 2D strain requires acquisition of high-quality images with clear endocardial definition, at

Appropriate depth Appropriate depth, including the mitral annulus in diastole and effective visualization of the leaflet insertion point at annulus. This usually requires inclusion of up to half of the left atrium.

Appropriate sector width

Wide sector, to include (a) the entire apex in diastole, (b) some RV to ensure capture of the entire septum, and (c) epi-cardial borders of the anterior and lateral walls. A wide sec-tor may sacrifice frame rate, but this is acceptable if it re-mains in the range of 40–70 frames/second.

Appropriate gain setting

Slight overgain helps checking endocardial tracking.

Avoid foreshortening

Foreshortening may cause over-estimation of apical strain.

Clear endocardial border

Appropriate endocardial defini-tion is needed, as for perform-ing LV volumes and EF.

ECG tracing A good-quality ECG tracing is critical because QRS and T waves are often used to define beginning and end systole.

TABLE 1.2 Acquisition and Selection of Appropriate Image(s) for Global Longitudinal Strain Measurement

ECG, Electrocardiogram; EF, ejection fraction; LV, left ventricle; RV, right ventricle.


possible. Nonetheless, GLS measurements are feasible in about 90% of patients; as with 2D imaging generally, bigger problems arise from the constraints of lung disease than obesity.

Image ProcessingVendor-based variations in the assessment of strain are perceived as a major impediment to a general use of GLS in clinical practice. They arise almost completely from differences in postprocessing rather than differ-ences in imaging.22 Variations in strain measurements between vendors relate to different approaches to post-processing. Some vendors track the endocardium to provide endocardial strain, which is higher than mid-wall, epicardial, and total strain. Other sources of varia-tion include the automated measurement of peak versus end-systolic strain, which may differ if there is postsys-tolic shortening. Finally, different algorithms initiate the measurement process using different approaches, which also may contribute to variability. The European Asso-ciation of Cardiovascular Imaging (EACVI)/American Society of Echocardiography (ASE) Industry Task Force, involving major echocardiography societies and manu-facturers, has curtailed the degree of variation between manufacturers,23 which seems in the current era to be less than that reported in about 2010.24 It should be kept in mind that test-retest variation (as well as possible differences in software) lead to variations in standard parameters (including dimensions and EF) that exceed the current variations in GLS.

The second explanation for failure to use GLS is that it is perceived as technically challenging or requiring specialist knowledge. Although this was true during the development of deformation imaging from tissue Dop-pler, its derivation from speckle tracking has increased the feasibility of this measurement. Although there is a learning curve for trainees to attain the performance of experts (Fig. 1.6), this process appears to be short.25 Al-though it should not be used without knowledge of its limitations and potential pitfalls, this is true of any imaging modality.

GLS is load dependent, but this shortcoming is in common with all ejection phase indices. Although abso-lute GLS values .18% are normal, like any partition value, there is a “grey zone” of normal strain (16–18%) due in part to the effect of afterload on GLS as well as being age related26 and sex related.27

The first step in postprocessing is tracking the myo-cardium. Modern software will automatically identify

pAdj = 0.003

pAdj = 0.012

pAdj = 0.014

p trend < 0.001p overall = 0.007

p trend = 0.001p overall = 0.018

p trend = 0.001p overall = 0.019

0.0No

No

No Limited Intermed High

Limited Intermed High

Limited Intermed Highn = 23n = 10

MD

SD

CV

n = 12n = 13

n = 23n = 10n = 12n = 13

n = 23n = 10n = 12n = 13

1.0

2.0

3.0

0.0

0

5

10

15

20(%)

1.0

2.0

3.0

Fig. 1.6 The impact of experience on strain measurements. This study of readers, divided into four groups on the basis of their experience, calculated global longitudinal strain (GLS) from speckle strain analysis of nine cases of various degrees of im-age quality. Intraclass correlation coefficients (ICC), mean differ-ence (MD), standard deviation (SD), and coefficient variance (CV) were compared.25 With increasing experience, there was a significant trend toward lower MD, SD, and CV of GLS measure-ments. High, Highly experienced; Intermed, Intermediate ex-perience; Limited, Limited experience; No, No experience.


fiducial points at the mitral annulus and apex, and in-deed a fully automated approach has been described. All software gives an opportunity for user interaction, and checking the adequacy of tracking is an essential step. The second step of postprocessing is to designate the region of interest, and the most common pitfall is to include the pericardium within this region of interest by making this too wide. Inclusion of the pericardium leads to underestimation of deformation. The third step is the recognition of the beginning of systole, and a va-riety of algorithms have been used, including the up-stroke of the R wave and the opening of the aortic valve in the apical long-axis view. Attention to this is impor-tant because subsequent shortening will be calculated from this point. Figures 1.7, 1.8, 1.9, and 1.10 document the postprocessing sequence with a variety of vendors.28

Semiautomated measurements improve the speed and reproducibility of assessment of LV function29-33 us-ing a variety of methods, including EF and strain, but also require supervision by an expert.29-32 However, it is important to realize that although there is a learning curve for GLS measurement, echocardiographers with no experience in strain imaging have a high precision (intraclass correlation coefficient, 0.975; 95% confidence interval [CI], 0.912–0.998), similar to that of expert readers (0.996; 95% CI, 0.988–1.000; p5.0002).25 A fully automated assessment has been described for measure-ment of both LVEF and GLS.33 This kind of approach appears to be reliable for the recognition of normal function; however, it overestimates the number of stud-ies with dysfunction, probably because of the underesti-mation of actual parameter values with poor image quality. Thus, although an abnormal automated test warrants review by a human, the ability to have an initial screening process using an automated approach would be useful in a number of settings, including the detection of cancer treatment–related cardiac dysfunction,34,35 as well as the detection of asymptomatic LV systolic dys-function in the community.36-38 In these situations, an experienced GLS user is not always immediately avail-able for interpretation, and an automated assessment with high negative predictive value may be useful.

