Transposition of the great arteries in the fetus: assessment of the spatial relationships of the arterial trunks by four-dimensional echocardiography

Transcription

1 Ultrasound Obstet Gynecol 2008; 31: Published online in Wiley InterScience ( DOI: /uog.5276 Transposition of the great arteries in the fetus: assessment of the spatial relationships of the arterial trunks by four-dimensional echocardiography D. PALADINI*, P. VOLPE, G. SGLAVO*, M. VASSALLO*, V. DE ROBERTIS, M. MARASINI and M. G. RUSSO *Fetal Cardiology Unit, Department of Gynecology and Obstetrics, University Federico II of Naples and Department of Pediatric Cardiology, 2 nd University of Naples, Monaldi Hospital, Naples, Department of Obstetrics and Gynecology, Di Venere-Giovanni XXIII Hospital, Bari and Pediatric Cardiology IRCSS G. Gaslini, Genoa, Italy KEYWORDS: congenital heart disease; fetal echocardiography; prenatal diagnosis; STIC; ultrasound ABSTRACT Objective Coronary arterial abnormalities can be one of the few negative prognostic indicators in transposition of the great arteries (TGA), and their occurrence is related to the type of spatial relationship of the great arteries. The main objective of this study was to assess whether the use of the reconstructed en-face view with color Doppler imaging of the four cardiac valves can demonstrate the different types of spatial relationship of the arterial trunks in fetuses with TGA, in order to derive the risk of coronary abnormalities. A secondary end-point was the evaluation of the type of coronary arterial branching pattern. Methods Twenty-three fetuses with a confirmed diagnosis of TGA underwent four-dimensional (4D) echocardiography at gestational weeks. The en-face view of the four cardiac valves and color Doppler with high persistence were employed to assess the spatial relationships of the great arteries. In all cases, confirmation of the vessels arrangement and coronary arterial distribution was obtained at neonatal echocardiography and/or surgery. Results The spatial relationships of the great vessels was identified correctly in 20/23 (87%) cases. The aorta was found to be located anterior to and to the right of the pulmonary trunk in 13/23 (56.5%) cases and just anterior to the pulmonary artery in 6/23 (26.1%) cases; in the remaining four (17.4%) cases, the two vessels were side by side. With respect to the association between the spatial relationship of the great arteries and the occurrence of an unusual pattern of coronary arterial branching, five of the TGA fetuses had abnormal coronary arterial distribution. Conclusions Using 4D echocardiography with color Doppler, it is possible to define the spatial relationships of the great arteries in fetuses with TGA with a high degree of accuracy. This information can be used during counseling to predict the likelihood of abnormal coronary arterial distribution. Copyright 2008 ISUOG. Published by John Wiley & Sons, Ltd. INTRODUCTION Transposition of the great arteries (TGA) is defined by discordant ventriculoarterial and concordant atrioventricular alignment, with the aorta arising from the right ventricle and the pulmonary artery arising from the left ventricle. Its incidence at birth is 4.7% of all congenital heart diseases, while in prenatal series TGA accounts for 4 6% of these cases 1 3. The arterial switch operation is now considered the surgical treatment of choice for neonates with TGA and the overall surgical mortality is remarkably low in most centers 4. However, atypical patterns of origin and course of the coronary arteries are still associated with increased mortality rates, though not in all reports 4 7. Furthermore, the occurrence of coronary arterial abnormalities has been shown to be related to the spatial relationships of the arterial trunks 7,8.Until now, none of these important prognostic findings could be employed in customized prenatal counseling due to the impossibility of defining by two-dimensional (2D) ultrasound coronary arterial abnormalities in the fetus with TGA. However, it has been demonstrated recently that four-dimensional (4D) echocardiography (spatiotemporal image correlation (STIC)) allows visualization of the en-face view of the four cardiac valves 9. Correspondence to: Dr D. Paladini, Via Petrarca, 72, Naples, Italy ( paladini@unina.it) Accepted: 29 December 2007 Copyright 2008 ISUOG. Published by John Wiley & Sons, Ltd. ORIGINAL PAPER

