Strain imaging – from Scandinavian research to global deployment

Abstract

Knowledge about myocardial function is important for diagnosis, treatment and prediction of the majority of all cardiac diseases. Ejection fraction (EF) by echocardiography has been the preferred diagnostic tool for these purposes, but do have some important limitations. Strain imaging has emerged as a relatively new and important echocardiographic method that will give cardiologists incremental and additional information to EF in several important diseases affecting the myocardium. This paper will give the readers a brief overview on how and when the clinicians can use strain imaging by echocardiography in their daily practice.

Keywords:

• Strain imaging
• myocardial function
• ejection fraction
• ischemic heart disease
• cardiomyoapathies
• arrhythmias

Introduction

Myocardial function is essential for prognosis of almost all cardiac diseases. The clinically most widely used tool for assessing left ventricular (LV) function is ejection fraction (EF) by echocardiography, despite all its inherent weaknesses. The search for a better echocardiographic tool to assess LV function has therefore been extensive. One approach to overcome the limitations of EF has been indirect measures of LV function based on hemodynamic assumptions. However, a more direct measure of myocardial deformation – strain – has emerged as an excellent method that could overcome some of EF’s limitations. Strain imaging has been a reality since 1998 and this paper will present the history, evolution and clinical advantages of myocardial strain.

Strain imaging – the beginning

The first principles of myocardial strain were published in 1972 by Mirsky et al.,[1] and were demonstrated by echocardiography in 1974 by Quinones et al. using M-mode.[2] This method, however, was cumbersome and never in clinical use.

An excellent team of engineers and cardiologists from The Norwegian University of Science and Technology in Trondheim, Norway described in 1998 how myocardial strain could be calculated from tissue Doppler imaging (TDI).[3] This algorithm utilized the gradient between neighboring myocardial velocities along the longitudinal axis of the LV to calculate the shortening and lengthening of the myocardium. Strain is reported as percentage lengthening and shortening. Systolic shortening results in negative strains, and systolic lengthening results in positive strains as assessed from the long axis views. Otto Smiseth and his research team understood the potential importance of this technique as an objective method for quantifying regional myocardial function. They performed the first validation studies using sonomicrometry and tagged MRI,[4,5] and started several studies to learn more about myocardial mechanics and strain imaging.[6–8]

Several questions were raised, however, about the TDI based technique regarding clinical reproducibility. The clinical use was not particularly user friendly, and had several limitations such as angle dependency and significant noise. The more recent speckle tracking technique facilitated the clinical breakthrough for the use of strain imaging.[9–11] Speckles are myocardial “footprints” generated by the interference of ultrasound beams in the myocardium and specialized software can track these speckles and calculate the distance between them throughout the cardiac cycle. This implies that the speckle tracking technique is more accurate than the TDI method, and just as important is the user friendliness of the vendors’ software. Strain imaging allows for regional assessment of function. Function is commonly assessed in 16 or 18 LV segments and the average of this regional function is presented as global longitudinal strain (GLS). This is a very robust parameter and has shown excellent ability to express myocardial dysfunction when EF is relatively preserved.[12,13] Nevertheless, assessment of LV function in patients should start with more traditional techniques like visual evaluation, EF and M-mode.[14] However, numerous cardiac diseases can affect the myocardium without any obvious and easily recognizable signs of dysfunction by traditional echocardiographic methods.[15,16] The appropriate clinical strategy should therefore be to add speckle-tracking analyses if traditional echocardiographic methods give the impression of normal function, despite clinical suspicion of myocardial disease.

Ischemia

The clinical presentation of ischemia varies from silent ischemia, stable angina pectoris to acute coronary syndrome (ACS) and death. ACS comprises non-ST elevation (NSTE)-ACS and ST-elevation myocardial infarction (STEMI). ST-elevation typically represents coronary occlusion requiring acute reperfusion therapy, while patients with suspected NSTE-ACS have more heterogeneous findings on coronary angiography. Assessment of global and regional myocardial function by strain echocardiography adds important diagnostic and prognostic information to ECG findings in patients with myocardial ischemia.

