- Email: [email protected]
Cyclic Variation of Integrated Backscatter: Dependence of Time Delay on the Echocardiographic View Used and the Myocardial Segment Analyzed
Cyclic Variation of Integrated Backscatter: Dependence of Time Delay on the Echocardiographic View Used and the Myocardial Segment Analyzed Ann E. Finch-Johnston, MA, Hiie M. Gussak, MD, Joel Mobley, PhD, Mark R. Holland, PhD, Olivera Petrovic, MD, Julio E. Pérez, MD, and James G. Miller, PhD, St Louis, Missouri
To determine the influence of myocardial anisotropy in ultrasonic tissue characterization, we measured the time delay (and magnitude) of the cyclic variation of myocardial integrated backscatter from specific segments visualized in the 4 standard transthoracic echocardiographic views. The cyclic variation data in 10 myocardial regions were obtained from analyses of 2-dimensional integrated backscatter images from 23 healthy subjects. Resultant values (mean ± SD) for the time delay were as follows: parasternal long-axis view: 1.08 ± 0.17 (septum) and 1.00 ± 0.14 (posterior
wall); parasternal short-axis view: 1.03 ± 0.16 (anterior septum), 1.03 ± 0.14 (posterior wall), 2.22 ± 0.71 (lateral wall), and 1.65 ± 0.66 (posterior septum); apical 4-chamber view: 1.08 ± 0.31 (septum) and 2.20 ± 0.79 (lateral wall); and apical 2-chamber view: 1.68 ± 0.62 (inferior wall) and 2.04 ± 0.72 (anterior wall). Hence, results of this study indicate that myocardial ultrasonic characterization that uses the cyclic variation is influenced by the echocardiographic view and the specific segment of the left ventricle. (J Am Soc Echocardiogr 2000;13:9-17.)
INTRODUCTION We and others have demonstrated that ultrasonic backscatter can characterize structural and functional alterations in the myocardium of patients with cardiomyopathy,1-7 coronary artery disease,8-14 or diabetes without overt heart disease.15-18 One method of quantifying such alterations clinically is the use of the cardiac cycle–dependent variation of myocardial integrated backscatter.8 Quantification of the cyclic variation of integrated backscatter based solely upon measurements of the “apparent” magnitude without regard to its “phase” (time delay) can significantly underestimate the true magnitude of cyclic variation.19 Furthermore, the characterization of specific pathologies can be significantly enhanced when both the magnitude and corresponding time delay are measured.17 The objective of the present study was to determine the dependence of the normalized time delay and magnitude of cyclic variation of integrated From the Department of Physics and the Cardiovascular Division, Washington University, St Louis, Mo. This work is supported in part by NIH Grants HL40302 and HL53461. Reprint requests: Julio E. Pérez, MD, Campus Box 8086, Cardiovascular Division, Washington University, 660 S Euclid Avenue, St Louis, MO 63110. Copyright © 2000 by the American Society of Echocardiography. 0894-7317/2000 $12.00 + 0 27/1/102695
backscatter on the transthoracic echocardiographic view used and specific left ventricular (LV) myocardial segment investigated.The range of normal values for the time delay and magnitude of cyclic variation must be defined for all echocardiographic views and specific LV myocardial segments before the cyclic variation can be reliably used as a clinical diagnostic tool for a specific patient population. Our approach was to obtain standard transthoracic images from a series of healthy human subjects and perform an automated analysis of the measured cyclic variation to provide operator-independent measurements of the time delay and magnitude in 10 of the 16 LV myocardial segments recognized in transthoracic imaging. The measurement of the echocardiographic view–dependence of the time delay of cyclic variation reported in this study further extends and complements previous studies that have reported the view-dependence of the apparent magnitude of cyclic variation.20-24
METHODS Subjects Twenty-three healthy human volunteers without history of cardiovascular disease (mean age 31 ± 8 years, range 19 to 54 years) were recruited for participation in this study.
9
10 Finch-Johnston et al
A
B
Figure 1 A, Diagram illustrating the definition of the magnitude of cyclic variation of hypothetical data, and B, the corresponding normalized time delay. ECG, Electrocardiogram.
Data Acquisition Two-dimensional integrated backscatter images of the left ventricle were obtained from standard transthoracic echocardiography views (parasternal long- and short-axis, apical 2- and 4-chamber) with use of a SONOS 1500 (Hewlett-Packard Co, Andover, Mass) ultrasonic imaging system operating in the Acoustic Densitometry (AD) acquisition mode. Use of the AD imaging mode permits the acquisition, display, and subsequent analyses of real-time integrated backscatter images on the basis of a previously described method.1,10,12,25,26 Images were obtained with a 2.5-MHz phased-array probe with the transmit power and time gain compensation values set to optimize image quality for each of the 4 echocardiographic views. Care was taken to ensure that the level of backscattered signal (image brightness) from the midmyocardial regions remained well within the dynamic range of the imaging system over the entire cardiac cycle. Lateral gain compensation was not used in this study. For each echocardiographic view, digital cineloops of 60 consecutive integrated backscatter image frames (approximately 2 cardiac cycles) were acquired at a rate of 30 frames/s and were stored on optical disk for subsequent analysis.
Journal of the American Society of Echocardiography January 2000
Data Analysis Analysis of integrated backscatter images. The integrated backscatter image data, corresponding to each view from each subject, were analyzed by using the AD analysis package available on the SONOS 1500 imaging system. An elliptical region of interest was placed in the mid myocardium of specific myocardial segments in each view. The size of the region of interest used in analyzing each segment was chosen to be the largest size available that would fit within the myocardium, avoiding the endocardial and epicardial boundaries. For each myocardial segment analyzed, the region of interest was carefully positioned to avoid the bright endocardial and epicardial specular reflections and was tracked throughout the entire cineloop (60 frames), keeping the region of interest within the mid myocardium. This was accomplished by adjusting the position of the region of interest in every frame such that approximately the same 2-dimensional area of myocardium was sampled over the heart cycle. Myocardial segments in which the tissue was poorly visualized were not analyzed. For each myocardial segment, the AD analysis package provided a mean integrated backscatter value, representing those values within the region of interest, for each image frame. Thus a data recording that represented the cyclic variation of integrated backscatter values as a function of frame number (time) was generated for each myocardial segment. Because the images were acquired with a frame rate of 30 Hz, the resulting cyclic variation data was sampled in 33-ms time increments. For subsequent analyses of the cyclic variation data, the LV end-diastolic and end-systolic frames were identified as defined by the American Society of Echocardiography.27 End-diastole was defined as the frame at or before mitral valve closure, or the largest cavity area if no valves were visible. End-systole was defined as the frame preceding mitral valve opening, or the smallest cavity area if no valves were visible. The systolic interval, as determined in this study, was the time difference between the end-diastolic and end-systolic frames. Analysis of cyclic variation data. Methods for acquiring and estimating the cyclic variation with real-time integrated backscatter imaging have been previously described.1,10,12,25,26 In characterizing the cyclic variation, we report the magnitude and the corresponding time delay normalized to the duration of the systolic interval.28,29 We define the magnitude of cyclic variation as the difference between the average peak and average nadir values of integrated backscatter as illustrated in Figure 1, A. The corresponding normalized time delay of cyclic variation is expressed as a dimensionless ratio obtained by dividing the time interval from end-diastole to the nadir of the integrated backscatter trace by the systolic interval as shown in Figure 1, B. These parameters were calculated by using a previously reported automated analysis algo-
Journal of the American Society of Echocardiography Volume 13 Number 1
Finch-Johnston et al 11
A
B
Figure 2 Measured values of the magnitude and normalized time delay for specified regions of interest in the parasternal long-axis view (A), and parasternal short-axis view (B). Magnitude values are shown on the left axis and normalized time delay values on the right axis.
rithm.28,29 This algorithm uses cross-correlation and Fourier decomposition of a model function to provide operator-independent estimates of the time delay and magnitude. Implementation of the automated analysis algorithm requires the end-diastolic and end-systolic frame numbers (systolic interval) to be supplied in addition to the corresponding integrated backscatter values as a function of frame number obtained with the AD analysis package (typically over a couple complete cardiac cycles). From the input end-diastolic and end-systolic frame numbers (timing information), a model function representing the anticipated periodicity of the cyclic variation waveform was constructed. This model function was then crosscorrelated with the measured cyclic variation of integrated
backscatter waveform. The time shift applied to the model function that maximized the correlation with the measured cyclic variation waveform was then used to identify the time at which the nadir of the cyclic variation occurred. The time interval from end-diastole to the nadir of the integrated backscatter trace was then used to compute the normalized time delay as described above. Because this cross-correlation technique was often applied over several heart cycles and the average systolic interval was used, the effects of beat-to-beat variations should have been reduced. Both the magnitude and the time delay parameters characterizing the cyclic variation were calculated from the same cardiac cycles. Statistical methods for comparisons. The measured val-
Journal of the American Society of Echocardiography January 2000
12 Finch-Johnston et al
A
B
Figure 3 Measured values of the magnitude and normalized time delay for specified regions of interest in the apical 4-chamber view (A), and the apical 2-chamber view (B). Magnitude values are shown on the left axis and normalized time delay values on the right axis.
ues for the time delay and magnitude of cyclic variation corresponding to specific myocardial segments were compared by using a paired, 2-tailed Student t test with Microsoft Excel (Microsoft Corporation, Redmond, Wash) and the corresponding P value for each of the comparisons was given. Comparisons of the mean values for which P was less than .05 were considered significant.