2 272 Paladini et al. The main end-point of this study was to assess whether the use of the reconstructed en-face view with color Doppler imaging of the four cardiac valves can demonstrate the different types of spatial relationships of the arterial trunks in fetuses diagnosed with TGA, in order to predict the risk of coronary arterial abnormalities. A secondary end-point was the evaluation of the type of coronary arterial branching pattern, as evident from postnatal catheterization or surgery, with respect to type of spatial arrangement of the great arteries. METHODS This was a prospective observational study conducted in two institutions (University Federico II of Naples and Di Venere Hospital, Bari) from January 2004 to December The study population included 23 consecutive fetuses with isolated TGA (no major associated cardiac defect) referred to one of the two centers at gestational weeks for diagnosis and management. During the same period, a total of 496 fetuses with congenital heart disease (CHD) was seen in the two centers. The procedure used was developed in an initial set of 10 normal fetuses in which it was confirmed that the normal relationships of the great vessels could be demonstrated. All cases underwent 2D and 4D echocardiography carried out with a Voluson 730 Expert (General Electric, Kretztechnik, Zipf, Austria) ultrasound machine. In each case, for the purposes of this study, a color Doppler volume was acquired at diagnosis and/or at follow-up. A total of 28 volumes was available for assessment. Care was taken in acquiring the volume to meet the following criteria: 1) the apical four-chamber view was used as a reference plane for acquisition, and 2) a high persistence setting was used for color Doppler, to ensure that both atrioventricular and ventriculoarterial flow could be displayed at the same time. The acquisition angle ranged between 25 and 30, according to the gestational age and the distance of the fetus from the transducer. The acquisition time ranged from 10 to 15 s. The 4D volume datasets were Figure 1 Four-dimensional echocardiography (spatio-temporal image correlation): en-face (coronal) view of the four cardiac valves. In the left panel, the correct position of the region of interest (ROI) window across the atrioventricular plane is demonstrated: the ROI should encircle only the four valve orifices, and the green bar should be positioned on the atrial side. In the right panel, the corresponding glass-body rendered image is shown: in this case, the rare side-by-side spatial arrangement of the great vessels is displayed in a case of transposition of the great arteries, with the aortic valve (A) on the right, anterior to the tricuspid valve (TV), and the pulmonary valve (P) on the left, anterior to the mitral valve (MV).

3 4D echocardiography in TGA 273 then processed offline with dedicated software (4Dviewer 5.0, General Electric). Post-processing included the following steps: standardization of the image appearance in the window of the acquisition plane, following the procedure described elsewhere 10. The rendering function was then activated and the region of interest (ROI) box was placed across the atrioventricular plane, with the green reference line on the atrial side (Figure 1). If the two arterial orifices could not be separated clearly in this view, the green line was positioned on the ventricular side. The transparent mode was selected for visualization of the grayscale component of the image and the surface mode for the color Doppler component. Once a good rendering of the four valve orifices was obtained, the cineloop sequence of the rendered glassbody image was reviewed frame by frame and the best frame showing the relationships of the atrioventricular and the ventriculoarterial valves, identified on the basis of color flow mapping, was identified and selected for the study. The classification used to define the spatial relationships of the great arteries was that illustrated in the postmortem study by Massoudy et al. 7, which includes five different variants. These are, in decreasing order of frequency: 1) aorta anterior to and to the right of the pulmonary trunk; 2) aorta directly anterior to the pulmonary artery; 3) side by side, with the aorta on the right and the pulmonary trunk on the left; 4) aorta anterior to and to the left of the pulmonary artery; 5) aorta posterior to and to the right of the pulmonary artery (Figure 2). The diagnosis of TGA was confirmed in all cases at neonatal echocardiography and/or surgery. The spatial relationships of the great arteries and the type of coronary arterial distribution were gathered from catheterization Figure 2 Spatial relationships of the great arteries in a normal heart (a) and in transposition of the great arteries (TGA) (b d) as demonstrated in the en-face view of the four cardiac valves using spatio-temporal image correlation. In the normal heart (a) the aorta is seen wedged in between the two atrioventicular orifices and the pulmonary artery lies just anterior to the aorta. The most common arrangement seen in this series of TGA fetuses was with the aorta anterior to and to the right of the pulmonary artery (b); the second most common arrangement was with the aorta just anterior to the pulmonary artery (c); the rarest fetal variant was the side-by-side arrangement (d), with the two vessels parallel to the two atrioventricular orifices, with the aorta on the right, in front of the tricuspid valve (TV), and the pulmonary artery on the left, anterior to the mitral valve (MV). A, aortic valve; P, pulmonary valve.

