International Journal of Cardiology 176 (2014) 360–366
Contents lists available at ScienceDirect
International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard
Three-dimensional speckle strain echocardiography is more accurate and efficient than 2D strain in the evaluation of left ventricular function Ting-Yan Xu a,1, Jing Ping Sun b,1, Alex Pui-wai Lee b, Xing Sheng Yang b, Zhiqing Qiao b, Xiuxia Luo b, Fang Fang b, Yan Li a, Cheuk-man Yu b,⁎, Ji-Guang Wang a,⁎⁎ a b
The Shanghai Institute of Hypertension, Ruijin Hospital, Shanghai Jiaotong University School of Medicine, Shanghai, China Division of Cardiology, S.H. Ho Cardiovascular and Stroke Centre, Prince of Wales Hospital, Chinese University of Hong Kong, Hong Kong, China
a r t i c l e
i n f o
Article history: Received 3 April 2014 Accepted 5 July 2014 Available online 12 July 2014 Keywords: Two-dimensional tracking echocardiography Three-dimensional tracking echocardiography Left ventricular strain
a b s t r a c t Background: Two-dimensional speckle tracking echocardiography (2DSTE) has been used widely in research, but rarely in clinical practice because data acquisition and analysis are time-consuming. By reducing the acquisition and analysis time, 3-dimensional STE may improve clinical impact. We investigated the feasibility of 3DSTE myocardial deformation, with comparison to 2DSTE. Methods: Transthoracic 3DSTE and 2DSTE were performed in 230 adults (138 men, 51 ± 14 years, and 142 hypertension, 10 heart failure and 78 normotensive subjects). The variables of LV deformation were analyzed using EchoPAC software. Results: The 3D LV longitudinal (LS) analysis was feasible in 84.9% of the study subjects, which was lower than the 2D analysis (97.2%). The success rates for circumferential strain (CS) and radial strain (RS) were similar between the 2D and 3D techniques. All magnitude of strains measured by 2DSTE and 3DSTE were significantly correlated. The magnitude of 3D LS and CS was lower, but the 3D RS is higher than that of 2DSTE (−18.5 ± 2.8 vs. −21.2 ± 3.5; 20.8 ± 4.1 vs. 21.7; and 50.0 ± 11.2 vs. 37.7 ± 12.6, respectively). Strains measured by 3DSTE exhibited stronger correlation with LV ejection fraction (EF) than that by 2DSTE. In inter- and intra-observer reproducibility for 3D LS, CS, RS and AS were acceptable. The mean time of analysis for LV volume, EF and strains was 116 s by 3DSTE, which was significantly shorter than that by 2DSTE (5 min, P b 0.0001). Conclusions: Three-dimensional STE is feasible and reproducible in the estimation of LV function, requires substantially less time than 2DSTE and is a more feasible technique for LV function assessment in clinical practice. © 2014 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Two-dimensional speckle tracking (2DSTE) is a novel technique of cardiac imaging for quantifying complex cardiac motion based on frame-to-frame tracking of ultrasonic speckles in gray scale 2D images. The deformation data are obtained by automatic measurement of the change of distance between 2 points of a left ventricular (LV) segment during the cardiac cycle. 2DSTE is a relatively angle-independent technology that can measure global and regional strains, strain rate, displacement, and velocity in longitudinal, radial, and circumferential directions [1]. 2DSTE has been used widely in clinical research settings and was found to be clinically useful in the assessment of cardiac systolic and ⁎ Corresponding author at: Institute of Vascular Medicine, Li Ka Shing Institute of Health Sciences, Division of Cardiology, Heart Education and Research Training (HEART) Centre, Department of Medicine and Therapeutics, Prince of Wales Hospital, The Chinese University of Hong Kong, Shatin, N.T., Hong Kong. Tel.: +852 2632 3594; fax: + 852 2637 5643. ⁎⁎ Corresponding author. E-mail address: [email protected] (C. Yu). 1 Co-first author.
http://dx.doi.org/10.1016/j.ijcard.2014.07.015 0167-5273/© 2014 Elsevier Ireland Ltd. All rights reserved.
diastolic function, ischemia [2–4], dyssynchrony [5,6], and other cardiac conditions [7]. However, it has not been widely adopted in clinical practice partly because data acquisition and analysis are time-consuming. Three-dimensional speckle tacking (3DSTE) was recently introduced as a new technique that can be used for strain analysis using volumetric data set of the LV. By reducing the time of data acquisition and analysis, 3DSTE might improve diagnostic usefulness [8]. The inter-vendor agreement of parameters of 3D speckle tracking measured with different vendors was poor [9]. Reference values should be specific for each system and baseline and follow-up data in longitudinal studies should be obtained using the same 3DSTE platform. The aim of the present study was to investigate the feasibility of 3DSTE in assessing LV longitudinal, circumferential, and radial strains, and to compare the data analyzed by 3DSTE and 2DSTE, respectively. 2. Methods 2.1. Study population From September 2012 to June 2013, we enrolled 230 consecutive patients. The exclusion criterion was the presence of a poor acoustic window, atrial fibrillation or patient
T.-Y. Xu et al. / International Journal of Cardiology 176 (2014) 360–366 unwillingness. The Ethics Committee of Ruijin Hospital, Shanghai Jiaotong University School of Medicine, The Prince of Wales Hospital, The Chinese University of Hong Kong approved the protocol. 2.2. Transthoracic 2D and 3D data acquisitions The patients were examined in the left lateral decubitus position using a Vivid E9 commercial ultrasound scanner (E9, GE Health Care, Milwaukee, WI) with phased-array transducers (M5S-D and 4V-D). Two-dimensional data acquisitions were obtained from parasternal long-axis and short-axis views and the three standard apical views. For each view, three consecutive cardiac cycles were recorded during quiet respiration. Grayscale recordings were optimized for LV evaluation at a frame rate of ≥50 frame/s. The mitral inflow velocity was obtained from the apical 4-chamber view by placing a pulsed-wave Doppler sample volume between mitral leaflet tips during diastole. The cursor of continuous wave Doppler was put between LV outflow and inflow tracts, and a tracing of LV outflow tract and inflow waves was recorded from the apical 5-chamber view. Tissue Doppler imaging (TDI) was performed to measure myocardial velocities. Pulsed-wave sample volume was placed at the septal corner of the mitral valve annulus. Early diastolic (E′) myocardial velocity was recorded. The 3D data acquisition was obtained in an adjustable volume divided into six subchamber views (Fig. 1). The acquisition of sub-volumes from sub-chamber views was steered electronically using the ultrasound system, with the transducer kept in a stable position. The acquisition was triggered to the R wave of the electrocardiogram on consecutive heartbeats. To ensure correct spatial registration of each sub-volume, four-beat acquisitions were obtained in an end-expiratory breath-holding position lasting for 6 to 8 s, at a frame rate of 25–50 frame/s. Care was taken to include the entire LV cavity and myocardium, including the epicardium, and to ensure an optimal temporal and spatial resolution. 2.3. Analysis of 2D and Doppler parameters Left atrial diameters, left ventricular diastolic diameters (LVDD), interventricular septal and posterior wall thickness (IVSTD and PWTD) were measured according to the guidelines of the American Society of Echocardiography [10]. The left ventricular volumes were measured by biplane 2D Simpson's method. The peak early (E) and late (A) transmitral flow velocities and deceleration time of E velocity were measured; the ratio of early-to-late peak velocities (E/A) was calculated. From TDI, the early diastolic (E′) at the septal corner of the mitral valve annulus was measured; the E/E′ ratio was calculated [11].