RESULTS Parasternal Views The locations of regions of interest and the resulting average measurements of time delay and magnitude
of cyclic variation for specific segments visualized in the parasternal views are illustrated in Figure 2.Values for the time delay and magnitude for the 2 regions analyzed in the parasternal long-axis view (midseptal and midposterior wall) are shown in Figure 2, A, and values for the 4 regions analyzed in the parasternal short-axis view (anterior septum, lateral wall, posterior wall, and posterior septum) are depicted in Figure 2, B.The measured values of the time delay for both of the myocardial segments analyzed in the parasternal long-axis view were similar (ie, relatively short). However, the values for the time delay for the segments analyzed in the parasternal short-axis view were relatively longer for the lateral wall and posterior septum compared with the values for the anterior
Journal of the American Society of Echocardiography Volume 13 Number 1
Finch-Johnston et al 13
septum and posterior wall. (P < .001, lateral wall compared with anterior septum; P = .02, posterior septum compared with anterior septum; P < .001, lateral wall compared with posterior wall; P < .001, posterior septum compared with posterior wall.) Apical Views The locations of the regions of interest and the resulting average measurements of time delay and magnitude of cyclic variation for specific myocardial segments visualized in the apical views are illustrated in Figure 3. Values for the 2 regions analyzed in the apical 4-chamber view (midseptal and midlateral wall) are shown in Figure 3, A, and the values for the 2 regions analyzed in the apical 2-chamber view (midinferior wall and midanterior wall) are shown in Figure 3, B.The values of the time delay corresponding to the midlateral wall in the 4-chamber view and the midinferior wall and midanterior wall in the 2chamber view were relatively long compared with the value measured for the mid septum in the 4chamber view. (P was less than .001 for each of the time delay comparisons: midlateral wall compared with the mid septum in the 4-chamber view; midinferior wall in the 2-chamber view compared with the mid septum in the 4-chamber view; midanterior wall in the 2-chamber view compared with the mid septum in the 4-chamber view.) Comparison of Myocardial Segments A comparison of the time delay and magnitude of cyclic variation values for comparable myocardial segments are given in Table 1. This table provides a direct comparison between specific myocardial segments representing roughly the same myocardial region when imaged in different echocardiographic views (ie, approximately the same region and perfused by the same coronary distribution). The results of this comparison indicate that the measured values of the time delay and magnitude for comparable myocardial segments were the same (ie, not significantly different) in all but one of the comparisons. Only the magnitude of cyclic variation from the posterior septal segment in the parasternal short-axis view exhibited a significant difference (ie, P < .05) when compared with the magnitude from the inferior wall when imaged in the apical 2-chamber view.
DISCUSSION Quantitative echocardiography with ultrasonic tissue characterization methods complement conventional 2-dimensional echocardiography by providing
Figure 4 Relationship between the predominant myocardial fiber orientation and the direction of the insonifying ultrasonic beams for the parasternal short-axis view.
an approach for directly evaluating the physical state of the myocardium.1-18,30,31 The underlying hypothesis is that pathologic changes of myocardial structure and function result in alterations in the fundamental physical properties of tissue that can be quantified with indexes based on frequency-dependent ultrasonic attenuation and backscatter.1,8-10,12,31-34 Our laboratory introduced the concepts of magnitude and phase to characterize the cyclic variation of myocardial integrated backscatter and introduced the parameter of phase-weighted magnitude which we11,35 and others14,36 have used successfully. However, we advocate reporting both the magnitude and the time delay of cyclic variation, as exemplified in a study that detected changes in backscatter from patients with insulin-dependent diabetes mellitus17 and identified otherwise latent changes in myocardial structure that accompany the evolution of insulin-dependent diabetes mellitus. Although the cyclic variation is dependent in part on regional myocardial contractile function,35,37 the definitive mechanisms responsible for the cyclic variation of integrated backscatter have not been elucidated. We hypothesize that the anisotropy of both ultrasonic backscatter and attenuation (ie, their dependence upon the angle of the incident ultrasonic beam relative to the direction of the myocardial fibers) plays a significant role.32,38-42 Previous studies have demonstrated that the inherent anisotropy of myocardium can significantly influence quantitative ultrasonic measurements. Specifically, integrated backscatter values were largest when the insonifying ultrasound was perpendicular to the predominant myocardial fiber orientation,38-40,43 whereas the ultrasonic attenuation was the largest when the insonifying ultrasound was parallel to the myofibers.44,45 These inherent anisotropic properties of the heart have significant effects in the
Journal of the American Society of Echocardiography January 2000
14 Finch-Johnston et al
Table 1 Cyclic variation of integrated backscatter for comparable myocardial segments in different views Region of interest
Normalized time delay ± SD
Comparison P value (delay)
Magnitude ± SD (dB)
Comparison P value (magnitude)
SAX anterior septum LAX septum SAX anterior septum A4C septum LAX septum A4C septum SAX posterior wall LAX posterior wall SAX lateral wall A4C lateral wall SAX lateral wall A2C anterior wall A4C lateral wall A2C anterior wall SAX post septum A2C inferior wall
1.03 ± 0.16 1.08 ± 0.17 1.03 ± 0.16 1.08 ± 0.31 1.08 ± 0.17 1.08 ± 0.31 1.03 ± 0.14 1.00 ± 0.14 2.22 ± 0.71 2.20 ± 0.79 2.22 ± 0.71 2.04 ± 0.72 2.20 ± 0.79 2.04 ± 0.72 1.65 ± 0.66 1.68 ± 0.62
.62
4.3 ± 2.3 4.4 ± 1.4 4.3 ± 2.3 4.7 ± 2.0 4.4 ± 1.4 4.7 ± 2.0 6.0 ± 1.5 6.1 ± 2.0 2.5 ± 1.2 2.8 ± 1.3 2.5 ± 1.2 2.3 ± 0.8 2.8 ± 1.3 2.3 ± 0.8 2.5 ± 1.0 3.7 ± 1.8
.80
.62 .89 .27 .18 .11 .49 .80
.26 .51 .78 .67 .73 .08 .01
SAX, Parasternal short-axis; LAX, parasternal long-axis; A4C, apical 4-chamber; A2C, apical 2-chamber.