4 274 Paladini et al. and/or surgical files. With one exception, all neonates were alive and well at the time of writing. RESULTS The spatial relationships of the great arteries could always be demonstrated in the small preliminary series of 10 normal cases and were identified correctly in 20/23 (87%) cases with TGA (Table 1). Neonatally, the aorta was found to be located anterior to and to the right of the pulmonary trunk in 13/23 (56.5%) cases and just anterior to the pulmonary artery in 6/23 (26.1%) cases; in the remaining four (17.4%) cases, the vessels were side by side (Figure 2d). The other two rarer variants (aorta anterior to and to the left of the pulmonary artery; aorta posterior to and to the right of the pulmonary artery) were not encountered in this fetal series. As evident from Table 1, in all three misdiagnosed cases, the aorta was thought to be anterior to and to the right of the pulmonary artery, while it was actually directly anterior. It is noteworthy that there was an associated ventricular septal defect in all three cases and an aortic coarctation in two (Table 1). The agreement with postnatal findings was therefore 87% (20/23). In one case, characterized by a large ventricular septal defect with tricuspid valve straddling, the abnormal atrioventricular plane could be identified in the rendered image as well as on 2D imaging (Figure 3). There were five cases of abnormal coronary arterial distribution: one of the 13 with the aorta anterior to and to the right of the pulmonary artery; three of the six with the aorta anterior to the pulmonary artery, and one of the four with the vessels side by side (Table 1). DISCUSSION The abnormalities of the coronary arteries which have been demonstrated to complicate, in various but not all series, the surgical approach to TGA include: intramural origin of a coronary artery, tangential origin and course of a coronary artery, and dual and single sinus origin 4 7. These abnormalities are too subtle to be recognized directly on fetal echocardiography; therefore, until now, it has not been possible to include this type of information in prenatal multidisciplinary counseling for fetuses with TGA. However, it has been demonstrated that the incidence of abnormal coronary arterial distribution in individuals Table 1 Spatial relationship of the great arteries and coronary arterial branching pattern in a series of 23 consecutive cases of prenatally detected transposition of the great arteries GA Spatial arrangement of the great arteries Coronary arterial distribution (weeks) Additional anomalies Prenatal Postnatal Pattern Description 22 VSD, AoCo Ao-AR Ao-Ant Normal 21 No Ao-AR Ao-AR Abnormal LAD from S1, RCA + Cx from S2 28 VSD S-by-S S-by-S Abnormal Cx + LAD from S3 and coursing behind PA, RCA from S2 26 VSD S-by-S S-by-S Normal 21 VSD Ao-Ant Ao-Ant Normal 28 VSD Ao-Ant Ao-Ant Normal 26 VSD, AoCo Ao-AR Ao-Ant Abnormal Cx from S2, LAD + RCA from S1 21 No Ao-AR Ao-AR Normal 22 VSD Ao-AR Ao-Ant Abnormal LAD from S1, RCA + Cx from S2 24 VSD Ao-Ant Ao-Ant Abnormal LAD + infund. from S1, RCA + common trunk bifurcating in Cx + LAD from S2 21 AoCo S-by-S S-by-S Normal 21 No Ao-AR Ao-AR Normal 23 No Ao-AR Ao-AR Normal 21 VSD, tricuspid straddling, AAI S-by-S S-by-S Normal 23 PV dysplasia Ao-AR Ao-AR Normal 21 No Ao-AR Ao-AR Normal 28 AoCo Ao-AR Ao-AR Normal 26 No Ao-AR Ao-AR Normal 19 No Ao-AR Ao-AR Normal 33 No Ao-AR Ao-AR Normal 24 No Ao-AR Ao-AR Normal 22 No Ao-AR Ao-AR Normal 25 No Ao-AR Ao-AR Normal AAI, aortic arch interruption; Ao-Ant, aorta directly anterior to the pulmonary artery; AoCo, aortic coarctation; Ao-AR, aorta anterior to and to the right of the pulmonary trunk; Cx, circumflex coronary artery; GA, gestational age; infund., infundibular coronary branch; LAD, left anterior descending coronary artery; PA, pulmonary artery; PV, pulmonary vein; RCA, right coronary artery; S-by-S, vessels side by side, with the aorta on the right and the pulmonary trunk on the left; S1/2/3, aortic sinus 1/2/3; VSD, ventricular septal defect.

5 4D echocardiography in TGA 275 Figure 3 Transposition of the great arteries with ventricular septal defect and straddling of the tricuspid valve. (a) On two-dimensional imaging, the large inlet septal defect (arrowhead) and the straddling of the tricuspid valve (arrow) were clearly seen on the four-chamber view; (b) in the en-face view of the four cardiac valves using spatio-temporal image correlation, the discrepant size of the two atrioventricular orifices could be demonstrated together with the type of spatial relationship of the great arteries. A, aortic valve; LV, left ventricle; MV, mitral valve; P, pulmonary valve; RA, right atrium; TV, tricuspid valve. with TGA is related to the type of spatial relationships of the arterial trunks 7,8. The rationale of our study was to assess the possibility of gathering this additional prognostic information in utero using 4D ultrasound. In particular, our aim was to test the possibility that the coronal plane represented by the en-face view of the four cardiac valves reconstructed from 4D volume datasets of the fetal heart might allow clear definition of the various types of spatial arrangement of the vessels in TGA, in order to derive the probability of coronary arterial pattern abnormalities. Our data demonstrate that, by using the en-face view with color Doppler of the cardiac valves, it is possible to define with good accuracy the spatial relationships of the arterial trunks in normal and abnormal fetal hearts. In our study of 23 fetuses with TGA, this type of diagnostic approach yielded an 87% (20/23) concordance with the postnatal diagnosis. Regarding the relative incidence of the five types of spatial relationship of the great arteries, our data were reasonably similar to those reported in a postmortem study by Massoudy et al. 7 and in an echocardiographic study by Pasquini et al. 8 (Table 2), considering the 10- and 20-to-1 ratio, respectively, in the sample populations (200 and 409 vs. 23). In particular, the arrangement with the aorta anterior to and to the right of the pulmonary artery was the most common arrangement in all three series (comprising 74% of cases in the series of Massoudy et al. and 73% in that of Pasquini et al., and 57% in our series). Also, the incidences of the second and third most common arrangements, namely aorta anterior to the pulmonary artery and vessels side by side, were reasonably comparable (11% and 10% vs. 26% and 11% and 12% vs. 17%, respectively). Hence, 4D echocardiography can be employed reliably to assess the spatial relationships of the great vessels in CHD characterized by an abnormal ventriculoarterial junction. There are two technical issues which need to be addressed: the procedure for assignment of the four cardiac valves and the use of color Doppler. Regarding the cardiac valves, it is not possible to differentiate between mitral and tricuspid valves and between aortic and pulmonary ones from their appearance on a rendered glass-body image: this task should be accomplished while assessing the heart on multiplanar imaging, before switching to the coronal en-face view, using the green markers appearing on the ROI window in the different planes (Figure 1). Only in this way is it then possible to be sure to tag the four valves correctly on the rendered en-face coronal view. Secondly, regarding the use of color Doppler, in the Methods section we stated that the persistence of the color signal was high to allow us to visualize both the atrioventricular orifices (in red) and the ventriculoarterial ones (in blue) simultaneously on therenderedimage.wedecidedtouseasingleframe to allow us to show on a single still image the four valves at the same time. However, the spatial relationships of the great arteries can also be analyzed adequately if the persistence of the color signal is not high, the only difference being that the assessment should comprise evaluation of the whole cineloop sequence rather than a single frame. We emphasize this so that the reader is not misled into thinking that it is impossible to apply this technique retrospectively to personal archives just because the persistence of the color Doppler was not as high as

