Global and regional left ventricular (LV) function are two of the most common echocardiographic evaluations in daily clinical practice and are essential for the clinical management of patients with various cardiovascular diseases. Global chamber indices such as LV volumes and ejection fraction are used for determining prognosis of patients with ischaemic heart disease, as well as when to plan surgical interventions such as valve replacement and implantation of a biventricular pacemaker or intracardiac defibrillator. In addition, LV regional wall motion is regularly evaluated in rest or under stress for diagnosis of myocardial ischaemia and viability.
However, in daily clinical practice, the echocardiographic evaluation of LV function is still frequently performed by visual estimation. Currently, the clinical evaluation of regional LV function through visual assessment of myocardial wall displacement and thickening remains largely qualitative. Besides potential errors, such as cardiac translation and tethering, which may limit accurate assessment of regional wall motion, the assessment itself requires extensive skill and experience of the echocardiographer.1 Nonetheless, even among expert readers using contemporary echocardiographic equipment, the observer reliability of visual assessment of regional wall motion abnormalities (WMAs) is relatively low.2,3 Although, for global LV function assessment, more objective quantitative measurements can be obtained with M-mode and 2D echocardiography (2DE), these techniques are limited by their assumptions of symmetrical LV geometry, which are unreliable, especially in patients with ventricles that are dilated, aneurysmatic or with segmental WMA.4
In recent years, efforts have been made to overcome these limitations through development of more automated systems for quantification of global and regional LV function, such as tissue Doppler imaging (TDI) and 2D speckle tracking echocardiography (2DSTE).5–7 Although a highly automated quantitative assessment of the left ventricle that is both accurate and reliable would be ideal, most of these approaches are currently impractical for daily clinical evaluation due to time-consuming and labour-intensive acquisition and analysis. Moreover, these techniques are still significantly limited by their 2D nature, which not only introduces issues such as foreshortening and signal noise, but ultimately may only provide a poor facsimile of true LV mechanics, which are intrinsically 3D.
A significant step forward has been the development of 3D echocardiography (3DE).8 Although the technique initially had a slow clinical acceptance – primarily due to poor image quality and consequently limited applicability – with the current state of technology, substantial evidence now exists for the superiority of 3DE over 2DE in both accuracy and reproducibility of LV volumes and ejection fraction measurements when compared with magnetic resonance imaging (MRI) as the reference technique.9 Mannaerts et al.10 described a novel 3D sphericity index that was able to prospectively identify myocardial infarction patients with a high risk of LV remodelling more accurately than other clinical, electrocardiographic and echocardiographic parameters. However, studies evaluating the diagnostic accuracy of 3DE to detect ischaemic heart disease through assessment of regional WMAs at rest and during stress testing demonstrated mixed results.11–15 One of many reasons for this is likely to be that, although 3DE eliminates many limitations of 2DE – such as high operator dependency and limited number of cross-sections available for evaluation – the semi-automated endocardial border detection algorithm cannot differentiate between passive motion and true myocardial deformation, significantly limiting its clinical applicability for regional WMAs.
Recently, new technological advances have provided the opportunity to combine speckle tracking imaging with 3DE to address this limitation and offer new possibilities for analysis of the left ventricle: 3D speckle tracking echocardiography (3DSTE) offers a fast and highly automated quantitative assessment of the left ventricle in three dimensions during the entire cardiac cycle, providing not only global chamber indices such as volumes and ejection fraction, but also regional myocardial indices such as strain and rotation, as well as parameters of LV dyssynchrony (see Figure 1).
Quantitative Assessment of Left Ventricular Size and Function
Several recent studies have demonstrated the usefulness of 3DSTE for quantitative evaluation of LV size and function. First of all, the average time for image acquisition and analysis by 3DSTE was found to be significantly shorter than by 2DSTE. The feasibility was much higher with 3DSTE, demonstrating it to be a practical and time-efficient tool for daily clinical routine.16,17
A study conducted by Nesser et al.18 was the first to validate the use of 3DSTE for LV volume measurements and demonstrated superior accuracy and reproducibility of 3DSTE over 2DSTE when compared with MRI. Furthermore, it was recently demonstrated that 3DE direct volumetric and speckle-tracking methods give comparable and reproducible quantification of LV and left atrial (LA) volumes and function, making interchangeable application a viable option in daily clinical practice.19 Moreover, excellent observer and test-retest reproducibility support the use of 3DSTE for routine evaluation of LV volumes and ejection fraction.20 Of the myocardial strain indices evaluated, measurements of circumferential strain demonstrated good-to-excellent reproducibility, whereas moderate-to-good reproducibility was found for longitudinal and radial strain measurements. Additional validation of 3DSTE for assessment of strain and rotation was provided in animal studies in vitro and in vivo,21,22 while similar results were found for assessment of global and regional parameters of LV strain, displacement, rotation and twist in patients and healthy subjects.16,17,23–26 Accumulated results demonstrated better quantitative discrimination between normal and abnormal segments, with larger magnitude of displacement in normal segments, with 3DSTE compared with 2DSTE strain parameters, likely reflecting better registration of all components of LV deformation within the 3D dataset. In addition, clear differences were demonstrated in the magnitude of strain between different segments, regions and levels of the left ventricle, both in healthy volunteers and patients. This observation, which demonstrates functional non-uniformity of both the normal and diseased left ventricle, deserves attention, as it may have implications for the clinical validity of regional LV function assessment by strain. The potential impact of functional non-uniformity on the performance of the various 3DSTE strain parameters in different settings, such as acute myocardial infarction or stress echocardiography, remains to be determined in future studies.