DisplayAfter imaging data have been processed, they need to be displayed for interpretation, either as a quantitative measure, waveforms, or a parametric display. The large spectrum of potential measurements is daunting and

View selection

Tracing

Trackingquality

Baseline21%, follow-up 15.8%

Fig. 1.7 Steps for global longitudinal strain (GLS) measure-ment using Automated Function Imaging (AFI; GE Medical Systems).28 Selection of apical views: The apical long axis is the first view for AFI. Go to the “Measurement” menu to choose AFI and begin with the APLAX view. This enables a manual cross-validation of end systole with aortic valve closure.

Tracing: Tracing with AFI occurs automatically in end systole, though manual marking of fiducial landmarks (two annulus and apex) may be necessary with technically difficult images. These points should be made based on the endocardial border rather than myocardium or pericardium. The automated definition of end systole is based on anticipated duration from the QRS wave or the end of the T wave.

Tracking quality: This is the most important step in strain analy-sis. The overlay between the software’s detection of the endo-cardial border and the actual border should be assessed visu-ally. Automatic displays of the adequacy of tracking are based on statistical metrics regarding edge detection within seg-ments, but there is no substitute for visual assessment. If necessary, the border and the width of the region of interest may be adjusted. A broader width may include pericardium causing underestimation of GLS. Strain profiles are often used to check tracking quality. In a ventricle that appears normal, the strain profiles should be roughly similar in timing, morphology, and magnitude—major regional disparities should lead to fur-ther efforts to check tracking.

Integration: AFI gives peak systolic strain in a parametric (bull’s eye) display, but this is only possible if all three views had similar cycle length. In this case, follow-up GLS is reduced (215.9%) and deteriorated from baseline (221%).


to identify aortic valve closure. It has a stronger rela-tion to stroke volume and EF than contractility.

4. Postsystolic strain: the shortening that occurs after the end of systolic ejection, sometimes expressed as postsystolic index (PSI, the ratio between postsys-tolic and total strain). This signal may be a marker of ischemia and viability (Fig. 1.11).

5. Systolic strain rate: This has been reported to be lin-early related to indices of myocardial contractility, such as dP/dt.

may have been a contributor to the slow uptake of the modality.

The most useful magnitude parameters are 1. Peak (maximal) strain: This is the peak value during

the entire cardiac cycle; this is the most widely used, usually averaged across all LV segments as GLS.

2. Peak systolic strain: This is the peak value during systole.

3. End-systolic strain: This necessitates definition of end systole, usually based on the use of spectral Doppler

View selection

Tracing

Trackingquality

Integration

Fig. 1.8 Steps for global longitudinal strain (GLS) measure-ment using AutoStrain (Image Arena, TomTec Imaging Systems).28 Selection of images: AutoStrain by Image Arena, a vendor-neutral platform, is used by importing studies in DICOM format. Double click on the selected view (apical 4 chamber (4CH), 2 chamber (2CH) and 3 chamber (3CH)) to move the im-ages into the workspace.

Tracing: The entire left ventricular (LV) endocardial border is traced automatically, starting and finishing with the mitral valve insertion points. Timing of end systole is automated based on the minimum LV cavity volume, which can also be defined manually by measuring aortic valve closure time using pulsed wave Doppler of the LV outflow track (LVOT).

Tracking quality: The assessment of tracking quality involves visual assessment, and adjustment of the end-diastolic and end-systolic endocardial linear contour may be required.

Integration: After tracking three apical views, regional strains are combined to derive GLS in each view. This integration does not require three views of similar cycle length.

Viewselection

Tracing

Trackingquality

Integration

Fig. 1.9 Steps for global longitudinal strain (GLS) measure-ment using Automated Cardiac Motion Quantification (aCMQ; Philips, Best, Netherlands).28 Selection of images: After acquisition of appropriate images, aCMQ is activated for image processing. This software does not require a particular view to begin.

Tracing: aCMQ also allows an automated tracing or semiauto-mated by marking three fiducial landmarks (two annulus and apex) on the endocardial border in technically difficult images.

Tracking quality: Similarly, the overlay between the software’s detection of the endocardial border and the actual border should be assessed visually. If necessary, the border and the width of region of interest (ROI) may be adjusted. The ability to make this adjustment on segmental ROI is helpful to avoid the pericardial layer and papillary muscle; failure to do this may lead to underestimation or overestimation of GLS.

Integration: After strain profiles are created, the global results tab provides a parametric (bull’s eye) display of peak systolic strain (221%), which is normal.


imaging. As with all timing measurements in echocar-diography, this provides a higher degree of variability than the assessment of magnitude, and the use of speckle strain leads to potential undersampling. Nonetheless, increased LV mechanical dispersion (LVMD) (Fig. 1.12) assessed by speckle-tracking strain has been shown to be associated with ventricular arrhythmias for about a de-cade.39,40 A recent meta-analysis has shown the optimal LVMD cutpoint is between 60 and 70 ms (Fig. 1.13). The predictive value of LVMD was superior to that of LVEF or GLS, and LVMD is on average 20 ms greater in groups with arrhythmias than those without; each 10-ms incre-ment of LVMD was independently associated with ven-tricular arrhythmia events (HR, 1.19; 95% CI, 1.09–1.29; P ,.01).41 Timing parameters have also been used in the assessment of LV synchrony, but their use has been dwarfed by the assessment of magnitude.

