Towards assessment of left ventricular mechanics in true three dimensions

Accurate quantification of left ventricular (LV) dimensions and systolic function is crucial for the optimal clinical management of patients with cardiovascular disease. The LV ejection fraction (LVEF) is used as a major threshold for implantation of an internal automatic defibrillator or biventricular pacemaker, whereas LV volumes are key to selecting the optimal timing for surgical valve replacement. Moreover, both the LVEF and LV volumes are important for the prognosis of patients with ischaemic heart disease. Currently, two-dimensional (2D) echocardiography is the most frequently used imaging technique for quantification of LV volumes and LVEF. Visual assessment of the LVEF is commonly used in routine clinical practice, but it requires significant expertise and remains a subjective estimation. The modified Simpson's method calculates LV volumes by tracing the endocardial boundaries in the apical two- and four-chamber views at end-diastole and end-systole, and subsequently the LVEF is derived. However, this method is time consuming and its accuracy is limited by foreshortened apical views, inaccuracies in endocardial border detection, or LV geometric assumptions.1 In search of an automatic, accurate, and reproducible method to quantify LV volumes and LVEF, real-time three-dimensional echocardiography (RT3DE) is an important step forward in cardiac ultrasound. Multiple studies have demonstrated the superior accuracy and reproducibility of RT3DE over 2D echocardiography for assessment of LV volumes and LVEF.2,3 In addition, assessment of the LVEF and LV volumes by RT3DE correlated well with magnetic resonance imaging (MRI), whereas 2D echocardiography showed only a modest correlation with MRI.2,3 Several RT3DE methods to measure LV volumes and LVEF have been proposed, with direct volumetric quantification being the most accurate approach, since this is not affected by LV foreshortening and it is independent of geometric assumptions.4 However, beyond the assessment of LV function by simple endocardial excursion, tissue velocity and deformation analysis provides a more challenging echocardiographic approach to characterize global and regional myocardial function. The use of tissue velocity and deformation imaging has been applied increasingly in the fields of ventricular function, myocardial ischaemia and viability, valvular heart disease, and heart failure.5–8 Myocardial deformation imaging may be preferred over velocity assessment in patients with ischaemic heart disease and previous myocardial infarction, since deformation imaging may permit differentiation between passive motion (as can occur in infarcted tissue) and active deformation (which will discriminate between scar tissue and viable myocardium). Deformation or strain assessment was initially derived from tissue Doppler imaging, but had the disadvantage of being dependent on the ultrasound beam angle insonation, resulting in reduced applicability and diminished reproducibility. Recently, 2D speckle tracking has been introduced as a technique to measure multidirectional myocardial strain which is independent of the ultrasound beam angle insonation, and may overcome the limitations of strain derived from tissue Doppler imaging. Previous studies have validated 2D speckle tracking against sonomicrometry and tagged MRI.9,10 Recent technical developments have contributed to increasing use of 2D strain imaging in different clinical scenarios, permitting automated measurements of regional and global strain as well as myocardial rotation. In addition, the feasibility of 2D speckle tracking imaging for automatic measurements of LV volumes and LVEF has been demonstrated.11 With the use of automatic detection of the endocardial boundary at an end-diastolic frame, the software traces the endocardium throughout the cardiac cycle, and LV time–volume curves are derived. LV end-diastolic and end-systolic volumes are calculated as the maximum and minimum volume, and the LVEF is derived. Although promising, this technique showed a significant underestimation of LV end-diastolic volume and LVEF when compared with RT3DE and MRI.11 This underestimation is most probably related to the fact that the methodology relies upon 2D imaging, whereas 3D approaches (MRI and RT3DE) are not hampered by LV foreshortening and geometric assumptions. Another limitation of 2D speckle tracking is the out-of-plane motion of the speckles during the cardiac cycle, resulting in noise and suboptimal tracking, which leads to less accurate quantification of LV volumes and LVEF. Therefore, bearing in mind the complex 3D architecture of the left ventricle, myocardial motion, displacement, and deformation would be ideally derived from full-volume grey-scale data sets. Three-dimensional speckle tracking may overcome all the aforementioned limitations, providing more accurate assessment of regional and global LV function. Recent experimental and clinical data have demonstrated the feasibility of novel 3D speckle tracking methodology to measure regional LV strain and rotation.12 Nesser and co-workers13 have presented the first step towards the assessment of LV mechanics in true three dimensions. In 43 patients with varying LVEFs, 2D and 3D speckle tracking measurements of LV volumes were compared side by side and validated against MRI (as the gold standard). LV volumes were calculated by 2D speckle tracking as previously described,11 whereas assessment of 3D speckle tracking of LV volumes was performed with the use of a novel algorithm for pyramidal RT3DE data. From LV volume–time curves obtained with these methods, LV end-diastolic and end-systolic volumes were derived as the minimum and the maximum volumes. With MRI, the LV volumes were calculated by the disk area summation method, tracing the endocardial border in every short axis slice at end-diastole and end-systole. 2D and 3D speckle tracking measurements of LV volumes were feasible in all patients, with a relatively short post-processing time. However, whereas 2D speckle tracking imaging underestimated LV end-diastolic and end-systolic volumes with wide limits of agreement as compared with MRI, 3D speckle tracking measurements showed only a small bias for LV end-diastolic volume (−16 mL) and end-systolic volume (−1 mL). Importantly, 3D speckle tracking enabled automated quantification of the LV volume during the cardiac cycle and eliminated the need for endocardial border tracing in multiple consecutive frames. This approach results in a shorter time being needed for post-processing and analysis. One of the main limitations of this novel technology is the relatively high variability in measurements of LV volumes. Future technical improvements in image quality and temporal resolution of RT3DE may reduce the current variability of LV 3D speckle tracking measurements. Finally, additional studies to validate the assessment of regional LV mechanics (displacement, velocity, strain, and rotation) with 3D speckle tracking imaging are warranted in order to provide further insight into the true 3D ventricular physiology and to implement this technique in routine clinical practice. Conflict of interest: none declared.