|Year : 2017 | Volume
| Issue : 1 | Page : 18-23
Three-dimensional versus two-dimensional strain for the assessment of myocardial function: A case series
Monica Vinesh Dillikar1, Ashwin Venkateshvaran2, Banajit Barooah1, Reeta Varyani1, Prayaag Kini1, PK Dash1, Srikanth Sola1
1 Department of Cardiology, Sri Sathya Sai Institute of Higher Medical Sciences, Bengaluru, Karnataka, India
2 Department of Cardiovascular Sciences, Karolinska Institute, Stockholm, Sweden
|Date of Web Publication||7-Apr-2017|
Department of Cardiology, Sri Sathya Sai Institute of Higher Medical Sciences, EPIP Area, Whitefield, Bengaluru - 560 066, Karnataka
Source of Support: None, Conflict of Interest: None
Introduction: Two-dimensional (2D) strain assessment is an important diagnostic and prognostic tool in various clinical conditions, particularly coronary artery disease (CAD). However, these measurements are limited in that the information is obtained in only a single plane (2D). Three-dimensional (3D) strain tracks the myocardium in all 3D, potentially overcoming the limitation of 2D strain. The objective of this study was to establish normal values for 3D strain in a population of healthy, normal controls and to compare these values with 2D strain values. In addition, we sought to evaluate the utility of 3D strain in patients with known or suspected CAD. Methods: We conducted a prospective study at a single major tertiary care center. Individuals were recruited for the study and divided into two groups: a normal control group and a CAD group. Global longitudinal strain (GLS) and global circumferential and global radial strain were calculated by both 2D and 3D strain methods. In addition, 3D was used to calculate area strain. Results: We enrolled a total of 43 individuals (20 normal control group, mean age 33 ± 2.7 years, and 23 CAD group, mean age 57 ± 2.8 years, 80% male). Values for 3D strain were consistently lower for GLS and global circumferential strain in both groups compared with 2D measurements. In the control group, the mean 2D GLS was −20 ± 1.6% versus −17.5 ± 1.5% for 3D GLS (P < 0.001). Similarly, the mean 2D circumferential strain was −17.7 ± 2.3% versus −15.6 ± 2.1% for 3D circumferential strain (P < 0.001). Combining both groups, the sensitivity of GLS for CAD was 80% for 2D versus 93% for 3D. Similar findings were seen for global circumferential strain (sensitivity 87% for 2D vs. 100% for 3D). However, the sensitivity of 3D global radial strain was lower (93% for 2D vs. 47% for 3D). 3D strain data were acquired in a shorter time span compared with 2D (2.2 ± 1 min vs. 3 ± 1 min). Conclusions: 3D strain assessment of longitudinal and circumferential strain is similar but mildly reduced compared with 2D techniques, with similar sensitivity for CAD. Radial strain measurements by 3D, however, are not accurate and correlate poorly with 2D.
Keywords: Assessment of myocardial function, longitudinal and circumferential strain, three-dimensional versus two-dimensional strain
|How to cite this article:|
Dillikar MV, Venkateshvaran A, Barooah B, Varyani R, Kini P, Dash P K, Sola S. Three-dimensional versus two-dimensional strain for the assessment of myocardial function: A case series. J Indian Acad Echocardiogr Cardiovasc Imaging 2017;1:18-23
|How to cite this URL:|
Dillikar MV, Venkateshvaran A, Barooah B, Varyani R, Kini P, Dash P K, Sola S. Three-dimensional versus two-dimensional strain for the assessment of myocardial function: A case series. J Indian Acad Echocardiogr Cardiovasc Imaging [serial online] 2017 [cited 2017 Jun 24];1:18-23. Available from: http://www.jiaecho.org/text.asp?2017/1/1/18/204060
| Introduction|| |
Two-dimensional (2D) strain using speckle tracking echocardiography (ECG) has recently been shown to be highly accurate in identifying patients with chronic stable angina who have hemodynamically significant coronary artery disease (CAD). Several studies have found that various strain parameters – global longitudinal strain (GLS), postsystolic shortening, and postsystolic strain index, – are both accurate and have high interobserver reproducibility in this patient population.,,, However, these measurements are limited in that the information obtained is obtained in only a single plane (2D) whereas myocardial deformation occurs in all planes three-dimensional (3D).
3D strain is a new modality which tracks the myocardium in all 3 dimensions potentially overcoming the limitation of 2D strain. Several studies have suggested that 3D strain may be more accurate than 2D strain when using sonomicrometry at the reference standard as strain values are measured at end systolic frame as compared to the 2D strain values which is from peak systolic frame., In addition, global deformation indices can be obtained in a shorter span of time by obtaining a single large 3D volume. Finally, 3D strain exclusively allows for the measurement of “area strain (AS),” defined as longitudinal strain × circumferential strain, providing additional information on global myocardial function. However, data on 3D strain are still very limited, and it has not been studied in the Indian population.