361
2.4. Analysis of 2DSTE and 3DSTE parameters All data of speckle tracking were analyzed off-line using a dedicated automated software (EchoPAC PC, Version 112; GE Health Care, Milwaukee, WI). For the 2D longitudinal speckle tracking analysis, three endocardial markers were placed in an end-diastolic frame at apical 4-, 2- and 3-chamber views, respectively. The software automatically tracked the contour of endocardium to cover the myocardial thickness of the entire LV wall. Adequate tracking could be verified in real time and corrected by adjusting the region of interest or by manually correcting the contour to ensure optimal tracking. Two-dimensional longitudinal strain was recorded in apical views. The mean LV global longitudinal strains were calculated for the 18 segments (6 basal, 6 mid, and 6 apical) in relation to the strain magnitude at aortic valve closure. Longitudinal systolic deformation was characterized as shortening. Using the strain definition, systolic indices yielded a negative value. Circumferential and radial strains were obtained from parasternal short-axis views. The sampling points were placed manually along the endocardium at LV base, middle and apex views during end-systole. A region of interest (ROI) was then generated by the software to cover the myocardial thickness of the LV wall. The LV wall was divided into 6 segments at short axis views. The ROI was adjusted manually to ensure that the inner margin tracing of LV endocardial border included the entire thickness of the LV myocardium. The software subsequently identified tissue speckles and tracked their movement frame-by-frame throughout the cardiac cycle. The data of 3DSTE was analyzed using automated software allowing a manual adjustment for myocardial border detection. The LV volume and ejection fraction (EF) and mass were presented as volume curve and values; the peak systolic strain values from each of the LV myocardial segments were presented as Bull's Eye figures (LV divided into 17 segments, Figs. 1 and 2). The correlations between LV ejection fraction and systolic parameters by 3D and 2D speckle tracking techniques were analyzed. The analysis time to obtain all the data including LV volumes, EF and strains by 3D and 3DSTE was calculated and compared with that for the 2D and 2DSTE analyses from 20 randomly selected cases by two experienced investigators.
2.5. Inter- and intra-observer reproducibility To assess intra-observer variability of offline analyses, 3D echocardiographic data from 68 randomly selected patients were analyzed twice by the same operator, with a 3-day interval between the two analyses. A second observer blinded to the results of first investigator repeated the measurements to assess interobserver variability.
Fig. 1. The analysis views and volume curve and the results of left ventricular (LV) end-diastolic and end-systolic volumes, ejection fraction and mass analyzed by three-dimensional speckle tracking echocardiography (3DSTE).
362
T.-Y. Xu et al. / International Journal of Cardiology 176 (2014) 360–366
Fig. 2. A. The figures showed the curves and results of left ventricular (LV) segmental and global longitudinal (left) and circumferential (right) strains. B. The figures showed the curves and results of left ventricular (LV) segmental and global radial (left), area (right) strains. 2.6. Statistical methods All statistical analyses were performed using SPSS for Windows version 20.0 (SPSS, Inc., Chicago, IL). The means (±standard deviations) and numbers (percentage) were compared using the Student's t test and the χ2 test, respectively. Pearson's correlation analysis was performed for the comparison between the 3DSTE and 2DSTE parameters and for the correlations of the 3DSTE and 2DSTE parameters with LVEF. Agreements between 3D and 2D strains were also assessed using the Bland–Altman method. Intra- and inter-observer variability was assessed using the reliability analysis and expressed as intra-class correlation coefficient. Intra- and inter-observer variability was calculated as the absolute difference divided by the average of the two measurements analyzed by the first operator and two measurements by the two operators. Statistical significance was a P-value b 0.05.
manual initialization and software running time. The adjustment time for the 3DSTE analysis was dependent on the quality of imaging, similarly as the 2D methods. The computation of LVEF by the bi-plane 2D Simpson's method needed to open 4- and 2-chamber views and to trace the endocardium 4 times, and required 151 ± 14 s. The LVM measurement required 14 ± 1 s. The 2DSTE analysis needed a measurement on the apical 4-, 2- and
Table 1 Clinical and conventional echocardiographic characteristics of the study population (n = 230).
3. Results
Characteristic
Population = 230 (normal, n = 78; hypertension, n = 142; CHF, n = 10)
3.1. Characteristics of the study population
Age, years Men, n (%) Body surface area, m2 Office systolic blood pressure, mm Hg Office diastolic blood pressure, mm Hg Heart rate, beat/min Interventricular septum thickness, mm Left ventricular end-diastole diameter, mm Left ventricular posterior wall thickness, mm Mitral inflow E/A ratio E/E′ ratio
51 ± 14 138 (60) 1.7 ± 0.2 151 ± 31 91 ± 20 70 ± 12 9.0 ± 2.1 46.0 ± 5.3 8.5 ± 1.7 1.2 ± 0.5 10.7 ± 6.5
The present analysis included a total of 230 subjects. The clinical and echocardiographic characteristics of the study population are shown in Table 1. 3.2. Time for the 2D, 2DSTE and 3DSTE analyses The mean time consumed for offline analysis of the 3D data set (LV volume, EF, mass and strains) in 20 subjects was 116 ± 8 s, including
Values are mean ± standard deviation or number of subjects (%).
T.-Y. Xu et al. / International Journal of Cardiology 176 (2014) 360–366
3-chamber views for longitudinal strain and parasternal basal, middle and apical views for circumferential and radial strains, and required 131 ± 17 s. Altogether, the measurement of LV volume, EF, and strains using the 2D technique required approximately 300 s, which was 5 times more than the 3D technique (Table 2). 3.3. Comparison between 2DSTE and 3DSTE analyses The analysis for the longitudinal strain by 3DSTE was feasible in 84.9% of the study subjects, which was significantly lower than that by 2DSTE (97.2%, P b 0.0001). However, the success rates for circumferential and radial strains did not statistically differ between 3DSTE and 2DSTE (89.7% vs. 90.9%, P = 0.1; 93.7% vs. 92.4%, P = 0.1, respectively). Table 3 shows the LV structural and functional measurements of the 2D and 3D echocardiography and the correlation coefficients for these LV parameters between the 2D and 3D techniques. The correlation coefficients were larger for the structural measurements, such as LV volume, than for the functional measurements, such as LVEF and strains. Nonetheless, as demonstrated in Fig. 3, all the strains measured by the 2DSTE and 3DSTE techniques were significantly correlated (P b 0.0001). In further analyses, we studied the correlation of various 2DSTE and 3DSTE strains with LVEF (biplane vs. 2D and 3D methods vs. 3D; Table 4). In general, the correlation coefficients for strains and LVEF measurement on the 3DSTE were greater than that on the 2DSTE. 3.4. Reproducibility of the 3DSTE strains The analysis for the intra-observer and inter-observer variability for 3DSTE strains is presented in Table 5. The mean percentage variability ranged from 4.8% to 7.9%. The intra-class correlation coefficients ranged from 0.90 to 0.95 with the lower boundary of all 95% confidence intervals greater than 0.84. 4. Discussion Our study showed that strain measurement by 3DSTE was feasible in 85–94% of the subjects and that the reproducibility of 3DSTE strains appeared to be acceptable and suitable for clinical use. The LV global strains by the 3DSTE and 2DSTE analyses were significantly correlated. The strains estimated by 3DSTE significantly correlated with LVEF and the correlation coefficients were greater than that for 2DSTE. The time consumed for analysis by 3DSTE was 5 times less than that needed for 2D Simpson's method for LVEF and 2DSTE for strain analysis. This is important for routine strain application in clinical practice. 4.1. Comparison of 3D and 2D strain parameters Our study showed that the LV volume, EF and mass derived from 3DSTE data set correlated well with that obtained by the traditional 2D methods. The correlation for longitudinal strain was better than that for circumferential and radial strains. These observations were consistent with the results of previous studies [12–14]. The results of LV global peak magnitude of strains varied between different studies. In Reant et al.'s study, the global longitudinal strain values measured by 3DSTE were lower than 2DSTE measurements by 1.3% [12]. However, they did not compare circumferential and radial
363
Table 3 Correlation between LV parameters estimated by 3D and 2D echocardiography techniques. Parameter
3D method
Left ventricular enddiastolic volume, ml Left ventricular endsystolic volume, ml Left ventricular ejection fraction, % Longitudinal Basal strain Middle Apical Global Circumferential Basal strain Middle Apical Global Radial strain Basal Middle Apical Global
2D method
Correlation coefficient
P value
89 ± 30
86 ± 33
0.93
b0.0001
33 ± 19
31 ± 21
0.94
b0.0001
64.3 ± 5.1
65.6 ± 4.9
0.67
b0.0001
0.40 0.49 0.36 0.66 0.29 0.36 0.32 0.43 0.33 0.38 0.32 0.37
b0.0001 b0.0001 b0.0001 b0.0001 b0.0001 b0.0001 b0.0001 b0.0001 b0.0001 b0.0001 b0.0001 b0.0001
−15.8 −19.4 −19.7 −18.5 −18.6 −21.4 −21.3 −20.8 41.0 52.2 60.2 50.0
± ± ± ± ± ± ± ± ± ± ± ±
4.1 4.5 3.6 2.8 3.7 3.5 4.6 4.1 11.5 14.1 17.0 11.2
−19.5 −21.2 −16.0 −21.2 −19.0 −20.6 −24.6 −21.7 37.9 46.9 30.3 37.7
± ± ± ± ± ± ± ± ± ± ± ±
3.0 3.0 2.5 3.5 5.6 4.1 6.3 5.2 15.9 20.2 18.4 12.6
Values are mean ± standard deviation, unless indicated otherwise.