6 276 Paladini et al. Table 2 Spatial relationship of the great arteries vs. abnormal coronary arterial branching in cases with transposition of the great arteries: comparison of fetal data from this series (n = 23) with necropsy data from Massoudy et al. 7 (n = 200) and neonatal data from Pasquini et al. 8 (n = 409) Incidence (n (%)) Abnormal coronary arterial patterns (n (%)) Great arterial spatial relationships Massoudy et al. 7 Pasquini et al. 8 This series Massoudy et al. 7 Pasquini et al. 8 This series* Ao anterior to and to right of PA 148 (74) 299 (73) 13 (57) 37 (25) 31 (10) 1 (8) Ao directly anterior to PA 22 (11) 41 (10) 6 (26) 5 (23) 1 (2) 3 (50) Vessels side-by-side 23 (11.5) 48 (12) 4 (17) 14 (61) 45 (94) 1 (25) Ao anterior to and to left of PA 5 (2.5) 15 (4) 0 3 (60) 4 (27) Ao posterior to and to right of PA 2 (1) 3 (1) 0 2 (100) 3 (100) *See Table 1 for description of abnormal coronary arterial branching patterns in this series. With the aorta on the right and the pulmonary trunk on the left. Ao, aorta; PA, pulmonary artery. that used in this study. Retrospective assessment of any color Doppler 4D volume dataset is feasible, provided that the orientation of the four-chamber view in relation to the ultrasound beam was parallel or nearly so when the volume was acquired, in order to ensure that there is indeed color signal both in the inflows and in the outflows. The analysis of the incidence of coronary arterial branching abnormalities according to type of spatial arrangement of the great arteries gave very different results. There were also significant discrepancies between the postmortem series of Massoudy et al. 7 and the echocardiographic one of Pasquini et al. 8 (Table 2), which were related at least in part to the extremely different sample populations. Our study, which represents a third type of cohort, a fetal one, was limited by the small number of cases (23 cases with TGA, five of which had an abnormal coronary arterial distribution pattern). However, despite these differences and limitations, our figures were fairly similar to those reported by Pasquini et al. 8, at least as far as the arrangements with the aorta anterior and to the right of the pulmonary artery are concerned (Table 2). The fact that the rare types of spatial relationship of the great arteries (aorta anterior to and to the left of, and posterior to and to the right of, the pulmonary artery) were not encountered in our series was certainly due to the relatively small sample size. In fact, these arrangements, which were associated with the highest risk of coronary arterial distribution abnormalities in other studies (Table 2), accounted for only 3% and 5% of the cases in these series 7,8. The objective of this report, however, was to demonstrate that it is possible using 4D echocardiography to derive the risk of coronary arterial distribution abnormalities for fetuses diagnosed with TGA, in order to aid prenatal counseling. In particular, it should be possible to reassure the majority of couples whose fetus has been diagnosed with a TGA showing anatomy that is associated with a low risk of coronary arterial abnormalities. At the same time, we might be able to illustrate the possibly increased surgical risks associated with a TGA showing a spatial arrangement in which an abnormal coronary arterial branching pattern is more likely. REFERENCES 1. Epidemiology of Congenital Heart Disease. The Baltimore Washington Infant Study Ferencz C, Rubin DJ, Loffredo AC, Magee AC (eds). Perspectives in Pediatric Cardiology, Vol. 4. Mount Kisco Futura Publishing Co: New York, Paladini D, Russo MG, Teodoro A, Pacileo G, Capozzi G, Martinelli P, Nappi C, Calabrò R. Prenatal diagnosis of congenital heart disease in the Naples area during the years the experience of a joint fetal pediatric cardiology unit. Prenat Diagn 2002; 22: Allan LD, Sharland GK, Milburn A, Lockhart SM, Groves AM, Anderson RH, Cook AC, Fagg NL. Prospective diagnosis of 1006 consecutive cases of congenital heart disease in the fetus. J Am Coll Cardiol 1994; 23: Pasquali SK, Hasselblad V, Li JS, Kong DF, Sanders SP. Coronary artery pattern and outcome of arterial switch operation for transposition of the great arteries. A meta-analysis. Circulation 2002; 106: de Leval MR, Carthey J, Wright DJ, Farewell VT, Reason JT. Human factors and cardiac surgery: a multicenter study. J Thorac Cardiovasc Surg 2000; 119: Wernovsky G, Mayer JE Jr, Jonas RA, Hanley FL, Blackstone EH, Kirklin JW, Castañeda AR. Factors influencing early and late outcome of the arterial switch operation for transposition of the great arteries. J Thorac Cardiovasc Surg 1995; 109: Massoudy P, Baltalarli A, de Laval MR, Cook A, Neudorf U, Derrick G, McCarthy KP, Anderson RH. Anatomic variability in coronary arterial distribution with regard to the arterial switch procedure. Circulation 2002; 106: Pasquini L, Sanders SP, Parness IA, Wernovsky G, Mayer JE Jr, Van der Velde ME, Spevak PJ, Colan SD. Coronary echocardiography in 406 patients with d-loop transposition of the great arteries. J Am Coll Cardiol 1994; 24: Chaoui R, Hoffmann J, Heling KH. Three-dimensional (3D) and 4D color Doppler fetal echocardiography using spatiotemporal image correlation (STIC). Ultrasound Obstet Gynecol 2004; 23: Paladini D. Standardization of on-screen fetal heart orientation prior to storage of spatio-temporal image correlation (STIC) volume datasets [Editorial]. Ultrasound Obstet Gynecol 2007; 29: