Three-Dimensional Echocardiography: Use of Additional Spatial Data for Measuring Left Ventricular Mass

As described in this issue of the Mayo Clinic Proceedings (pages 205 to 211), Shub and colleagues studied the effect of age, sex, and body size on left ventricular (LV) mass. They estimated LV mass by using two-dimensional (2D) image-guided one-dimensional (1D) M-mode echocardiographic measurements. Previously, Devereux and associates1Devereux RB Lutas EM Casale PN Kligfield P Eisenberg RR Hammond IW et al.Standardization of M-mode echocardiographic left ventricular anatomic measurements.J Am Coll Cardiol. 1984; 4: 1222-1230Abstract Full Text PDF PubMed Scopus (727) Google Scholar and Gardin and coworkers2Gardin JM Henry WL Savage DD Ware JH Burn C Borer JS Echocardiographic measurements in normal subjects: evaluation of an adult population without clinically apparent heart disease.J Clin Ultrasound. 1979; 7: 439-447Crossref PubMed Scopus (275) Google Scholar used unguided 1D M-mode echocardiography to obtain similar measurements. Undoubtedly, Shub and colleagues chose their method as an improvement over the blind, unguided method; however, the regression equation used to obtain their results was the same as that used previously. Although use of an image to guide this measurement should improve its accuracy and reproducibility, the method still has limitations and could be further improved. In the method of Shub and associates, a parasternal short-axis image of the left ventricle obtained just below the tips of the mitral leaflets was used to position the 1D M-mode measurement of the LV chamber and walls. By this means, they could center the 1D measurement fairly precisely and accurately across the ventricle in the plane of the image. Nevertheless, some questions remain about the position and orientation of the image itself. How far is it below the leaflet tips, and is it angled or rotated relative to the ventricular long axis? Could such variations be significant? To explore these questions, we studied positioning of standard 2D images by experienced technicians in the clinical environment and found that the short-axis image plane is consistently angulated across the ventricle; thus, the measured LV dimension is lengthened to a variable degree.3King DL Harrison MR King Jr, DL Gopal AS Kwan OL DeMaria AN Ultrasound beam orientation during standard two-dimensional imaging: assessment by three-dimensional echocardiography.J Am Soc Echocardiogr. 1992; 5: 569-576PubMed Scopus (81) Google Scholar Accordingly, we infer that variation of the position and orientation of the short-axis image, as used by Shub and colleagues, could contribute substantially to technical differences in the measurements and that minimizing this variation would improve the accuracy and reproducibility of the results obtained by their method. Three-dimensional (3D) echocardiography provides the means for such improvement. In 2D echocardiography, the precise position and orientation of the real-time image are unknown. Operators must use their knowledge, the visualized anatomy, and their proprioceptive sense of transducer motion to estimate image position and orientation on the basis of mental reconstruction of ventricular anatomy. The position and orientation of the real-time 2D image cannot be measured. In addition, no means exist for visualizing anatomic landmarks orthogonal to the real-time image—that is, for viewing the nonvisualized dimension. No independent visual display of the 3D spatial relationship of the real-time image to a series of 2D images is available to guide the operator. These deficiencies are major limitations of 2D echocardiography. In contrast, 3D echocardiography overcomes both of these deficiencies by registering transducer position and orientation in an independent 3D spatial coordinate system and by creating a “line of intersection” display.4King DL Al-Banna SJ Larach DL A new three-dimensional random scanner for ultrasonic-computergraphic imaging of the heart.in: White DN Barnes R Ultrasound in Medicine. Vol 2. Plenum, New York1976: 363-368Google Scholar, 5King DL King Jr, DL Shao MY-C Three-dimensional spatial registration and interactive display of position and orientation of real-time ultrasound images.J Ultrasound Med. 1990; 9: 525-532Crossref PubMed Scopus (137) Google Scholar Our 3D echocardiographic system is composed of a conventional real-time scanner, an acoustic spatial locator, and a personal computer. This system assigns a set of 3D (x, y, and z) spatial coordinates to each acquired real-time image and thereby measures its position and orientation in 3D space. We have developed our unique line of intersection display to guide data acquisition by visualizing the 3D position and orientation of a series of 2D images. To create the display, the 3D system acquires and saves in computer memory an initial reference image—typically, a parasternal long-axis image. A subsequent real-time short-axis image will intersect the reference long-axis image and create a line of intersection, the line common to both images. The spatial coordinates of each image are known; thus, rapid and repeated computation and display of the line of intersection in each image are possible. As the