Background
Myocardial strain is useful in the evaluation of left ventricular function using high–frame rate two-dimensional (2D) speckle-tracking echocardiography (STE). Three-dimensional (3D) STE allows 3D measurement of myocardial deformation, which is potentially more accurate, because it is not affected by through-plane motion. However, the low frame rates of 3D STE are a potential limitation that has not been studied to date. Whereas with 2D STE, high frame rates are necessary because speckles may move out of the imaging plane, it was hypothesized that because they always remain within the scan volume, they should be tracked with 3D STE, even if frame rates are considerably lower.
Methods
Twenty-seven subjects were studied, including 16 normal volunteers and 11 patients with nonischemic dilated cardiomyopathy, who underwent 2D (frame rate, 62 ± 9 frames/sec) and 3D echocardiographic imaging. In normal subjects, 3D imaging was performed at four different frame rates, achieved by varying the number of beats used for full-volume acquisition (six, four, two, and one). In the patients with dilated cardiomyopathy, 3D imaging was performed using a four-beat acquisition. The principal components of strain and the corresponding strain rates were calculated in 16 myocardial segments and averaged. Both 2D and 3D images were analyzed using TomTec software to avoid analysis-related differences.
Results
In normal subjects, strain and strain rate values were the same for 3D STE with six-beat and four-beat full-volume data sets, corresponding to 25 and 18 frames/sec, respectively. In contrast, 3D STE with one-beat and two-beat data sets, corresponding to 5 and 10 frames/sec, respectively, resulted in significantly lower values. Strain and strain rate values derived from six-beat and four-beat 3D data sets were not significantly lower than 2D STE–derived values, indicating that there was no loss of information due to lower frame rates. In patients with dilated cardiomyopathy, both 2D STE–derived and 3D STE–derived strain values were significantly reduced compared with normal hearts. The differences between 2D STE–derived and 3D STE–derived strain values echoed those noted in the normal subjects.
Conclusions
Three-dimensional speckle-tracking echocardiographic assessment of myocardial deformation is not compromised by low frame rates when derived from 18 or 25 frames/sec data sets but is underestimated with lower frame rates.
Myocardial strain measured using Doppler tissue imaging and high–frame rate two-dimensional (2D) speckle-tracking echocardiography (STE) has been shown to be useful in evaluating left ventricular (LV) function. Although real-time three-dimensional (3D) echocardiography has in the past decade evolved as a useful modality for imaging cardiac chambers and valves, and its ability to provide accurate estimates of LV volume throughout the cardiac cycle is well established, only a few studies have attempted 3D assessment of LV deformation using 3D measurements of myocardial strain. These studies have indicated that 3D STE may be more accurate than 2D STE in this context, because it is not affected by through-plane motion.
However, the relatively low frame rates of 3D STE are a potential limitation that has not been studied to date. Specifically, there is a concern that 3D STE–derived strain measurements may be underestimated because the frame rates of real-time 3D echocardiography are not sufficiently high to accurately capture all phases of the cardiac cycle. Although it is understandable that with 2D STE, high frame rates are necessary because speckles may move out of the imaging plane and thus be lost, 3D STE may be able to track them even at considerably lower frame rates, as long as they remain within the 3D echocardiographic scan volume. Accordingly, we hypothesized that comparable LV strain measurements could be obtained by 3D STE, despite the lower frame rates.
The present study was designed to test this hypothesis by evaluating the effects of frame rate on the 3D speckle-tracking echocardiographic measurements of LV strain and strain rate. The specific aims of this study were (1) to compare measurements of the principal components of LV systolic strain and strain rate obtained by 3D STE from full-volume data sets with different frame rates, (2) to determine the minimal frame rate that would result in strain and strain rate values comparable with those obtained on 2D STE, and (3) to test whether this minimal frame rate would also result in strain values on 3D STE comparable to those on 2D STE in patients with reduced LV function.
Methods
Subjects and Study Design
First, to achieve goals 1 and 2 above, we studied 16 normal volunteers (eight men; mean age, 45 ± 14 years; age range, 23–64 years) with good 2D and 3D echocardiographic image quality. Each subject underwent transthoracic 2D and 3D echocardiographic imaging at different frame rates to determine the minimal frame rate that would not result in a loss of strain information. Subsequently, to achieve goal 3, we studied a group of 11 patients with reduced LV function due to nonischemic dilated cardiomyopathy (DCM) (mean LV ejection fraction, 22 ± 9%; four men; mean age, 59 ± 11 years), in whom 3D STE–derived strain values obtained with the minimal frame rate determined in the normal subjects were compared with strain values on 2D STE to confirm that this frame rate was also sufficient to measure strain in dilated ventricles. To avoid analysis-related differences, 2D and 3D speckle-tracking echocardiographic analyses were both performed using software from the same manufacturer.
Echocardiographic Imaging
Imaging was performed by an experienced cardiac sonographer using an iE33 imaging system and an X5 matrix-array transducer (Phillips Medical Systems, Andover, MA) with the subject in the left lateral decubitus position. Two-dimensional imaging included apical two-chamber, three-chamber, and four-chamber views as well as a short-axis view obtained at the midpapillary level. Two-dimensional imaging settings were selected to maximize frame rate, resulting in a mean frame rate of 62 ± 9 frames/sec (range, 43–88 frames/sec). During 3D imaging, to ensure the inclusion of the entire left ventricle within the pyramidal scan volume, data sets were acquired using wide-angle full-volume mode during a single breath hold. To allow 3D imaging at four different frame rates, the number of beats used for full-volume acquisition was varied, including one, two, four, and six beats, corresponding to mean frame rates of 5 ± 1, 10 ± 1, 18 ± 2, and 25 ± 3 frames/sec, respectively ( Figure 1 ). The rest of the 3D imaging settings were kept unchanged throughout image acquisition.
Three-Dimensional Image Analysis
Three-dimensional images were analyzed using premarket, prototype 3D speckle-tracking echocardiographic software (4D LV Analysis; TomTec Imaging Systems, Unterschleissheim, Germany). After the LV long axis was manually aligned in the three apical views (four, three, and two chamber), the software automatically identified the LV endocardial border ( Figure 2 ), while including the papillary muscles in the LV cavity, and tracked it throughout the cardiac cycle, resulting in a dynamic cast of the LV cavity. Endocardial contours were manually adjusted when necessary to optimize boundary position and tracking. Finally, time curves of the three principal components of segmental myocardial strain as well as the corresponding strain rates were obtained using the standard 16-segment model. In addition, for each segment, a time curve of 3D strain was calculated as the vector sum of the longitudinal and circumferential components. All time curves were interpolated, resulting in effective temporal resolutions of 150 to 200 samples/sec. For each strain component, peak segmental strain and strain rate values were averaged over the 16 segments, resulting in mean 3D strain and strain rate values, as well as mean longitudinal, radial, and circumferential peak strain and strain rate values. In the normal subjects, these analyses were repeated for the one-beat, two-beat, four-beat, and six-beat full-volume data sets by a reader who was blinded to all prior measurements.
Two-Dimensional Image Analysis
Two-dimensional images were analyzed to quantify segmental myocardial strain throughout the cardiac cycle using the 2D speckle-tracking echocardiographic software (LV Analysis; TomTec Imaging Systems). Apical two-chamber, three-chamber, and four-chamber views were used to obtain longitudinal and radial strain components and the corresponding strain rates in 16 myocardial segments (six basal, six midventricular, and four apical segments), whereas the short-axis view was used to measure the circumferential component of strain and strain rate in six segments. Peak segmental values were averaged for each strain component, resulting in mean longitudinal, radial, and circumferential peak strain and strain rate values. Two-dimensional measurements were performed by the same reader, who was blinded to the 3D measurements.
Analysis of Regional Variability
To determine the effects of the frame rates of 3D echocardiographic data sets on regional variability in peak strain measurements, we calculated the standard deviation of the 16 segments in percentage of their mean measured value. This analysis was performed for the 3D speckle-tracking echocardiographic measurements at each of the four frame rates, as well as the 2D speckle-tracking echocardiographic measurements obtained in the 16 normal subjects.
Statistical Analysis
For each peak strain and strain rate value, one-way analysis of variance for independent samples was used to detect statistically significant differences between 3D STE with one-beat, two-beat, four-beat, and six-beat full-volume data sets. Then, paired two-tailed Student’s t tests were used to test differences between each type of acquisition (one, two, and four beats) against the six-beat acquisition (i.e., the highest frame rate). Finally, paired two-tailed Student’s t tests were also used to test differences between the six-beat acquisition and the 2D STE–derived values, as well as the differences between the normal subjects and the patients with DCM. P values < .05 were considered significant.
Results
Figure 3 shows examples of average segmental strain and strain rate time curves obtained in a study subject. In each subject, the 3D strain time curves were similar in their shape to an LV volume curve. Of note, the radial strain curve was inverted (i.e., radial strain was positive throughout the cardiac cycle), reflecting a systolic increase in myocardial thickness, while longitudinal and circumferential strain values were negative, reflecting systolic shortening of the ventricle in these two dimensions.
Figure 4 shows the average segmental peak 3D, longitudinal, circumferential, and radial components of strain and strain rate values obtained by 3D STE from the one-beat, two-beat, four-beat, and six-beat full-volume data sets. Peak values of all strain components and the corresponding strain rates were the same for 3D STE with six-beat and four-beat full-volume data sets, corresponding to 25 and 18 frames/sec, respectively. In contrast, 3D STE with one-beat and two-beat data sets, corresponding to 5 and 10 frames/sec, resulted in significantly lower strain and strain rate values.
Compared with 2D STE–derived strain and strain rate values, longitudinal and radial components obtained from six-beat and four-beat data sets using 3D STE were equal to or higher than the values on 2D STE ( Figure 5 ), indicating that there was no loss of information due to lower frame rates. Circumferential strain and strain rate were lower for 3D STE, but the difference did not reach statistical significance.
Table 1 shows the results of the regional variability analysis for the normal ventricles. Although for all strain components, regional variability was similar for the four-beat and six-beat acquisitions, it progressively increased with decreasing frame rate. For longitudinal and circumferential strain values, irrespective of frame rate, the variability was higher for 3D STE compared with 2D STE. In contrast, for radial strain, the regional variability was identical for 2D and the four-beat and six-beat 3D acquisitions.
| Strain | Six-beat 3D STE (25 ± 3 frames/sec) | Four-beat 3D STE (18 ± 2 frames/sec) | Two-beat 3D STE (10 ± 1 frames/sec) | One-beat 3D STE (5 ± 1 frames/sec) | 2D STE (62 ± 9 frames/sec) |
|---|---|---|---|---|---|
| 3D strain | 36% | 34% | 43% | 74% | — |
| Longitudinal | 39% | 41% | 44% | 85% | 22% |
| Circumferential | 54% | 50% | 63% | 93% | 39% |
| Radial | 48% | 45% | 54% | 71% | 48% |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