The objective of this study was to establish normal values for 3D strain in a population of healthy, normal controls and to compare these values with 2D strain values. In addition, we sought to evaluate the utility of 3D strain in patients with known or suspected CAD.
| Methods|| |
We conducted a prospective study at a single major tertiary care center. Individuals were recruited for the study and divided into two arms: a normal control arm and a CAD arm. Healthy individuals were included if they were between the ages of 20–50 years, had no comorbidities, nonsmokers, and without any previous history of CAD. For individuals with suspected or known CAD, we included adults age >18 years with chronic stable angina scheduled for invasive coronary angiography and with normal left ventricular (LV) systolic function and no regional wall motion abnormalities (RWMAs) on ECG. Individuals in both the control and CAD arms were excluded if they had poor image quality (two or more segments not evaluable on transthoracic ECG), inability to breathe hold for the 3D images, or irregular heart rhythm. All individuals gave written informed consent in their native language. The study was approved by the Institutional Ethics Committee.
Transthoracic echocardiographic examinations were performed with commercially available machine (Vivid E9 XD Clear echo machine, GE Healthcare, Horton, Norway) equipped with M4S and 4V-phased array matrix transducer and BT13 software. 2D and 3D datasets were acquired in sequence with the patient in the left lateral decubitus position. Standard 2D grayscale images of the three standard apical views (four-chamber, two-chamber, and apical long axis) and three parasternal short axis views (basal, mid papillary muscle, and apical levels) were obtained. To optimize speckle tracking, frame rate was set at 56–90 frames/s for 2D strain and a frame rate of 25 frame/s for 3D strain analysis was maintained throughout the study. Subsequently, four-beat ECG-gated LV full volumes were acquired from the apical with breath held during end expiration. The data were stored in a raw format and then analyzed offline (EchoPac version 113, GE Healthcare, Horten, Norway). The quality of each acquisition was verified in each patient by ensuring optimal imaging of entire LV wall at each level with adequate frame rate.
Two-dimensional strain analysis and measurements
For 2D strain, the endocardial borders were traced at end systolic frame in three apical views (long axis, four-chamber, and two-chamber) for longitudinal strain and three parasternal short axis views (basal, mid papillary, and apical levels) to obtain radial and circumferential strain. The software then automatically tracked myocardial motion and any poorly traced segments were rejected. If a rejected segment was present, the region of interest was then readjusted by manually adjusting the endocardial trace. Segments that could not be satisfactorily evaluated due to suboptimal image quality or artifact were excluded from analysis. The region of interest size was kept the same for each patient.
Three-dimensional strain analysis and measurements
For 3D strain, a large volume acquisition from the apical window was used. Semi-automated contouring of the endocardial and epicardial borders of the left ventricle was done at end diastole and end systole, with manual adjustments made as necessary to ensure accurate tracking.
The LV was divided into 17 3D segments using standard segmentation. The following parameters of global myocardial deformation (Lagrangian strain) were measured: longitudinal strain, circumferential strain, radial strain, and AS. In addition, we reported LV volume, mass, and ejection fraction. The deformation parameters were reported as color-coded polar maps and time strain curves of an LV 17-segment model. The time taken for each method (2D vs. 3D) was recorded for comparison.
Invasive coronary angiography was performed in the study group with suspected CAD (Philips Allura clarity FD-10 C Lab or Artis Zee Siemens Lab). Diameters of reference and stenotic coronary arteries were measured by a computer-assisted quantitative method.
Data analysis was performed using the SPSS version 19.0 manufactured by IBM. All values were presented as a mean ± standard deviation for continuous variables and as a percentage of total patients for categorical variables. For all analysis, P ≤ 0.05 was considered statistically significant.
| Results|| |
We enrolled a total of 43 patients [Table 1]. Out of these, 20 were healthy normal controls and 23 were patients with CAD. In the normal control arm, the mean age was 33 ± 2.7 years, and 95% were male. In the CAD arm, the mean age was 57 ± 2.8 years, and 80% were male. All individuals in both groups had normal LV ejection fraction without RWMA. However, eight of the patients in the CAD arm were excluded due to poor image quality and/or persisting stitch artifact on 3D volume data. A total of 15 patients were analyzed in the CAD arm.
