- Email: [email protected]
Accuracy and Reproducibility of Real-Time Three-Dimensional Echocardiography for Assessment of Right Ventricular Volumes and Ejection Fraction in Children
Accuracy and Reproducibility of Real-Time Three-Dimensional Echocardiography for Assessment of Right Ventricular Volumes and Ejection Fraction in Children Xiuzhang Lu, Vyacheslav Nadvoretskiy, Liping Bu, Alan Stolpen, Nancy Ayres, Ricardo H. Pignatelli, John P. Kovalchin, Michelle Grenier, Berthold Klas, and Shuping Ge, MD, Houston, Texas; Iowa City, Iowa; and Munich, Germany
Background: Measurement of right ventricular (RV) volumes and ejection fraction (EF) by two-dimensional echocardiography has limited accuracy and reproducibility because of the complex RV geometry. Objectives: This study sought to validate real-time three-dimensional echocardiography (RT3DE) using a disk summation method for assessment of RV volumes and RVEF in children by comparing it with magnetic resonance imaging (MRI) measurements. Methods: A total of 20 children (mean age 10.6 ⫾ 2.8 years) were studied. Transthoracic RT3DE was performed using a RT3DE system to acquire full-volume RT3DE data sets from apical windows and data were processed offline using a software package. RV end-systolic volume and end-diastolic volume (EDV) were measured using a disk summation method by manually tracing the endocardial borders. RVEF was calculated as: RVEF ⫽ (EDV ⫺ end-systolic volume)/EDV ⫻ 100%. All participants also underwent MRI studies for comparison of RV indexes. Results: Of the 20 children, 3 were excluded because of poor or incomplete RV images (two RT3DE and one MRI study). For the remaining 17 children, good correlation and agreement between RT3DE and MRI were found (RVEDV: r ⫽ 0.98, P ⬍ .001, mean difference ⫽ ⫺7.0 ⫾ 9.0 mL, P ⬍ .01; RV end-systolic volume: r ⫽ 0.96, P ⬍ .001, mean difference ⫽ ⫺3.2 ⫾ 7.1 mL, P ⬎ .05; RVEF: r ⫽ 0.89, P ⬍ .001, mean difference ⫽ ⫺0.3 ⫾ 7.1%, P ⬎ .05). The intraobserver and the interobserver variabilities ranged from ⫺1.1% to 5.8%. Conclusion: Measurement of RV volumes and EF by RT3DE is feasible, accurate, and reproducible in children compared with MRI measurements.
The right ventricle (RV) plays a critical role in congenital heart diseases in children, and RV dysfunction develops in many patients and leads to considerable morbidity and mortality. RV volumes and ejection fraction (EF) are the most widely used indexes to assess RV function; therefore, reliable determination of these RV indexes is essential for prognosis, guiding therapy, and longitudinal follow-up in these patients.1-4
From the Baylor College of Medicine and Texas Children’s Hospital, Houston, Texas (X.L., V.N., N.A., R.H.P., J.P.K., M.G., S.G.); University of Iowa, Iowa City, Iowa (L.B., A.S.); TomTec Imaging Systems, Munich, Germany (B.K.); and Texas Heart Institute/St Luke’s Episcopal Hospital, Houston, Texas (S.G.). Reprint requests: Shuping Ge, MD, The Lillie Frank Abercrombie Section of Cardiology, Department of Pediatrics, Baylor College of Medicine and Texas Children’s Hospital, 6621 Fannin, MC 19345-C, Houston, TX 77030 (E-mail: [email protected]). 0894-7317/$34.00 Copyright 2008 by the American Society of Echocardiography. doi:10.1016/j.echo.2007.05.009
84
Two-dimensional (2D) echocardiography (2DE) is the mainstay for analysis of RV function. However, 2DE only provides a qualitative assessment of RV volumes and systolic function. Quantitative 2DE methods have limited accuracy and reproducibility because of the complex RV geometry.5-14 First described by Dekker et al15 in 1974, three-dimensional (3D) echocardiography (3DE) has been validated for assessment of cardiac anatomy and function.16-21 More recently, real-time 3DE (RT3DE) is available for investigational and clinical use.21,22 The feasibility, accuracy, and reproducibility of the quantification of left ventricular (LV) volumes and function using RT3DE have been validated in adults and children.23-25 However, the feasibility, accuracy, and reproducibility of measuring RV volumes and EF by RT3DE in children are unknown. The purpose of this study was to validate RT3DE for evaluation of RV volumes and RVEF in children and to determine the feasibility, accuracy, and reproducibility of this technology using a disk summation method by comparing the results with magnetic resonance imaging (MRI) measurements, a technique currently considered as the clinical gold standard for this purpose.
Journal of the American Society of Echocardiography Volume 21 Number 1
Lu et al
85
METHODS Study Participants This prospective study was approved by our institutional review board. A total of 20 children (10 male and 10 female) with no history of cardiovascular disease volunteered and were enrolled after written informed consent was obtained for each study child. The mean age was 10.6 ⫾2.8 years, ranging from 6 to 18 years. All children had a screening 2DE. One child was found incidentally to have an isolated moderate-sized secundum atrial septal defect and the rest of the children had normal cardiac anatomy by 2DE. All participants were in sinus rhythm. All were included in the study as per institutional review board approved protocol. Before the study, the weight, height, heart rate, and blood pressures were obtained for each child. The children were placed in left recumbent position and electrocardiogram (ECG) was connected. RT3DE Data Acquisition and Analysis RT3DE was performed from apical window and views using a 2- to 4-MHz X4 matrix-array transducer connected to a RT3DE system (Sonos 7500, Philips Medical Systems, Andover, MA). The X4 matrix-array transducer houses 2880 elements so that a pyramid volumetric data set can be acquired. The images were first optimized using biplane mode, a simultaneous display of modified 4- and 2-chamber views, to visualize the RV inflow and outflow tracts. Subsequently, a full-volume 3D data set with a field of view of 93 ⫻ 84 degrees was acquired. The full-volume data set is compiled from 4 30- ⫻ 60-degree subvolumes with ECG triggering and the acquisition takes 4 cardiac cycles. The resolution of this RT3DE is from approximately 0.7 ⫻ 0.7 ⫻ 0.5 mm to 1.2 ⫻ 1.2 ⫻ 0.8 mm as measured by voxel size. The children were asked to hold their breath during actual acquisition. After each study, all data sets were saved onto hard drive and subsequently downloaded to rewritable compact disks for offline analysis. All RT3DE images were processed offline using a dedicated analysis software package (4D Echo-View, TomTec Imaging Systems, Munich, Germany). The 3D data set was displayed in a 4-tile image screen (Figure 1). Using the navigation tools and by manipulating the x, y, and z planes of the data set, the RV images were displayed in short-axis views from the apex to the base for measurement of RV volumes. Three reference views, ie, modified 4-chamber, 2-chamber, and dynamic short-axis views (Figure 1), were used to assist in identification of the RV inflow, trabecular, and outflow portions for endocardial border tracing. The RV volume measurement started from the tricuspid annulus by a set of short-axis disks parallel to the tricuspid annulus, and ended at the RV apex. The RV end-systolic and end-diastolic endocardial borders were manually traced in 5-mm contiguous slices. Large trabeculations and papillary muscles were not included in chamber volumes. These volumetric slices (planimetered area ⫻ slice thickness) were summated based on the Simpson’s principle to provide the RV volumes in end systole and end diastole (Figure 2), and the RVEF was calculated as: RVEF (%) ⫽ (end-diastolic volume [EDV] ⫺ end-systolic volume [ESV])/EDV ⫻ 100%. MRI Data For each study participant, MRI was performed on the same day as the RT3DE. Participants were imaged on a 1.5-T scanner (CV/I, GE Medical Systems, Milwaukee, WI) equipped with a 4-element torso coil. The participants were placed in a supine position and their blood pressure and heart rate were measured. From 8 to 12 contiguous short-axis slices through the LV and RV were acquired using a