In addition to strain curves, a variety of parametric displays can be used to summarize magnitude, space, and time. Examples are the “bull’s eye” map or a parametric curved M-mode display in which strain rates are com-pared in several segments. These M-mode displays are very effective for the measurement of differences in timing of contraction and relaxation, and this assessment is fa-cilitated by marking two mechanical events: aortic valve closure and mitral valve opening. We mark the time of aortic valve closure from the spectral Doppler display of LV outflow. Mitral valve opening timing is marked on the transmitral Doppler image as the beginning of the E wave.

DEFINING THE NORMAL RANGENormal myocardium shortens by approximately 20%. However, definition of the normal range provides some specific challenges. First, it cannot accurately be identi-fied from normal echocardiograms gathered in a clinical echocardiography laboratory because these individuals undergo testing for the elucidation of symptoms. Prefer-ably, normal populations (i.e., comprising normal vol-unteers), specifically recruited as healthy referent groups, might serve for this purpose. Meta-analysis of GLS in different patient subsets has defined the distribution of mean values in these populations but has not accurately defined the variance in normal GLS.42,43 The difficulty in defining the lower limit of normality has been one factor that has delayed uptake of GLS into clinical practice.

The normal range of GLS has been addressed in a re-cent individual patient meta-analysis of reported results after 2011, to supersede the previous meta-analysis43 and

6. Diastolic strain rate: The diastolic profile reflects two phases, passive and active filling.2 However, the timing and magnitude of this measurement is nonuniform.

7. Isovolumic contraction and relaxation velocity waves: These may invert from negative to positive during ischemia.As well as assessing the magnitude of contraction,

timing parameters can be assessed using deformation

Viewselection

Tracing

Trackingquality

Integration

Fig. 1.10 Steps for global longitudinal strain (GLS) measure-ment using Velocity Vector Imaging (Siemens, Mountain View, CA)28 Selection of images: After acquisition of appropriate apical images, a right click on the selected view permits selec-tion of the velocity vector imaging (VVI) function. This does not require a specific initial view. The definition of end systole in VVI is based on electrocardiogram (QRS and T wave).

Tracing: Manual tracing the entire left ventricular (LV) endocar-dial contour is necessary, and care should be taken to begin and end the tracing on the mitral valve leaflet insertion points. Inclusion of the papillary muscles may cause overestimation of GLS. Tracing is often done in end systole, and manual adjust-ment of timing of tracing is possible.

Tracking quality: Visual assessment of tracking quality in VVI can be challenging compared to AFI or aCMQ because it is a linear contour. “Endo vector” can be turned on or off to facili-tate this process visually. During processing, slight “over gain” of the images may optimize endocardial tracking, and freezing and “stepping through” freeze-frame images is essential.

Integration: After tracking three apical views, regional strains are combined to derive GLS based on either peak or peak systolic strain. The bull’s eye can be displayed in recent versions of the software and does not require three views of similar cycle length.


*

*

** *

**

****

*

*

*

**

**

*

* * *

*

*

*

*

*

*

*

*

*

Max

imal

sys

tolic

stra

in r

ate

(s–1

)12

80

60

40

20

8

4

0 2000 4000dP/dtMAX (mm Hg • s-1)

6000

Sys

tolic

str

ain

(%)

Baseli

ne

Ischa

emia

Stunn

ing

Baseli

ne

Recov

ery

10 d

ays

DobutaminedP/dtMAX (mm Hg • s-1)0 2000 4000 6000 S

ysto

lic th

icke

ning

(%

)

DobutamineIsc

haem

ia

Stunn

ing

Baseli

ne

Recov

ery

Recov

ery

10 d

ays

Dobutamine

10

6

2

80

60

40

20

80

60

40

20

80

1010

6

2

60

40

20

80

60

40

20

80

60

40

20

90

70

50

Max

imal

sys

tolic

stra

in r

ate

(s–1

)S

ysto

lic s

trai

n (%

)

Max

imal

sys

tolic

stra

in r

ate

(s–1

)S

ysto

lic s

trai

n (%

)S

ysto

lic th

icke

ning

(%

)

Max

imal

sys

tolic

stra

in r

ate

(s–1

)S

ysto

lic s

trai

n (%

)S

ysto

lic th

icke

ning

(%

)

y = 0.001x + 2.4 R = 0.85 P <.001

Y = –5.10-06x 2 +0.04x +10

R = 0.7 p < 0.01

HR < 135HR ≥135

HR < 135HR ≥ 135

2.5 5 10 202.5 5 10 202.5 5 10 20

6

2

A Posterior wall B Septum

Fig. 1.11 Use of strain and strain rate to identify ischemia. In this closed-chest pig model of dobutamine re-sponse after 30 minutes of hypoperfusion, ischemia is most readily identified with strain rate. The normal response of strain to increasing dobutamine dose (and therefore heart rate [HR]) is that it reaches a plateau as left ventricular (LV) cavity size decreases, whereas the normal response of strain rate to dobutamine is to show a continuing increment.