real-time image is moved, the recomputed line of intersection also moves in both images, and the changing relationship of the real-time image to the reference image is evident. This additional information provides the operator with a visual means to position the real-time image accurately with respect to otherwise nonvisualized anatomic landmarks now shown in the orthogonal reference image. By using the line of intersection displayed in the long-axis reference image, one can carefully position the short-axis image without angulation across the ventricle. This approach achieves a standardized, accurate, reproducible, and documented M-mode measurement of the LV correctly positioned in two orthogonal image planes, rather than just the single plane of the real-time image. The line of intersection display also provides, for the first time, a means to assess the performance of experienced operators in positioning standard 2D echocardiographic views.3King DL Harrison MR King Jr, DL Gopal AS Kwan OL DeMaria AN Ultrasound beam orientation during standard two-dimensional imaging: assessment by three-dimensional echocardiography.J Am Soc Echocardiogr. 1992; 5: 569-576PubMed Scopus (81) Google Scholar In a recent study of 85 examinations that included a total of 340 echocardiographic images obtained by 11 experienced operators at 3 institutions, we assessed actual image positioning achieved relative to defined optimal positioning in standard long-axis, short-axis, and apical two- and four-chamber views. Images were considered optimal if they were within ±5 mm of displacement and ±15 degrees of rotation-angulation of the defined optimal position. Overall, only 24% of images were optimally positioned relative to both displacement and angulation-rotation. The parasternal long-axis image was optimal two-thirds of the time, and two-thirds of the optimal images were the parasternal long-axis view. The parasternal short-axis image was displaced or angulated (or both) in 93% of the examinations. From this study, we concluded that 2D echocardiography does not achieve a reasonably consistent optimal positioning of standard 2D imaging views; hence, measurements determined from these views are likely to be suboptimal. Use of the line of intersection display to guide positioning of these images resulted in a threefold improvement.3King DL Harrison MR King Jr, DL Gopal AS Kwan OL DeMaria AN Ultrasound beam orientation during standard two-dimensional imaging: assessment by three-dimensional echocardiography.J Am Soc Echocardiogr. 1992; 5: 569-576PubMed Scopus (81) Google Scholar In another study, we assessed the reproducibility of the 1D measurement of the LV minor dimension determined by using the short-axis image and the line of intersection display for guidance.6King DL Harrison MR King Jr, DL Gopal AS Martin RP DeMaria AN Improved reproducibility of left atrial and left ventricular measurements by guided three-dimensional echocardiography.J Am Coll Cardiol. 1992; 20: 1238-1245Abstract Full Text PDF PubMed Scopus (102) Google Scholar We conducted this study because, in the assessment of standard 2D images, we found that one-third of the parasternal long-axis images were displaced medially or laterally by more than 5 mm from the center of the ventricle.3King DL Harrison MR King Jr, DL Gopal AS Kwan OL DeMaria AN Ultrasound beam orientation during standard two-dimensional imaging: assessment by three-dimensional echocardiography.J Am Soc Echocardiogr. 1992; 5: 569-576PubMed Scopus (81) Google Scholar This finding suggested that measurements commonly made from this view—the LV and left atrial (LA) dimensions—might also be suboptimal because of variation in the position of the image plane. We assessed the reproducibility of the LV and LA dimensions in a study of 30 patients at 3 institutions by using 3 independent pairs of experienced operators. They made the LA and LV measurements from the standard unguided 2D parasternal long-axis image and from a 3D line of intersection-guided parasternal short-axis image. The latter was similar to the short-axis image used by Shub and colleagues but carefully positioned by using the line of intersection display to be orthogonal to the left ventricular long axis in the same plane as the 2D measurement. The conventional unguided 2D examination was associated with an interobserver variability of 9.1% for the LV measurement, whereas the guided 3D echocardiographic method decreased interobserver variability to 3.1%.6King DL Harrison MR King Jr, DL Gopal AS Martin RP DeMaria AN Improved reproducibility of left atrial and left ventricular measurements by guided three-dimensional echocardiography.J Am Coll Cardiol. 1992; 20: 1238-1245Abstract Full Text PDF PubMed Scopus (102) Google Scholar The standard error was 5.6 mm for the unguided measurement and 2.0 mm for the guided measurement. Thus, with unguided examinations, the same LV dimension would differ by more than 5.6 mm in one-third of the cases. We concluded that 3D-guided imaging can significantly improve the reproducibility of these measurements and provide the basis for improved serial assessment of patients. If this method were used to assess LV mass, one would expect a similar improvement in reproducibility. That such variation can be significant is apparent when one realizes that a single-millimeter change of the ventricular chamber dimension can typically result in a change of approximately 8 g of mass when it is estimated by the 1D echocardiographic method in subjects with normal-size ventricles.1Devereux RB Lutas EM Casale PN Kligfield P Eisenberg RR Hammond IW et al.Standardization of M-mode echocardiographic left ventricular anatomic measurements.J Am Coll Cardiol. 1984; 4: 1222-1230Abstract Full Text PDF PubMed Scopus (727) Google Scholar 2D echocardiographic methods of estimating LV mass have also been developed.7Schiller NB Shah PM Crawford M DeMaria A Devereux R Feigenbaum H et al.Recommendations for quantitation of the left ventricle by two-dimensional echocardiography.J Am Soc Echocardiogr. 1989; 2: 358-367PubMed Google Scholar They are not, however, in general use because they do not provide sufficient improvement to justify their technical difficulty. As previously mentioned, the 2D echocardiographic methods are adversely affected by use of geometric assumptions and errors in image plane positioning. In addition, these methods use apical views of the ventricle, which are commonly foreshortened because of rib interference. 3D echocardiography now offers an alternative method of estimating LV mass—namely, 3D ventricular surface reconstruction. With use of a polyhedral surface reconstruction algorithm, the endocardial and epicardial surfaces of the ventricle can be reconstructed in three dimensions by a series of triangular tiles, and their volume can be calculated by subdividing it into a series of tetrahedrons.8Gopal AS King DL Katz J Boxt LM King Jr, DL Shao MY-C Three-dimensional echocardiographic volume computation by polyhedral surface reconstruction: in vitro validation and comparison to magnetic resonance imaging.J Am Soc Echocardiogr. 1992; 5: 115-124PubMed Google Scholar This method has proved to be accurate in vitro8Gopal AS King DL Katz J Boxt LM King Jr, DL Shao MY-C Three-dimensional echocardiographic volume computation by polyhedral surface reconstruction: in vitro validation and comparison to magnetic resonance imaging.J Am Soc Echocardiogr. 1992; 5: 115-124PubMed Google Scholar and in vivo.9Gopal AS Keller AM Rigling R King Jr, DL King DL Left ventricular volume and endocardial surface area by three-dimensional echocardiography: comparison with two-dimensional echocardiography and nuclear magnetic resonance imaging in normal subjects.J Am Coll Cardiol. 1993; 22: 258-270Abstract Full Text PDF PubMed Scopus (169) Google Scholar The myocardial volume is calculated by subtracting the endocardial from the epicardial volume, and the myocardial mass is determined by multiplying the myocardial volume by its density. Determination of LV mass of fixed hearts in vitro was associated with a standard error of 2.9 g and a root mean square percent error of 3.5%.10Gopal AS, Keller AM, Shen Z, Sapin PM, Schroeder KM, King DL Jr, et al. Three-dimensional echocardiography: in vitro and in vivo validation of left ventricular mass and comparison to conventional echocardiographic methods. J Am Coll Cardiol [in press]Google Scholar In normal human subjects, 3D echocardiographic determination of LV mass at end-systole in comparison with determination of LV mass by magnetic resonance imaging has also proved to be accurate, with a standard error of 9.2 g (r = 0.93; P<0.001).10Gopal AS, Keller AM, Shen Z, Sapin PM, Schroeder KM, King DL Jr, et al. Three-dimensional echocardiography: in vitro and in vivo validation of left ventricular mass and comparison to conventional echocardiographic methods. J Am Coll Cardiol [in press]Google Scholar No significant difference from magnetic resonance imaging was noted with use of the F test. A comparison with 2D echocardiographic determination of LV mass at end-diastole and end-systole with use of the truncated ellipsoid formula yielded 2D echocardiographic standard errors of 21.8 and 24.5 g, respectively, 2 to 3 times those with 3D echocardiography.10Gopal AS, Keller AM, Shen Z, Sapin PM, Schroeder KM, King DL Jr, et al. Three-dimensional echocardiography: in vitro and in vivo validation of left ventricular mass and comparison to conventional echocardiographic methods. J Am Coll Cardiol [in press]Google Scholar Thus, 3D echocardiography and polyhedral surface reconstruction offer an alternative method of measuring LV mass that may be more accurate than the 1D and 2D echocardiographic methods and may also, by eliminating geometric assumptions, be applicable to the study of LV mass in patients with abnormally shaped ventricles. In addition, 3D echocardiography provides a means of viewing the nonvisualized dimension of standard real-time images and thus, through application of the line of intersection display, can be used to guide the placement of the real-time image for making single dimensional measurements. Furthermore, it is able to make single dimensional measurements between image planes—that is, between two anatomic structures that cannot be visualized simultaneously. Consequently, we expect that 3D echocardiography will, in the future, contribute substantially to improved accuracy and reproducibility of clinical and research studies of LV mass.