strains between 3DSTE and 2DSTE. Perez de Isla et al. studied 30 cases using the Toshiba Medical Systems [13], and found that 3DSTE provided a complete radial and longitudinal LV strain assessment, similar to 2DSTE. In another study, Saito et al. examined longitudinal, circumferential, and radial 3DSTE strains in 46 healthy subjects using the Artida system [14]. The magnitude of longitudinal strain was significantly lower on 3DSTE than on 2DSTE analysis, whereas the circumferential strain was greater on the 3D than 2D analysis. Our study showed that the magnitude of 3D LV longitudinal and circumferential strains was lower than that of the corresponding 2D strains (− 19 ± 4 vs. − 21 ± 4%, P b 0.0001), and that the 3D radial strain was greater than 2D radial strain (50 ± 11 vs. 38 ± 13%, P b 0.0001); the magnitude of circumferential strain was smaller when measured by 3D than 2D method (20 ± 4 vs. 22 ± 5%, P = 0.04). The longitudinal strain measured by 3DSTE was consistently smaller than 2DSTE in different studies, which may be explained by the twisting of the heart and out-of-plane rotation of myocardial segments on the 3DSTE imaging. The conflicting results for the circumferential and radial strains might be explained by relatively small sample size of previous studies and by methodological difference among various systems and manufacturers. This problem was recently stressed and the solution certainly requires collaboration between manufacturers. The success rate for the analysis of longitudinal strain by 3DSTE was lower than 2DSTE (85% vs. 97%). According to our experience, it is technically more demanding to acquire 3D data set, because the entire LV volume needs to be captured through a single imaging window. In 2DSTE, apical views are acquired separately. Also, because the width of screen is limited by the frame rate in 3DSTE technique, the LV apex out-of-plane part may result in error values. We used parasternal basal, middle and apical short axis views to analyze LV global circumferential and radial strains for 2DSTE. These three short axis views were easier to obtain in the 3D data set, which might
Table 2 Time of analysis by two- (2DSTE) and three-dimensional speckle tracking echocardiography (3DSTE) imaging for left ventricular volume, ejection fraction, mass and strains. Variable
2DSTE analyses (time, seconds)
Longitudinal strain Circumferential and radial strains Left ventricular volume and ejection fraction Volumes and strains
68.0 63.3 151.2 296.3
Values are mean ± standard deviation.
± ± ± ±
9.2 8.2 13.5 30.0
3DSTE analyses (time, seconds)
P (2DE vs. 3DE)
50.5 ± 6.4
b0.0001
41.6 ± 4.6 116.2 ± 7.9
b0.0001 b0.0001
364
T.-Y. Xu et al. / International Journal of Cardiology 176 (2014) 360–366
Fig. 3. Correlation and agreement between the two- (2DSTE) and three-dimensional speckle tracking echocardiography (3DSTE) for longitudinal strain (panels A and B, respectively), circumferential strain (panels C and D, respectively), and radial strain (panels E and F, respectively).
Table 4 Correlation between left ventricular ejection fraction and parameters of left ventricular deformation estimated by two- (2DSTE) and three dimensional speckle tracking echocardiography (3DSTE). Correlation coefficient
Longitudinal strain Circumferential strain Radial strain Area strain Twist
P value
3DSTE
2DSTE
3DSTE
2DSTE
0.51 0.42 0.55 0.53 –
0.36 0.22 0.06 – 0.26
b0.0001 b0.0001 b0.0001 b0.0001 –
b0.0001 0.002 0.42 – b0.0001
explain why the success rates for circumferential and radial strains by 2DSTE and 3DSTE were similar. Previous studies have compared the analysis time between 3DSTE and 2DSTE, but did not take into account LV volume and EF. In our study, the time consumed for the analysis of 3DSTE was 60 s, which was 5 times less than that for the 2D analysis. In our view, this is an important advantage of 3DSTE compared with 2DSTE in clinical practice. 4.2. Comparison of 3D strains with LVEF Because LVEF is the reference index in terms of LV systolic performance evaluation, we attempted to compare various 3D strains with
T.-Y. Xu et al. / International Journal of Cardiology 176 (2014) 360–366 Table 5 Intra- and inter-observer reproducibility of the three-dimensional speckle tracking echocardiographic (3DSTE) strains. Strain
Longitudinal Circumferential Radial Area
Coefficient of variation (%)
InterIntraInterIntraInterIntraInterIntra-
6.4 5.2 7.9 6.5 7.8 6.3 5.4 4.8
Intra-class correlation coefficient
0.91 0.95 0.93 0.90 0.92 0.93 0.93 0.94
95% confidence interval Lower bound
Upper bound
0.85 0.91 0.88 0.84 0.87 0.88 0.88 0.90
0.95 0.97 0.96 0.94 0.95 0.95 0.96 0.96
365
addition, the LV wall was divided into 18 segments in the 2DSTE analysis, but 17 segments in the 3DSTE analysis, which might also influence the comparison between the two techniques.