STEMI

ST-elevation usually indicates coronary artery occlusion requiring urgent reperfusion and it is obvious that the primary diagnostic tool is ECG.[17] Strain echocardiography has therefore a limited role in the acute diagnostics of STEMI, but can be used in the assessment of infarct size. It is therefore an important prognostic tool in both acute [17,18] (Figure 1) and chronic clinical settings.[19] A relationship between GLS and heart failure has been demonstrated even in patients with acute myocardial infarction and preserved ejection fraction (EF <40%).[17] Gjesdal et al. has shown that GLS was an excellent marker of myocardial infarct size [19] and that territorial strain was a specific marker of the infarcted coronary artery in patients with chronic ischemic heart disease after STEMI. Peak systolic strain can also differentiate between non-infarcted, transmural-infarcted and subendocardial-infarcted segments.[11]

Figure 1. Speckle tracking strain analysis from day 1 after STEMI with culprit lesion in the left anterior descending coronary artery, displaying (A) peak systolic longitudinal strain in the apical long-axis view of −8.7%, (B) peak systolic longitudinal strain in the four-chamber view of −10.0%, (C) peak systolic longitudinal strain in the two-chamber view of −11.8%, and (D) the corresponding strain map with systolic strain values provided in each myocardial segment and with a parametric code, with strong red colors representing systolic shortening and blue colors depicting systolic lengthening of a segment. GLS was 10.2% in this patient. With permission from: Munk et al.[18]

Strain imaging provides therefore prognostic information in both patients with acute myocardial infarction and in patients with chronic heart failure after myocardial infarction over the whole spectrum of systolic function.

NSTE-ACS

Although, ST-elevation has high specificity to predict acute coronary occlusion, ECG in general has limited ability to detect acute coronary occlusion.[20] About 30% of patients with acute coronary occlusion do not demonstrate ST-elevation and are diagnosed with NSTE-ACS. These patients may develop extensive myocardial damage, although criteria for acute reperfusion therapy are not fulfilled.[21] Acute coronary occlusion induces LV systolic dysfunction that can be assessed by echocardiography. Strain echocardiography has been demonstrated to be superior to visual assessment of wall motion in the detection and quantification of regional systolic function.[7] Eek et al.[22] demonstrated that all indices of LV global and regional function were impaired in patients with acute coronary occlusion, compared with those with a patent infarct related artery. Importantly, strain imaging could identify NSTE-ACS patients with acute coronary occlusion, who may benefit from urgent reperfusion therapy,[21] and consequently improved prognosis could be achieved in these patients. The technique can therefore have an important role in the early assessment of patients with suspected acute ischemia, when ECG and traditional echocardiographic methods fail.[15]

Prediction of recovery after myocardial infarction

Acute myocardial ischemia is frequently accompanied by reduced regional myocardial systolic function caused by several factors including ongoing ischemia, stunning/hibernation, or irreversible necrosis. These factors frequently overlap and the level of contribution of each of these factors may vary. Viable and necrotic myocardium are indistinguishable in the initial phase of ischemia, characterized by reduced systolic function. Viable myocardium may recover systolic function after reperfusion therapy, while necrotic myocardium will not recover. Prediction of recovery, and consequently prediction of long-term systolic function, may therefore help to identify patients at high risk of adverse events including heart failure.

It was originally suggested by angiographic studies that post-systolic (post-ejection) shortening (PSS) indicates actively contracting and therefore viable myocardium. Strain echocardiography is very suitable for quantifying PSS noninvasively and might facilitate the clinical use in determining myocardial viability. Skulstad et al.[6] showed that PSS associated with systolic hypokinesis or akinesis, indicating actively contracting and, therefore, potentially viable myocardium. PSS in combination with dyskinesis, however, was a nonspecific marker of severe ischemia. The clinical value of PSS in patients with acute myocardial ischemia was studied by Eek et al.[23] They found that PSS was a robust independent predictor of recovery of systolic function in patients undergoing successful revascularization after NSTEMI (Figure 2). Evaluation of PSS might therefore have a role in the prediction of the potential benefit of reperfusion therapy in terms of LV systolic function, in patients where there is uncertainty concerning the effect of revascularization. Haugaa et al. showed that strain imaging could predict outcome after myocardial infarction. Mortality and the incidence of ventricular arrhythmias were studied in a prospective study of almost 600 patients after myocardial infarction.[12,13] GLS was a better marker of mortality compared to EF, and they showed that a new marker, mechanical dispersion was an excellent marker of ventricular arrhythmias. Mechanical dispersion is calculated as the standard deviation of time to peak strain in 16 LV segments. A pronounced mechanical dispersion will reflect myocardial contraction heterogeneity and can be explained as a regional contraction dyssynchrony. Mechanical dispersion has been shown to have predictive value for ventricular arrhythmias in a variety of cardiac diseases as explained below.[12,13]

Figure 2. Example of strain curves from the apical long-axis view, from a patient with an occluded left circumflex artery at baseline. Substantial post-systolic shortening is observed in the basal- and mid-inferolateral segments (yellow and cyan traces). After successful revascularization, normal systolic function is observed in the same segments at follow-up. With permission from: Eek et al.[23]