Figure 1 Offline analysis of real-time three-dimensional echocardiographic data using disk summation algorithm for right ventricular (RV) volume calculations at end diastole. Top left, Four-chamber view of RV. Top right, Two-chamber view of RV perpendicular to 4-chamber view. Bottom left, Series of shortaxis slices were used to trace RV endocardial borders to derive RV volumes and ejection fraction (EF). Tricuspid annulus, apex, interventricular septum, and free wall in other panels were used as references for measurements of RV indexes. Bottom right, Cine short-axis image displayed in this panel to add to 4- and 2-chamber views as references for border identification and tracing. Color figure online. segmented 2D true steady-state free precession cine gradient echo sequence during consecutive 8- to 10-second breath holds at end expiration (Figure 3). Twenty cardiac phases were acquired at each location. The slice thickness was 8 mm, the field of view was 30 to 40 cm, depending on patient size, and the matrix was 224 ⫻ 160. The image resolution (as measured by voxel dimension) varied from approximately 1.3 ⫻ 1.9 ⫻ 8 mm to 1.8 ⫻ 2.5 ⫻ 8 mm. MRI data were then transferred to an offline workstation (Advantage Windows, GE Medical Systems) and analyzed using vendorprovided software. The RV endocardial borders at end systole and end diastole were manually traced on every short-axis slice and RV volumes were calculated by summation of the product (planimetered area ⫻ slice thickness) of all slices.26 Statistical Analysis All numeric values are expressed as mean ⫾ SD. The RV volumes and EF measured by RT3DE and MRI were analyzed using the MRI measurements as the reference standard. To examine the correlation and agreement between the two sets of measurements, Pearson regression and Bland and Altman27 analyses were used. Statistical significance was defined as P less than .05. All statistical analyses were done using software (Excel, Microsoft, Seattle, WA) with statistical analysis add-ons. Intraobserver and Interobserver Variability for Measurements by RT3DE To determine the intraobserver variability, one observer analyzed and measured the RT3DE data twice at 4-week intervals, blinded to the results of the first measurements. A second observer also analyzed and measured the RT3DE data independently, blinded to the results
86
Lu et al
Journal of the American Society of Echocardiography January 2008
Figure 3 Analysis of magnetic resonance imaging data. Series of contiguous short-axis images of right ventricle (RV) from base (top left) to apex (bottom right) with slice thickness of 8 mm were obtained. Endocardial borders were traced manually at end systole and end diastole for calculation of RV volumes and ejection fraction. Figure 2 Real-time three-dimensional echocardiography representation of end-diastolic (top) and end-systolic (bottom) right ventricular volumes. Color figure online.
of the first observer, to determine the interobserver variability. Both observers were blinded to results of MRI measurements. The intraobserver and interobserver variabilities were expressed as percentile differences (difference between the measurements/mean of the measurements ⫻ 100%). RESULTS Of the 20 participants enrolled in the study, two were excluded because of obvious subvolume mismatch artifacts for RT3DE and one was excluded because of insufficient slices through base of heart to calculate RV volumes for MRI. During the study, one child was found incidentally to have a moderate-sized secundum atrial septal defect, but this patient was included in the study. Among the 17 participants, there was no statistically significant difference between the systolic (P ⫽ .33) and diastolic (P ⫽ .16) blood pressures or the heart rate (P ⫽ .09) before RT3DE and MRI studies. The mean values of RV volumes and RVEF by RT3DE and MRI were: 109.6 ⫾ 36.2 versus 116.7 ⫾ 41.2 mL for RVEDV; 45.1 ⫾ 16.5 versus 47.8 ⫾ 19.8 mL for RVESV; and 58.2 ⫾ 8.5 versus 58.6 ⫾ 9.3% for RVEF. The correlation and agreement analyses of the RV volumes and RVEF between RT3DE and MRI measurements are depicted in Figures 4 to 6 for RVESV, RVEDV, and RVEF, respectively. Close correlation between the two methods was observed: r ⫽ 0.96 for RVESV, r ⫽ 0.98 for RVEDV, and r ⫽ 0.89 for RVEF. However, there was a small underestimation of RVEDV (mean difference ⫽ ⫺7.0 ⫾ 9.0 mL, P ⬍ .01) by RT3DE as values of the indexes increased. The mean difference for RVESV (mean difference ⫽ ⫺3.2 ⫾ 7.1 mL, P ⬎ .05) and RVEF (mean difference ⫽0.3 ⫾ 4.1%, P ⫽ .75) were not statistically significant between RT3DE and MRI measurements.
As an internal control for the accuracy of MRI volume measurements in our study, LV and RV volumes during end systole and end diastole were measured by MRI and used to determine LVSV and RVSV. There was good correlation (r ⫽ 0.93, standard error of the estimate ⫽ 10.66 mL, P ⬍ .001) and agreement (mean difference ⫽ ⫺1.35 ⫾ 10.54 mL) between LVSV and RVSV measurements by MRI when the subject with atrial septal defect was excluded. Intraobserver variabilities for RV volumes and EF measurements were generally smaller than the corresponding interobserver variabilities: ⫺2.1% ⫾ 5.3% vs. ⫺5.4% ⫾ 9.2% for RVEDV; ⫺1.1% ⫾ 3.2% vs. 2.5% ⫾ 6.5% for RVESV; and 2.3% ⫾ 3.8% vs. 5.8% ⫾ 6.1% for RVEF, respectively. DISCUSSION To our knowledge, this is the first study to validate the feasibility, accuracy, and reproducibility of RT3DE for measuring RV volumes and RVEF using a disk summation method in children in comparison with MRI measurements. Although RT3DE tends to underestimate RVEDV volumes, the underestimation was small. RVESV and RVEF measurements correlated and agreed well with MRI measurements. Advantages of Measurements of RV Volumes by RT3DE Two-dimensional echocardiographic methods to assess ventricular volumes and EF are highly dependent on ventricular geometry. It has been shown that 2DE is reasonably accurate for LV indexes but is limited when it is applied to RV volume measurements, secondary to the complex geometry of the RV.7,10,14,28 Because a 3DE data set comprises the entire RV volume, our study demonstrated that a series of consecutive slices from base to the apex of the heart can be obtained, collated, and summed to derive RV volumes using a disk summation method based on the Simpson’s principle. RVEDV and ESV can be calculated by tracing the endocardial borders, similar to MRI. Compared with the previous technologies,29-35 the RT3DE used in this study has several important techni-
Journal of the American Society of Echocardiography Volume 21 Number 1
Figure 4 Pearson regression and Bland-Altman analyses of right ventricular end-systolic volume (RVESV) measurements by real-time three-dimensional echocardiography (RT3DE) and magnetic resonance imaging (MRI). SEE, Standard error of estimate. Color figure online.
Figure 5 Pearson regression and Bland-Altman analyses of right ventricular end-diastolic volume (RVEDV) measurements by real-time three-dimensional echocardiography (RT3DE) and magnetic resonance imaging (MRI). SEE, Standard error of estimate. Color figure online.
cal advances. First, RT3DE provides high-resolution RT3DE volumetric images for assessment of cardiac anatomy and function in real time. Second, a complete full-volume 3D data set to encompass the entire heart can be acquired in 4 consecutive cardiac cycles with ECG trigging. Finally, the full-volume data set can be viewed, analyzed, and measured offline in numerous perspectives and views to provide an avenue to systematic qualitative and quantitative assessment of the anatomy and function by gray scale and Doppler images and data.
Lu et al
87
Figure 6 Pearson regression and Bland-Altman analyses of right ventricular ejection fraction (RVEF) measurements by real-time three-dimensional echocardiography (RT3DE) and magnetic resonance imaging (MRI). SEE, Standard error of estimate. Color figure online. RT3DE Versus MRI for Assessment of RV Volumes In recent years, MRI scanners and imaging protocols have been developed and modern MRI systems have high spatial resolution and excellent image quality. The use of MRI to quantitate RV volumes and mass has been established and has been validated for its accuracy and excellent reproducibility. It is currently considered to be the clinical gold standard.36-39 The method of RV volume assessment by RT3DE using a disk summation method is similar to MRI in our study. Although RT3DE is limited by the acoustic windows, RT3DE has several important features that may be advantageous. First, RT3DE data acquisition is more rapid than MRI. A full-volume RT3DE data set for measurement of all RV indexes takes 4 cardiac cycles to acquire and the analysis time is similar to MRI. The total time for quantifying RV indexes is shorter for RT3DE than MRI in our study (about 25 minutes for RT3DE vs 40 minutes for MRI). Second, RT3DE instrumentation is less expensive and more portable to the bedside than MRI. Third, although RV volume measurement can be performed on a series of short-axis images using this method, the long-axis images, ie, both 4- and 2-chamber views, can also be shown on a same 4-tile image screen simultaneously as reference images to ascertain accurate endocardial border tracing. Therefore, RT3DE may offer an attractive first-line imaging modality for accurate assessment of RV indexes if the acoustic window is adequate. It may also provide a useful method for assessing patients who are critically ill, claustrophobic, unable or unwilling to perform prolonged breath holding, or have other contraindications to MRI such as a cardiac pacemaker. There is a trend of a small underestimation of the RVEDV volumes by RT3DE compared with MRI. The result is consistent with the previous comparative studies using MRI, 3DE, and angiography in animals with known true LV and RV volumes. MRI and 3DE measurements appeared reasonably accurate but angiography was associated with significant error of estimation (mean difference ⫺1.9 ⫾ 3.3% for MRI, 9.3 ⫾ 6.3% for 3DE, and 57.9 ⫾ 40.1% for angiography, respectively).40 The current RT3DE technology has improved image resolution and accuracy of RV volume measure-
88
Lu et al
ments. In our study, the RT3DE and MRI data were not obtained simultaneously. Although there were no significant differences in blood pressures and heart rate before the RT3DE and MRI studies, the RV indices may be affected by prolonged breath holding and sympathetic stimulation caused by claustrophobia during the MRI studies. Such effects may partly account for the differences between RT3DE and MRI measurements. Technically, more pronounced RV trabeculations may influence identification of endocardial borders especially at the apex of RV. This may also contribute to the underestimation of RV indexes by RT3DE in this study. Study Limitations First, only children from 6 to 18 years were enrolled in this study. Younger infants who cannot perform breath holding and cooperate with the MRI study may need sedation and possibly endotracheal intubation, posing a significant ethical challenge to enroll younger healthy participants. Second, the only available RT3DE transducer has an operating frequency of 2 to 4 MHz. Our preliminary experience and that of others41 have shown that the image quality is adequate for younger infants as well. For older children and adults or patients with suboptimal acoustic windows, the addition of harmonic imaging and/or intravenous contrast agents may be required to enhance the endocardial border delineation. Third, current ECG trigger is necessary for a full-volume data set from 4 subvolumes, which may potentially cause data discontinuity between the 4 individual data sets secondary to breathing, patient movement, or arrhythmia. However, such potential artifacts did not seem to be a significant problem in the majority of children in this study. Fourth, because of RV anatomy and current resolution of the RT3DE, the RV border tracing, especially at the apex,anterior free wall, and outflow tract, can be challenging in some participants. A slightly modified view with emphasis of the RV, especially with an anterior angulation of the transducer to visualize the RV outflow tract, may often optimize 3D RV data acquisition and analysis. The 4-chamber, 2-chamber, and cine short-axis views can be used as references to facilitate RV endocardial border detection on each short-axis image. A learning curve is also needed to improve operator’s ability for RV border identification. Fifth, in this study we did not calculate the RV mass because the RV wall is thin and it was difficult to trace the epicardial borders reliably by current RT3DE techniques. Finally, the measurement of RV indexes by the current RT3DE and commercially available offline software package is still somewhat painstaking. Automated algorithms for measurement of RV volumes may facilitate the clinical use of this technology. Conclusions In conclusion, this prospective study demonstrated that measurement of RV volumes and EF by RT3DE is feasible, accurate, and reproducible in children and is comparable with MRI measurements. Further investigation in other population, especially in children and adults with congenital and acquired heart diseases that affect RV volumes and function, is needed to validate this method as a clinical tool to assess RV volumes and EF and to monitor and evaluate the efficacy of therapy. REFERENCES 1. Graham TP Jr, Bernard YD, Mellen BG, et al. Long-term outcome in congenitally corrected transposition of the great arteries: a multi-institutional study. J Am Coll Cardiol 2000;36:255-61.