P value
5. Conclusions
b0.0001 b0.0001 b0.0001 b0.0001 b0.0001 b0.0001 b0.0001 b0.0001
LVEF. As previous studies demonstrated using the 2DSTE approach, LV global longitudinal strain was significantly correlated with LVEF [15]. The relation between these two parameters is not linear, because heart failure patients with preserved EF may exhibit altered longitudinal contraction along with normal EF [16]. This may be accounted for by the fact that in the sub-endocardium, the fibers are roughly longitudinally oriented, with an angle of about 80° relative to the circumferential direction. The angle decreases toward the mid-wall, where the fibers are oriented in the circumferential direction. The angle decreases further to an oblique orientation of about 60° in the sub-epicardium. The above myocardial structure broadly determines the components of myocardial deformation. The subendocardial region contributes predominantly to the longitudinal mechanics of LV, whereas the mid-wall and the sub-epicardium contribute predominantly to the rotational motion [17]. The subendocardial layer is vulnerable to the effect of pressure and ischemia that are more commonly present in the elderly or hypertensive patients [18]. Our study revealed better correlations of LVEF with longitudinal, circumferential, radial and area strains by 3DSTE, than by 2DSTE. In a larger study, Kleijn et al. [19] evaluated 140 consecutive patients using the Artida system to investigate the reliability of area strain. Normal reference values for global and segmental area strains were obtained in 56 healthy subjects. These investigators observed a good correlation between global area strain and LVEF. Compared with this previous study, the correlation coefficient in our study was smaller; probably because our participants had normal LVEF, and hence our LVEF data had a narrow range of variation. Compared with previous studies, we investigated a larger population and hence could have a more valid comparison of the four strain components between 2D and 3D echocardiography. Our study also compared LV volume, EF, and mass by the 2D and 3D methods, with good correlations between the two techniques (r = 0.79–0.92) and less analysis time by 3DSTE than 2DSTE (116 vs. 296 s). 4.3. Potential clinical significance of 3DSTE strains As shown previously, 2D analysis of myocardial deformations was able to detect early systolic function abnormalities in patients with normal LVEF but already abnormal myocardial deformation mechanics. Moreover, 2D strain assessment is useful for the evaluation of prognosis, dyssynchrony, and treatment response in patients with heart failure [20–23]. However, routine clinical use of 2DSTE is limited because it is time-consuming. Our results showed that 3DSTE analysis will be a better option as it has a better representation of LV deformation and requires less time for obtaining strain measurements. 4.4. Limitations Since we studied on most of the population with normal LVEF, our results may not be generalized to patients with reduced LVEF. In
Our study showed that 3DSTE is a simple, feasible, and reproducible method to measure longitudinal, circumferential, and radial strains. The discordant results between 3DSTE and 2DSTE might be explained by the 3D cardiac motion that has been ignored in the current 2DSTE analysis. 3DSTE is a promising technique, which has the advantage of being significantly less time-consuming in analyzing real myocardial deformation, LV volume, EF and mass, as compared with the conventional 2D methods. Conflicts of interest There are no conflicts of interest. Acknowledgments The present study was financially supported by grants from the National Natural Science Foundation of China (grants 81201099). References [1] Biswas M, Sudhakar S, Nanda NC, et al. Two- and three-dimensional speckle tracking echocardiography: clinical applications and future directions. Echocardiography 2013;30:88–105. [2] Reant P, Labrousse L, Lafitte S, et al. Experimental validation of circumferential, longitudinal, and radial 2-dimensional strain during dobutamine stress echocardiography in ischemic conditions. J Am Coll Cardiol 2008;51:149–57. [3] Amundsen BH, Helle-Valle T, Edvardsen t, et al. Noninvasive myocardial strain measurement by speckle tracking echocardiography: validation against sonomicrometry and tagged magnetic resonance imaging. J Am Coll Cardiol 2006;47:789–93. [4] Helle-Valle T, Crosby J, Edvardsen t, et al. New noninvasive method for assessment of left ventricular rotation: speckle tracking echocardiography. Circulation 2005;112:3149–56. [5] Helm RH, Leclercq C, Faris OP, et al. Cardiac dyssynchrony analysis using circumferential versus longitudinal strain: implications for assessing cardiac resynchronization. Circulation 2005;111:2760–7. [6] Suffoletto MS, Dohi K, Cannesson M, Saba S, Gorcsn 3rd J. Novel specking radial strain from routine black-and-white echocardiographic images to quantify dyssynchrony and predict response to cardiac resynchronization therapy. Circulation 2006;113:960–8. [7] Serri K, Reant P, Lafitte M, et al. Global and regional myocardial function quantification by two-dimensional strain: application in hypertrophic cardiomyopathy. J Am Coll Cardiol 2006;47:1175–81. [8] Perez de Isla L, vivas D, zamorano J. Three-dimensional speckle tracking. Curr Cardiovasc Imaging Rep 2008;1:25–9. [9] Badano LP, Cucchini U, Muraru D, Al Nono O, Sarais C, Iliceto S. Use of threedimensional speckle tracking to assess left ventricular myocardial mechanics: inter-vendor consistency and reproducibility of strain measurements. Eur Heart J Cardiovasc Imaging 2013 Mar;14(3):285–93. [10] Sahn D, De Maria A, Kisslo J, Weyman A. The committee on m-mode standardization of the American Society of Echocardiography: results of a survey of echocardiographic measurements. Circulation 1978;58:1072–83. [11] Ommen SR, Nishimura RA, Appleton CP, et al. Clinical utility of Doppler echocardiography and tissue Doppler imaging in the estimation of left ventricular filling pressures: a comparative simultaneous Doppler-catheterization study. Circulation 2000;102:1788–94. [12] Reant P, Barbot L, Touche C, et al. Evaluation of global left ventricular systolic function using three-dimensional echocardiography speckle-tracking strain parameters. J Am Soc Echocardiogr 2012 Jan;25(1):68–79. [13] Perez de Isla L, Balcones CV, Fernandez-Golfin C, et al. Three-dimensional-wall motion tracking: a new and faster tool for myocardial strain assessment: comparison with two-dimensional-wall motion tracking. J Am Soc Echocardiogr 2009;22:325–30. [14] Saito K, Okura H, Watanabe N, et al. Comprehensive evaluation of left ventricular strain using speckle tracking echocardiography in normal adults: comparison of three-dimensional and two-dimensional approaches. J Am Soc Echocardiogr 2009;21:1025–30. [15] Delgado V, Mollema SA, Ypenburg C, et al. Relation between global left ventricular longitudinal strain assessed with novel automated function imaging and biplane left ventricular ejection fraction in patients with coronary artery disease. J Am Soc Echocardiogr 2008;21(11):1244–50.
366
T.-Y. Xu et al. / International Journal of Cardiology 176 (2014) 360–366
[16] Wang J, Khoury DS, Yue Y, Torre-Amione G, Nagueh SF. Preserved left ventricular twist and circumferential deformation, but depressed longitudinal and radial deformation in patients with diastolic heart failure. Eur Heart J 2008;29:1283–9. [17] Kocica MJ, Corno AF, Carreras-Costa F, et al. The helical ventricular myocardial band: global, three-dimensional, functional architecture of the ventricular myocardium. Eur J Cardiothorac Surg Apr 2006;29(Suppl. 1):S21–40. [18] Sun JP, Lee APW, Wu CQ, et al. Quantification of left ventricular regional myocardial function using two-dimensional speckle tracking echocardiography in healthy volunteers—a multi-center study. Int J Cardiol 2013 Jul 31;167(2):495–501. [19] Kleijn SA, Aly MF, Terwee CB, van Rossum AC, Kamp O. Three-dimensional speckle tracking echocardiography for automatic assessment of global and regional left ventricular function based on area strain. J Am Soc Echocardiogr 2011;24:314–21.
[20] Lafitte S, Perlant M, Reant P, et al. Impact of impaired myocardial deformations on exercise tolerance and prognosis in patients with asymptomatic aortic stenosis. Eur J Echocardiogr 2009;10:414–9. [21] Ernande L, Rietzschel ER, Bergerot C, et al. Impaired myocardial radial function in asymptomatic patients with type 2 diabetes mellitus: a speckle-tracking imaging study. J Am Soc Echocardiogr 2010;23:1266–72. [22] Reant P, Lafitte S, Bougteb H, et al. Effect of catheter ablation for isolated paroxysmal atrial fibrillation on longitudinal and circumferential left ventricular systolic function. Am J Cardiol 2009;103:232–7. [23] Mignot A, Donal E, Zaroui A, et al. Global longitudinal strain as a major predictor of cardiac events in patients with depressed left ventricular function: a multicenter study. J Am Soc Echocardiogr 2010;23:1019–24.
Contents lists available at ScienceDirect
International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard
Three-dimensional speckle strain echocardiography is more accurate and efficient than 2D strain in the evaluation of left ventricular function Ting-Yan Xu a,1, Jing Ping Sun b,1, Alex Pui-wai Lee b, Xing Sheng Yang b, Zhiqing Qiao b, Xiuxia Luo b, Fang Fang b, Yan Li a, Cheuk-man Yu b,⁎, Ji-Guang Wang a,⁎⁎ a b
The Shanghai Institute of Hypertension, Ruijin Hospital, Shanghai Jiaotong University School of Medicine, Shanghai, China Division of Cardiology, S.H. Ho Cardiovascular and Stroke Centre, Prince of Wales Hospital, Chinese University of Hong Kong, Hong Kong, China
a r t i c l e
i n f o
Article history: Received 3 April 2014 Accepted 5 July 2014 Available online 12 July 2014 Keywords: Two-dimensional tracking echocardiography Three-dimensional tracking echocardiography Left ventricular strain
a b s t r a c t Background: Two-dimensional speckle tracking echocardiography (2DSTE) has been used widely in research, but rarely in clinical practice because data acquisition and analysis are time-consuming. By reducing the acquisition and analysis time, 3-dimensional STE may improve clinical impact. We investigated the feasibility of 3DSTE myocardial deformation, with comparison to 2DSTE. Methods: Transthoracic 3DSTE and 2DSTE were performed in 230 adults (138 men, 51 ± 14 years, and 142 hypertension, 10 heart failure and 78 normotensive subjects). The variables of LV deformation were analyzed using EchoPAC software. Results: The 3D LV longitudinal (LS) analysis was feasible in 84.9% of the study subjects, which was lower than the 2D analysis (97.2%). The success rates for circumferential strain (CS) and radial strain (RS) were similar between the 2D and 3D techniques. All magnitude of strains measured by 2DSTE and 3DSTE were significantly correlated. The magnitude of 3D LS and CS was lower, but the 3D RS is higher than that of 2DSTE (−18.5 ± 2.8 vs. −21.2 ± 3.5; 20.8 ± 4.1 vs. 21.7; and 50.0 ± 11.2 vs. 37.7 ± 12.6, respectively). Strains measured by 3DSTE exhibited stronger correlation with LV ejection fraction (EF) than that by 2DSTE. In inter- and intra-observer reproducibility for 3D LS, CS, RS and AS were acceptable. The mean time of analysis for LV volume, EF and strains was 116 s by 3DSTE, which was significantly shorter than that by 2DSTE (5 min, P b 0.0001). Conclusions: Three-dimensional STE is feasible and reproducible in the estimation of LV function, requires substantially less time than 2DSTE and is a more feasible technique for LV function assessment in clinical practice. © 2014 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Two-dimensional speckle tracking (2DSTE) is a novel technique of cardiac imaging for quantifying complex cardiac motion based on frame-to-frame tracking of ultrasonic speckles in gray scale 2D images. The deformation data are obtained by automatic measurement of the change of distance between 2 points of a left ventricular (LV) segment during the cardiac cycle. 2DSTE is a relatively angle-independent technology that can measure global and regional strains, strain rate, displacement, and velocity in longitudinal, radial, and circumferential directions [1]. 2DSTE has been used widely in clinical research settings and was found to be clinically useful in the assessment of cardiac systolic and ⁎ Corresponding author at: Institute of Vascular Medicine, Li Ka Shing Institute of Health Sciences, Division of Cardiology, Heart Education and Research Training (HEART) Centre, Department of Medicine and Therapeutics, Prince of Wales Hospital, The Chinese University of Hong Kong, Shatin, N.T., Hong Kong. Tel.: +852 2632 3594; fax: + 852 2637 5643. ⁎⁎ Corresponding author. E-mail address: [email protected] (C. Yu). 1 Co-first author.
http://dx.doi.org/10.1016/j.ijcard.2014.07.015 0167-5273/© 2014 Elsevier Ireland Ltd. All rights reserved.
diastolic function, ischemia [2–4], dyssynchrony [5,6], and other cardiac conditions [7]. However, it has not been widely adopted in clinical practice partly because data acquisition and analysis are time-consuming. Three-dimensional speckle tacking (3DSTE) was recently introduced as a new technique that can be used for strain analysis using volumetric data set of the LV. By reducing the time of data acquisition and analysis, 3DSTE might improve diagnostic usefulness [8]. The inter-vendor agreement of parameters of 3D speckle tracking measured with different vendors was poor [9]. Reference values should be specific for each system and baseline and follow-up data in longitudinal studies should be obtained using the same 3DSTE platform. The aim of the present study was to investigate the feasibility of 3DSTE in assessing LV longitudinal, circumferential, and radial strains, and to compare the data analyzed by 3DSTE and 2DSTE, respectively. 2. Methods 2.1. Study population From September 2012 to June 2013, we enrolled 230 consecutive patients. The exclusion criterion was the presence of a poor acoustic window, atrial fibrillation or patient
T.-Y. Xu et al. / International Journal of Cardiology 176 (2014) 360–366 unwillingness. The Ethics Committee of Ruijin Hospital, Shanghai Jiaotong University School of Medicine, The Prince of Wales Hospital, The Chinese University of Hong Kong approved the protocol. 2.2. Transthoracic 2D and 3D data acquisitions The patients were examined in the left lateral decubitus position using a Vivid E9 commercial ultrasound scanner (E9, GE Health Care, Milwaukee, WI) with phased-array transducers (M5S-D and 4V-D). Two-dimensional data acquisitions were obtained from parasternal long-axis and short-axis views and the three standard apical views. For each view, three consecutive cardiac cycles were recorded during quiet respiration. Grayscale recordings were optimized for LV evaluation at a frame rate of ≥50 frame/s. The mitral inflow velocity was obtained from the apical 4-chamber view by placing a pulsed-wave Doppler sample volume between mitral leaflet tips during diastole. The cursor of continuous wave Doppler was put between LV outflow and inflow tracts, and a tracing of LV outflow tract and inflow waves was recorded from the apical 5-chamber view. Tissue Doppler imaging (TDI) was performed to measure myocardial velocities. Pulsed-wave sample volume was placed at the septal corner of the mitral valve annulus. Early diastolic (E′) myocardial velocity was recorded. The 3D data acquisition was obtained in an adjustable volume divided into six subchamber views (Fig. 1). The acquisition of sub-volumes from sub-chamber views was steered electronically using the ultrasound system, with the transducer kept in a stable position. The acquisition was triggered to the R wave of the electrocardiogram on consecutive heartbeats. To ensure correct spatial registration of each sub-volume, four-beat acquisitions were obtained in an end-expiratory breath-holding position lasting for 6 to 8 s, at a frame rate of 25–50 frame/s. Care was taken to include the entire LV cavity and myocardium, including the epicardium, and to ensure an optimal temporal and spatial resolution. 2.3. Analysis of 2D and Doppler parameters Left atrial diameters, left ventricular diastolic diameters (LVDD), interventricular septal and posterior wall thickness (IVSTD and PWTD) were measured according to the guidelines of the American Society of Echocardiography [10]. The left ventricular volumes were measured by biplane 2D Simpson's method. The peak early (E) and late (A) transmitral flow velocities and deceleration time of E velocity were measured; the ratio of early-to-late peak velocities (E/A) was calculated. From TDI, the early diastolic (E′) at the septal corner of the mitral valve annulus was measured; the E/E′ ratio was calculated [11].