Finally, our own group recently evaluated the clinical applicability of a novel parameter called area strain to reliably quantify global and regional LV function and accurately identify WMAs when compared with current echocardiographic standards.26 Area strain represents the change in endocardial surface area during the cardiac cycle and quantifies it by giving the percentage change in area from its original dimensions at end-diastolic time.
In a large population of patients with various diagnoses of heart disease and a wide range of LV function, area strain had excellent correlations with independently derived LV ejection fraction and wall motion score index, good agreement with visual assessment of regional wall motion by expert echocardiographers, as well as good-to-excellent observer and test-retest reproducibility.
All these studies provide evidence that 3DSTE offers a comprehensive quantitative assessment of the left ventricle, including accurate and reproducible measurements of volumes as well as global and regional myocardial function, all within one fast and feasible analysis. Additional studies are needed to demonstrate the clinical value of 3DSTE-derived parameters of global and regional LV function for early identification and prevention of cardiovascular disease. However, it has been established that 3DSTE-derived measurements of LV volumes and ejection fraction are reliable enough for patient management and follow-up, including evaluation of treatment effect in clinical trials. Furthermore, normal reference values need to be determined in a large healthy population stratified according to age in order to establish diagnostic thresholds to distinguish between what should be considered normal and what constitutes global or regional LV dysfunction. Finally, one caveat recently addressed by Gayat et al.27 is that, similar to other imaging techniques, inter-vendor variability exists for quantitative deformation parameters by 3DSTE, despite acceptable observer and test-retest reproducibility. This should be solved by further standardisation in acquisition and analysis methodologies between vendors in the near future.
Quantitative Assessment of Left Ventricular Dyssynchrony
Concerning the dimension of time, LV dyssynchrony has emerged as an important factor for decreased LV function and a therapeutic target for cardiac resynchronisation therapy (CRT) in patients with advanced drug-refractory heart failure.31–34 However, 30 to 40% of patients were considered non-responders in clinical trials, depending on which measure of clinical responsiveness were used.32,35,36 In search for better predictors of response to CRT, many single-centre studies have focused on assessment of mechanical LV dyssynchrony by echocardiography. Responders have indeed been identified before device implantation by several 2DE and TDI parameters of LV dyssynchrony.34-37
Unfortunately, the Predictors of response to CRT (PROSPECT) trial demonstrated a relatively low yield and poor inter-observer variability for all studied parameters within a multicentre setting, making them unfit for general clinical use.38 Moreover, it was concluded that none of the echocardiographic measures of dyssynchrony had enough predictive value to be recommended as an additional selection criterion for CRT.
Furthermore, the ability of 2D parameters to define the underlying pathophysiological substrate might be poor. As mentioned, myocardial deformation is a complex 3D process and accurate assessment of LV dyssynchrony will therefore likely need a similar approach. In this regard, 3DE seems ideally suited to not only evaluate the effects of biventricular pacing therapy through measurement of changes in volumes and ejection fraction, but also to comprehensively assess LV dyssynchrony.39,40 Many studies have compared the assessment of LV dyssynchrony by 3DE with that by other echocardiographic techniques and imaging modalities such as TDI and MRI.41–44 Not surprisingly, their assessments of dyssynchrony are not interchangeable due to striking differences found in both the presence and extent of LV dyssynchrony. The lack of a clear gold standard of LV dyssynchrony assessment seems to preclude identification of the most appropriate technique. In the end, however, LV dyssynchrony is primarily evaluated with the aim of providing better prediction of response to CRT. In that regard, the results of studies that have determined the accuracy of 3DE-derived LV dyssynchrony to predict response to CRT are very promising,40,43,45–51 as are the first results of the use of 3DSTE for this purpose.52 Moreover, two studies demonstrated that the site of latest mechanical activation could be identified with 3DSTE in a manner not previously possible with 2DE or 3DE techniques, which only focus on endocardial inward motion, potentially improving identification of the preferred site for LV pacing.53,54
Up until now, the ability to predict response to CRT has been evaluated with mechanical dyssynchrony indices that are solely based on measurement of differences in segmental timing of mechanical activation, without taking global and regional myocardial contractility into account. Although the presence of mechanical dyssynchrony is undoubtedly essential to derive benefit from CRT, conversely, ischaemia and scar tissue are important factors known to negatively influence response to CRT.55,56
Not only does the extent of myocardial scarring seem important,56–59 but also its presence at the site of preferred pacing.55,56,59–61 Beside the assessment of LV dyssynchrony, evaluation of myocardial viability might, therefore, be a useful addition to the selection of eligible CRT candidates.
Accordingly, many studies have evaluated the potential role of different modalities to assess the extent of viable myocardium and scar tissue in these patients, including the Single photon emission computerised tomography (SPECT) study56,59,62 and delayed contrast-enhanced MRI studies.55,57 However, although many agree that such a multifactor evaluation will likely lead to the best prediction of response to CRT, a multimodality assessment is clearly not preferable. Furthermore, most of these modalities have limitations in their clinical use – for example, they are still not widely available, are labour-intensive, costly and often disliked by patients. In contrast, due to its low cost, ease of use and continually growing number of applications, the clinical use of 3DE has broadened significantly in recent years.
3DSTE integrates mechanical dyssynchrony and myocardial viability within one simple, fast and comprehensive 3D assessment of LV function. Thebault et al.63 evaluated multiple LV dyssynchrony parameters during optimisation of CRT, including a global performance index that provides graphical information on the amount of global LV strain and dyssynchrony. More interestingly, however, the sophisticated wall motion tracking software can provide segmental time-strain curves, which give detailed quantitative data on LV strain and dyssynchrony during the entire cardiac cycle for all 16 segments. Not only can the timing of contraction of each segment be measured, but also the amount of strain each segment contributes to global systolic LV function. Segments that are hypokinetic to akinetic due to ischaemic cardiomyopathy will contribute little to global LV function because of severely reduced or absent myocardial contraction. Time-strain curves of these segments exhibit minimal contraction during the cardiac cycle, irrespective of a long delay in myocardial contraction. The area under the segmental time-strain curves may serve as a correction factor, in that values of segments with severe WMA are low, regardless of the time until contraction (see Figure 2). It is likely that, when these segments are resynchronised by biventricular pacing, the area under time-strain curve cannot increase substantially and thus cardiac function might fail to improve. This approach might provide additional information in assessment of response to CRT compared with parameters that are solely based on measurement of segmental timing.
We propose a novel quantitative 3DE parameter of mechanical dyssynchrony based on the combination of segmental timing and myocardial contraction that may have additional value in predicting benefit to CRT, but particularly in the equally important identification of the large number of non-responders among patients with ischaemic cardiomyopathy. On a regional level, taking into account the amount of contraction at the site of latest contraction could also prove to be useful in determining the optimal location for LV lead placement. Future studies will have to demonstrate proof of concept for this proposed method, but the ability of 3DSTE to quantitatively evaluate global and regional LV function, including dyssynchrony and myocardial viability, within one analysis undoubtedly holds great promise for the continued evaluation of patients with heart failure.
Algorithms for 3D border detection and speckle tracking have become increasingly sophisticated in recent years, giving more automated assessment of global LV size and function wider acceptance in routine practice. In contrast, a complete transition from visual to automatic regional WM assessment is currently not indicated, nor do the results of current studies support such a transition at this time. However, it is acknowledged by cardiologists and industry alike that more objective quantitative measures of regional LV function are warranted, and it is expected that further technological advances in automated tracking software will provide the necessary tools in the near future. This will significantly decrease the operator- and observer-dependency of the technique, making it more consistent in the hands of echocardiographers with varying levels of expertise. Furthermore, promising novel parameters of global and regional LV function will need further validation and strenuous testing in prospective studies to gain general acceptance and facilitate a change in paradigm from semi-quantitative visual assessment of wall motion to true quantitative assessment of LV mechanics. Finally, further improvements in temporal resolution and image quality, as well as the possibility of a wider sector angle for adequate acquisition of severely dilated ventricles often present in heart failure patients, will greatly enhance LV dyssynchrony assessment and provide the opportunity for more comprehensive evaluation of diastolic LV function beyond mere LA volume measurement – that is, assessment of diastolic strain-rate and untwisting. It is clear, however, that important steps are currently being made towards a fast and fully automated evaluation of LV function that will ultimately have a lasting effect on our everyday clinical practice.
Quantification of global and regional LV function has significantly improved in the last few decades with regards to accuracy and reproducibility. Moreover, advances in echocardiographic technology have made it possible to quantitatively assess not only LV volumes and ejection fraction, but all LV mechanics, including 3D strain and its components, area strain, as well as rotation, twist and torsion in all four dimensions. It is anticipated that reliable fully automated echocardiographic assessment of global and regional LV function will become available for routine use in the near future, reaffirming the position of echocardiography as the most clinically applicable imaging tool for LV quantification.