Journal of the American Society of Echocardiography January 2008
2. Carr-White GS, Koh TW, Bhabra MS, et al. Right ventricular function following pulmonary autograft surgery. Circulation 1997;96:I-430. 3. Tulevski II, Dodge-Khatami A, Groenink M, et al. Right ventricular function in congenital cardiac disease: noninvasive quantitative parameters for clinical follow-up. Cardiol Young 2003;13:397-403. 4. Piran S, Veldtman G, Siu S, Webb GD, Liu PP. Heart failure and ventricular dysfunction in patients with single or systemic right ventricles. Circulation 2002;105:1189-94. 5. Silverman NH, Hudson S. Evaluation of right ventricle volume and ejection fraction in children by two-dimensional echocardiography. Pediatr Cardiol 1983;4:197-203. 6. Kaul S, Tei C, Hopkins JM, et al. Assessment of right ventricular function using two-dimensional echocardiography. Am Heart J 1984;107:526-31. 7. Levine RA, Gibson TC, Aretz T, et al. Echocardiographic measurement of right ventricular volume. Circulation 1984;69:497-505. 8. Foale R, Nihoyannopoulos P, McKenna W, et al. Echocardiographic measurement of the normal adult right ventricle. Br Heart J 1986;56:3344. 9. Helbing WA, Bosch HG, Maliepaard C, et al. Comparison of echocardiographic methods with magnetic resonance imaging for assessment of right ventricular function in children. Am J Cardiol 1995;76:589-94. 10. Apfel HD, Solowiejczyk DE, Printz BF, et al. Feasibility of a two-dimensional echocardiographic method for the clinical assessment of right ventricular volume and function in children. J Am Soc Echocardiogr 1996;9:637-45. 11. Munoz RA, Marcus E, Colan SD, Stanchak L, van der Velde M, Wessel DL. Accurate two-dimensional echocardiographic evaluation of right ventricular shape and volume using a combination of Fourier analysis and Simpson’s biplane method: an in vitro analysis. Circulation 1997;96:I470. 12. Denslow S, Wiles HB. Right ventricular volumes revisited: a simple model and simple formula for echocardiographic determination. J Am Soc Echocardiogr 1998;11:864-73. 13. Kovalova S, Necas J, Cerbak R, et al. Echocardiographic volumetry of the right ventricle. Eur J Echocardiogr 2005;6:15-23. 14. Bleeker GB, Steendijk P, Holman ER, et al. Assessing right ventricular function: the role of echocardiography and complementary technologies. Heart 2006;92:i19-26. 15. Dekker DL, Piziali RL, Dong E Jr. A system for ultrasonically imaging the human heart in three dimensions. Comput Biomed Res 1974;7:544-53. 16. Ge S, Warner JG Jr, Fowle KM, et al. Morphology and dynamic change of discrete subaortic stenosis can be imaged and quantified with threedimensional transesophageal echocardiography. J Am Soc Echocardiogr 1997;10:713-6. 17. Abraham TP, Warner JG Jr, Kon ND, et al. Feasibility, accuracy, and incremental value of intraoperative three-dimensional transesophageal echocardiography in valve surgery. Am J Cardiol 1997;80:1577-82. 18. Ge S, Warner JG Jr, Abraham TP, et al. Three-dimensional surface area of the aortic valve orifice by three-dimensional echocardiography: clinical validation of a novel index for assessment of aortic stenosis. Am Heart J 1998;136:1042-50. 19. Magni G, Marx G, Vogel M, et al. Confirmation of 3-dimensional echocardiographic findings in congenital heart disease: comparison with surgery, anatomic specimens, and anatomic changes pre and post surgery. J Am Soc Echocardiogr 1996;9:373. 20. Ludomirsky A, Silberbach M, Kenny A, et al. Superiority of rotational scan reconstruction strategies for transthoracic 3-dimensional real-time echocardiographic studies in pediatric patients with congenital heart disease. J Am Coll Cardiol 1994;23:169A. 21. Wang XF, Deng YB, Nanda NC, et al. Live three-dimensional echocardiography: imaging principles and clinical application. Echocardiography 2003;20:593-604. 22. Sugeng L, Weinert L, Lammertin G, et al. Utility of real-time 3-D transthoracic echocardiography in the evaluation of mitral valve disease [abstract]. J Am Coll Cardiol 2003;41:429. 23. Bu L, Munns S, Zhang H, et al. Rapid full volume data acquisition by real-time three-dimensional echocardiography for assessment of left ven-
Journal of the American Society of Echocardiography Volume 21 Number 1
24.
25.
26. 27. 28.
29.
30.
31.
32.
tricular indexes in children: a validation study compared with magnetic resonance imaging. J Am Soc Echocardiogr 2005;18:299-305. Jekins C, Bricknell K, Hanekom L, et al. Reproducibility and accuracy of echocardiographic measurements of left ventricular parameters using real-time three-dimensional echocardiography. J Am Coll Cardiol 2004; 44:878-86. Annemien E, Danielle RV, Boudewijin J, et al. Real-time transthoracic three-dimensional echocardiographic assessment of left ventricular volume and ejection fraction in congenital heart disease. J Am Coll Cardiol 2006;19:1-6. Pennell DJ. Ventricular volume and mass by CMR. J Cardiovasc Magn Reson 2002;4:507-13. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307-10. Aebischer NM, Czegledy F. Determination of right ventricular volume by two-dimensional echocardiography with a crescentic model. J Am Soc Echocardiogr 1989;2:110-8. Jiang L, Siu SC, Handschumacher M, et al. Myocardial imaging: threedimensional echocardiography: in vivo validation for right ventricular volume and function. Circulation 1994;89:2342-50. Cao Q, Delabaise A, Sugeng L, Krauss M, Vannan M, Schwartz S. Measurement of right ventricular volume and depiction of right ventricular anatomy in volume-rendered tissue-depiction three-dimensional echocardiography. Circulation 1994;90:II66. Gutberlet M, Dittrich S, Hosten N, Lange PE. Comparison of magnetic resonance imaging and 3-dimensional echocardiography in the in vivo assessment of RV volume and mass. Circulation 1996;94:I-416. Nesser HJ, Tkalec W, Niel J, Masani N, Pandian N. Accurate quantitation of right ventricular size and function by voxel-based 3-dimensional echocardiography in patients with normal and abnormal ventricles (3DE): comparison with magnetic resonance imaging. J Am Coll Cardiol 1996;27:149A.
Lu et al
89
33. Shiota T, Jones M, Ota T, et al. Application of a new real-time threedimensional method for evaluating right ventricular stroke volume. Circulation 1997;96:I-328. 34. Fujimoto S, Mizuno R, Nakagawa Y, et al. Estimation of the right ventricular volume and ejection fraction by transthoracic three-dimensional echocardiography: a validation study using magnetic resonance imaging. Int J Card Imaging 1998;14:385-90. 35. Vogel M, Gutberlet M, Dittrich S, et al. Comparison of transthoracic three dimensional echocardiography with magnetic resonance imaging in the assessment of right ventricular volume and mass. Heart 1997;78:127-30. 36. Mogelvang J, Stubgaard M, Thomsen C, et al. Evaluation of right ventricular volumes measured by magnetic resonance imaging. Eur Heart J 1988;9:529-33. 37. Jauhiainen T, Jarvinen VM, Hekali PE. Evaluation of methods for MR imaging of human right ventricular heart volumes and mass. Acta Radiol 2002;43:587-92. 38. Alfakih K, Plein S, Bloomer T, et al. Comparison of right ventricular volume measurements between axial and short axis orientation using steady-state free precession magnetic resonance imaging. J Magn Reson Imaging 2003;18:25-32. 39. Grothues F, Moon JC, Bellenger NG, et al. Interstudy reproducibility of right ventricular volumes, function, and mass with cardiovascular magnetic resonance. Am Heart J 2004;147:218-23. 40. Heusch A, Koch JA, Krogmann ON, Korbmacher B, et al. Volumetric analysis of the right and left ventricle in a porcine heart model: comparison of three-dimensional echocardiography, magnetic resonance imaging and angiocardiography. Eur J Ultrasound 1999;9:245-55. 41. Wong PC, Gao H, Sklansky M. Real-time volume-rendered transthoracic three-dimensional echocardiography for the evaluation of pediatric heart disease [abstract]. Circulation 2003;108:677.
Background: Measurement of right ventricular (RV) volumes and ejection fraction (EF) by two-dimensional echocardiography has limited accuracy and reproducibility because of the complex RV geometry. Objectives: This study sought to validate real-time three-dimensional echocardiography (RT3DE) using a disk summation method for assessment of RV volumes and RVEF in children by comparing it with magnetic resonance imaging (MRI) measurements. Methods: A total of 20 children (mean age 10.6 ⫾ 2.8 years) were studied. Transthoracic RT3DE was performed using a RT3DE system to acquire full-volume RT3DE data sets from apical windows and data were processed offline using a software package. RV end-systolic volume and end-diastolic volume (EDV) were measured using a disk summation method by manually tracing the endocardial borders. RVEF was calculated as: RVEF ⫽ (EDV ⫺ end-systolic volume)/EDV ⫻ 100%. All participants also underwent MRI studies for comparison of RV indexes. Results: Of the 20 children, 3 were excluded because of poor or incomplete RV images (two RT3DE and one MRI study). For the remaining 17 children, good correlation and agreement between RT3DE and MRI were found (RVEDV: r ⫽ 0.98, P ⬍ .001, mean difference ⫽ ⫺7.0 ⫾ 9.0 mL, P ⬍ .01; RV end-systolic volume: r ⫽ 0.96, P ⬍ .001, mean difference ⫽ ⫺3.2 ⫾ 7.1 mL, P ⬎ .05; RVEF: r ⫽ 0.89, P ⬍ .001, mean difference ⫽ ⫺0.3 ⫾ 7.1%, P ⬎ .05). The intraobserver and the interobserver variabilities ranged from ⫺1.1% to 5.8%. Conclusion: Measurement of RV volumes and EF by RT3DE is feasible, accurate, and reproducible in children compared with MRI measurements.
The right ventricle (RV) plays a critical role in congenital heart diseases in children, and RV dysfunction develops in many patients and leads to considerable morbidity and mortality. RV volumes and ejection fraction (EF) are the most widely used indexes to assess RV function; therefore, reliable determination of these RV indexes is essential for prognosis, guiding therapy, and longitudinal follow-up in these patients.1-4
From the Baylor College of Medicine and Texas Children’s Hospital, Houston, Texas (X.L., V.N., N.A., R.H.P., J.P.K., M.G., S.G.); University of Iowa, Iowa City, Iowa (L.B., A.S.); TomTec Imaging Systems, Munich, Germany (B.K.); and Texas Heart Institute/St Luke’s Episcopal Hospital, Houston, Texas (S.G.). Reprint requests: Shuping Ge, MD, The Lillie Frank Abercrombie Section of Cardiology, Department of Pediatrics, Baylor College of Medicine and Texas Children’s Hospital, 6621 Fannin, MC 19345-C, Houston, TX 77030 (E-mail: [email protected]). 0894-7317/$34.00 Copyright 2008 by the American Society of Echocardiography. doi:10.1016/j.echo.2007.05.009
84
Two-dimensional (2D) echocardiography (2DE) is the mainstay for analysis of RV function. However, 2DE only provides a qualitative assessment of RV volumes and systolic function. Quantitative 2DE methods have limited accuracy and reproducibility because of the complex RV geometry.5-14 First described by Dekker et al15 in 1974, three-dimensional (3D) echocardiography (3DE) has been validated for assessment of cardiac anatomy and function.16-21 More recently, real-time 3DE (RT3DE) is available for investigational and clinical use.21,22 The feasibility, accuracy, and reproducibility of the quantification of left ventricular (LV) volumes and function using RT3DE have been validated in adults and children.23-25 However, the feasibility, accuracy, and reproducibility of measuring RV volumes and EF by RT3DE in children are unknown. The purpose of this study was to validate RT3DE for evaluation of RV volumes and RVEF in children and to determine the feasibility, accuracy, and reproducibility of this technology using a disk summation method by comparing the results with magnetic resonance imaging (MRI) measurements, a technique currently considered as the clinical gold standard for this purpose.
Journal of the American Society of Echocardiography Volume 21 Number 1
Lu et al
85
METHODS Study Participants This prospective study was approved by our institutional review board. A total of 20 children (10 male and 10 female) with no history of cardiovascular disease volunteered and were enrolled after written informed consent was obtained for each study child. The mean age was 10.6 ⫾2.8 years, ranging from 6 to 18 years. All children had a screening 2DE. One child was found incidentally to have an isolated moderate-sized secundum atrial septal defect and the rest of the children had normal cardiac anatomy by 2DE. All participants were in sinus rhythm. All were included in the study as per institutional review board approved protocol. Before the study, the weight, height, heart rate, and blood pressures were obtained for each child. The children were placed in left recumbent position and electrocardiogram (ECG) was connected. RT3DE Data Acquisition and Analysis RT3DE was performed from apical window and views using a 2- to 4-MHz X4 matrix-array transducer connected to a RT3DE system (Sonos 7500, Philips Medical Systems, Andover, MA). The X4 matrix-array transducer houses 2880 elements so that a pyramid volumetric data set can be acquired. The images were first optimized using biplane mode, a simultaneous display of modified 4- and 2-chamber views, to visualize the RV inflow and outflow tracts. Subsequently, a full-volume 3D data set with a field of view of 93 ⫻ 84 degrees was acquired. The full-volume data set is compiled from 4 30- ⫻ 60-degree subvolumes with ECG triggering and the acquisition takes 4 cardiac cycles. The resolution of this RT3DE is from approximately 0.7 ⫻ 0.7 ⫻ 0.5 mm to 1.2 ⫻ 1.2 ⫻ 0.8 mm as measured by voxel size. The children were asked to hold their breath during actual acquisition. After each study, all data sets were saved onto hard drive and subsequently downloaded to rewritable compact disks for offline analysis. All RT3DE images were processed offline using a dedicated analysis software package (4D Echo-View, TomTec Imaging Systems, Munich, Germany). The 3D data set was displayed in a 4-tile image screen (Figure 1). Using the navigation tools and by manipulating the x, y, and z planes of the data set, the RV images were displayed in short-axis views from the apex to the base for measurement of RV volumes. Three reference views, ie, modified 4-chamber, 2-chamber, and dynamic short-axis views (Figure 1), were used to assist in identification of the RV inflow, trabecular, and outflow portions for endocardial border tracing. The RV volume measurement started from the tricuspid annulus by a set of short-axis disks parallel to the tricuspid annulus, and ended at the RV apex. The RV end-systolic and end-diastolic endocardial borders were manually traced in 5-mm contiguous slices. Large trabeculations and papillary muscles were not included in chamber volumes. These volumetric slices (planimetered area ⫻ slice thickness) were summated based on the Simpson’s principle to provide the RV volumes in end systole and end diastole (Figure 2), and the RVEF was calculated as: RVEF (%) ⫽ (end-diastolic volume [EDV] ⫺ end-systolic volume [ESV])/EDV ⫻ 100%. MRI Data For each study participant, MRI was performed on the same day as the RT3DE. Participants were imaged on a 1.5-T scanner (CV/I, GE Medical Systems, Milwaukee, WI) equipped with a 4-element torso coil. The participants were placed in a supine position and their blood pressure and heart rate were measured. From 8 to 12 contiguous short-axis slices through the LV and RV were acquired using a
Figure 1 Offline analysis of real-time three-dimensional echocardiographic data using disk summation algorithm for right ventricular (RV) volume calculations at end diastole. Top left, Four-chamber view of RV. Top right, Two-chamber view of RV perpendicular to 4-chamber view. Bottom left, Series of shortaxis slices were used to trace RV endocardial borders to derive RV volumes and ejection fraction (EF). Tricuspid annulus, apex, interventricular septum, and free wall in other panels were used as references for measurements of RV indexes. Bottom right, Cine short-axis image displayed in this panel to add to 4- and 2-chamber views as references for border identification and tracing. Color figure online. segmented 2D true steady-state free precession cine gradient echo sequence during consecutive 8- to 10-second breath holds at end expiration (Figure 3). Twenty cardiac phases were acquired at each location. The slice thickness was 8 mm, the field of view was 30 to 40 cm, depending on patient size, and the matrix was 224 ⫻ 160. The image resolution (as measured by voxel dimension) varied from approximately 1.3 ⫻ 1.9 ⫻ 8 mm to 1.8 ⫻ 2.5 ⫻ 8 mm. MRI data were then transferred to an offline workstation (Advantage Windows, GE Medical Systems) and analyzed using vendorprovided software. The RV endocardial borders at end systole and end diastole were manually traced on every short-axis slice and RV volumes were calculated by summation of the product (planimetered area ⫻ slice thickness) of all slices.26 Statistical Analysis All numeric values are expressed as mean ⫾ SD. The RV volumes and EF measured by RT3DE and MRI were analyzed using the MRI measurements as the reference standard. To examine the correlation and agreement between the two sets of measurements, Pearson regression and Bland and Altman27 analyses were used. Statistical significance was defined as P less than .05. All statistical analyses were done using software (Excel, Microsoft, Seattle, WA) with statistical analysis add-ons. Intraobserver and Interobserver Variability for Measurements by RT3DE To determine the intraobserver variability, one observer analyzed and measured the RT3DE data twice at 4-week intervals, blinded to the results of the first measurements. A second observer also analyzed and measured the RT3DE data independently, blinded to the results
86
Lu et al
Journal of the American Society of Echocardiography January 2008
Figure 3 Analysis of magnetic resonance imaging data. Series of contiguous short-axis images of right ventricle (RV) from base (top left) to apex (bottom right) with slice thickness of 8 mm were obtained. Endocardial borders were traced manually at end systole and end diastole for calculation of RV volumes and ejection fraction. Figure 2 Real-time three-dimensional echocardiography representation of end-diastolic (top) and end-systolic (bottom) right ventricular volumes. Color figure online.
of the first observer, to determine the interobserver variability. Both observers were blinded to results of MRI measurements. The intraobserver and interobserver variabilities were expressed as percentile differences (difference between the measurements/mean of the measurements ⫻ 100%). RESULTS Of the 20 participants enrolled in the study, two were excluded because of obvious subvolume mismatch artifacts for RT3DE and one was excluded because of insufficient slices through base of heart to calculate RV volumes for MRI. During the study, one child was found incidentally to have a moderate-sized secundum atrial septal defect, but this patient was included in the study. Among the 17 participants, there was no statistically significant difference between the systolic (P ⫽ .33) and diastolic (P ⫽ .16) blood pressures or the heart rate (P ⫽ .09) before RT3DE and MRI studies. The mean values of RV volumes and RVEF by RT3DE and MRI were: 109.6 ⫾ 36.2 versus 116.7 ⫾ 41.2 mL for RVEDV; 45.1 ⫾ 16.5 versus 47.8 ⫾ 19.8 mL for RVESV; and 58.2 ⫾ 8.5 versus 58.6 ⫾ 9.3% for RVEF. The correlation and agreement analyses of the RV volumes and RVEF between RT3DE and MRI measurements are depicted in Figures 4 to 6 for RVESV, RVEDV, and RVEF, respectively. Close correlation between the two methods was observed: r ⫽ 0.96 for RVESV, r ⫽ 0.98 for RVEDV, and r ⫽ 0.89 for RVEF. However, there was a small underestimation of RVEDV (mean difference ⫽ ⫺7.0 ⫾ 9.0 mL, P ⬍ .01) by RT3DE as values of the indexes increased. The mean difference for RVESV (mean difference ⫽ ⫺3.2 ⫾ 7.1 mL, P ⬎ .05) and RVEF (mean difference ⫽0.3 ⫾ 4.1%, P ⫽ .75) were not statistically significant between RT3DE and MRI measurements.
As an internal control for the accuracy of MRI volume measurements in our study, LV and RV volumes during end systole and end diastole were measured by MRI and used to determine LVSV and RVSV. There was good correlation (r ⫽ 0.93, standard error of the estimate ⫽ 10.66 mL, P ⬍ .001) and agreement (mean difference ⫽ ⫺1.35 ⫾ 10.54 mL) between LVSV and RVSV measurements by MRI when the subject with atrial septal defect was excluded. Intraobserver variabilities for RV volumes and EF measurements were generally smaller than the corresponding interobserver variabilities: ⫺2.1% ⫾ 5.3% vs. ⫺5.4% ⫾ 9.2% for RVEDV; ⫺1.1% ⫾ 3.2% vs. 2.5% ⫾ 6.5% for RVESV; and 2.3% ⫾ 3.8% vs. 5.8% ⫾ 6.1% for RVEF, respectively. DISCUSSION To our knowledge, this is the first study to validate the feasibility, accuracy, and reproducibility of RT3DE for measuring RV volumes and RVEF using a disk summation method in children in comparison with MRI measurements. Although RT3DE tends to underestimate RVEDV volumes, the underestimation was small. RVESV and RVEF measurements correlated and agreed well with MRI measurements. Advantages of Measurements of RV Volumes by RT3DE Two-dimensional echocardiographic methods to assess ventricular volumes and EF are highly dependent on ventricular geometry. It has been shown that 2DE is reasonably accurate for LV indexes but is limited when it is applied to RV volume measurements, secondary to the complex geometry of the RV.7,10,14,28 Because a 3DE data set comprises the entire RV volume, our study demonstrated that a series of consecutive slices from base to the apex of the heart can be obtained, collated, and summed to derive RV volumes using a disk summation method based on the Simpson’s principle. RVEDV and ESV can be calculated by tracing the endocardial borders, similar to MRI. Compared with the previous technologies,29-35 the RT3DE used in this study has several important techni-
Journal of the American Society of Echocardiography Volume 21 Number 1
Figure 4 Pearson regression and Bland-Altman analyses of right ventricular end-systolic volume (RVESV) measurements by real-time three-dimensional echocardiography (RT3DE) and magnetic resonance imaging (MRI). SEE, Standard error of estimate. Color figure online.
Figure 5 Pearson regression and Bland-Altman analyses of right ventricular end-diastolic volume (RVEDV) measurements by real-time three-dimensional echocardiography (RT3DE) and magnetic resonance imaging (MRI). SEE, Standard error of estimate. Color figure online.
cal advances. First, RT3DE provides high-resolution RT3DE volumetric images for assessment of cardiac anatomy and function in real time. Second, a complete full-volume 3D data set to encompass the entire heart can be acquired in 4 consecutive cardiac cycles with ECG trigging. Finally, the full-volume data set can be viewed, analyzed, and measured offline in numerous perspectives and views to provide an avenue to systematic qualitative and quantitative assessment of the anatomy and function by gray scale and Doppler images and data.
Lu et al
87
Figure 6 Pearson regression and Bland-Altman analyses of right ventricular ejection fraction (RVEF) measurements by real-time three-dimensional echocardiography (RT3DE) and magnetic resonance imaging (MRI). SEE, Standard error of estimate. Color figure online. RT3DE Versus MRI for Assessment of RV Volumes In recent years, MRI scanners and imaging protocols have been developed and modern MRI systems have high spatial resolution and excellent image quality. The use of MRI to quantitate RV volumes and mass has been established and has been validated for its accuracy and excellent reproducibility. It is currently considered to be the clinical gold standard.36-39 The method of RV volume assessment by RT3DE using a disk summation method is similar to MRI in our study. Although RT3DE is limited by the acoustic windows, RT3DE has several important features that may be advantageous. First, RT3DE data acquisition is more rapid than MRI. A full-volume RT3DE data set for measurement of all RV indexes takes 4 cardiac cycles to acquire and the analysis time is similar to MRI. The total time for quantifying RV indexes is shorter for RT3DE than MRI in our study (about 25 minutes for RT3DE vs 40 minutes for MRI). Second, RT3DE instrumentation is less expensive and more portable to the bedside than MRI. Third, although RV volume measurement can be performed on a series of short-axis images using this method, the long-axis images, ie, both 4- and 2-chamber views, can also be shown on a same 4-tile image screen simultaneously as reference images to ascertain accurate endocardial border tracing. Therefore, RT3DE may offer an attractive first-line imaging modality for accurate assessment of RV indexes if the acoustic window is adequate. It may also provide a useful method for assessing patients who are critically ill, claustrophobic, unable or unwilling to perform prolonged breath holding, or have other contraindications to MRI such as a cardiac pacemaker. There is a trend of a small underestimation of the RVEDV volumes by RT3DE compared with MRI. The result is consistent with the previous comparative studies using MRI, 3DE, and angiography in animals with known true LV and RV volumes. MRI and 3DE measurements appeared reasonably accurate but angiography was associated with significant error of estimation (mean difference ⫺1.9 ⫾ 3.3% for MRI, 9.3 ⫾ 6.3% for 3DE, and 57.9 ⫾ 40.1% for angiography, respectively).40 The current RT3DE technology has improved image resolution and accuracy of RV volume measure-
88
Lu et al
ments. In our study, the RT3DE and MRI data were not obtained simultaneously. Although there were no significant differences in blood pressures and heart rate before the RT3DE and MRI studies, the RV indices may be affected by prolonged breath holding and sympathetic stimulation caused by claustrophobia during the MRI studies. Such effects may partly account for the differences between RT3DE and MRI measurements. Technically, more pronounced RV trabeculations may influence identification of endocardial borders especially at the apex of RV. This may also contribute to the underestimation of RV indexes by RT3DE in this study. Study Limitations First, only children from 6 to 18 years were enrolled in this study. Younger infants who cannot perform breath holding and cooperate with the MRI study may need sedation and possibly endotracheal intubation, posing a significant ethical challenge to enroll younger healthy participants. Second, the only available RT3DE transducer has an operating frequency of 2 to 4 MHz. Our preliminary experience and that of others41 have shown that the image quality is adequate for younger infants as well. For older children and adults or patients with suboptimal acoustic windows, the addition of harmonic imaging and/or intravenous contrast agents may be required to enhance the endocardial border delineation. Third, current ECG trigger is necessary for a full-volume data set from 4 subvolumes, which may potentially cause data discontinuity between the 4 individual data sets secondary to breathing, patient movement, or arrhythmia. However, such potential artifacts did not seem to be a significant problem in the majority of children in this study. Fourth, because of RV anatomy and current resolution of the RT3DE, the RV border tracing, especially at the apex,anterior free wall, and outflow tract, can be challenging in some participants. A slightly modified view with emphasis of the RV, especially with an anterior angulation of the transducer to visualize the RV outflow tract, may often optimize 3D RV data acquisition and analysis. The 4-chamber, 2-chamber, and cine short-axis views can be used as references to facilitate RV endocardial border detection on each short-axis image. A learning curve is also needed to improve operator’s ability for RV border identification. Fifth, in this study we did not calculate the RV mass because the RV wall is thin and it was difficult to trace the epicardial borders reliably by current RT3DE techniques. Finally, the measurement of RV indexes by the current RT3DE and commercially available offline software package is still somewhat painstaking. Automated algorithms for measurement of RV volumes may facilitate the clinical use of this technology. Conclusions In conclusion, this prospective study demonstrated that measurement of RV volumes and EF by RT3DE is feasible, accurate, and reproducible in children and is comparable with MRI measurements. Further investigation in other population, especially in children and adults with congenital and acquired heart diseases that affect RV volumes and function, is needed to validate this method as a clinical tool to assess RV volumes and EF and to monitor and evaluate the efficacy of therapy. REFERENCES 1. Graham TP Jr, Bernard YD, Mellen BG, et al. Long-term outcome in congenitally corrected transposition of the great arteries: a multi-institutional study. J Am Coll Cardiol 2000;36:255-61.
Journal of the American Society of Echocardiography January 2008
2. Carr-White GS, Koh TW, Bhabra MS, et al. Right ventricular function following pulmonary autograft surgery. Circulation 1997;96:I-430. 3. Tulevski II, Dodge-Khatami A, Groenink M, et al. Right ventricular function in congenital cardiac disease: noninvasive quantitative parameters for clinical follow-up. Cardiol Young 2003;13:397-403. 4. Piran S, Veldtman G, Siu S, Webb GD, Liu PP. Heart failure and ventricular dysfunction in patients with single or systemic right ventricles. Circulation 2002;105:1189-94. 5. Silverman NH, Hudson S. Evaluation of right ventricle volume and ejection fraction in children by two-dimensional echocardiography. Pediatr Cardiol 1983;4:197-203. 6. Kaul S, Tei C, Hopkins JM, et al. Assessment of right ventricular function using two-dimensional echocardiography. Am Heart J 1984;107:526-31. 7. Levine RA, Gibson TC, Aretz T, et al. Echocardiographic measurement of right ventricular volume. Circulation 1984;69:497-505. 8. Foale R, Nihoyannopoulos P, McKenna W, et al. Echocardiographic measurement of the normal adult right ventricle. Br Heart J 1986;56:3344. 9. Helbing WA, Bosch HG, Maliepaard C, et al. Comparison of echocardiographic methods with magnetic resonance imaging for assessment of right ventricular function in children. Am J Cardiol 1995;76:589-94. 10. Apfel HD, Solowiejczyk DE, Printz BF, et al. Feasibility of a two-dimensional echocardiographic method for the clinical assessment of right ventricular volume and function in children. J Am Soc Echocardiogr 1996;9:637-45. 11. Munoz RA, Marcus E, Colan SD, Stanchak L, van der Velde M, Wessel DL. Accurate two-dimensional echocardiographic evaluation of right ventricular shape and volume using a combination of Fourier analysis and Simpson’s biplane method: an in vitro analysis. Circulation 1997;96:I470. 12. Denslow S, Wiles HB. Right ventricular volumes revisited: a simple model and simple formula for echocardiographic determination. J Am Soc Echocardiogr 1998;11:864-73. 13. Kovalova S, Necas J, Cerbak R, et al. Echocardiographic volumetry of the right ventricle. Eur J Echocardiogr 2005;6:15-23. 14. Bleeker GB, Steendijk P, Holman ER, et al. Assessing right ventricular function: the role of echocardiography and complementary technologies. Heart 2006;92:i19-26. 15. Dekker DL, Piziali RL, Dong E Jr. A system for ultrasonically imaging the human heart in three dimensions. Comput Biomed Res 1974;7:544-53. 16. Ge S, Warner JG Jr, Fowle KM, et al. Morphology and dynamic change of discrete subaortic stenosis can be imaged and quantified with threedimensional transesophageal echocardiography. J Am Soc Echocardiogr 1997;10:713-6. 17. Abraham TP, Warner JG Jr, Kon ND, et al. Feasibility, accuracy, and incremental value of intraoperative three-dimensional transesophageal echocardiography in valve surgery. Am J Cardiol 1997;80:1577-82. 18. Ge S, Warner JG Jr, Abraham TP, et al. Three-dimensional surface area of the aortic valve orifice by three-dimensional echocardiography: clinical validation of a novel index for assessment of aortic stenosis. Am Heart J 1998;136:1042-50. 19. Magni G, Marx G, Vogel M, et al. Confirmation of 3-dimensional echocardiographic findings in congenital heart disease: comparison with surgery, anatomic specimens, and anatomic changes pre and post surgery. J Am Soc Echocardiogr 1996;9:373. 20. Ludomirsky A, Silberbach M, Kenny A, et al. Superiority of rotational scan reconstruction strategies for transthoracic 3-dimensional real-time echocardiographic studies in pediatric patients with congenital heart disease. J Am Coll Cardiol 1994;23:169A. 21. Wang XF, Deng YB, Nanda NC, et al. Live three-dimensional echocardiography: imaging principles and clinical application. Echocardiography 2003;20:593-604. 22. Sugeng L, Weinert L, Lammertin G, et al. Utility of real-time 3-D transthoracic echocardiography in the evaluation of mitral valve disease [abstract]. J Am Coll Cardiol 2003;41:429. 23. Bu L, Munns S, Zhang H, et al. Rapid full volume data acquisition by real-time three-dimensional echocardiography for assessment of left ven-
Journal of the American Society of Echocardiography Volume 21 Number 1
24.
25.
26. 27. 28.
29.
30.
31.
32.
tricular indexes in children: a validation study compared with magnetic resonance imaging. J Am Soc Echocardiogr 2005;18:299-305. Jekins C, Bricknell K, Hanekom L, et al. Reproducibility and accuracy of echocardiographic measurements of left ventricular parameters using real-time three-dimensional echocardiography. J Am Coll Cardiol 2004; 44:878-86. Annemien E, Danielle RV, Boudewijin J, et al. Real-time transthoracic three-dimensional echocardiographic assessment of left ventricular volume and ejection fraction in congenital heart disease. J Am Coll Cardiol 2006;19:1-6. Pennell DJ. Ventricular volume and mass by CMR. J Cardiovasc Magn Reson 2002;4:507-13. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307-10. Aebischer NM, Czegledy F. Determination of right ventricular volume by two-dimensional echocardiography with a crescentic model. J Am Soc Echocardiogr 1989;2:110-8. Jiang L, Siu SC, Handschumacher M, et al. Myocardial imaging: threedimensional echocardiography: in vivo validation for right ventricular volume and function. Circulation 1994;89:2342-50. Cao Q, Delabaise A, Sugeng L, Krauss M, Vannan M, Schwartz S. Measurement of right ventricular volume and depiction of right ventricular anatomy in volume-rendered tissue-depiction three-dimensional echocardiography. Circulation 1994;90:II66. Gutberlet M, Dittrich S, Hosten N, Lange PE. Comparison of magnetic resonance imaging and 3-dimensional echocardiography in the in vivo assessment of RV volume and mass. Circulation 1996;94:I-416. Nesser HJ, Tkalec W, Niel J, Masani N, Pandian N. Accurate quantitation of right ventricular size and function by voxel-based 3-dimensional echocardiography in patients with normal and abnormal ventricles (3DE): comparison with magnetic resonance imaging. J Am Coll Cardiol 1996;27:149A.
Lu et al
89
33. Shiota T, Jones M, Ota T, et al. Application of a new real-time threedimensional method for evaluating right ventricular stroke volume. Circulation 1997;96:I-328. 34. Fujimoto S, Mizuno R, Nakagawa Y, et al. Estimation of the right ventricular volume and ejection fraction by transthoracic three-dimensional echocardiography: a validation study using magnetic resonance imaging. Int J Card Imaging 1998;14:385-90. 35. Vogel M, Gutberlet M, Dittrich S, et al. Comparison of transthoracic three dimensional echocardiography with magnetic resonance imaging in the assessment of right ventricular volume and mass. Heart 1997;78:127-30. 36. Mogelvang J, Stubgaard M, Thomsen C, et al. Evaluation of right ventricular volumes measured by magnetic resonance imaging. Eur Heart J 1988;9:529-33. 37. Jauhiainen T, Jarvinen VM, Hekali PE. Evaluation of methods for MR imaging of human right ventricular heart volumes and mass. Acta Radiol 2002;43:587-92. 38. Alfakih K, Plein S, Bloomer T, et al. Comparison of right ventricular volume measurements between axial and short axis orientation using steady-state free precession magnetic resonance imaging. J Magn Reson Imaging 2003;18:25-32. 39. Grothues F, Moon JC, Bellenger NG, et al. Interstudy reproducibility of right ventricular volumes, function, and mass with cardiovascular magnetic resonance. Am Heart J 2004;147:218-23. 40. Heusch A, Koch JA, Krogmann ON, Korbmacher B, et al. Volumetric analysis of the right and left ventricle in a porcine heart model: comparison of three-dimensional echocardiography, magnetic resonance imaging and angiocardiography. Eur J Ultrasound 1999;9:245-55. 41. Wong PC, Gao H, Sklansky M. Real-time volume-rendered transthoracic three-dimensional echocardiography for the evaluation of pediatric heart disease [abstract]. Circulation 2003;108:677.