Left ventricular ejection fraction in children measured by three-dimensional echocardiography using a new transthoracic integrated 3D-probe

Transcription

1 European Heart Journal (1998) 19, Article No. hj Left ventricular ejection fraction in children measured by three-dimensional echocardiography using a new transthoracic integrated 3D-probe A comparison with equilibrium radionuclide angiography P. Acar*, C. Maunoury, T. Antonietti, D. Bonnet*, D. Sidi* and J. Kachaner* Services de *Cardiologie Pédiatrique et de Médecine Nucléaire, Hôpital Necker/Enfants-malades, Paris, France Background Three-dimensional echocardiography allows calculation of left ventricular ejection fraction without geometric assumption on the ventricular shape. Our aim was to validate this technique in a paediatric population with distorted ventricles. Methods Twenty-one patients aged 6 months to 17 years underwent equilibrium radionuclide angiography and three-dimensional echocardiography. Fourteen patients had dilated cardiomyopathy and seven had univentricular hearts. A new, easy to handle, transthoracic rotational probe was used and motion artefacts were limited during the rotation (3 intervals with ECG and respiratory gating). Left ventricular volumes and ejection fraction were calculated using the Simpson s rule with 12 slices. Results Three-dimensional echocardiography correlated well with equilibrium radionuclide angiography for ejection fraction measurement (r=0 90; the mean difference between the two methods being 3 8 6%). Intra-observer and interobserver variabilities for 3D echocardiography were 2 4% and 4 5%. Conclusions Three-dimensional echocardiography is an accurate, non-invasive, and reproducible methods to measure left ventricular ejection fraction in children. (Eur Heart J 1998; 19: ) Key Words: Three-dimensional echocardiography, left ventricular ejection fraction, children. Introduction Revision submitted 23 March 1998, and accepted 30 March Correspondence: Dr Philippe Acar, Service de Cardiologie Pédiatrique, Hôpital Necker Enfants-malades 149, rue de Sèvres Paris cedex 15, France X/98/ $18.00/0 Left ventricular ejection fraction is an important index in children with cardiomyopathy or complex congenital heart disease to assess cardiac function. It is of great help in both the management and follow-up of these patients [1,2]. Equilibrium radionuclide angiography is an accepted and validated method for left ventricular ejection fraction measurement [3]. However, a peripheral venous line is required and it exposes the patient to radiation. Two-dimensional echocardiography may assess left ventricular volumes and function but implies geometric assumptions which are not applicable to distorted left ventricular cavities such as dyskinetic dilated cardiomyopathy or univentricular hearts. Three-dimensional (3D) echocardiography allows the reconstruction of the true geometry of the left ventricle [4,5], allowing the quantification of left ventricular volumes and ejection fraction without geometric assumptions [6]. Accurate estimates of left ventricular ejection fraction by 3D echocardiography in adults have been reported [7]. The aim of the present study was to determine the accuracy of 3D echocardiography in measuring the left ventricular ejection fraction in children in comparison with equilibrium radionuclide angiography using a novel transthoracic integrated 3D-probe adapted to small children. Study population Twenty-one patients with abnormal left ventricles were studied. Their age ranged from 6 months to 17 years (mean: 9 5 years) and their weight from 5 to 66 kg (mean: kg). Fourteen patients had a dilated cardiomyopathy and seven had a univentricular heart palliated by a Fontan-type operation. In each patient, the equilibrium radionuclide angiographic and 3D echocardiographic studies were performed on the same day for left ventricular ejection fraction calculation The European Society of Cardiology