All global strain parameters were consistently lower when analyzed by 3D versus 2D methods [Table 2]a. In the normal control arm, the mean 2D GLS was −20 ± 1.6% versus −17.5 ± 1.5% for 3D GLS (P < 0.001) [Table 2]b. Similarly, the mean 2D circumferential strain was −17.7 ± 2.3% versus −15.6 ± 2.1% for 3D circumferential strain (P < 0.001) [Table 2]c. The 2D radial strain was +52 ± 5.3% versus + 41.4 ± 6.2% for 3D radial strain (P < 0.001) [Table 2]d. Similar reductions in 3D versus 2D strain values were seen in the CAD group [Table 3]a,[Table 3]b,[Table 3]c,[Table 3]d.
|Table 2a: Strain deformation parameters in healthy normal subjects (n=20)|
Click here to view
Combining both groups, the sensitivity of GLS for CAD was 80% for 2D versus 93% for 3D. Similar findings were seen for global circumferential strain (sensitivity 87% for 2D vs. 100% for 3D). However, the sensitivity of 3D global radial strain was lower (93% for 2D vs. 47% for 3D).
The time required for analysis of strain data was lower using 3D versus 2D (2.2 ± 1 min vs. 3 ± 1 min). The number of steps for 3D strain analysis was also lower as compared to 2D strain analysis [Table 4].
|Table 2c: Summary statistics for circumferential strain in normal patients|
Click here to view
| Discussion|| |
In this study, we have specifically focused on the comparison of 2D and 3D strain in assessing the longitudinal, circumferential, and radial strain and the utility of AS in clinical practice [Figure 1]. Through this study, we found that 3D strain measurements of GLS and global circumferential strain correlates well with 2D strain measurements in both normal controls as well as patients with CAD. However, the 3D values are typically 2%–4% lower than those obtained by 2D. After analyzing, we found that all the 3D strain can estimate the longitudinal and circumferential strain with acceptable accuracy [Figure 2]. However, the radial strain remains difficult, often bearing no correlation with the 2D and 3D strain value.
|Figure 1: (a) Left ventricular display set with two-dimensional strain analysis in 17-segment model along with Bull's eye map in a normal, healthy individual. (b) Corresponding left ventricular data display with right panel displaying the derived time-strain curves, the left panel and central panel displays three apical (4CH, 2CH, 3CH) and three short axis (apex, mid papillary, and basal), respectively. Seventeen-segment Bull's eye map is shown in right lower panel. (c) left ventricular display of area strain of −29% in of the same normal, healthy individual|
Click here to view
|Figure 2: (a) Left ventricular display set with two-dimensional longitudinal strain in 17-segment model and Bull's eye map in a patient with anterior wall infraction. Note the markedly reduced strain in apical segments and arrow in red pointing to the significant postsystolic strain. Corresponding three-dimensional strain analysis curves of longitudinal strain along with bulls eye map depicting the reduced strain the distal apical region. The right panel displaying the derived time-strain curves, the left panel and central panel displays three apical (4CH, 2CH, 3CH), and three short axis (apex, mid papillary and basal), respectively. Seventeen-segment Bull's eye map is shown in right lower panel, depicting reduced strain values in apical region. (b) Corresponding display two-dimensional and three-dimensional circumferential strain for different segment along with Bull's eye plot of above patient. The overall circumferential strain is reduced with red arrow pointing to the significant postsystolic strain in the anterior septum region, represented by yellow curve line in the panel. In three-dimensional strain analysis, the right panel displaying the derived time-strain curves, the left panel and central panel displays three apical (4CH, 2CH, 3CH) and three short axis (apex, mid papillary, and basal), respectively. Seventeen-segment Bull's eye map is shown in right lower panel, depicting overall reduced circumferential strain values with positive circumferential strain values in apical region. (c) Corresponding display of two-dimensional radial strain waveforms showing markedly reduced radial strain, the red arrow pointing to postsystolic strain. In three-dimensional strain analysis, the right panel displaying the derived time-strain curves, the left panel and central panel displays three apical (4CH, 2CH, 3CH) and three short axis (apex, mid papillary and basal), respectively. Seventeen-segment Bull's eye map is shown in right lower panel, depicting reduced radial strain values especially in apical region. (d) Left ventricular display of area strain of −10% in the patient with anterior wall infraction|
Click here to view
Several recent studies have proved that 3D strain analysis technique is trustworthy for evaluating global LV function.,,,
Reant et al. and Hayat et al. showed good correlations with 3D strain analysis global strain and convectional parameters of LV systolic function like ejection fraction. Moreover, 3D global longitudinal has been shown to be significantly lower in patients with ischemic heart disease and in hypertension as compared to the global circumferential strain which is reduced in patients with ischemic heart disease.
Jasaityte et al. had demonstrated that the global strain values specially the 3D longitudinal and circumferential strain values were lower in patients with ischemic heart disease. This did not hold true in our study where we had patients with CAD having normal range of 2D and 3D GLS.
Abate et al. had performed 3D strain analysis in patients with acute myocardial infraction undergoing primary percutaneous coronary intervention. A cutoff value of −11.1% for segmental 3D longitudinal strain had 92% sensitivity and 91% specificity for predicting the functional improvement at 6 months. In addition, 3D GLS was found to have incremental value over clinical and convectional echocardiographic variables in predicting global LV function improvement.
3D strain imaging technique allows measurement of area tracking expressed as AS.
AS is quantified by percentage of deformation in the LV endocardial surface area. It is a parameter integrating longitudinal and circumferential strain, providing a more global and comprehensive evaluation of LV systolic function.
Seo et al. had inferred that because AS integrates the two-directional component of LV myocardial deformation, it could reduce the tracking error and emphasize on the magnitude of deformation.
Regardless of the physiopathological disease like CAD, hypertension, diabetes, which can cause indirect or direct macro/micro-vascular abnormalities, endocardium is more susceptible to ischemia. Thus, it can be reasonable to expect that deteriorated LV function can be detected at the earlier stage by measuring the AS.
With all the published comparative study, the reported limits of agreement between the 2 techniques were ±5% with respect to bias.
3D strain analysis has very good reproducibility and less time-consuming making it very attractive. However, technical factors such as the need of good quality volumetric data without artifacts/dropouts make it challenging. In addition, the patient should not have any rhythm abnormalities and also should be capable of holding breath for at least four cardiac cycles to allow ECG gating. Intervendor variability on strain values still remains a challenging area. Furthermore, it cannot measure the postsystolic shortening/postsystolic index which can be analyzed with 2D strain.
There are several limitations of our study which must be noted. First, our sample size was small (n = 43), and our results may have been different with a larger sample size or diverse patient population. Second, the majority of our patients were young (33 ± 2.7 years) and mostly male reflecting the type of CAD patients seen at our institution where echo windows are adequate.
| Conclusions|| |
3D strain is less cumbersome to acquire and faster to analyze than 2D methods strain. However, image quality may be limited in some patients, and careful attention to acquisition techniques is need. Finally, 3D GLS and global circumferential strain correlates well with 2D methods, whereas radial stain has only poor correlation. Further studies are needed for validation of this technique across various patient populations.
We thank Prof G. K. Padmashree for all the statistical help provided in all the phases of the study.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Jasaityte R, Heyde B, D'hooge J. Current state of three-dimensional myocardial strain estimation using echocardiography. J Am Soc Echocardiogr 2013;26:15-28.
Wen H, Liang Z, Zhao Y, Yang K. Feasibility of detecting early left ventricular systolic dysfunction using global area strain: A novel index derived from three-dimensional speckle-tracking echocardiography. Eur J Echocardiogr 2011;12:910-6.
Badano LP, Cucchini U, Muraru D, Al Nono O, Sarais C, Iliceto S. Use of three-dimensional speckle tracking to assess left ventricular myocardial mechanics: inter-vendor consistency and reproducibility of strain measurements. Eur Heart J Cardiovasc Imaging. 2013;14:285-93.
Reant P, Barbot L, Touche C, Dijos M, Arsac F, Pillois X, et al.
Evaluation of global left ventricular systolic function using three-dimensional echocardiography speckle-tracking strain parameters. J Am Soc Echocardiogr 2012;25:68-79.
Sahn D, Ashraf M, Balbach T, Desrochers K. Oregon Health and science university: A new 3D strain method for processing 4D echo images can delineate regional myocardial dysfunction: Validation against sonomicrometry. JACC 2011: 57. E707.
Saito K, Okura H, Watanabe N, Hayashida A, Obase K, Imai K, 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;22:1025-30.
Pérez de Isla L, Balcones DV, Fernández-Golfín C, Marcos-Alberca P, Almería C, Rodrigo JL, 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.
Hayat D, Kloeckner M, Nahum J, Ecochard-Dugelay E, Dubois-Randé JL, Jean-François D, et al.
Comparison of real-time three-dimensional speckle tracking to magnetic resonance imaging in patients with coronary heart disease. Am J Cardiol 2012;109:180-6.
Abate E, Hoogslag GE, Antoni ML, Nucifora G, Delgado V, Holman ER, et al.
Value of three-dimensional speckle-tracking longitudinal strain for predicting improvement of left ventricular function after acute myocardial infarction. Am J Cardiol 2012;110:961-7.
Seo Y, Ishizu T, Enomoto Y, Sugimori H, Aonuma K. Endocardial surface area tracking for assessment of regional LV wall deformation with 3D speckle tracking imaging. JACC Cardiovasc Imaging 2011;4:358-65.
[Figure 1], [Figure 2]
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5], [Table 6], [Table 7], [Table 8], [Table 9], [Table 10]