How to exclude ischemia

Significant coronary artery stenosis may cause persistently reduced longitudinal LV function. Furthermore, ischemic myocardium with impaired active contraction will stretch when LV pressure increases during early systole and delay start of systolic shortening. Smedsrud et al.[24] demonstrated that the duration of early systolic stretch by strain echocardiography in stable coronary artery disease (CAD) patients was a strong and independent predictor of significant disease (Figure 3). The duration of early systolic stretch was superior to peak systolic longitudinal strain in the detection of significant CAD. This novel parameter could aid in the identification of significant CAD in patients at rest, without the need for a stress protocol. Patients with suspected CAD typically undergo coronary angiography, however, almost 50% of these patients have normal or non-stenotic coronary arteries. A completely normal GLS will in most cases be an accurate measure for ruling out severe CAD.[25]

Figure 3. Strain curves from a patient with significant coronary artery stenosis (A). Strain curves from an individual without coronary artery disease (B). With permission from: Smedsrud et al.[24]

Strain echocardiography prediction of outcome

Prediction of sudden cardiac death (SCD)

In a number of cardiac disorders, the ability of GLS to predict cardiovascular outcome is superior to LVEF. The accurate information about regional function and timing from strain echocardiography has opened new arenas for investigation of cardiac diseases.

One of the most challenging tasks in cardiology is the prediction of SCD. Myocardial strain echocardiography can improve risk stratification of SCD in individuals with relatively preserved myocardial function. The principle of heterogeneous myocardial contraction as a marker of ventricular arrhythmias, mechanical dispersion, was shown in the long QT syndrome (LQTS). LQTS is a cardiac ion channel disease, and has been considered a purely electrical disease with high risk of ventricular arrhythmias and SCD. Studies by TDI and speckle tracking strain echocardiography showed that mechanical dispersion in LQTS was associated with risk of ventricular arrhythmias [26] and that these patients had a subclinical decrease in both diastolic and systolic function.[27]

Further studies have tested the principle of mechanical dispersion in other patient groups, e.g. after myocardial infarction [13,17,26] and in patients with non-ischemic cardiomyopathy,[28] showing that pronounced mechanical dispersion > approximately 70 ms is associated with increased risk of ventricular arrhythmias (Figure 4). Whether electrical dispersion is caused by electrical dispersion, mechanical dysfunction or a mix of electrical and mechanical causes, remains to be addressed.

Figure 4. Left panel shows time to peak strain in an HCM patient without ventricular arrhythmias. Mechanical dispersion is 50 ms. Right panel shows time to peak strain in an HCM patient with ventricular arrhythmias. Mechanical dispersion is pronounced with 90 ms. With permission from: Haland et al.[30]

GLS and mechanical dispersion are valuable for assessment of systolic function and risk of ventricular arrhythmias supplementary to standard means, and seem to be of particular value in patients with relatively preserved systolic function.

Strain can be used to help diagnosis, to monitor disease progression and interestingly can also be used to improve risk stratification for ventricular arrhythmias in several cardiac diseases. Different diagnoses require different approaches, but combined with the conventional echocardiography, strain imaging will improve the diagnostic information and risk stratification.

Strain echocardiography in cardiomyopathies

Hypertrophic cardiomyopathy (HCM)

Hypertrophic cardiomyopathy is characterized by a heterogeneous clinical expression with increased risk of SCD from ventricular arrhythmias.[29] Systolic function assessed by EF typically remains normal during the initial disease stages of HCM and in only a few cases deteriorate at advanced disease. EF is therefore considered inadequate for evaluation of e.g. cardiac transplantation in HCM. The reason why GLS appears to be more sensitive than EF to detect myocardial dysfunction may in part reflect the limited ability of EF to assess systolic function in ventricles with hypertrophy. Haland et al.[30] and others demonstrated reduced systolic function by GLS in HCM patients compared to healthy individuals despite similar EF. GLS was also shown to be a marker of ventricular arrhythmias in HCM, and was related to the presence of fibrosis detected by LGE-CMR. GLS seems to be a more appropriate parameter for evaluating systolic function than EF in patients with HCM.

In HCM patients, various approaches have been presented for risk stratification of ventricular arrhythmias. The most commonly used are the risk calculation from the European guidelines [31] and the American guidelines.[32] However, the selection of patients for an implantable cardioverter defibrillator (ICD) as primary prevention remains challenging. Recent studies have shown that mechanical dispersion might help risk stratification in HCM patients.[30] Mechanical dispersion was higher in HCM than in healthy individuals and an independent predictor of arrhythmias in HCM patients. Furthermore, they demonstrated a relationship between mechanical dispersion and myocardial fibrosis detected by LGE-CMR, which may explain the profound relation of ventricular arrhythmias and mechanical dispersion.

Arrhythmogenic right ventricular cardiomyopathy

ARVC is an inherited cardiomyopathy leading to fibro fatty replacement of cardiomyocytes and predominantly affecting the RV. To make the diagnosis of ARVC, the 2010 Task Force Criteria (TFC 2010) combine data from different categories including imaging, electrical parameters from ECG and Holter monitoring, family history, genetic testing, and tissue properties.[33] Imaging is important in ARVC diagnosis, including both echocardiography (echo) and cardiac magnetic resonance (CMR) for detecting structural and functional abnormalities. Being a progressive cardiomyopathy, repeated cardiac imaging is needed in ARVC patients to diagnose the disease, to follow disease progression and importantly, for risk assessment of life threatening ventricular arrhythmias.[34] ARVC is a common cause of VT and sudden death in athletes. Recent studies have shown that athletic activity aggravates and accelerates disease progression.[35]

The quantitative analysis of RV function is difficult due to the complex RV anatomy and to its load dependency. New echocardiographic techniques have increased the performance of conventional echocardiography. RV function can be assessed by speckle tracking echocardiography providing a quantitative measure of RV function. The use of speckle tracking strain in ARVC patients was first described by Sarvari et al.[16] RV strain by different techniques seems to improve diagnosis, differential diagnosis and risk stratification in ARVC.[16,36,37] Mechanical dispersion of the RV (Figure 5) was also a significant marker of ventricular arrhythmias and may be used as an additional marker of early disease and for the difficult task of risk stratification of arrhythmic events in so far asymptomatic ARVC mutation positive family members.[16,35]

Figure 5. Mechanical dispersion in a healthy individual (left panel), an asymptomatic mutation carrier (mid panel) and an arrhythmogenic right ventricular cardiomyopathy patient with recurrent arrhythmias (right panel). Horizontal white arrow indicates contraction duration defined as the time from onset R to maximum myocardial shortening. Vertical arrows indicate the timing of maximum myocardial shortening in each segment. Right panel shows more pronounced mechanical dispersion. Modified from Sarvari et al.[16]

Figure 6. Kaplan–Meier analyses showing unfavorable outcome in CRT patients with dyssynchrony after CRT implantation. With permission from: Haugaa et al.[39]

An important differential diagnosis is RVOT-VT, which is characterized by ventricular premature complexes and ventricular tachycardia (VT) from the RV outflow tract. In contrast to ARVC, RVOT-VT is supposed to be a relatively benign condition, with generally well tolerated ventricular arrhythmias.[38] However, the RVOT area may also be origin of VT in patients with ARVC and in early stages of ARVC the distinction to RVOT-VT may be challenging. The treatment and prognosis however, differ substantially and an incorrect diagnosis may be devastating. Any findings of regional RV hypokinesia and dyskinesia by echo or CMR in addition to RVOT dilatation make the diagnosis of ARVC more probable [36] and prognosis more severe. Strain echocardiography can help discrimination between these entities. Saberniak et al. showed that a pronounced mechanical dispersion was a marker of ARVC disease in contrast to RVOT-VT.[36] The results may be explained by the fact that mechanical dispersion may have the ability to detect subtle changes in contraction and might be explained by diffuse cell necrosis causing electrical and myocardial remodeling in ARVC in contrast to RVOT-VT.

DCM – strain in prediction of outcome and in CRT evaluation

Strain echocardiography has been used to evaluate dyssynchrony in patients eligible for cardiac resynchronization therapy (CRT). Although mechanical dyssynchrony has not been proven to be useful to select patients for CRT, strain measurements have shown interesting findings in the concept of dyssynchrony. A number of studies have shown that pre-existing mechanical dyssynchrony is an important prerequisite for achieving CRT response. However, it has not been shown that mechanical dyssynchrony is superior to the electrical dyssynchrony, the QRS interval, to predict response to CRT.

In prediction of ventricular arrhythmias, however, mechanical dyssynchrony may have an important role. Recently, it was shown that persistent dyssynchrony, i.e. dyssynchrony that did not improve by CRT, was a risk marker for ventricular arrhythmias after CRT implantation (Figure 6).[39]

Furthermore, the presence of mechanical dispersion confirmed the results from previous dyssynchrony studies. Patients with an improved mechanical dispersion after CRT implantation had lower incidence of ventricular arrhythmias and better outcome compared with patients with unaltered or increase in mechanical dispersion after CRT implantation.[40] Therefore, dyssynchrony/mechanical dispersion should not be used as markers for CRT response before CRT implantation, but may be used as marker of future arrhythmic events if present after CRT implantation.