361
2.4. Analysis of 2DSTE and 3DSTE parameters All data of speckle tracking were analyzed off-line using a dedicated automated software (EchoPAC PC, Version 112; GE Health Care, Milwaukee, WI). For the 2D longitudinal speckle tracking analysis, three endocardial markers were placed in an end-diastolic frame at apical 4-, 2- and 3-chamber views, respectively. The software automatically tracked the contour of endocardium to cover the myocardial thickness of the entire LV wall. Adequate tracking could be verified in real time and corrected by adjusting the region of interest or by manually correcting the contour to ensure optimal tracking. Two-dimensional longitudinal strain was recorded in apical views. The mean LV global longitudinal strains were calculated for the 18 segments (6 basal, 6 mid, and 6 apical) in relation to the strain magnitude at aortic valve closure. Longitudinal systolic deformation was characterized as shortening. Using the strain definition, systolic indices yielded a negative value. Circumferential and radial strains were obtained from parasternal short-axis views. The sampling points were placed manually along the endocardium at LV base, middle and apex views during end-systole. A region of interest (ROI) was then generated by the software to cover the myocardial thickness of the LV wall. The LV wall was divided into 6 segments at short axis views. The ROI was adjusted manually to ensure that the inner margin tracing of LV endocardial border included the entire thickness of the LV myocardium. The software subsequently identified tissue speckles and tracked their movement frame-by-frame throughout the cardiac cycle. The data of 3DSTE was analyzed using automated software allowing a manual adjustment for myocardial border detection. The LV volume and ejection fraction (EF) and mass were presented as volume curve and values; the peak systolic strain values from each of the LV myocardial segments were presented as Bull's Eye figures (LV divided into 17 segments, Figs. 1 and 2). The correlations between LV ejection fraction and systolic parameters by 3D and 2D speckle tracking techniques were analyzed. The analysis time to obtain all the data including LV volumes, EF and strains by 3D and 3DSTE was calculated and compared with that for the 2D and 2DSTE analyses from 20 randomly selected cases by two experienced investigators.
2.5. Inter- and intra-observer reproducibility To assess intra-observer variability of offline analyses, 3D echocardiographic data from 68 randomly selected patients were analyzed twice by the same operator, with a 3-day interval between the two analyses. A second observer blinded to the results of first investigator repeated the measurements to assess interobserver variability.
Fig. 1. The analysis views and volume curve and the results of left ventricular (LV) end-diastolic and end-systolic volumes, ejection fraction and mass analyzed by three-dimensional speckle tracking echocardiography (3DSTE).
362
T.-Y. Xu et al. / International Journal of Cardiology 176 (2014) 360–366
Fig. 2. A. The figures showed the curves and results of left ventricular (LV) segmental and global longitudinal (left) and circumferential (right) strains. B. The figures showed the curves and results of left ventricular (LV) segmental and global radial (left), area (right) strains. 2.6. Statistical methods All statistical analyses were performed using SPSS for Windows version 20.0 (SPSS, Inc., Chicago, IL). The means (±standard deviations) and numbers (percentage) were compared using the Student's t test and the χ2 test, respectively. Pearson's correlation analysis was performed for the comparison between the 3DSTE and 2DSTE parameters and for the correlations of the 3DSTE and 2DSTE parameters with LVEF. Agreements between 3D and 2D strains were also assessed using the Bland–Altman method. Intra- and inter-observer variability was assessed using the reliability analysis and expressed as intra-class correlation coefficient. Intra- and inter-observer variability was calculated as the absolute difference divided by the average of the two measurements analyzed by the first operator and two measurements by the two operators. Statistical significance was a P-value b 0.05.
manual initialization and software running time. The adjustment time for the 3DSTE analysis was dependent on the quality of imaging, similarly as the 2D methods. The computation of LVEF by the bi-plane 2D Simpson's method needed to open 4- and 2-chamber views and to trace the endocardium 4 times, and required 151 ± 14 s. The LVM measurement required 14 ± 1 s. The 2DSTE analysis needed a measurement on the apical 4-, 2- and
Table 1 Clinical and conventional echocardiographic characteristics of the study population (n = 230).
3. Results
Characteristic
Population = 230 (normal, n = 78; hypertension, n = 142; CHF, n = 10)
3.1. Characteristics of the study population
Age, years Men, n (%) Body surface area, m2 Office systolic blood pressure, mm Hg Office diastolic blood pressure, mm Hg Heart rate, beat/min Interventricular septum thickness, mm Left ventricular end-diastole diameter, mm Left ventricular posterior wall thickness, mm Mitral inflow E/A ratio E/E′ ratio
51 ± 14 138 (60) 1.7 ± 0.2 151 ± 31 91 ± 20 70 ± 12 9.0 ± 2.1 46.0 ± 5.3 8.5 ± 1.7 1.2 ± 0.5 10.7 ± 6.5
The present analysis included a total of 230 subjects. The clinical and echocardiographic characteristics of the study population are shown in Table 1. 3.2. Time for the 2D, 2DSTE and 3DSTE analyses The mean time consumed for offline analysis of the 3D data set (LV volume, EF, mass and strains) in 20 subjects was 116 ± 8 s, including
Values are mean ± standard deviation or number of subjects (%).
T.-Y. Xu et al. / International Journal of Cardiology 176 (2014) 360–366
3-chamber views for longitudinal strain and parasternal basal, middle and apical views for circumferential and radial strains, and required 131 ± 17 s. Altogether, the measurement of LV volume, EF, and strains using the 2D technique required approximately 300 s, which was 5 times more than the 3D technique (Table 2). 3.3. Comparison between 2DSTE and 3DSTE analyses The analysis for the longitudinal strain by 3DSTE was feasible in 84.9% of the study subjects, which was significantly lower than that by 2DSTE (97.2%, P b 0.0001). However, the success rates for circumferential and radial strains did not statistically differ between 3DSTE and 2DSTE (89.7% vs. 90.9%, P = 0.1; 93.7% vs. 92.4%, P = 0.1, respectively). Table 3 shows the LV structural and functional measurements of the 2D and 3D echocardiography and the correlation coefficients for these LV parameters between the 2D and 3D techniques. The correlation coefficients were larger for the structural measurements, such as LV volume, than for the functional measurements, such as LVEF and strains. Nonetheless, as demonstrated in Fig. 3, all the strains measured by the 2DSTE and 3DSTE techniques were significantly correlated (P b 0.0001). In further analyses, we studied the correlation of various 2DSTE and 3DSTE strains with LVEF (biplane vs. 2D and 3D methods vs. 3D; Table 4). In general, the correlation coefficients for strains and LVEF measurement on the 3DSTE were greater than that on the 2DSTE. 3.4. Reproducibility of the 3DSTE strains The analysis for the intra-observer and inter-observer variability for 3DSTE strains is presented in Table 5. The mean percentage variability ranged from 4.8% to 7.9%. The intra-class correlation coefficients ranged from 0.90 to 0.95 with the lower boundary of all 95% confidence intervals greater than 0.84. 4. Discussion Our study showed that strain measurement by 3DSTE was feasible in 85–94% of the subjects and that the reproducibility of 3DSTE strains appeared to be acceptable and suitable for clinical use. The LV global strains by the 3DSTE and 2DSTE analyses were significantly correlated. The strains estimated by 3DSTE significantly correlated with LVEF and the correlation coefficients were greater than that for 2DSTE. The time consumed for analysis by 3DSTE was 5 times less than that needed for 2D Simpson's method for LVEF and 2DSTE for strain analysis. This is important for routine strain application in clinical practice. 4.1. Comparison of 3D and 2D strain parameters Our study showed that the LV volume, EF and mass derived from 3DSTE data set correlated well with that obtained by the traditional 2D methods. The correlation for longitudinal strain was better than that for circumferential and radial strains. These observations were consistent with the results of previous studies [12–14]. The results of LV global peak magnitude of strains varied between different studies. In Reant et al.'s study, the global longitudinal strain values measured by 3DSTE were lower than 2DSTE measurements by 1.3% [12]. However, they did not compare circumferential and radial
363
Table 3 Correlation between LV parameters estimated by 3D and 2D echocardiography techniques. Parameter
3D method
Left ventricular enddiastolic volume, ml Left ventricular endsystolic volume, ml Left ventricular ejection fraction, % Longitudinal Basal strain Middle Apical Global Circumferential Basal strain Middle Apical Global Radial strain Basal Middle Apical Global
2D method
Correlation coefficient
P value
89 ± 30
86 ± 33
0.93
b0.0001
33 ± 19
31 ± 21
0.94
b0.0001
64.3 ± 5.1
65.6 ± 4.9
0.67
b0.0001
0.40 0.49 0.36 0.66 0.29 0.36 0.32 0.43 0.33 0.38 0.32 0.37
b0.0001 b0.0001 b0.0001 b0.0001 b0.0001 b0.0001 b0.0001 b0.0001 b0.0001 b0.0001 b0.0001 b0.0001
−15.8 −19.4 −19.7 −18.5 −18.6 −21.4 −21.3 −20.8 41.0 52.2 60.2 50.0
± ± ± ± ± ± ± ± ± ± ± ±
4.1 4.5 3.6 2.8 3.7 3.5 4.6 4.1 11.5 14.1 17.0 11.2
−19.5 −21.2 −16.0 −21.2 −19.0 −20.6 −24.6 −21.7 37.9 46.9 30.3 37.7
± ± ± ± ± ± ± ± ± ± ± ±
3.0 3.0 2.5 3.5 5.6 4.1 6.3 5.2 15.9 20.2 18.4 12.6
Values are mean ± standard deviation, unless indicated otherwise.