echocardiographic images that are obtained from standard views.46-48 Hence, because the angle of insonification between an incident ultrasonic beam and the local myofiber orientation for a specific region of the heart is echocardiographic view– dependent, we postulated that the time delay of cyclic variation would be view-dependent as well. Measurements of the cyclic variation of myocardial segments obtained from the parasternal longaxis view, with the insonifying beam approximately perpendicular to myofibers in all segments, exhibit minimal anisotropic effects.* However, in other standard transthoracic echocardiography views, the angle between the insonifying beam and the predominant myofibers is dependent on the myocardial segment of interest. In the parasternal short-axis imaging plane, the midmyocardial fibers exhibit primarily a circumferential arrangement,52,53 as indicated in Figure 4. Thus the midmyocardial fibers are predominantly perpendicular to the insonifying ultrasonic beams in the anterior septal and posterior wall regions and predominantly parallel in the posterior septal and lateral wall regions.The relationship between the orientation of the myocardial fibers and the insonifying ultrasonic beam is more complex in the apical views, especially near the apex of the heart. The echocardiographic view–dependence of the measured magnitude of cyclic variation of integrated backscatter for specific myocardial segments has been reported previously.20-24
*1,3,4,10,12,15,36,49-51
In the present study, measurements of the time delay and magnitude of cyclic variation for the parasternal views demonstrate a strong view-dependence, as indicated in Figure 2. As described above, in the mid myocardium surrounding the left ventricle at the midpapillary level, a significant fraction of the myocardial fibers run circumferentially in the transverse plane, with the largest deviations out of the transverse plane occurring near the endocardial and epicardial surfaces.39,42,52,54-56 In spite of these transmural changes in fiber direction, the ultrasonic beam is primarily perpendicular to the predominant fiber direction in the septum and posterior wall within the imaging plane of the parasternal long-axis view.48 Myocardial segments in which the insonifying ultrasonic beam was primarily perpendicular to the predominant myocardial fibers yielded time delays close to unity with corresponding magnitudes of cyclic variation consistent with those previously reported in healthy subjects (4 to 6 dB).1,30 Myocardial segments imaged in those regions of the shortaxis view in which the ultrasonic beam propagated more nearly parallel to the predominant myocardial fibers (lateral wall and posterior septum) were characterized as having substantially longer values of time delays and somewhat smaller magnitudes of cyclic variation. A comparison of the measured values of cyclic variation for myocardial segments visualized in both parasternal echocardiographic views (anterior septum and posterior wall) show that they are quite comparable (Table 1). For these myocardial segments the insonifying beam is approximately perpendicular to the local myofiber direction in both views. Values of the magnitude of cyclic variation of inte-
Journal of the American Society of Echocardiography Volume 13 Number 1
grated backscatter obtained from the long-axis view (septum and posterior wall) in this study are consistent with those of previous studies.1,49 The values obtained for the normalized time delay, however, are slightly larger than those previously reported (mid septum: 1.08 ± 0.17 in this study compared with a value of 0.84 ± 0.09 reported previously; midposterior wall: 1.00 ± 0.14 in this study compared with a value of 0.85 ± 0.12 reported previously).57 This difference may be the result, in part, of technical limitations in the rate of sampling of integrated backscatter values used in this study, compared with previous M-mode–based studies. In this study, a value for the mean integrated backscatter within a specific region of interest was determined once per frame (ie, once every 33 ms [for a 30 frame/s image display rate]. In previous M-mode–based studies, the effective sampling rate was much larger with sampling periods of around 5 ms. Hence, in the present study, there could be an approximate 10% to 20% uncertainty in the measurement of the systolic interval, depending on the timing of image frame acquisition relative to the true systolic interval. Furthermore, another contributing factor is that in previous studies, estimates of the normalized time delay were measured as the ratio of the time from the R wave to the integrated backscatter nadir, which was normalized by the QTinterval on the basis of information from the ECG tracing, whereas in our study the systolic interval was more accurately based on mechanical events guided from the 2-dimensional images. This latter approach results in both timing intervals (ie, the start of systole to integrated backscatter nadir and the systolic interval) being measured from the same starting time (end-diastolic frame), which tends to give a larger ratio and therefore a larger time delay. Measurements of the time delay and magnitude of myocardial cyclic variation in the apical 4- and 2chamber views, in which the orientation of the myocardial fibers with respect to the insonifying ultrasonic beam is more complex, also show a dependence on the myocardial segment analyzed as indicated in Figure 3. A comparison of the cyclic variation values for the lateral wall, imaged in the apical 4-chamber view, with the corresponding values for the lateral wall, imaged in the parasternal shortaxis view, reveals comparable values for time delay and magnitude (Table 1). Measurements of the cyclic variation in this region from these two views exhibited relatively long time delays with relatively small magnitudes. The model of fiber orientation would suggest that the ultrasonic beam was relatively parallel to the predominant fiber direction in the mid myocardium of the lateral wall when imaged in the
Finch-Johnston et al 15
parasternal short-axis view (Figure 4) and was perpendicular to the predominant fiber direction in the mid myocardium of the lateral wall of the apical 4chamber view. However, these results for the lateral wall indicate that the cyclic variation of the lateral wall of the 4-chamber view exhibits values more consistent with parallel insonication.We hypothesize that this may be a result of the relatively complex fiber orientation in the apical views and, perhaps, may be related to the transmural change in fiber orientation in the lateral wall, resulting in fibers (or fiber components) within the image plane of the apical views. Quite similar values of time delay and magnitude were found for the anterior wall of the apical 2-chamber view, the lateral wall of the parasternal short-axis view, and the lateral wall of the apical 4chamber view (Table 1). A comparison of the cyclic variation obtained from the septum of the apical 4-chamber view with that from the anterior septum of the parasternal short-axis view and the septum of the parasternal long-axis view revealed quite similar values for time delay and magnitude in each case (Table 1).The values from the septum of the apical 4-chamber view exhibited a relatively short time delay with a corresponding large magnitude of cyclic variation.These results are consistent with those obtained for perpendicular insonication with respect to the myofibers, as in the case for the corresponding segments in the parasternal views. Hence, these results, along with the observation that the septum is often much more easily visualized than the lateral wall in the apical 4-chamber view in a clinical echocardiographic exam, suggest that significant differences may exist in fiber orientation in the septum when compared with the fiber architecture of the lateral wall as imaged in the apical 4-chamber view. Summary Results of this study indicate that myocardial ultrasonic characterization that uses the time delay of cyclic variation is influenced by the echocardiographic view employed and the specific segment of the left ventricle analyzed. This is consistent with measurements of the view dependence of the magnitude of cyclic variation reported by others.20-24 In regions of myocardium where the angle of insonication relative to local myofiber orientation is nearly parallel, relatively long time delays and small magnitudes are to be expected in comparison with the corresponding results for regions with perpendicular insonication. Furthermore, measurements of cyclic variation may be affected not only by the complex
Journal of the American Society of Echocardiography January 2000
16 Finch-Johnston et al
fiber geometry but also by the effects of translation, rotation, and twist of the heart during cardiac systole. Comparison of the time delay and magnitude of cyclic variation values for comparable myocardial segments indicate that the measured values were the same in all but one of the comparisons. Although the healthy subjects we studied represent a relatively young population, this systematic mapping of the time delay (and magnitude) of the cyclic variation of integrated backscatter provides a first step toward the development of a basis for delineating the measurements obtained from patients with altered LV segmental function. In addition, it establishes a framework for the interpretation of alterations in acoustic properties (eg,those resulting from ischemia, infarction, fibrosis, the passage of microbubbles in contrast echocardiography) of specific myocardial segments visualized from any of the 4 standard transthoracic windows. We would like to thank Kirk D. Wallace and Steven L. Baldwin for their contributions toward data analysis and Barbara L. Roman and Debbie Taylor for their assistance in manuscript preparation.
9.
10.
11.
12.
13.
14.
15.
16.
REFERENCES 1. Vered Z, Barzilai B, Mohr GA, et al. Quantitative ultrasonic tissue characterization with real-time integrated backscatter imaging in normal human subjects and in patients with dilated cardiomyopathy. Circulation 1987;76:1067-73. 2. Goens MB, Karr SS, Martin GR. Cyclic variation of integrated ultrasound backscatter: normal and abnormal myocardial patterns in children. J Am Soc Echocardiogr 1996;9:616-21. 3. Naito J, Masuyama T, Tanouchi J, et al. Analysis of transmural trend of myocardial integrated ultrasound backscatter for differentiation of hypertrophic cardiomyopathy and ventricular hypertrophy due to hypertension. J Am Coll Cardiol 1994;24:517-24. 4. Naito J, Masuyama T, Mano T, et al. Dobutamine stress ultrasonic myocardial tissue characterization in patients with dilated cardiomyopathy. J Am Soc Echocardiogr 1996;9:4709. 5. Dibello V, Pedrinelli R, Giorgi D, et al. Ultrasonic videodensitometric analysis of two different models of left ventricular hypertrophy—athletes heart and hypertension. Hypertension 1997;29:937-44. 6. Zoni A, Regolisti G, Aschieri D, Borghetti A. Myocardial ultrasonic tissue characterization in patients with different types of left ventricular hypertrophy—a videodensitometric approach. J Am Soc Echocardiogr 1997;10:74-82. 7. Ciliberto GR, Pingitore A, Mangiavacchi M, Alberti A, Paterni M, Picano E. The clinical value of blunting of cyclic gray level variation for the detection of acute cardiac rejection: a two-dimensional, Doppler, and videodensitometric ultrasound study. J Am Soc Echocardiogr 1996;9:306-13. 8. Barzilai B, Madaras EI, Sobel BE, Miller JG, Perez JE. Effects of myocardial contraction on ultrasonic backscatter before
17.
18.
19.
20.
21.
22.
23.
24.
and after ischemia. Am J Physiol 1984;247(3 Pt 2):H478H483. Fitzgerald PJ, McDaniel MD, Rolett EL, Strohbehn JW, James DH. Two-dimensional ultrasonic tissue characterization: backscatter power, endocardial wall motion, and their phase relationship for normal, ischemic, and infarcted myocardium. Circulation 1987;76:850-9. Vered Z, Mohr GA, Barzilai B, et al. Ultrasound integrated backscatter tissue characterization of remote myocardial infarction in human subjects. J Am Coll Cardiol 1989;13:8491. Wickline SA, Thomas LJ III, Miller JG, Sobel BE, Perez JE. Sensitive detection of the effects of reperfusion on myocardium by ultrasonic tissue characterization with integrated backscatter. Circulation 1986;74:389-400. Milunski MR, Mohr GA, Perez JE, et al. Ultrasonic tissue characterization with integrated backscatter: acute myocardial ischemia, reperfusion, and stunned myocardium in patients. Circulation 1989;80:491-503. Saeian K, Rhyne TL, Sagar KB. Ultrasonic tissue characterization for diagnosis of acute myocardial infarction in the coronary care unit. Am J Cardiol 1994;74:1211-5. Takiuchi S, Ito H, Iwakura K, et al. Ultrasonic tissue characterization predicts myocardial viability in early stage of reperfused acute myocardial infarction. Circulation 1998;97:35662. Perez JE, McGill JB, Santiago JV, et al. Abnormal myocardial acoustic properties in diabetic patients and their correlation with the severity of disease. J Am Coll Cardiol 1992;19: 1154-6. DiBello V, Talarico L, Picano E, et al. Increased echodensity of myocardial wall in the diabetic heart: an ultrasound tissue characterization study. J Am Coll Cardiol 1995;25:1408-15. Wagner RF, Wear KA, Perez JE, McGill JB, Schechtman KB, Miller JG. Quantitative assessment of myocardial ultrasound tissue characterization through receiver operating characteristic analysis of Bayesian classifiers. J Am Coll Cardiol 1995;25: 1706-11. Dibello V, Giampietro O, Matteucci E, et al. Ultrasonic videodensitometric analysis in type 1 diabetic myocardium. Coron Artery Dis 1996;7:895-901. Finch-Johnston AE, Gussak HM, Mobley J, et al. Dependence of “apparent” magnitude on the time delay of cyclic variation of myocardial backscatter. Ultrasound Med Biol 1999;25:759-62. Bouki KP, Lange A, Palka P, et al. Regional variations of ultrasonic integrated backscatter in normal and myopathic left ventricles—a new multi-view approach. Eur Heart J 1996;17:1747-55. Lange A, Moran CM, Palka P, Fenn LN, Sutherland GR, McDicken WN. The variation of integrated backscatter in human hearts in differing ultrasonic transthoracic views. J Am Soc Echocardiogr 1995;8:830-8. Dittrich HC, Wheeler K, Kwan OL, et al. Quantitation of integrated myocardial backscatter from two-dimensional echo images: influence of beam orientation upon cyclic variation. J Am Soc Echocardiogr 1993;3:S30. Vandenberg BF, Rath L, Shoup TA, Kerber RE, Collins SM, Skorton DJ. Cyclic variation of ultrasound backscatter in normal myocardium is view dependent: clinical studies with a real-time backscatter imaging system. J Am Soc Echocardiogr 1989;2:308-14. Vandenberg BF, Stuhlmuller JE, Rath L, et al. Diagnosis of recent myocardial infarction with quantitative backscatter
Journal of the American Society of Echocardiography Volume 13 Number 1
25.
26.
27.
28.
29.
30.
31. 32.
33.
34.
35.
36.
37.
38.
39.
40.
imaging: preliminary studies. J Am Soc Echocardiogr 1991; 4:10-8. Thomas LJ, Wickline SA, Pérez JE, Sobel BE, Miller JG. A real-time integrated backscatter measurement system for quantitative tissue characterization. IEEE Trans Ultrason Ferroelect Freq Control UFFC 1986;33:27-32. Barzilai B, Thomas LJ III, Glueck RM, et al. Detection of remote myocardial infarction with quantitative real-time ultrasonic characterization. J Am Soc Echocardiogr 1988;1: 179-86. Schiller NB, Shah PM, Crawford M, et al. Recommendations for quantitation of the left ventricle by two-dimensional echocardiography. J Am Soc Echocardiogr 1989;2:358-68. Mohr GA, Vered Z, Barzilai B, Perez JE, Sobel BE, Miller JG. Automated determination of the magnitude and time delay (“phase”) of the cardiac cycle dependent variation of myocardial ultrasonic integrated backscatter. Ultrason Imaging 1989;11:245-59. Mobley J, Banta CE, Gussak HM, Perez JE, Miller JG. Clinical tissue characterization: online determination of magnitude and time delay of myocardial backscatter. Video Journal of Echocardiography 1995;5:40-8. Perez JE, Holland MR, Barzilai B, et al. Ultrasonic characterization of cardiovascular tissue. In: Cardiac imaging: a companion to Braunwald’s heart disease. Skorton DJ, Schelbert HR, Wolf GL, Brundage BH, Skorton DJ, Schelbert HR, Wolf GL, Brundages BH, editors. 2nd ed. Philadelphia: WB Saunders Co; 1996. p. 606-22. Miller JG, Perez JE, Sobel BE. Ultrasonic characterization of myocardium. Prog Cardiovasc Dis 1985;XXVIII:85-110. O’Brien WD Jr, Sagar KB, Warltier DC, Rhyne TL. Acoustic propagation properties of normal, stunned, and infarcted myocardium: morphological and biochemical determinants. Circulation 1995;91:154-60. Rijsterborgh H, Mastik F, Lancee CT, et al. Ultrasonic myocardial integrated backscatter and myocardial wall thickness in animal experiments. Ultrasound Med Biol 1990;16: 29-36. Mimbs JW, O’Donnell M, Bauwens D, Miller JG, Sobel BE. The dependence of ultrasonic attenuation and backscatter on collagen content in dog and rabbit hearts. Circ Res 1980;47: 49-58. Wickline SA, Thomas LJ III, Miller JG, Sobel BE, Perez JE. The dependence of myocardial ultrasonic integrated backscatter on contractile performance. Circulation 1985;72:18392. Vitale DF, Bonow RO, Gerundo G, et al. Alterations in ultrasonic backscatter during exercise-induced myocardial ischemia in humans. Circulation 1995;92:1452-7. Wickline SA, Thomas LJ III, Miller JG, Sobel BE, Perez JE. A relationship between ultrasonic integrated backscatter and myocardial contractile function. J Clin Invest 1985;76:215160. Madaras EI, Perez J, Sobel BE, Mottley JG, Miller JG. Anisotropy of the ultrasonic backscatter of myocardial tissue: II. Measurements in vivo. J Acoust Soc Am 1988;83:762-9. Verdonk ED, Wickline SA, Miller JG. Anisotropy of ultrasonic velocity and elastic properties in normal human myocardium. J Acoust Soc Am 1992;92:3039-50. Hoffmeister BK, Wong AK, Verdonk ED, Wickline SA, Miller JG. Comparison of the anisotropy of apparent integrated ultrasonic backscatter from fixed human tendon and fixed human myocardium. J Acoust Soc Am 1995;97:1307-13.
Finch-Johnston et al 17
41. O’Brien PD, O’Brien WD Jr, Rhyne TL, Warltier DC, Sagar KB. Hypertrophy/heart failure/ventricular dysfunction: relation of ultrasonic backscatter and acoustic propagation properties to myofibrillar length and myocardial thickness. Circulation 1995;91:171-5. 42. Verdonk ED, Hoffmeister BK, Wickline SA, Miller JG. Anisotropy of the slope of ultrasonic attenuation in formalin fixed human myocardium. J Acoust Soc Am 1996;99:3837-43. 43. Mottley JG, Miller JG. Anisotropy of the ultrasonic backscatter of myocardial tissue: I. Theory and measurements in vitro. J Acoust Soc Am 1988;83:755-61. 44. Brandenburger GH, Klepper JR, Miller JG, Snyder DL. Effects of anisotropy in the ultrasonic attenuation of tissue on computed tomography. Ultrason Imaging 1981;3:113-43. 45. Mottley JG, Miller JG. Anisotropy of the ultrasonic attenuation in soft tissues: measurements in vitro. J Acoust Soc Am 1990;88:1203-10. 46. Holland MR, Wilkenshoff UM, Finch-Johnston AE, Handley SM, Perez JE, Miller JG. Effects of myocardial fiber orientation in echocardiography: quantitative measurements and computer simulation of the regional dependence of backscattered ultrasound in the parasternal short-axis view. J Am Soc Echocardiogr 1998;11:929-37. 47. Recchia D, Miller JG, Wickline SA. Quantification of ultrasonic anisotropy in normal myocardium with lateral gain compensation of two-dimensional integrated backscatter images. Ultrasound Med Biol 1993;19:497-505. 48. Aygen M, Popp RL. Influence of the orientation of myocardial fibers on echocardiographic images. Am J Cardiol 1987; 60:147-52. 49. Stuhlmuller JE, Skorton DJ, Burns TL, Melton HEJ, Vandenberg BF. Reproducibility of quantitative backscatter echocardiographic imaging in normal subjects. Am J Cardiol 1992;69:542-6. 50. Vitale DF, Bonow RO, Calabro R, et al. Myocardial ultrasonic tissue characterization in pediatric and adult patients with hypertrophic cardiomyopathy. Circulation 1996;94: 2826-30. 51. Feinberg MS, Gussak HM, Davila-Roman VG, Baumann CM, Miller JG, Perez JE. Dissociation between wall thickening of normal myocardium and cyclic variation of backscatter during inotropic stimulation. Am J Cardiol 1996;77:515-20. 52. Streeter DD Jr, Spotnitz HM, Patel DP, Ross J Jr, EH Sonnenblick. Fiber orientation in the canine left ventricle during diastole and systole. Circ Res 1969;XXIV:339-47. 53. Greenbaum RA, Ho SY, Gibson DG, Becker AE, Anderson RH. Left ventricular fibre architecture in man. Br Heart J 1981;45:248-63. 54. Streeter DD Jr, Hanna WT. Engineering mechanics for successive states in canine left ventricular myocardium. II. Fiber angle and sarcomere length. Circ Res 1973;33:656-64. 55. Recchia D, Hall C, Shepard RK, Miller JG. Mechanisms of the view-dependence of ultrasonic backscatter from normal myocardium. IEEE Trans Ultrason Ferroelec Freq Contr 1995;42:91-8. 56. Wickline SA, Verdonk ED, Miller JG. Three-dimensional characterization of human ventricular myofiber architecture by ultrasonic backscatter. J Clin Invest 1991;88:438-46. 57. Loomis JF Jr, Waggoner AD, Schechtman KB, Miller JG, Sobel BE, Perez JE. Ultrasonic integrated backscatter 2-d imaging: evaluation of M-mode guided acquisition and immediate analysis in 55 consecutive patients. J Am Soc Echocardiogr 1990;3:255-65.