Standard deviation of timeto peak negative strain

Time to peak negative strain200

–24

–16

–8

0

8

16

400 600 800

Fig. 1.12 Contractile dispersion. The individual strain curves of this left ventricle show a different time course, measurable as time to peak strain. The standard deviation of this distribution is a reflection of heterogeneity of contraction and a marker of arrhythmogenicity.


systems.24 These results are similar to the vendors’ normal range for GLS in the meta-analysis (General Electric [220.7% 6 2.4%], Philips [220.1% 6 2.4%], and Sie-mens [220.0% 6 2.7%]), with the greater normal value with the TomTec system (222.2% 6 2.7%, P ,.01 vs other vendors) likely reflecting the measurement of endo-cardial strain. Strain has been reported to be less in women than men, presumably related to the smaller female heart; in people without cardiovascular disease or traditional risk factors, the absolute GLS difference between men and women is .1%.27,45

Although the use of single measurements is interpre-table, albeit with some ambiguity in the range of 216% to 218%, it is generally easier to use GLS with each patient acting as one’s own control (e.g., in sequential follow-up for potential cardiotoxicity).46 The estimation of myocardial work (MW), derived from strain, may overcome the afterload dependence of this parameter47 by integrating myocardial strain and afterload, which provides a parameter that can be compared under dif-ferent loading conditions.47,48 This measurement, vali-dated in animal models, links the process of myocardial shortening to intraventricular pressure within the left ventricle, which is approximated from estimation of BP.

reflect the evolution of speckle-tracking technology, which has seen a reduction of vendor variability since the previous report.44 The results were based on the data pro-vided by 8 of 25 publications, involving 2396 patients of mean age 42 years (range 18–92), with body surface area 1.7 6 0.2 m2. The normal range for GLS was 221.0 6 2.6%, so that more negative than 218% is definitely nor-mal, and strain that is less negative than 216% is almost certainly abnormal, with 2.8% of samples from normal people being less than this. The range between 216% and 218% is ambiguous, probably abnormal, but normal measurements may be in this category due to increased afterload and older age. The latter may reflect the influence of subclinical disease on strain, keeping in mind the nor-mal ranges are not necessarily defined in individuals tested to ensure that they do not have undiagnosed coronary disease, for example. Whatever the cause, age-based varia-tion of GLS seems to be relevant in individuals over age 60 years, who show impairment of GLS, compared with pa-tients under age 60 years (19.9% 6 2.9% vs 21.1% 6 2.6%, P ,.01) (Fig. 1.14). Normal ranges for GLS also varied with common clinical covariates such as weight and blood pressure (BP). Table 1.3 lists the average and lower limits of GLS in the same normal individuals scanned with multiple

1.0

0.8

0.6

0.4

0.2

0.0

RO

C A

UC

Cutoff of mechanical dispersion (ms)

0 20 40 60 80 100

100

80

60

40

20

0

Sen

sitiv

ity (

%)


0 20 40 60 80 100

100

80

60

40

20

0

Acc

urac

y (%

)Cutoff of mechanical dispersion (ms)

0 20 40 60 80 100

100

80

60

40

20

0

Spe

cific

ity (

%)


0 20 40 60 80 100

Fig. 1.13 Cutoffs for identification of mechanical dispersion. The sensitivity, specificity, accuracy, and area under the receiver operator characteristic curves (AUC ROCs) of cutoffs in recent studies of mechanical dis-persion for prediction of serious ventricular arrhythmias.


ReckefussNottinTakigikuTakigikuTakigikuFineMarharajKayaKockabayMorisCaselliCharfeddiChengChengCongFunMentingSugimotoKotwica

20.60016.70021.30018.90019.90017.30017.28020.58021.50021.23019.40022.99022.00020.20020.34017.30020.80022.50022.00020.166

0.2170.5140.1150.1380.1310.1830.2930.6220.1270.1130.3250.5190.1470.1660.5420.5390.1610.1150.3720.348

0.0470.2650.0130.0190.0170.0340.0860.3860.0160.0130.1060.2690.0220.0280.2940.2900.0260.0130.1380.121

20.17515.69221.07418.63019.64416.94116.70619.36221.25121.00918.76221.97321.71219.87419.27716.24520.48522.27421.27119.484

21.02517.70821.52619.17020.15617.65917.85421.79821.74921.45120.03824.00722.28820.52621.40318.35521.11522.72622.72920.848

n =

59

(20.

2 ±

1.8

)

n =

237

(20

.7 ±

2.3

)

n =

299

(21

.1 ±

2.4

)

n =

327

(21

.0 ±

2.5

)

n =

315

(21

.3 ±

2.4

)

n =

238

(21

.5 ±

2.9

)

n =

221

(21

.2 ±

2.5

)

n =

190

(21

.2 ±

2.7

)

n =

190

(21

.3 ±

3.0

)

n =

153

(20

.4 ±

2.5

)

n =

104

(19

.7 ±

3.2

)

n =

42

(19.

4 ±

2.8

)

n =

16

(17.

7 ±

2.3

)

StandardMean Variance limitlimiterror

Lower Upper

30G

loba

l lon

gitu

dina

l str

ain

(%)

25

20

15

1018-20 21-25 26-30 31-35 36-40 41-45 46-50

Age Group

0.00 12.50 25.00

51-55 56-60 61-65 66-70 71-75 75+

B

A

Study name Statistics for each study Mean and 95% Cl

Fig. 1.14 Normal strain and age. These age-based ranges, derived from an individual-patient meta-analysis of recent studies, show a reduction of strain with age. CI, Confidence interval.


wall and interventricular septum) is more robust than the exclusive delineation of RV free wall. However, RV defor-mation limited to RV free wall is more representative for RV contractile function than the four-chamber RV strain (i.e., including also interventricular septum).

LA strain (and by analogy, RA strain) evaluates res-ervoir, conduit, and booster function (see Fig. 1.15).

There are definitely attractions in using MW under cir-cumstances en BP may change sequentially, such as dur-ing follow-up for potential cardiotoxicity or in combi-nation with exercise. However, an important limitation of this new parameter is the very large variation of its normal range, which likely leads to some overlap be-tween normal and pathologic states.

One problem of defining a normal range is that it focuses attention on a single parameter. This is particu-larly problematic when segmental strain rather than GLS is being measured. The reliability of GLS is to a significant degree a reflection of the benefits of averaging, which re-duces the impact of noise in one or two individual seg-ments. Segmental strain is therefore more susceptible to noise and random variability. Assessment of the strain signal is an important means of providing quality control because (1) some strain morphologies can be recognized as nonphysiologic, and (2) postsystolic shortening is a specific signal of myocardial ischemia (see Fig. 1.11).

STRAIN MEASUREMENTS IN CARDIAC CHAMBERS OTHER THAN THE LEFT VENTRICLEMyocardial strain can be measured in all of the cardiac chambers. LV and RV strains are always measured by triggering on the R wave. Left atrial (LA) and right atrial (RA) strains can be assessed using R-wave triggering, as well as triggering on the P wave. Examples of strain curves in each cardiac chamber are provided in Figures 1.15 and 1.16 and discussed in later sections.

The right ventricle and, specifically, left and right atria are thin walled; therefore, special care should be taken when defining the region of interest to ensure adequate tracking. In normal hearts, the RV free wall has higher absolute longitudinal strain values than the left ventricle. For technical reasons, tracking (including both RV free

A

B

RESERVOIR BOOSTER

CONDUIT

RESERVOIR

CONDUIT

BOOSTER

Fig. 1.15 Components of the atrial strain curve. The components of the atrial strain profile are reservoir strain (the entire excur-sion), conduit, and booster strains. (A) QRS as a reference point. (B) P wave as a reference point.

GE Tomtec Siemens Philips

Mean 21.0 21.5 20.0 18.8Lower limit (2SD) 17.1 17.5 16.4 15.2

TABLE 1.3 Global Longitudinal Strain Measurements in the Same Population with Recent Software

From Farsalinos KE, Daraban AM, Ünlü S, et al. Head-to-head comparison of global longitudinal strain measurements among nine different vendors. J Am Soc Echocardiogr. 2015;28:1171-1181.e2.GE, General Electric; SD, standard deviation.


CLINICAL APPLICATIONSGLS is more sensitive for the detection of early change in LV function than LVEF49-51 and is a robust marker of subclinical LV dysfunction, which can be used for prog-nostic or diagnostic assessment. GLS can provide impor-tant predictive value for adverse events in patients in a variety of situations and has superior predictive value to LVEF for risk stratification.52-55 In a meta-analysis in which it was compared with EF, a standard deviation change of GLS provided approximately 1.5 times the prognostic information obtained from EF, irrespective of whether EF was preserved or reduced.54 The diagnostic value of GLS is based on its ability to identify subclinical LV dysfunction in a number of situations, including the early stages of heart failure (including cardiotoxicity,

These functions are relevant modulators of ventricular filling and cardiovascular performance. The reservoir function corresponds to the collection of venous inflow during ventricular systole and is dependent both on atrial characteristics (relaxation and stiffness) and ven-tricular contraction. The conduit function is associated with blood passage to the ventricle during early diastole and influenced by ventricular relaxation and atrial after-load. The active booster function reflects atrial intrinsic contractility but is modulated by ventricular compliance.

LA size (which is influenced by atrial pressure and pulmonary venous pressure) may be an important in-fluence on strain. The assessment of LA compliance is analogous to the measurement of LV work and com-pensates LA strain for LA volume.

B

A

C

Fig. 1.16 Strain profiles in different chambers. (A) Three apical views of the left ventricle with bull’s eye; (B) right ventricular free wall; (C) right atrium.


LV function assessment

Heart failure HF riskfactors? SBHF

CTRCDmonitoring

LVhypertrophy

Asymptomaticvalve disease

HFpEF HFrEF AS AR MR

RV function assessment Atrial function assessment

SuspectedCAD

PulmonaryHTN

Valvulardisease

Diastolicfunction

Risk of AF orrecurrence

post ablation

LV-GLS and regional strain RV-GLS and FWS LARS and LACS

Additive value to LVEF Superior to FAC, TAPSE, etc

PhenotypingRisk predictionMechanicaldispersionRegional scar

Recognitionof SBHF

Reduction ispredictive ofLVEF declineand CTRCD

Risk predictionDetection offibrosis

Assist DxIdentify LVfibrosisRiskprediction

Gooddiscriminatoryability inabsence ofRWMA or LVEFreduction

Management implications—LV

Trial patientselectionCRT leadplacement

Potentialreclassificationof SBHF

Cardiologyconsultation

Potential earlierinterventionand increasedsurveillance

Disease-specificMxPotential forHCM arrhythmicrisk guidelines

ExpeditedCAD testing

RV functionis maindriver ofoutcome

RVD and PHTare markers ofloss of reservedue to VHD

Management implications—RV

Pulmonaryvasodilatorresponse

Timing ofvalvular heartdiseaseintervention

Management implications—LA

Assessmentof diastolicdysfunction

Riskevaluationand intensityof follow-up

Superior to LA volume

LAV doesnot reverseremodel,LA strainmay bepreferable

AtriopathyprecedesAF

Central Illustration 1.1 The addition value of strain (LVGLS, RVGLS, atrial strain) to standard measures for left ventricular (LV), right ventricular (RV), and atrial assessment. In most instances, the incremental value of strain is based on its ability to identify subclinical dysfunction in a number of situations, including the early stages of disease.1

diabetes, hypertension, and obesity) and valvular disease (Central Illustration).1 Recent studies have also reported that subclinical LV systolic dysfunction, defined as a reduced GLS, was more frequent than that defined as a reduced LVEF and that this is a powerful and independent predictor of cardiovascular events in a community-based cohort.53,56 Evidence exists that LV untwisting measured by speckle tracking provides in-cremental value in predicting adverse outcome both in the preclinical disease and clinically overt heart failure (with preserved EF).57-59

Strain has emerged as a valuable tool in the assessment of atrial and ventricular function, in part because of the limitations of the existing modalities for assessment of the function of these chambers. Finally, strain has been used to characterize the underlying disease process affecting the

myocardium. These indications are based on the regional distribution of strain within the heart as an analog for the assessment of myocardial tissue characterization, For example, specific patterns of regional inhomogeneity of deformation have been employed in the distinction of amyloid disease from other forms of hypertrophy and in the recognition of hypertrophic cardiomyopathy. The use of strain for these purposes is confounded by the fact that it is a functional marker, potentially allowing functional change to either hide or exaggerate changes that are at-tributable to normal myocardium.

CONCLUSIONAssessment of cardiac (especially LV) function is es-sential to important decision making in cardiology.


9. Kagiyama N, Sugahara M, Crago EA, et al. Neurocardiac injury assessed by strain imaging is associated with in-hospital mortality in patients with subarachnoid hemor-rhage. JACC Cardiovasc Imaging. 2020;13:535-546.

10. Flachskampf FA, Blankstein R, Grayburn PA, et al. Global longitudinal shortening: a positive step towards reducing confusion surrounding global longitudinal strain. JACC Cardiovasc Imaging. 2019;12:1566-1567.

11. Hung CL, Verma A, Uno H, et al. Longitudinal and cir-cumferential strain rate, left ventricular remodeling, and prognosis after myocardial infarction. J Am Coll Cardiol. 2010;56:1812-1822.

12. Esch BT, Warburton DER. Left ventricular torsion and recoil: implications for exercise performance and cardiovascular disease. J Appl Physiol. 2009;106:362.

13. Buckberg G, Hoffman JIE, Mahajan A, et al. Cardiac mechanics revisited. Circulation. 2008;118:2571.

14. Weidemann F, Jamal F, Sutherland GR, et al. Myocardial function defined by strain rate and strain during altera-tions in inotropic states and heart rate. Am J Physiol Heart Circ Physiol. 2002;283:H792-H799.

15. Stokke TM, Hasselberg NE, Smedsrud MK, et al. Geome-try as a confounder when assessing ventricular systolic function: comparison between ejection fraction and strain. J Am Coll Cardiol. 2017;70:942-954.

16. Leitman M, Lysyansky P, Sidenko S, et al. Two-dimensional strain-a novel software for real-time quantitative echocar-diographic assessment of myocardial function. J Am Soc Echocardiogr. 2004;17:1021-1029.

17. Leitman M, Lysiansky M, Lysyansky P, et al. Circumfer-ential and longitudinal strain in 3 myocardial layers in normal subjects and in patients with regional left ventricular dysfunction. J Am Soc Echocardiogr. 2010;23:64-70.

18. Chan J, Hanekom L, Wong C, et al. Differentiation of subendocardial and transmural infarction using two- dimensional strain rate imaging to assess short-axis and long-axis myocardial function. J Am Coll Cardiol. 2006;48:2026-2033.

19. Thavendiranathan P, Negishi T, Cote MA, et al. Single versus standard multiview assessment of global longitu-dinal strain for the diagnosis of cardiotoxicity during cancer therapy. JACC Cardiovasc Imaging. 2018;11:1109-1118.

20. Collier P, Phelan D, Klein A. A test in context: myocardial strain measured by speckle-tracking echocardiography. J Am Coll Cardiol. 2017;69:1043-1056.

21. Negishi K, Negishi T, Kurosawa K, et al. Practical guid-ance in echocardiographic assessment of global longitu-dinal strain. JACC Cardiovasc Imaging. 2015;8:489-492.

22. Negishi K, Lucas S, Negishi T, et al. What is the primary source of discordance in strain measurement between

Although all clinicians rightly depend on LVEF for this purpose, there are fundamental problems with the reliability and precision of these measurements. More-over, LVEF served us well in the era of HF with reduced EF, but the emergence of HF with preserved EF as the predominant form of HF60 necessitates a new approach, as does the need to identify subclinical dis-ease and follow patients sequentially. GLS provides a solution to these limitations, and the use of strain per-mits quantification of the function of other chambers. In most laboratories, GLS is still not yet part of the clinical routine for a number of reasons. It has been perceived as being time consuming (but adds only a couple of minutes to a study), and although there is a learning curve, this is short.61,62 The other chapters of this book should inform the wider use of this impor-tant measurement.

REFERENCES 1. Potter E, Marwick TH. Assessment of left ventricular

function by echocardiography: the case for routinely add-ing global longitudinal strain to ejection fraction. JACC Cardiovasc Imaging. 2018;11:260-274.

2. Buchalter MB, Weiss JL, Rogers WJ, et al. Noninvasive quantification of left ventricular rotational deformation in normal humans using magnetic resonance imaging myocardial tagging. Circulation. 1990;81:1236.

3. Maier SE, Fischer SE, McKinnon GC, et al. Evaluation of left ventricular segmental wall motion in hypertrophic cardiomyopathy with myocardial tagging. Circulation. 1992;86:1919.

4. Yeon SB, Reichek N, Tallant BA, et al. Validation of in vivo myocardial strain measurement by magnetic resonance tagging with sonomicrometry. J Am Coll Cardiol. 2001;38:555-561.

5. MacGowan GA, Shapiro EP, Azhari H, et al. Noninvasive measurement of shortening in the fiber and cross-fiber directions in the normal human left ventricle and in idio-pathic dilated cardiomyopathy. Circulation. 1997;96:535.

6. Otterstad JE, Froeland G, St John Sutton M, et al. Accu-racy and reproducibility of biplane two-dimensional echocardiographic measurements of left ventricular dimensions and function. Eur Heart J. 1997;18:507-513.

7. Jenkins C, Bricknell K, Hanekom L, et al. Reproducibility and accuracy of echocardiographic measurements of left ventricular parameters using real-time three-dimensional echocardiography. J Am Coll Cardiol. 2004;44:878-886.

8. Park JJ, Park JB, Park JH, et al. Global longitudinal strain to predict mortality in patients with acute heart failure. J Am Coll Cardiol. 2018;71:1947-1957.


the FAST-EFs multicenter study. J Am Coll Cardiol. 2015;66:1456-1466.

34. Zamorano JL, Lancellotti P, Rodriguez Munoz D, et al. 2016 ESC position paper on cancer treatments and cardiovascular toxicity developed under the auspices of the ESC Committee for Practice Guidelines: the task force for cancer treatments and cardiovascular toxicity of the European Society of Cardiology (ESC). Eur Heart J. 2016;37:2768-2801.

35. Liu J, Banchs J, Mousavi N, et al. Contemporary role of echocardiography for clinical decision making in patients during and after cancer therapy. JACC Cardiovasc Imag-ing. 2018;11:1122-1131.

36. Wang TJ, Evans JC, Benjamin EJ, et al. Natural history of asymptomatic left ventricular systolic dysfunction in the community. Circulation. 2003;108:977-982.

37. Redfield MM, Jacobsen SJ, Burnett Jr JC, et al. Burden of systolic and diastolic ventricular dysfunction in the community: appreciating the scope of the heart failure epidemic. JAMA. 2003;289:194-202.

38. Yeboah J, Rodriguez CJ, Stacey B, et al. Prognosis of individuals with asymptomatic left ventricular systolic dysfunction in the multi-ethnic study of atherosclerosis (MESA). Circulation. 2012;126:2713-2719.

39. Haugaa KH, Amlie JP, Berge KE, et al. Transmural differ-ences in myocardial contraction in long-QT syndrome: mechanical consequences of ion channel dysfunction. Circulation. 2010;122:1355-1363.

40. Haugaa KH, Smedsrud MK, Steen T, et al. Mechanical dispersion assessed by myocardial strain in patients after myocardial infarction for risk prediction of ventricular arrhythmia. JACC Cardiovasc Imaging. 2010;3:247-256.

41. Kawakami H, Nerlekar N, Haugaa KH, et al. Prediction of ventricular arrhythmias with left ventricular mechani-cal dispersion: a systematic review and meta-analysis. JACC Cardiovasc Imaging. 2020;13:562-572.

42. Jashari H, Rydberg A, Ibrahimi P, et al. Normal ranges of left ventricular strain in children: a meta-analysis. Cardiovasc Ultrasound. 2015;13:37.

43. Yingchoncharoen T, Agarwal S, Popovic ZB, et al. Nor-mal ranges of left ventricular strain: a meta-analysis. J Am Soc Echocardiogr. 2013;26:185-191.

44. Yang H, Marwick TH, Fukuda N, et al. Improvement in strain concordance between two major vendors after the strain standardization initiative. J Am Soc Echocardiogr. 2015;28:642-648.e7.

45. Cheng S, Larson MG, McCabe EL, et al. Age- and s ex-based reference limits and clinical correlates of myocardial strain and synchrony: clinical perspective. Circ Cardiovasc Imaging. 2013;6:692.

46. Plana JC, Galderisi M, Barac A, et al. Expert consensus for multimodality imaging evaluation of adult patients

vendors: imaging or analysis? Ultrasound Med Biol. 2013;39:714-720.

23. Voigt J-U, Pedrizzetti G, Lysyansky P, et al. Definitions for a common standard for 2D speckle tracking echocardiog-raphy: consensus document of the EACVI/ASE/Industry Task Force to standardize deformation imaging. Eur Heart J Cardiovasc Imaging. 2015;16:1-11.

24. Farsalinos KE, Daraban AM, Ünlü S, et al. Head-to-head comparison of global longitudinal strain measurements among nine different vendors. J Am Soc Echocardiogr. 2015;28:1171-1181.e2.

25. Negishi T, Negishi K, Thavendiranathan P, et al. Effect of experience and training on the concordance and preci-sion of strain measurements. JACC Cardiovasc Imaging. 2017;10:518-522.

26. Zghal F, Bougteb H, Réant P, et al. Assessing global and regional left ventricular myocardial function in elderly patients using the bidimensional strain method. Echocar-diography. 2011;28:978-982.

27. Takigiku K, Takeuchi M, Izumi C, et al. Normal range of left ventricular 2-dimensional strain: Japanese Ultra-sound Speckle Tracking of the Left Ventricle (JUSTICE) study. Circ J. 2012;76:2623-2632.

28. Yang H, Wright L, Negishi T, et al. Research to practice: assessment of left ventricular global longitudinal strain for surveillance of cancer chemotherapeutic-related cardiac dysfunction. JACC Cardiovasc Imaging. 2018;11:1196-1201.

29. Cannesson M, Tanabe M, Suffoletto MS, et al. A novel two-dimensional echocardiographic image analysis sys-tem using artificial intelligence-learned pattern recogni-tion for rapid automated ejection fraction. J Am Coll Cardiol. 2007;49:217-226.

30. Liel-Cohen N, Tsadok Y, Beeri R, et al. A new tool for automatic assessment of segmental wall motion based on longitudinal 2D strain: a multicenter study by the Israeli Echocardiography Research Group. Circ Cardiovasc Imaging. 2010;3:47-53.

31. Blondheim DS, Friedman Z, Lysyansky P, et al. Use of an automatic application for wall motion classification based on longitudinal strain: is it affected by operator expertise in echocardiography? A multicentre study by the Israeli Echocardiography Research Group. Eur Heart J Cardiovasc Imaging. 2012;13:257-262.

32. Lee M, Chang SA, Cho EJ, et al. Role of strain values using automated function imaging on transthoracic echocardiography for the assessment of acute chest pain in emergency department. Int J Cardiovasc Imaging. 2015;31:547-556.

33. Knackstedt C, Bekkers SC, Schummers G, et al. Fully automated versus standard tracking of left ventricular ejection fraction and longitudinal strain:


55. Saito M, Negishi K, Eskandari M, et al. Association of left ventricular strain with 30-day mortality and readmission in patients with heart failure. J Am Soc Echocardiogr. 2015;28:652-666.

56. Yang H, Negishi K, Wang Y, et al. Echocardiographic screening for non-ischaemic stage B heart failure in the community. Eur J Heart Fail. 2016;18:1331-1339.

57. Przewlocka-Kosmala M, Marwick TH, Yang H, et al. Association of reduced apical untwisting with incident HF in asymptomatic patients with HF risk factors. JACC Cardiovasc Imaging. 2020;13:187-194.

58. Kosmala W, Przewlocka-Kosmala M, Marwick TH. Association of active and passive components of LV diastolic filling with exercise intolerance in heart failure with preserved ejection fraction: mechanistic insights from spironolactone response. JACC Cardiovasc Imaging. 2019;12:784-794.

59. Kosmala W, Przewlocka-Kosmala M, Rojek A, et al. Comparison of the diastolic stress test with a combined resting echocardiography and biomarker approach to pa-tients with exertional dyspnea: diagnostic and prognostic implications. JACC Cardiovasc Imaging. 2019;12:771-780.

60. Senni M, Tribouilloy CM, Rodeheffer RJ, et al. Congestive heart failure in the community: a study of all incident cases in Olmsted County, Minnesota, in 1991. Circulation. 1998;98:2282-2289.

61. Negishi K, Negishi T, Kurosawa K, et al. Practical guid-ance in echocardiographic assessment of global longitu-dinal strain. JACC Cardiovasc Imaging. 2015;8:489-492.

62. Negishi T, Negishi K, Thavendiranathan P, et al. Effect of experience and training on the concordance and preci-sion of strain measurements. JACC Cardiovasc Imaging. 2017;10:518-522.

during and after cancer therapy: a report from the American Society of Echocardiography and the Euro-pean Association of Cardiovascular Imaging. J Am Soc Echocardiogr. 2014;27:911-939.

47. Manganaro R, Marchetta S, Dulgheru R, et al. Echocar-diographic reference ranges for normal non-invasive myocardial work indices: results from the EACVI NORRE study. Eur Heart J Cardiovasc Imaging. 2019;20:582-590.

48. Russell K, Eriksen M, Aaberge L, et al. A novel clinical method for quantification of regional left ventricular pressure-strain loop area: a non-invasive index of myocardial work. Eur Heart J. 2012;33:724-733.

49. Helle-Valle T, Crosby J, Edvardsen T, et al. New noninva-sive method for assessment of left ventricular rotation: speckle tracking echocardiography. Circulation. 2005;112:3149-3156.

50. Nesbitt GC, Mankad S, Oh JK. Strain imaging in echo-cardiography: methods and clinical applications. Int J Cardiovasc Imaging. 2009;25:9-22.

51. Marwick TH. Methods used for the assessment of LV systolic function: common currency or tower of Babel? Heart. 2013;99:1078-1086.

52. Motoki H, Borowski AG, Shrestha K, et al. Incremental prognostic value of assessing left ventricular myocardial mechanics in patients with chronic systolic heart failure. J Am Coll Cardiol. 2012;60:2074-2081.

53. Russo C, Jin Z, Elkind MS, et al. Prevalence and prognos-tic value of subclinical left ventricular systolic dysfunc-tion by global longitudinal strain in a community-based cohort. Eur J Heart Fail. 2014;16:1301-1309.

54. Kalam K, Otahal P, Marwick TH. Prognostic implications of global LV dysfunction: a systematic review and meta-analysis of global longitudinal strain and ejection fraction. Heart. 2014;100:1673-1680.

1 - Strain imaging applications and techniques

Documents

Transcript of 1 - Strain imaging applications and techniques