strains between 3DSTE and 2DSTE. Perez de Isla et al. studied 30 cases using the Toshiba Medical Systems [13], and found that 3DSTE provided a complete radial and longitudinal LV strain assessment, similar to 2DSTE. In another study, Saito et al. examined longitudinal, circumferential, and radial 3DSTE strains in 46 healthy subjects using the Artida system [14]. The magnitude of longitudinal strain was significantly lower on 3DSTE than on 2DSTE analysis, whereas the circumferential strain was greater on the 3D than 2D analysis. Our study showed that the magnitude of 3D LV longitudinal and circumferential strains was lower than that of the corresponding 2D strains (− 19 ± 4 vs. − 21 ± 4%, P b 0.0001), and that the 3D radial strain was greater than 2D radial strain (50 ± 11 vs. 38 ± 13%, P b 0.0001); the magnitude of circumferential strain was smaller when measured by 3D than 2D method (20 ± 4 vs. 22 ± 5%, P = 0.04). The longitudinal strain measured by 3DSTE was consistently smaller than 2DSTE in different studies, which may be explained by the twisting of the heart and out-of-plane rotation of myocardial segments on the 3DSTE imaging. The conflicting results for the circumferential and radial strains might be explained by relatively small sample size of previous studies and by methodological difference among various systems and manufacturers. This problem was recently stressed and the solution certainly requires collaboration between manufacturers. The success rate for the analysis of longitudinal strain by 3DSTE was lower than 2DSTE (85% vs. 97%). According to our experience, it is technically more demanding to acquire 3D data set, because the entire LV volume needs to be captured through a single imaging window. In 2DSTE, apical views are acquired separately. Also, because the width of screen is limited by the frame rate in 3DSTE technique, the LV apex out-of-plane part may result in error values. We used parasternal basal, middle and apical short axis views to analyze LV global circumferential and radial strains for 2DSTE. These three short axis views were easier to obtain in the 3D data set, which might
Table 2 Time of analysis by two- (2DSTE) and three-dimensional speckle tracking echocardiography (3DSTE) imaging for left ventricular volume, ejection fraction, mass and strains. Variable
2DSTE analyses (time, seconds)
Longitudinal strain Circumferential and radial strains Left ventricular volume and ejection fraction Volumes and strains
68.0 63.3 151.2 296.3
Values are mean ± standard deviation.
± ± ± ±
9.2 8.2 13.5 30.0
3DSTE analyses (time, seconds)
P (2DE vs. 3DE)
50.5 ± 6.4
b0.0001
41.6 ± 4.6 116.2 ± 7.9
b0.0001 b0.0001
364
T.-Y. Xu et al. / International Journal of Cardiology 176 (2014) 360–366
Fig. 3. Correlation and agreement between the two- (2DSTE) and three-dimensional speckle tracking echocardiography (3DSTE) for longitudinal strain (panels A and B, respectively), circumferential strain (panels C and D, respectively), and radial strain (panels E and F, respectively).
Table 4 Correlation between left ventricular ejection fraction and parameters of left ventricular deformation estimated by two- (2DSTE) and three dimensional speckle tracking echocardiography (3DSTE). Correlation coefficient
Longitudinal strain Circumferential strain Radial strain Area strain Twist
P value
3DSTE
2DSTE
3DSTE
2DSTE
0.51 0.42 0.55 0.53 –
0.36 0.22 0.06 – 0.26
b0.0001 b0.0001 b0.0001 b0.0001 –
b0.0001 0.002 0.42 – b0.0001
explain why the success rates for circumferential and radial strains by 2DSTE and 3DSTE were similar. Previous studies have compared the analysis time between 3DSTE and 2DSTE, but did not take into account LV volume and EF. In our study, the time consumed for the analysis of 3DSTE was 60 s, which was 5 times less than that for the 2D analysis. In our view, this is an important advantage of 3DSTE compared with 2DSTE in clinical practice. 4.2. Comparison of 3D strains with LVEF Because LVEF is the reference index in terms of LV systolic performance evaluation, we attempted to compare various 3D strains with
T.-Y. Xu et al. / International Journal of Cardiology 176 (2014) 360–366 Table 5 Intra- and inter-observer reproducibility of the three-dimensional speckle tracking echocardiographic (3DSTE) strains. Strain
Longitudinal Circumferential Radial Area
Coefficient of variation (%)
InterIntraInterIntraInterIntraInterIntra-
6.4 5.2 7.9 6.5 7.8 6.3 5.4 4.8
Intra-class correlation coefficient
0.91 0.95 0.93 0.90 0.92 0.93 0.93 0.94
95% confidence interval Lower bound
Upper bound
0.85 0.91 0.88 0.84 0.87 0.88 0.88 0.90
0.95 0.97 0.96 0.94 0.95 0.95 0.96 0.96
365
addition, the LV wall was divided into 18 segments in the 2DSTE analysis, but 17 segments in the 3DSTE analysis, which might also influence the comparison between the two techniques.