To determine the influence of myocardial anisotropy in ultrasonic tissue characterization, we measured the time delay (and magnitude) of the cyclic variation of myocardial integrated backscatter from specific segments visualized in the 4 standard transthoracic echocardiographic views. The cyclic variation data in 10 myocardial regions were obtained from analyses of 2-dimensional integrated backscatter images from 23 healthy subjects. Resultant values (mean ± SD) for the time delay were as follows: parasternal long-axis view: 1.08 ± 0.17 (septum) and 1.00 ± 0.14 (posterior
wall); parasternal short-axis view: 1.03 ± 0.16 (anterior septum), 1.03 ± 0.14 (posterior wall), 2.22 ± 0.71 (lateral wall), and 1.65 ± 0.66 (posterior septum); apical 4-chamber view: 1.08 ± 0.31 (septum) and 2.20 ± 0.79 (lateral wall); and apical 2-chamber view: 1.68 ± 0.62 (inferior wall) and 2.04 ± 0.72 (anterior wall). Hence, results of this study indicate that myocardial ultrasonic characterization that uses the cyclic variation is influenced by the echocardiographic view and the specific segment of the left ventricle. (J Am Soc Echocardiogr 2000;13:9-17.)
INTRODUCTION We and others have demonstrated that ultrasonic backscatter can characterize structural and functional alterations in the myocardium of patients with cardiomyopathy,1-7 coronary artery disease,8-14 or diabetes without overt heart disease.15-18 One method of quantifying such alterations clinically is the use of the cardiac cycle–dependent variation of myocardial integrated backscatter.8 Quantification of the cyclic variation of integrated backscatter based solely upon measurements of the “apparent” magnitude without regard to its “phase” (time delay) can significantly underestimate the true magnitude of cyclic variation.19 Furthermore, the characterization of specific pathologies can be significantly enhanced when both the magnitude and corresponding time delay are measured.17 The objective of the present study was to determine the dependence of the normalized time delay and magnitude of cyclic variation of integrated From the Department of Physics and the Cardiovascular Division, Washington University, St Louis, Mo. This work is supported in part by NIH Grants HL40302 and HL53461. Reprint requests: Julio E. Pérez, MD, Campus Box 8086, Cardiovascular Division, Washington University, 660 S Euclid Avenue, St Louis, MO 63110. Copyright © 2000 by the American Society of Echocardiography. 0894-7317/2000 $12.00 + 0 27/1/102695
backscatter on the transthoracic echocardiographic view used and specific left ventricular (LV) myocardial segment investigated.The range of normal values for the time delay and magnitude of cyclic variation must be defined for all echocardiographic views and specific LV myocardial segments before the cyclic variation can be reliably used as a clinical diagnostic tool for a specific patient population. Our approach was to obtain standard transthoracic images from a series of healthy human subjects and perform an automated analysis of the measured cyclic variation to provide operator-independent measurements of the time delay and magnitude in 10 of the 16 LV myocardial segments recognized in transthoracic imaging. The measurement of the echocardiographic view–dependence of the time delay of cyclic variation reported in this study further extends and complements previous studies that have reported the view-dependence of the apparent magnitude of cyclic variation.20-24
METHODS Subjects Twenty-three healthy human volunteers without history of cardiovascular disease (mean age 31 ± 8 years, range 19 to 54 years) were recruited for participation in this study.
9
10 Finch-Johnston et al
A
B
Figure 1 A, Diagram illustrating the definition of the magnitude of cyclic variation of hypothetical data, and B, the corresponding normalized time delay. ECG, Electrocardiogram.
Data Acquisition Two-dimensional integrated backscatter images of the left ventricle were obtained from standard transthoracic echocardiography views (parasternal long- and short-axis, apical 2- and 4-chamber) with use of a SONOS 1500 (Hewlett-Packard Co, Andover, Mass) ultrasonic imaging system operating in the Acoustic Densitometry (AD) acquisition mode. Use of the AD imaging mode permits the acquisition, display, and subsequent analyses of real-time integrated backscatter images on the basis of a previously described method.1,10,12,25,26 Images were obtained with a 2.5-MHz phased-array probe with the transmit power and time gain compensation values set to optimize image quality for each of the 4 echocardiographic views. Care was taken to ensure that the level of backscattered signal (image brightness) from the midmyocardial regions remained well within the dynamic range of the imaging system over the entire cardiac cycle. Lateral gain compensation was not used in this study. For each echocardiographic view, digital cineloops of 60 consecutive integrated backscatter image frames (approximately 2 cardiac cycles) were acquired at a rate of 30 frames/s and were stored on optical disk for subsequent analysis.
Journal of the American Society of Echocardiography January 2000
Data Analysis Analysis of integrated backscatter images. The integrated backscatter image data, corresponding to each view from each subject, were analyzed by using the AD analysis package available on the SONOS 1500 imaging system. An elliptical region of interest was placed in the mid myocardium of specific myocardial segments in each view. The size of the region of interest used in analyzing each segment was chosen to be the largest size available that would fit within the myocardium, avoiding the endocardial and epicardial boundaries. For each myocardial segment analyzed, the region of interest was carefully positioned to avoid the bright endocardial and epicardial specular reflections and was tracked throughout the entire cineloop (60 frames), keeping the region of interest within the mid myocardium. This was accomplished by adjusting the position of the region of interest in every frame such that approximately the same 2-dimensional area of myocardium was sampled over the heart cycle. Myocardial segments in which the tissue was poorly visualized were not analyzed. For each myocardial segment, the AD analysis package provided a mean integrated backscatter value, representing those values within the region of interest, for each image frame. Thus a data recording that represented the cyclic variation of integrated backscatter values as a function of frame number (time) was generated for each myocardial segment. Because the images were acquired with a frame rate of 30 Hz, the resulting cyclic variation data was sampled in 33-ms time increments. For subsequent analyses of the cyclic variation data, the LV end-diastolic and end-systolic frames were identified as defined by the American Society of Echocardiography.27 End-diastole was defined as the frame at or before mitral valve closure, or the largest cavity area if no valves were visible. End-systole was defined as the frame preceding mitral valve opening, or the smallest cavity area if no valves were visible. The systolic interval, as determined in this study, was the time difference between the end-diastolic and end-systolic frames. Analysis of cyclic variation data. Methods for acquiring and estimating the cyclic variation with real-time integrated backscatter imaging have been previously described.1,10,12,25,26 In characterizing the cyclic variation, we report the magnitude and the corresponding time delay normalized to the duration of the systolic interval.28,29 We define the magnitude of cyclic variation as the difference between the average peak and average nadir values of integrated backscatter as illustrated in Figure 1, A. The corresponding normalized time delay of cyclic variation is expressed as a dimensionless ratio obtained by dividing the time interval from end-diastole to the nadir of the integrated backscatter trace by the systolic interval as shown in Figure 1, B. These parameters were calculated by using a previously reported automated analysis algo-
Journal of the American Society of Echocardiography Volume 13 Number 1
Finch-Johnston et al 11
A
B
Figure 2 Measured values of the magnitude and normalized time delay for specified regions of interest in the parasternal long-axis view (A), and parasternal short-axis view (B). Magnitude values are shown on the left axis and normalized time delay values on the right axis.
rithm.28,29 This algorithm uses cross-correlation and Fourier decomposition of a model function to provide operator-independent estimates of the time delay and magnitude. Implementation of the automated analysis algorithm requires the end-diastolic and end-systolic frame numbers (systolic interval) to be supplied in addition to the corresponding integrated backscatter values as a function of frame number obtained with the AD analysis package (typically over a couple complete cardiac cycles). From the input end-diastolic and end-systolic frame numbers (timing information), a model function representing the anticipated periodicity of the cyclic variation waveform was constructed. This model function was then crosscorrelated with the measured cyclic variation of integrated
backscatter waveform. The time shift applied to the model function that maximized the correlation with the measured cyclic variation waveform was then used to identify the time at which the nadir of the cyclic variation occurred. The time interval from end-diastole to the nadir of the integrated backscatter trace was then used to compute the normalized time delay as described above. Because this cross-correlation technique was often applied over several heart cycles and the average systolic interval was used, the effects of beat-to-beat variations should have been reduced. Both the magnitude and the time delay parameters characterizing the cyclic variation were calculated from the same cardiac cycles. Statistical methods for comparisons. The measured val-
Journal of the American Society of Echocardiography January 2000
12 Finch-Johnston et al
A
B
Figure 3 Measured values of the magnitude and normalized time delay for specified regions of interest in the apical 4-chamber view (A), and the apical 2-chamber view (B). Magnitude values are shown on the left axis and normalized time delay values on the right axis.
ues for the time delay and magnitude of cyclic variation corresponding to specific myocardial segments were compared by using a paired, 2-tailed Student t test with Microsoft Excel (Microsoft Corporation, Redmond, Wash) and the corresponding P value for each of the comparisons was given. Comparisons of the mean values for which P was less than .05 were considered significant.
RESULTS Parasternal Views The locations of regions of interest and the resulting average measurements of time delay and magnitude
of cyclic variation for specific segments visualized in the parasternal views are illustrated in Figure 2.Values for the time delay and magnitude for the 2 regions analyzed in the parasternal long-axis view (midseptal and midposterior wall) are shown in Figure 2, A, and values for the 4 regions analyzed in the parasternal short-axis view (anterior septum, lateral wall, posterior wall, and posterior septum) are depicted in Figure 2, B.The measured values of the time delay for both of the myocardial segments analyzed in the parasternal long-axis view were similar (ie, relatively short). However, the values for the time delay for the segments analyzed in the parasternal short-axis view were relatively longer for the lateral wall and posterior septum compared with the values for the anterior
Journal of the American Society of Echocardiography Volume 13 Number 1
Finch-Johnston et al 13
septum and posterior wall. (P < .001, lateral wall compared with anterior septum; P = .02, posterior septum compared with anterior septum; P < .001, lateral wall compared with posterior wall; P < .001, posterior septum compared with posterior wall.) Apical Views The locations of the regions of interest and the resulting average measurements of time delay and magnitude of cyclic variation for specific myocardial segments visualized in the apical views are illustrated in Figure 3. Values for the 2 regions analyzed in the apical 4-chamber view (midseptal and midlateral wall) are shown in Figure 3, A, and the values for the 2 regions analyzed in the apical 2-chamber view (midinferior wall and midanterior wall) are shown in Figure 3, B.The values of the time delay corresponding to the midlateral wall in the 4-chamber view and the midinferior wall and midanterior wall in the 2chamber view were relatively long compared with the value measured for the mid septum in the 4chamber view. (P was less than .001 for each of the time delay comparisons: midlateral wall compared with the mid septum in the 4-chamber view; midinferior wall in the 2-chamber view compared with the mid septum in the 4-chamber view; midanterior wall in the 2-chamber view compared with the mid septum in the 4-chamber view.) Comparison of Myocardial Segments A comparison of the time delay and magnitude of cyclic variation values for comparable myocardial segments are given in Table 1. This table provides a direct comparison between specific myocardial segments representing roughly the same myocardial region when imaged in different echocardiographic views (ie, approximately the same region and perfused by the same coronary distribution). The results of this comparison indicate that the measured values of the time delay and magnitude for comparable myocardial segments were the same (ie, not significantly different) in all but one of the comparisons. Only the magnitude of cyclic variation from the posterior septal segment in the parasternal short-axis view exhibited a significant difference (ie, P < .05) when compared with the magnitude from the inferior wall when imaged in the apical 2-chamber view.
DISCUSSION Quantitative echocardiography with ultrasonic tissue characterization methods complement conventional 2-dimensional echocardiography by providing
Figure 4 Relationship between the predominant myocardial fiber orientation and the direction of the insonifying ultrasonic beams for the parasternal short-axis view.
an approach for directly evaluating the physical state of the myocardium.1-18,30,31 The underlying hypothesis is that pathologic changes of myocardial structure and function result in alterations in the fundamental physical properties of tissue that can be quantified with indexes based on frequency-dependent ultrasonic attenuation and backscatter.1,8-10,12,31-34 Our laboratory introduced the concepts of magnitude and phase to characterize the cyclic variation of myocardial integrated backscatter and introduced the parameter of phase-weighted magnitude which we11,35 and others14,36 have used successfully. However, we advocate reporting both the magnitude and the time delay of cyclic variation, as exemplified in a study that detected changes in backscatter from patients with insulin-dependent diabetes mellitus17 and identified otherwise latent changes in myocardial structure that accompany the evolution of insulin-dependent diabetes mellitus. Although the cyclic variation is dependent in part on regional myocardial contractile function,35,37 the definitive mechanisms responsible for the cyclic variation of integrated backscatter have not been elucidated. We hypothesize that the anisotropy of both ultrasonic backscatter and attenuation (ie, their dependence upon the angle of the incident ultrasonic beam relative to the direction of the myocardial fibers) plays a significant role.32,38-42 Previous studies have demonstrated that the inherent anisotropy of myocardium can significantly influence quantitative ultrasonic measurements. Specifically, integrated backscatter values were largest when the insonifying ultrasound was perpendicular to the predominant myocardial fiber orientation,38-40,43 whereas the ultrasonic attenuation was the largest when the insonifying ultrasound was parallel to the myofibers.44,45 These inherent anisotropic properties of the heart have significant effects in the
Journal of the American Society of Echocardiography January 2000
14 Finch-Johnston et al
Table 1 Cyclic variation of integrated backscatter for comparable myocardial segments in different views Region of interest
Normalized time delay ± SD
Comparison P value (delay)
Magnitude ± SD (dB)
Comparison P value (magnitude)
SAX anterior septum LAX septum SAX anterior septum A4C septum LAX septum A4C septum SAX posterior wall LAX posterior wall SAX lateral wall A4C lateral wall SAX lateral wall A2C anterior wall A4C lateral wall A2C anterior wall SAX post septum A2C inferior wall
1.03 ± 0.16 1.08 ± 0.17 1.03 ± 0.16 1.08 ± 0.31 1.08 ± 0.17 1.08 ± 0.31 1.03 ± 0.14 1.00 ± 0.14 2.22 ± 0.71 2.20 ± 0.79 2.22 ± 0.71 2.04 ± 0.72 2.20 ± 0.79 2.04 ± 0.72 1.65 ± 0.66 1.68 ± 0.62
.62
4.3 ± 2.3 4.4 ± 1.4 4.3 ± 2.3 4.7 ± 2.0 4.4 ± 1.4 4.7 ± 2.0 6.0 ± 1.5 6.1 ± 2.0 2.5 ± 1.2 2.8 ± 1.3 2.5 ± 1.2 2.3 ± 0.8 2.8 ± 1.3 2.3 ± 0.8 2.5 ± 1.0 3.7 ± 1.8
.80
.62 .89 .27 .18 .11 .49 .80
.26 .51 .78 .67 .73 .08 .01
SAX, Parasternal short-axis; LAX, parasternal long-axis; A4C, apical 4-chamber; A2C, apical 2-chamber.
echocardiographic images that are obtained from standard views.46-48 Hence, because the angle of insonification between an incident ultrasonic beam and the local myofiber orientation for a specific region of the heart is echocardiographic view– dependent, we postulated that the time delay of cyclic variation would be view-dependent as well. Measurements of the cyclic variation of myocardial segments obtained from the parasternal longaxis view, with the insonifying beam approximately perpendicular to myofibers in all segments, exhibit minimal anisotropic effects.* However, in other standard transthoracic echocardiography views, the angle between the insonifying beam and the predominant myofibers is dependent on the myocardial segment of interest. In the parasternal short-axis imaging plane, the midmyocardial fibers exhibit primarily a circumferential arrangement,52,53 as indicated in Figure 4. Thus the midmyocardial fibers are predominantly perpendicular to the insonifying ultrasonic beams in the anterior septal and posterior wall regions and predominantly parallel in the posterior septal and lateral wall regions.The relationship between the orientation of the myocardial fibers and the insonifying ultrasonic beam is more complex in the apical views, especially near the apex of the heart. The echocardiographic view–dependence of the measured magnitude of cyclic variation of integrated backscatter for specific myocardial segments has been reported previously.20-24
*1,3,4,10,12,15,36,49-51
In the present study, measurements of the time delay and magnitude of cyclic variation for the parasternal views demonstrate a strong view-dependence, as indicated in Figure 2. As described above, in the mid myocardium surrounding the left ventricle at the midpapillary level, a significant fraction of the myocardial fibers run circumferentially in the transverse plane, with the largest deviations out of the transverse plane occurring near the endocardial and epicardial surfaces.39,42,52,54-56 In spite of these transmural changes in fiber direction, the ultrasonic beam is primarily perpendicular to the predominant fiber direction in the septum and posterior wall within the imaging plane of the parasternal long-axis view.48 Myocardial segments in which the insonifying ultrasonic beam was primarily perpendicular to the predominant myocardial fibers yielded time delays close to unity with corresponding magnitudes of cyclic variation consistent with those previously reported in healthy subjects (4 to 6 dB).1,30 Myocardial segments imaged in those regions of the shortaxis view in which the ultrasonic beam propagated more nearly parallel to the predominant myocardial fibers (lateral wall and posterior septum) were characterized as having substantially longer values of time delays and somewhat smaller magnitudes of cyclic variation. A comparison of the measured values of cyclic variation for myocardial segments visualized in both parasternal echocardiographic views (anterior septum and posterior wall) show that they are quite comparable (Table 1). For these myocardial segments the insonifying beam is approximately perpendicular to the local myofiber direction in both views. Values of the magnitude of cyclic variation of inte-
Journal of the American Society of Echocardiography Volume 13 Number 1
grated backscatter obtained from the long-axis view (septum and posterior wall) in this study are consistent with those of previous studies.1,49 The values obtained for the normalized time delay, however, are slightly larger than those previously reported (mid septum: 1.08 ± 0.17 in this study compared with a value of 0.84 ± 0.09 reported previously; midposterior wall: 1.00 ± 0.14 in this study compared with a value of 0.85 ± 0.12 reported previously).57 This difference may be the result, in part, of technical limitations in the rate of sampling of integrated backscatter values used in this study, compared with previous M-mode–based studies. In this study, a value for the mean integrated backscatter within a specific region of interest was determined once per frame (ie, once every 33 ms [for a 30 frame/s image display rate]. In previous M-mode–based studies, the effective sampling rate was much larger with sampling periods of around 5 ms. Hence, in the present study, there could be an approximate 10% to 20% uncertainty in the measurement of the systolic interval, depending on the timing of image frame acquisition relative to the true systolic interval. Furthermore, another contributing factor is that in previous studies, estimates of the normalized time delay were measured as the ratio of the time from the R wave to the integrated backscatter nadir, which was normalized by the QTinterval on the basis of information from the ECG tracing, whereas in our study the systolic interval was more accurately based on mechanical events guided from the 2-dimensional images. This latter approach results in both timing intervals (ie, the start of systole to integrated backscatter nadir and the systolic interval) being measured from the same starting time (end-diastolic frame), which tends to give a larger ratio and therefore a larger time delay. Measurements of the time delay and magnitude of myocardial cyclic variation in the apical 4- and 2chamber views, in which the orientation of the myocardial fibers with respect to the insonifying ultrasonic beam is more complex, also show a dependence on the myocardial segment analyzed as indicated in Figure 3. A comparison of the cyclic variation values for the lateral wall, imaged in the apical 4-chamber view, with the corresponding values for the lateral wall, imaged in the parasternal shortaxis view, reveals comparable values for time delay and magnitude (Table 1). Measurements of the cyclic variation in this region from these two views exhibited relatively long time delays with relatively small magnitudes. The model of fiber orientation would suggest that the ultrasonic beam was relatively parallel to the predominant fiber direction in the mid myocardium of the lateral wall when imaged in the
Finch-Johnston et al 15
parasternal short-axis view (Figure 4) and was perpendicular to the predominant fiber direction in the mid myocardium of the lateral wall of the apical 4chamber view. However, these results for the lateral wall indicate that the cyclic variation of the lateral wall of the 4-chamber view exhibits values more consistent with parallel insonication.We hypothesize that this may be a result of the relatively complex fiber orientation in the apical views and, perhaps, may be related to the transmural change in fiber orientation in the lateral wall, resulting in fibers (or fiber components) within the image plane of the apical views. Quite similar values of time delay and magnitude were found for the anterior wall of the apical 2-chamber view, the lateral wall of the parasternal short-axis view, and the lateral wall of the apical 4chamber view (Table 1). A comparison of the cyclic variation obtained from the septum of the apical 4-chamber view with that from the anterior septum of the parasternal short-axis view and the septum of the parasternal long-axis view revealed quite similar values for time delay and magnitude in each case (Table 1).The values from the septum of the apical 4-chamber view exhibited a relatively short time delay with a corresponding large magnitude of cyclic variation.These results are consistent with those obtained for perpendicular insonication with respect to the myofibers, as in the case for the corresponding segments in the parasternal views. Hence, these results, along with the observation that the septum is often much more easily visualized than the lateral wall in the apical 4-chamber view in a clinical echocardiographic exam, suggest that significant differences may exist in fiber orientation in the septum when compared with the fiber architecture of the lateral wall as imaged in the apical 4-chamber view. Summary Results of this study indicate that myocardial ultrasonic characterization that uses the time delay of cyclic variation is influenced by the echocardiographic view employed and the specific segment of the left ventricle analyzed. This is consistent with measurements of the view dependence of the magnitude of cyclic variation reported by others.20-24 In regions of myocardium where the angle of insonication relative to local myofiber orientation is nearly parallel, relatively long time delays and small magnitudes are to be expected in comparison with the corresponding results for regions with perpendicular insonication. Furthermore, measurements of cyclic variation may be affected not only by the complex
Journal of the American Society of Echocardiography January 2000
16 Finch-Johnston et al
fiber geometry but also by the effects of translation, rotation, and twist of the heart during cardiac systole. Comparison of the time delay and magnitude of cyclic variation values for comparable myocardial segments indicate that the measured values were the same in all but one of the comparisons. Although the healthy subjects we studied represent a relatively young population, this systematic mapping of the time delay (and magnitude) of the cyclic variation of integrated backscatter provides a first step toward the development of a basis for delineating the measurements obtained from patients with altered LV segmental function. In addition, it establishes a framework for the interpretation of alterations in acoustic properties (eg,those resulting from ischemia, infarction, fibrosis, the passage of microbubbles in contrast echocardiography) of specific myocardial segments visualized from any of the 4 standard transthoracic windows. We would like to thank Kirk D. Wallace and Steven L. Baldwin for their contributions toward data analysis and Barbara L. Roman and Debbie Taylor for their assistance in manuscript preparation.
9.
10.
11.
12.
13.
14.
15.
16.
REFERENCES 1. Vered Z, Barzilai B, Mohr GA, et al. Quantitative ultrasonic tissue characterization with real-time integrated backscatter imaging in normal human subjects and in patients with dilated cardiomyopathy. Circulation 1987;76:1067-73. 2. Goens MB, Karr SS, Martin GR. Cyclic variation of integrated ultrasound backscatter: normal and abnormal myocardial patterns in children. J Am Soc Echocardiogr 1996;9:616-21. 3. Naito J, Masuyama T, Tanouchi J, et al. Analysis of transmural trend of myocardial integrated ultrasound backscatter for differentiation of hypertrophic cardiomyopathy and ventricular hypertrophy due to hypertension. J Am Coll Cardiol 1994;24:517-24. 4. Naito J, Masuyama T, Mano T, et al. Dobutamine stress ultrasonic myocardial tissue characterization in patients with dilated cardiomyopathy. J Am Soc Echocardiogr 1996;9:4709. 5. Dibello V, Pedrinelli R, Giorgi D, et al. Ultrasonic videodensitometric analysis of two different models of left ventricular hypertrophy—athletes heart and hypertension. Hypertension 1997;29:937-44. 6. Zoni A, Regolisti G, Aschieri D, Borghetti A. Myocardial ultrasonic tissue characterization in patients with different types of left ventricular hypertrophy—a videodensitometric approach. J Am Soc Echocardiogr 1997;10:74-82. 7. Ciliberto GR, Pingitore A, Mangiavacchi M, Alberti A, Paterni M, Picano E. The clinical value of blunting of cyclic gray level variation for the detection of acute cardiac rejection: a two-dimensional, Doppler, and videodensitometric ultrasound study. J Am Soc Echocardiogr 1996;9:306-13. 8. Barzilai B, Madaras EI, Sobel BE, Miller JG, Perez JE. Effects of myocardial contraction on ultrasonic backscatter before
17.
18.
19.
20.
21.
22.
23.
24.
and after ischemia. Am J Physiol 1984;247(3 Pt 2):H478H483. Fitzgerald PJ, McDaniel MD, Rolett EL, Strohbehn JW, James DH. Two-dimensional ultrasonic tissue characterization: backscatter power, endocardial wall motion, and their phase relationship for normal, ischemic, and infarcted myocardium. Circulation 1987;76:850-9. Vered Z, Mohr GA, Barzilai B, et al. Ultrasound integrated backscatter tissue characterization of remote myocardial infarction in human subjects. J Am Coll Cardiol 1989;13:8491. Wickline SA, Thomas LJ III, Miller JG, Sobel BE, Perez JE. Sensitive detection of the effects of reperfusion on myocardium by ultrasonic tissue characterization with integrated backscatter. Circulation 1986;74:389-400. Milunski MR, Mohr GA, Perez JE, et al. Ultrasonic tissue characterization with integrated backscatter: acute myocardial ischemia, reperfusion, and stunned myocardium in patients. Circulation 1989;80:491-503. Saeian K, Rhyne TL, Sagar KB. Ultrasonic tissue characterization for diagnosis of acute myocardial infarction in the coronary care unit. Am J Cardiol 1994;74:1211-5. Takiuchi S, Ito H, Iwakura K, et al. Ultrasonic tissue characterization predicts myocardial viability in early stage of reperfused acute myocardial infarction. Circulation 1998;97:35662. Perez JE, McGill JB, Santiago JV, et al. Abnormal myocardial acoustic properties in diabetic patients and their correlation with the severity of disease. J Am Coll Cardiol 1992;19: 1154-6. DiBello V, Talarico L, Picano E, et al. Increased echodensity of myocardial wall in the diabetic heart: an ultrasound tissue characterization study. J Am Coll Cardiol 1995;25:1408-15. Wagner RF, Wear KA, Perez JE, McGill JB, Schechtman KB, Miller JG. Quantitative assessment of myocardial ultrasound tissue characterization through receiver operating characteristic analysis of Bayesian classifiers. J Am Coll Cardiol 1995;25: 1706-11. Dibello V, Giampietro O, Matteucci E, et al. Ultrasonic videodensitometric analysis in type 1 diabetic myocardium. Coron Artery Dis 1996;7:895-901. Finch-Johnston AE, Gussak HM, Mobley J, et al. Dependence of “apparent” magnitude on the time delay of cyclic variation of myocardial backscatter. Ultrasound Med Biol 1999;25:759-62. Bouki KP, Lange A, Palka P, et al. Regional variations of ultrasonic integrated backscatter in normal and myopathic left ventricles—a new multi-view approach. Eur Heart J 1996;17:1747-55. Lange A, Moran CM, Palka P, Fenn LN, Sutherland GR, McDicken WN. The variation of integrated backscatter in human hearts in differing ultrasonic transthoracic views. J Am Soc Echocardiogr 1995;8:830-8. Dittrich HC, Wheeler K, Kwan OL, et al. Quantitation of integrated myocardial backscatter from two-dimensional echo images: influence of beam orientation upon cyclic variation. J Am Soc Echocardiogr 1993;3:S30. Vandenberg BF, Rath L, Shoup TA, Kerber RE, Collins SM, Skorton DJ. Cyclic variation of ultrasound backscatter in normal myocardium is view dependent: clinical studies with a real-time backscatter imaging system. J Am Soc Echocardiogr 1989;2:308-14. Vandenberg BF, Stuhlmuller JE, Rath L, et al. Diagnosis of recent myocardial infarction with quantitative backscatter
Journal of the American Society of Echocardiography Volume 13 Number 1
25.
26.
27.
28.
29.
30.
31. 32.
33.
34.
35.
36.
37.
38.
39.
40.
imaging: preliminary studies. J Am Soc Echocardiogr 1991; 4:10-8. Thomas LJ, Wickline SA, Pérez JE, Sobel BE, Miller JG. A real-time integrated backscatter measurement system for quantitative tissue characterization. IEEE Trans Ultrason Ferroelect Freq Control UFFC 1986;33:27-32. Barzilai B, Thomas LJ III, Glueck RM, et al. Detection of remote myocardial infarction with quantitative real-time ultrasonic characterization. J Am Soc Echocardiogr 1988;1: 179-86. Schiller NB, Shah PM, Crawford M, et al. Recommendations for quantitation of the left ventricle by two-dimensional echocardiography. J Am Soc Echocardiogr 1989;2:358-68. Mohr GA, Vered Z, Barzilai B, Perez JE, Sobel BE, Miller JG. Automated determination of the magnitude and time delay (“phase”) of the cardiac cycle dependent variation of myocardial ultrasonic integrated backscatter. Ultrason Imaging 1989;11:245-59. Mobley J, Banta CE, Gussak HM, Perez JE, Miller JG. Clinical tissue characterization: online determination of magnitude and time delay of myocardial backscatter. Video Journal of Echocardiography 1995;5:40-8. Perez JE, Holland MR, Barzilai B, et al. Ultrasonic characterization of cardiovascular tissue. In: Cardiac imaging: a companion to Braunwald’s heart disease. Skorton DJ, Schelbert HR, Wolf GL, Brundage BH, Skorton DJ, Schelbert HR, Wolf GL, Brundages BH, editors. 2nd ed. Philadelphia: WB Saunders Co; 1996. p. 606-22. Miller JG, Perez JE, Sobel BE. Ultrasonic characterization of myocardium. Prog Cardiovasc Dis 1985;XXVIII:85-110. O’Brien WD Jr, Sagar KB, Warltier DC, Rhyne TL. Acoustic propagation properties of normal, stunned, and infarcted myocardium: morphological and biochemical determinants. Circulation 1995;91:154-60. Rijsterborgh H, Mastik F, Lancee CT, et al. Ultrasonic myocardial integrated backscatter and myocardial wall thickness in animal experiments. Ultrasound Med Biol 1990;16: 29-36. Mimbs JW, O’Donnell M, Bauwens D, Miller JG, Sobel BE. The dependence of ultrasonic attenuation and backscatter on collagen content in dog and rabbit hearts. Circ Res 1980;47: 49-58. Wickline SA, Thomas LJ III, Miller JG, Sobel BE, Perez JE. The dependence of myocardial ultrasonic integrated backscatter on contractile performance. Circulation 1985;72:18392. Vitale DF, Bonow RO, Gerundo G, et al. Alterations in ultrasonic backscatter during exercise-induced myocardial ischemia in humans. Circulation 1995;92:1452-7. Wickline SA, Thomas LJ III, Miller JG, Sobel BE, Perez JE. A relationship between ultrasonic integrated backscatter and myocardial contractile function. J Clin Invest 1985;76:215160. Madaras EI, Perez J, Sobel BE, Mottley JG, Miller JG. Anisotropy of the ultrasonic backscatter of myocardial tissue: II. Measurements in vivo. J Acoust Soc Am 1988;83:762-9. Verdonk ED, Wickline SA, Miller JG. Anisotropy of ultrasonic velocity and elastic properties in normal human myocardium. J Acoust Soc Am 1992;92:3039-50. Hoffmeister BK, Wong AK, Verdonk ED, Wickline SA, Miller JG. Comparison of the anisotropy of apparent integrated ultrasonic backscatter from fixed human tendon and fixed human myocardium. J Acoust Soc Am 1995;97:1307-13.
Finch-Johnston et al 17
41. O’Brien PD, O’Brien WD Jr, Rhyne TL, Warltier DC, Sagar KB. Hypertrophy/heart failure/ventricular dysfunction: relation of ultrasonic backscatter and acoustic propagation properties to myofibrillar length and myocardial thickness. Circulation 1995;91:171-5. 42. Verdonk ED, Hoffmeister BK, Wickline SA, Miller JG. Anisotropy of the slope of ultrasonic attenuation in formalin fixed human myocardium. J Acoust Soc Am 1996;99:3837-43. 43. Mottley JG, Miller JG. Anisotropy of the ultrasonic backscatter of myocardial tissue: I. Theory and measurements in vitro. J Acoust Soc Am 1988;83:755-61. 44. Brandenburger GH, Klepper JR, Miller JG, Snyder DL. Effects of anisotropy in the ultrasonic attenuation of tissue on computed tomography. Ultrason Imaging 1981;3:113-43. 45. Mottley JG, Miller JG. Anisotropy of the ultrasonic attenuation in soft tissues: measurements in vitro. J Acoust Soc Am 1990;88:1203-10. 46. Holland MR, Wilkenshoff UM, Finch-Johnston AE, Handley SM, Perez JE, Miller JG. Effects of myocardial fiber orientation in echocardiography: quantitative measurements and computer simulation of the regional dependence of backscattered ultrasound in the parasternal short-axis view. J Am Soc Echocardiogr 1998;11:929-37. 47. Recchia D, Miller JG, Wickline SA. Quantification of ultrasonic anisotropy in normal myocardium with lateral gain compensation of two-dimensional integrated backscatter images. Ultrasound Med Biol 1993;19:497-505. 48. Aygen M, Popp RL. Influence of the orientation of myocardial fibers on echocardiographic images. Am J Cardiol 1987; 60:147-52. 49. Stuhlmuller JE, Skorton DJ, Burns TL, Melton HEJ, Vandenberg BF. Reproducibility of quantitative backscatter echocardiographic imaging in normal subjects. Am J Cardiol 1992;69:542-6. 50. Vitale DF, Bonow RO, Calabro R, et al. Myocardial ultrasonic tissue characterization in pediatric and adult patients with hypertrophic cardiomyopathy. Circulation 1996;94: 2826-30. 51. Feinberg MS, Gussak HM, Davila-Roman VG, Baumann CM, Miller JG, Perez JE. Dissociation between wall thickening of normal myocardium and cyclic variation of backscatter during inotropic stimulation. Am J Cardiol 1996;77:515-20. 52. Streeter DD Jr, Spotnitz HM, Patel DP, Ross J Jr, EH Sonnenblick. Fiber orientation in the canine left ventricle during diastole and systole. Circ Res 1969;XXIV:339-47. 53. Greenbaum RA, Ho SY, Gibson DG, Becker AE, Anderson RH. Left ventricular fibre architecture in man. Br Heart J 1981;45:248-63. 54. Streeter DD Jr, Hanna WT. Engineering mechanics for successive states in canine left ventricular myocardium. II. Fiber angle and sarcomere length. Circ Res 1973;33:656-64. 55. Recchia D, Hall C, Shepard RK, Miller JG. Mechanisms of the view-dependence of ultrasonic backscatter from normal myocardium. IEEE Trans Ultrason Ferroelec Freq Contr 1995;42:91-8. 56. Wickline SA, Verdonk ED, Miller JG. Three-dimensional characterization of human ventricular myofiber architecture by ultrasonic backscatter. J Clin Invest 1991;88:438-46. 57. Loomis JF Jr, Waggoner AD, Schechtman KB, Miller JG, Sobel BE, Perez JE. Ultrasonic integrated backscatter 2-d imaging: evaluation of M-mode guided acquisition and immediate analysis in 55 consecutive patients. J Am Soc Echocardiogr 1990;3:255-65.