2 1584 P. Acar et al. image the left ventricle and then rotated from 0 to 180 to encompass the entire left ventricular capacity. The gain was adjusted to optimize the endocardial definition. The 2D echocardiographic system was connected to a 3D system (Echo-scan TomTec GmbH). The steps to measure left ventricular ejection fraction were the following [9]. Figure 1 The transthoracic integrated three-dimensional probe. The step motor is integrated to the probe and thus avoids the use of any heavy external motor. The size of the probe ( mm) allows for ease of handling with reduced motion artefacts during the rotation. Methods Equilibrium radionuclide angiography Twenty minutes after an intravenous injection of stannous pyrophosphate, 370 to 1110 Mbq (10 30 mci) of Tc-99m were injected to label the patient s red blood cells. Equilibrium radionuclide angiography was performed 10 min later on a single head camera equipped with a low energy all-purpose parallel hole collimator. The patients were imaged in the left anterior oblique ( best septal ) and the left lateral position. Three standard ECG electrodes were placed in the right and left supraventricular areas close to the mid-clavicular lines and at the junction of the left lateral chest and the abdomen. The acquisition was gated to the best R wave. The average R-R interval was divided into 16 intervals. The matrix size format was The stop conditions were 5 million counts per view, i.e. an average imaging time of about 10 min per view. The data were processed for left ventricular ejection fraction calculation following a standard processing protocol, but not to calculate left ventricular volumes [3]. Echocardiographic study Instrumentation We used a newly developed transthoracic rotational probe (TomTec Integrated TTE 3E-Probe, Munich, Germany) [8]. It has a 3 5 MHz frequency and a 64- channel phased array (Fig. 1). The step motor was integrated to the probe which avoids the use of a heavy external motor. The size of the probe (28 mm 160 mm) allowed for ease of handling and reduced motion artefacts during the rotation. The transducer rotation axis was kept exactly in the image plane centre. The transducer was connected for 2D images to an echocardiographic system (Toshiba Sonolayer SSH-140A). The probe was placed in a sub-costal or apical position to Image acquisition The integrated 3D-probe rotation was controlled by a steering logic for image acquisition. Electrocardiogram and respiratory gating by thoracic-impedance measurement ensured optimal spatial and temporal registration of the cardiac images. R-R intervals were predetermined with a variability of 100 ms or less. Respiration was gated at the end-expiratory phases in older children (>5 years). Because of the very short respiratory phase and the high heart rate (>120 beats. min 1 ) which allowed quick acquisition, respiration was not gated in younger children. Sixty sequential cross sections of the left ventricle acquired every 3 were obtained, each during a complete cardiac cycle and stored in the computer memory. An average of two acquisitions per patient was performed and the one with minimal motion artefact was chosen for 3D measurements. Image processing The recorded images were formatted in their correct rotational sequence according to their ECG phase in volumetric data sets. Post-processing of the data sets was performed off line, using the system s analysis program (Echo-view, TomTec GmbH). To fill the gaps in the far fields, a trilinear cylindrical interpolation algorithm was used. Data analysis From the conical data volume, a long-axis of the left ventricle was selected. The end-diastolic (first frame before mitral valve closing) and the end-systolic (first frame before mitral opening) frames were selected Fig. 2). The left ventricle was automatically sliced into 12 equidistant parallel short axis views. The surface area of each cross section was measured by manual endocardial tracing and the volume of each slice calculated (Figs 3 and 4). The entire volume of the left ventricular capacity was automatically derived by summing the volumes of each slice according to Simpson s method. Left ventricular ejection fraction was calculated using the formula: LVEF=EDV ESV/EDV (EDV and ESV: end-diastolic and end-systolic volumes). Statistical analysis The data are presented as mean values SD. Interobserver and intra-observer variabilities for left ventricular ejection fraction measurements by 3D echocardiography were determined in all patients. Variability was expressed as a percent error, and determined as the difference between the two measurements

3 Left ventricular ejection fraction in children 1585 Figure 2 Three-dimensional view of the left ventricle. Anterior view showing the left ventricular cavity in diastole (left) and systole (right). Figure 3 The principle of left ventricular volume measurement. The long axis view of the left ventricle is used as reference. The left ventricle is sliced at equidistant intervals to generate a series of short-axis views. The surface area of each cross-section is measured by planimetry and the volume of each slice calculated. The volume of the entire left ventricle is obtained by summing the volumes of all slices according to the Simpson s rule. Ao: aortic root; LA: left atrium. divided by the mean value of the two measurements. Intra-observer variability was obtained by the same observer after an interval of more than 1 week. The left ventricular ejection fraction measured by 3D echocardiography was compared to the left ventricular ejection fraction measured by equilibrium radionuclide angiography using linear regression analysis. The difference between the two methods was reported using the Bland and Altman graph [10]. Results Three-dimensional echocardiography could be performed in all patients with adequate image quality for left ventricular ejection fraction calculations. The mean acquisition time was 85 s (ranging from 50 to 200 s depending on the patient s heart rate and respiratory variability) and no sedation was required even in small infants. Storage of data in the 3D system required