Limitations with the strain technique

The speckle tracking method is sensitive to acoustic shadowing or reverberations, which will influence on the measures of deformation and usually result in lower absolute strain than actual real myocardial deformation. Reduced signal quality and suboptimal tracking may create non-physiological strain traces. To avoid these problems, most vendors have added spatial smoothing and previous knowledge of physiologic LV function in theirs tracking algorithms. These corrections might, however, affect real tracking in one segment dependent and similar to neighboring segments. The GLS value might be incorrect when too many segmental strain values are encountering these problems.

Speckle tracking is also challenging in structures with thin walls, such as the RV free wall and the atria.

2D-STE is not fully angle independent and the resolution along and parallel to the ultrasound beam is superior to other directions. Longitudinal strain measurements are therefore particularly attractive with the best reproducibility.

All echocardiographic and Doppler measures experience differences among different vendors. A task force of the European Association of Cardiovascular Imaging (EACVI), American Society of Echocardiography (ASE) and most vendors of ultrasound machines has resulted in fewer differences in the assessment of LV function.[41] The algorithms used by the different speckle-tracking analysis software packages are based on a block-matching approach of the speckle patterns within the myocardium. However, the strain algorithms are somewhat varying. There are presently ongoing trials with the different scanners using the same formula for strain calculations.

Finally, changes in hemodynamic parameters such as preload, afterload and contractility may influence strain measurements as they influence other echocardiography measures.[14]

Strain imaging in future practice – new modalities

Layer-specific strain echocardiography

Recent strain echocardiographic methods based on layer specific deformation analysis of the myocardium may further improve diagnostics in patients with NSTE-ACS (Figure 7). The common perception is that LV myocardium comprises an endocardial, a mid-myocardial, and an epicardial layer, and the endocardium is most susceptible to ischemic damage. Novel software allows individual assessment of all layers. The results so far implies that the assessment of endocardial and mid-myocardial territorial strain have higher accuracy than epicardial strain, wall motion score index, and EF in the identification of patients with NSTE-ACS and CAD.[42] There is, however, debated if the measured difference between the layers is a product of geometric assumptions in the current software or true biological differences.

Figure 7. The automatic strain analysis in a patient with non-ST-segment elevation acute coronary syndrome with occluded circumflex (CX) artery in apical four-chamber view shows reduced color-coded endocardial strain values in the segments supplied by the CX artery on the left. The red line and the red arrowheads depict the border of the epicardium. Color-coding from yellow to green indicates strain from +30% to –30%. Yellow indicates preserved strain. Brown indicates areas with reduced strain. Strain curves for the six endocardial segments are displayed on the right. The curves representing the segments supplied by the CX artery show reduced strain values of –15% to –17% (white arrow). Endocardial global longitudinal strain was reduced in this patient to –15%. With permission from: Sarvari et al.[42].

3D strain echocardiography

Two-dimensional strain echocardiography is a well-validated and robust method to identify various cardiac pathologies and even subclinical myocardial alterations. However, the method has obvious limitations including through-plane motion. Two-dimensional strain echocardiography assesses myocardial motion in two planes, however, the heart moves in three dimensions and therefore real myocardial motion can theoretically be assessed in all directions by 3D strain echocardiography. Another theoretical advantage of 3D strain echocardiography is acquisition of full volume from one probe position, which significantly reduces examination time. On the other hand, the apparent shortcomings of 3D strain are image quality (including stitching artifacts), lack of definition of normal values, limited validation studies, low spatial and temporal resolution, and even higher variability among vendors than 2D strain.

3D strain echocardiography has a long way to clinical acceptance. Further technical development, improved image quality and processing power will unquestionably improve the general performance of 3D strain echocardiography and will probably bring out the theoretical advantages of the method in the future. Nevertheless, additional improvement and studies are necessary before it can be included in our daily practice.[43] Until then, 2D strain echocardiography is still the method of choice.

Summary

The clinical use of strain imaging is increasing and the understanding of how and when to use it, is well studied. The method is exceptionally well suitable to detect the initial stages of myocardial diseases and in particular where more traditional echocardiographic methods fail to demonstrate myocardial dysfunction. Research teams in the Scandinavian countries have been in the forefront of this clinical evolution.

Disclosure statement

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of this article.