P value
5. Conclusions
b0.0001 b0.0001 b0.0001 b0.0001 b0.0001 b0.0001 b0.0001 b0.0001
LVEF. As previous studies demonstrated using the 2DSTE approach, LV global longitudinal strain was significantly correlated with LVEF [15]. The relation between these two parameters is not linear, because heart failure patients with preserved EF may exhibit altered longitudinal contraction along with normal EF [16]. This may be accounted for by the fact that in the sub-endocardium, the fibers are roughly longitudinally oriented, with an angle of about 80° relative to the circumferential direction. The angle decreases toward the mid-wall, where the fibers are oriented in the circumferential direction. The angle decreases further to an oblique orientation of about 60° in the sub-epicardium. The above myocardial structure broadly determines the components of myocardial deformation. The subendocardial region contributes predominantly to the longitudinal mechanics of LV, whereas the mid-wall and the sub-epicardium contribute predominantly to the rotational motion [17]. The subendocardial layer is vulnerable to the effect of pressure and ischemia that are more commonly present in the elderly or hypertensive patients [18]. Our study revealed better correlations of LVEF with longitudinal, circumferential, radial and area strains by 3DSTE, than by 2DSTE. In a larger study, Kleijn et al. [19] evaluated 140 consecutive patients using the Artida system to investigate the reliability of area strain. Normal reference values for global and segmental area strains were obtained in 56 healthy subjects. These investigators observed a good correlation between global area strain and LVEF. Compared with this previous study, the correlation coefficient in our study was smaller; probably because our participants had normal LVEF, and hence our LVEF data had a narrow range of variation. Compared with previous studies, we investigated a larger population and hence could have a more valid comparison of the four strain components between 2D and 3D echocardiography. Our study also compared LV volume, EF, and mass by the 2D and 3D methods, with good correlations between the two techniques (r = 0.79–0.92) and less analysis time by 3DSTE than 2DSTE (116 vs. 296 s). 4.3. Potential clinical significance of 3DSTE strains As shown previously, 2D analysis of myocardial deformations was able to detect early systolic function abnormalities in patients with normal LVEF but already abnormal myocardial deformation mechanics. Moreover, 2D strain assessment is useful for the evaluation of prognosis, dyssynchrony, and treatment response in patients with heart failure [20–23]. However, routine clinical use of 2DSTE is limited because it is time-consuming. Our results showed that 3DSTE analysis will be a better option as it has a better representation of LV deformation and requires less time for obtaining strain measurements. 4.4. Limitations Since we studied on most of the population with normal LVEF, our results may not be generalized to patients with reduced LVEF. In
Our study showed that 3DSTE is a simple, feasible, and reproducible method to measure longitudinal, circumferential, and radial strains. The discordant results between 3DSTE and 2DSTE might be explained by the 3D cardiac motion that has been ignored in the current 2DSTE analysis. 3DSTE is a promising technique, which has the advantage of being significantly less time-consuming in analyzing real myocardial deformation, LV volume, EF and mass, as compared with the conventional 2D methods. Conflicts of interest There are no conflicts of interest. Acknowledgments The present study was financially supported by grants from the National Natural Science Foundation of China (grants 81201099). References [1] Biswas M, Sudhakar S, Nanda NC, et al. Two- and three-dimensional speckle tracking echocardiography: clinical applications and future directions. Echocardiography 2013;30:88–105. [2] Reant P, Labrousse L, Lafitte S, et al. Experimental validation of circumferential, longitudinal, and radial 2-dimensional strain during dobutamine stress echocardiography in ischemic conditions. J Am Coll Cardiol 2008;51:149–57. [3] Amundsen BH, Helle-Valle T, Edvardsen t, et al. Noninvasive myocardial strain measurement by speckle tracking echocardiography: validation against sonomicrometry and tagged magnetic resonance imaging. J Am Coll Cardiol 2006;47:789–93. [4] Helle-Valle T, Crosby J, Edvardsen t, et al. New noninvasive method for assessment of left ventricular rotation: speckle tracking echocardiography. Circulation 2005;112:3149–56. [5] Helm RH, Leclercq C, Faris OP, et al. Cardiac dyssynchrony analysis using circumferential versus longitudinal strain: implications for assessing cardiac resynchronization. Circulation 2005;111:2760–7. [6] Suffoletto MS, Dohi K, Cannesson M, Saba S, Gorcsn 3rd J. Novel specking radial strain from routine black-and-white echocardiographic images to quantify dyssynchrony and predict response to cardiac resynchronization therapy. Circulation 2006;113:960–8. [7] Serri K, Reant P, Lafitte M, et al. Global and regional myocardial function quantification by two-dimensional strain: application in hypertrophic cardiomyopathy. J Am Coll Cardiol 2006;47:1175–81. [8] Perez de Isla L, vivas D, zamorano J. Three-dimensional speckle tracking. Curr Cardiovasc Imaging Rep 2008;1:25–9. [9] Badano LP, Cucchini U, Muraru D, Al Nono O, Sarais C, Iliceto S. Use of threedimensional speckle tracking to assess left ventricular myocardial mechanics: inter-vendor consistency and reproducibility of strain measurements. Eur Heart J Cardiovasc Imaging 2013 Mar;14(3):285–93. [10] Sahn D, De Maria A, Kisslo J, Weyman A. The committee on m-mode standardization of the American Society of Echocardiography: results of a survey of echocardiographic measurements. Circulation 1978;58:1072–83. [11] Ommen SR, Nishimura RA, Appleton CP, et al. Clinical utility of Doppler echocardiography and tissue Doppler imaging in the estimation of left ventricular filling pressures: a comparative simultaneous Doppler-catheterization study. Circulation 2000;102:1788–94. [12] Reant P, Barbot L, Touche C, et al. Evaluation of global left ventricular systolic function using three-dimensional echocardiography speckle-tracking strain parameters. J Am Soc Echocardiogr 2012 Jan;25(1):68–79. [13] Perez de Isla L, Balcones CV, Fernandez-Golfin C, et al. Three-dimensional-wall motion tracking: a new and faster tool for myocardial strain assessment: comparison with two-dimensional-wall motion tracking. J Am Soc Echocardiogr 2009;22:325–30. [14] Saito K, Okura H, Watanabe N, et al. Comprehensive evaluation of left ventricular strain using speckle tracking echocardiography in normal adults: comparison of three-dimensional and two-dimensional approaches. J Am Soc Echocardiogr 2009;21:1025–30. [15] Delgado V, Mollema SA, Ypenburg C, et al. Relation between global left ventricular longitudinal strain assessed with novel automated function imaging and biplane left ventricular ejection fraction in patients with coronary artery disease. J Am Soc Echocardiogr 2008;21(11):1244–50.
366
T.-Y. Xu et al. / International Journal of Cardiology 176 (2014) 360–366
[16] Wang J, Khoury DS, Yue Y, Torre-Amione G, Nagueh SF. Preserved left ventricular twist and circumferential deformation, but depressed longitudinal and radial deformation in patients with diastolic heart failure. Eur Heart J 2008;29:1283–9. [17] Kocica MJ, Corno AF, Carreras-Costa F, et al. The helical ventricular myocardial band: global, three-dimensional, functional architecture of the ventricular myocardium. Eur J Cardiothorac Surg Apr 2006;29(Suppl. 1):S21–40. [18] Sun JP, Lee APW, Wu CQ, et al. Quantification of left ventricular regional myocardial function using two-dimensional speckle tracking echocardiography in healthy volunteers—a multi-center study. Int J Cardiol 2013 Jul 31;167(2):495–501. [19] Kleijn SA, Aly MF, Terwee CB, van Rossum AC, Kamp O. Three-dimensional speckle tracking echocardiography for automatic assessment of global and regional left ventricular function based on area strain. J Am Soc Echocardiogr 2011;24:314–21.
[20] Lafitte S, Perlant M, Reant P, et al. Impact of impaired myocardial deformations on exercise tolerance and prognosis in patients with asymptomatic aortic stenosis. Eur J Echocardiogr 2009;10:414–9. [21] Ernande L, Rietzschel ER, Bergerot C, et al. Impaired myocardial radial function in asymptomatic patients with type 2 diabetes mellitus: a speckle-tracking imaging study. J Am Soc Echocardiogr 2010;23:1266–72. [22] Reant P, Lafitte S, Bougteb H, et al. Effect of catheter ablation for isolated paroxysmal atrial fibrillation on longitudinal and circumferential left ventricular systolic function. Am J Cardiol 2009;103:232–7. [23] Mignot A, Donal E, Zaroui A, et al. Global longitudinal strain as a major predictor of cardiac events in patients with depressed left ventricular function: a multicenter study. J Am Soc Echocardiogr 2010;23:1019–24.