4 1586 P. Acar et al. Figure 4 Three-dimensional reconstruction of the left ventricle. On the left, the left ventricle is viewed from an antero-posterior standpoint and the different slices used for the volume calculation are well defined. On the right, the ventricular cavity is viewed from above, the base at the surface and the apex in depth. LVEF by 3DE 70% y = 0.8x + 10 r = 0.90 P < n = LVEF by ERNA 3 min. Off-line left ventricular ejection fraction calculation required an average of 30 min for each patient. 3D echocardiographic data 70% Figure 5 Three-dimensional echocardiography vs radionuclide angiography. Linear regression of the left ventricular ejection fraction (LVEF) measured by threedimensional echocardiography (3DE) vs equilibrium radionuclide angiography (ERNA). The end-diastolic volume, end-systolic volume and left ventricular ejection fraction measured by 3D echocardiography with 12 slices are presented in Figs 5 and 6. There was no significant difference in the measurement obtained by the same observer in two different settings (intra-observer variability of 2 4 5%). There was no significant difference in the measurement of left ventricular ejection fraction obtained by the two independent observers (inter-observer variability of 4 5 6%). 25% mean + 2DS 0 5 mean 10 mean 2DS % Mean LVEF by ERNA and 3DE Figure 6 Bland and Altman graph. This graph shows the differences of each pair of measurements of the left ventricular ejection fraction (LVEF) by equilibrium radionuclide angiography (ERNA) and three-dimensional echocardiography (3DE). LVEF by ERNA LVEF by 3DE Comparison with equilibrium radionuclide angiography The values of left ventricular ejection fraction obtained by equilibrium radionuclide angiography and 3D echocardiography were compared and we found a good correlation in the left ventricular ejection fraction measurement by the two methods as shown in Fig. 5 (y=0 8x+10, r=0 90). The mean difference between the average values of left ventricular ejection fraction obtained by equilibrium radionuclide angiography and 3D echocardiography was 3 8 6% (Fig. 6). Discussion Ventricular volume measurement in children The standard method for evaluating the left ventricular volume and function has been cineangiography, an

5 Left ventricular ejection fraction in children 1587 invasive technique using ventricular outlines traced from cineangiograms [11]. Cardiac ultrasounds became an alternative to X rays in the 1970s using the M-mode approach. However, using one single plane of the left ventricle to estimate its volume is jeopardized by anomalies such as segmental dysfunction or paradoxical septal motion, frequently seen in patients with right ventricular septal anomalies as in the postoperative course or with cardiomyopathy with right ventricular pressure or volume load [12]. Furthermore, twodimensional echocardiography was said to be more accurate in measuring left ventricular volumes in children since it was able to multiply the ventricular views [13,14]. Whatever the algorithms used to calculate the left ventricular volumes, 2D echocardiography requires geometric assumptions on the left ventricular shape: the left ventricular cavity has to be assumed as a cone (Simpson s formula) or an ellipse (ellipsoid model) [15,16]. Furthermore, distortions of the left ventricular cavity are frequent in children with heart diseases either due to a congenital malformation such as a univentricular heart or to a dilated cardiomyopathy with its volume overload. The third dimension is required in these situations to assess the left ventricular volumes and ejection fraction. Magnetic resonance imaging is a satisfactory answer to this problem since it acquires contiguous, parallel slices with a high temporal resolution avoiding assumptions of ventricular shape [17]. However, this technique is cumbersome in young non-compliant patients. Three-dimensional echocardiography Three-dimensional echocardiography allows the calculation of left ventricular volumes and ejection fraction without any geometric assumption because it contains the entire left ventricular cavity and its accuracy has been well established in adults in comparison to other conventional methods including radionuclide angiography [18,19]. The left ventricular volumes are calculated from a series of real-time, parasternal, short-axis images acquired with a line of intersection display as a guide [20]. This line indicates the location of the short axis images obtained for volume computation. The left ventricular volume is computed from the traced endocardial boundaries of each short-axis section using a polyhedral surface reconstruction algorithm [21,22]. However, the need for suspending the ventilation during the image acquisition contra-indicates the study in children. We used another 3D echocardiographic system based on cross-sectional images acquired under respiratory and electrocardiography monitoring [23]. The short time needed for acquisition of the images (average of 80 s) is well-adapted to small infants. The main limitation had been the rotation of the probe which had to be mounted in a heavy, cylindrical, motorized transducer holder. The mechanical rotation of the probe was necessarily accompanied with motion making the last image (180 ) different from the first one (0 ). Therefore, spatial interpolation created artefacts in the volumetric data sets and the resolution of the 3D reconstructions was poor. The transoesophageal approach optimizes 3D acquisition [24]. The probe is stable in the oesophagus and the motion artefacts are limited during the rotation. This approach is not practical in small children and small multiplane transducer assemblies are not yet available. We have tested the new transthoracic probe with an integrated motor. The size of the probe is very similar to a transthoracic probe; it allows for ease of handling and erases most of the motion artefacts during the rotation. The 3D reconstructions of the left ventricle had a good quality allowing us to measure the left ventricular cavity volume in all patients, including those with funny-shaped cavities. Off-line measurements and limitations The 3 5 MHz integrated 3D-probe greatly improved the 3D acquisition which could be performed by a single experimenter. Further, miniaturizing the transducer and increasing the frequency of the probe enhanced the quality of the images. If the acquisition time was facilitated, the volume measurement of the left ventricular cavity had to be post-processed and analysed. Choosing the right end-diastolic and end-systolic frames together with the manual endocardial tracing of each short axis slice are time-consuming and limit the routine use of 3D echocardiography. The development of a software able to recognize automatically the borders for area and volume analysis techniques would shorten the time needed for volume calculation. Equilibrium radionuclide angiography is a reliable method to assess left ventricular ejection fraction but not to estimate left ventricular volumes. Validation of 3D echocardiography for volume measurement is still needed. Three-dimensional echocardiography is a noninvasive method which only requires a good acoustic window, very common in a paediatric population. The small inter- and intra-observer variability makes the technique reproducible. Conclusions Three-dimensional echocardiography appears to be an accurate method to measure the left ventricular ejection fraction of various acquired or congenital heart diseases in children. It is non-invasive, reproducible and may be an improvement on equilibrium radionuclide angiography in calculating left ventricular ejection fraction in distorted ventricles, which is frequently the case in congenital heart disease. Validation of volume measurement by 3D is still requested. References [1] Matitiau A, Perez-Atayde A, Sanders SP et al. Infantile dilated cardiomyopathy. Circulation 1994; 90:

6 1588 P. Acar et al. [2] Mayer JE, Bridges ND, Lock JE, Hanley FL, Jonas RA, Castaneda AR. Factors associated with marked reduction in mortality for Fontan operations in patients with single ventricle. J Thorac Cardiovasc Surg 1992; 103: [3] Hains AD, Khawaja IA, Lahiri A, Raftery FB. Radionuclide left ventricular ejection fraction: a comparison of the three methods. Br Heart J 1987; 57: [4] Roelandt J, Salustri A, Vletter W, Nosir Y, Bruining N. Precordial multiplane echocardiography for dynamic anyplane, paraplane and three-dimensional imaging of the heart. Thoraxcenter 1994; 6: [5] Salustri A, Roelandt JR. Ultrasonic three-dimensional reconstruction of the heart. Ultrasound Med Biol 1995; 21: [6] Siu SC, Rivera JM, Guerrero JL et al. Three-dimensional echocardiography: in vivo validation for left ventricular volume and function. Circulation 1993; 88: [7] Nosir Y, Fioretti P, Vletter W et al. Accurate measurement of left ventricular ejection fraction by three-dimensional echocardiography. Circulation 1995; 94: [8] Pini R, Giannazzo G, Di Bari M et al. Transthoracic three-dimensional echocardiographic reconstruction of left and right ventricles: in vitro validation and comparison with magnetic resonance imaging. Am Heart J 1997; 133: [9] Pandian NG, Roelandt JR, Nanda NC et al. Dynamic three-dimensional echocardiography: methods and clinical potential. Echocardiography 1994; 11: [10] Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986; 1: [11] Dodge HT. Angiographic evaluation of ventricular function. N Engl J Med 1977; 296: [12] Kaye HH, Tynan M, Hunter S. Validity of echocardiographic estimates of left ventricular size and performance in infants and children. Br Heart J 1975; 37: [13] Silverman NH, Ports TA, Snider AR, Schiller NB, Carlsson E, Heilbron DC. Determination of left ventricular volume in children: echocardiographic and angiographic comparisons. Circulation 1980; 62: [14] Mercier JC, DiSessa TG, Jarmakani JM et al. Twodimensional echocardiographic assessment of left ventricular volumes and ejection fraction in children. Circulation 1982; 65: [15] Gueret P, Meerbaum S, Wyatt HL, Lang TW, Corday E. Two-dimensional echocardiographic quantitation of left ventricular volumes and ejection fraction. Importance of accounting for dyssynergy in short-axis reconstruction models. Circulation 1980; 62: [16] Schiller NB, Acquatella H, Ports TA et al. Left ventricular volume for paired biplane two-dimensional echocardiography. Circulation 1979; 60: [17] Boxt LM, Katz J, Kolb T, Czegledy FP, Barst RJ. Direct quantification of right and left ventricular volumes with nuclear magnetic resonance imaging in patients with primary pulmonary hypertension. J Am Coll Cardiol 1992; 19: [18] Kupferwasser I, Mohr-Kahaly S, Stahr P et al. Transthoracic three-dimensional echocardiographic volumetry of distorted left ventricles using rotational scanning. J Am Soc Echocardiogr 1997; 10: [19] Sapin PM, Schröder KM, Gopal AS, Smith MD, DeMaria AN, King DL. Comparison of two and three-dimensional echocardiography with cineventriculography for measurement of left ventricular volume in patients. J Am Coll Cardiol 1994; 24: [20] Moritz WE, Shreve PL. A system for locating points, lines and planes in space. IEEE Trans Instrum Measurement 1977; 26: [21] King 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: [22] Gopal AS, King DL, 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: [23] Vogel M, Lösch S. Dynamic three-dimensional echocardiography with a computed tomography imaging probe: initial clinical experience with transthoracic applications in infants and children with congenital heart defects. Br Heart J 1994; 71: [24] Pandian NG, Nanda NC, Schwartz SL et al. Threedimensional and four-dimensional transesophageal echocardiographic imaging of the heart and aorta in human using a computed tomographic probe. Echocardiography 1992; 9: