- Email: [email protected]
Magnetic resonance imaging of the great arteries in infants
International Elsevier
CARD10
Journal of Cardiology,
13
28 (1990) 73-85
01087
Magnetic resonance imaging of the great arteries in infants J.M. Parsons Departments
I, E.J. Baker ‘, A. Hayes ‘, E.J. Ladusans ‘, S.A. Qureshi M.N. Maisey * and M. Tynan *
of ’ Paediatric
‘, R.H. Anderson
Cardiology and 2 Radiological Sciences, Guy’s Hospital, London, U.K.; ’ Department National Heart and Lung Institute, London, U.K. (Received
20 September
1989; revision
accepted
Parsons JM, Baker EJ, Hayes A, Ladusans EJ, Qureshi Magnetic resonance imaging of the great arteries in infants.
2 February
SA, Anderson Int J Cardiol
3,
of Paediatrics,
1990)
RH, Maisey 1990;28:73-85.
MN,
Tynan
M.
Sixty infants, aged 1-48 (median 8) weeks, with suspected congenital heart disease underwent a morphological evaluation of the great arteries using magnetic resonance imaging at 1.5 Tesla. Cross-sectional echocardiography was performed in all infants, angiography in 33 and surgery in 44. Multiple sections, 5 mm thick and gated to the patients’ electrocardiogram were acquired in standard and oblique imaging planes. Ventriculo-arterial connexions were correctly identified in 54 infants (6 did not have intracardiac imaging performed) and an accurate description of the relationships of the great arteries was made in all. Magnetic resonance imaging clearly demonstrated normal and hypoplastic pulmonary arteries to the level of the first hilar branches and was better than echocardiography at confirming the presence or absence of central intrapericardial pulmonary arteries in 4 infants with pulmonary atresia. All parts of the thoracic aorta were accurately demonstrated and, in 23 infants with clinical suspicion of aortic coarctation, magnetic resonance images provided more information than echocardiography. Magnetic resonance imaging accurately demonstrates great arteries non-invasively supplementing echocardiographic and angiographic findings. In many cases, it replaces the need for invasive investigations. Key words:
Magnetic
resonance
imaging;
Great
arteries
Introduction Correspondence IO: Dr. E.J. Baker, Dept. of Paediatric Cardiology. Guy’s Hospital, St Thomas Street, London SE1 9RT. U.K. This research was supported by the British Heart Foundation. We would also like to acknowledge Sir Philip and Lady Harris, the Special Trustees of Guy’s Hospital and Philips Medical Systems who all helped in the funding of this project. Presented at the 10th Congress of the European Society of Cardiology, September 1988, Vienna.
0167-5273/90/$03.50
0 1990 Elsevier Science Publishers
Magnetic resonance imaging is a non-invasive technique which has been used to demonstrate congenital cardiac abnormalities [2-61. The majority of these studies have been performed in older children and detailed examination of infants has not been reported. In this age group, accurate diagnosis can usually be achieved non-invasively
B.V. (Biomedical
Division)
74
Patients and Methods
using cross-sectional echocardiography although abnormalities of the aorta and pulmonary arteries may not always be adequately defined [7-91. In this investigation, we have assessed the ability of magnetic resonance imaging to provide an accurate morphological description of great arteries in infants with suspected congenital heart disease. We have then attempted to define possible areas where this technique may be of benefit in their management.
TABLE
Patients Between April 1987 and October 1988, 60 infants with clinically suspected congenital heart disease were studied (Table 1). The median age of this group was 8 weeks (range l-48 weeks). Cross-sectional echocardiography had been performed in all of the infants prior to magnetic
1
Patient details. No.
Age
Diagnosis
No.
Age (wk)
Diagnosis
coarctation coarctation coarctation normal anomalous LCA coarctation, anomalous RSA multiple VSDs DORV, BTS pul. atresia, AP ~011. double aortic arch TAPVC normal VSD, PA band DORV ASD coarctation, VSD coarctation pul. atresia, VSD, AP ~011. hypoplastic aorta VSD normal VSD, coarctation pul. atresia, DILV, AP ~011. coarctation coarctation, DORV TGA, VSD Tetralogy of Fallot common arterial trunk coarctation, VSD CCTGA
31 32 33 34 35 36 3-l 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
1 8 2 20 4 7 36 3 29 44 2 1 4 9 8 2 2 48 44 16 12 1 4 12 8 2 4 40 16 16
normal coarctation normal TGA CCTGA, coarctation repair coarctation VSD VSD, recoarctation normal tricuspid atresia, BTS RAI, TAPVC, pul. atresia, AP ~011. normal AVSD coarctation HOCM, PS coarctation AVSD, TOF rhabdomyoma AVSD AVSD TGA, VSD TAPVC DORV coarctation DORV RAI, AVSD, DORV TAPVC coarctation coarctation
fwk) 1 2 3 4 5 6 7 a 9 10 11 12 13 14 1.5 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
3 3 1 2 8 1 12 8 3 12 8 1 16 12 26 6 40 16 12 6 1 16 16 12 4 20 4 3 8 48
wk = weeks; LCA = left coronary artery; RSA = right subclavian artery; VSD = ventricular septal defect; DORV = double outlet right ventricle; BTS = Blalock-Taussig shunt; pul. atresia = pulmonary atresia; AP ~011.= aorta-pulmonary collateral vessel; DILV = double inlet left ventricle; TGA = transposition of the great arteries; CCTGA = congenitally corrected transposition of the great arteries; RAI = right atrial isomerism; TAPVC = total anomalous pulmonary venous circulation; AVSD = atrioventricular septal defect; HOCM = hypertrophic obstructive cardiomyopathy; PS = pulmonary stenosis; TOF = tetralogy of Fallot.
resonance imaging. Angiography was subsequently performed in 33 infants and 44 infants have undergone surgery (including 4 infants who had operations before the study). Three infants have since died and autopsies have been performed. Six infants with suspected coarctation and one infant with suspected double aortic arch had equivocal cross-sectional echocardiographic examinations and were diagnosed as having normal hearts by magnetic resonance imaging. The diagnosis has since been confirmed with angiography in 3 infants and by a normal clinical course in the other four (mean follow-up 21.2 months). Methods Full informed consent was obtained from parents or guardians before starting the study. All the infants were examined with a 1.5 Tesla whole body imaging system (Gyroscan) manufactured by Philips Medical Systems. They were all sedated at the start of the procedure using chloral hydrate, 50-75 mg/kg. Standard electrocardiogram electrodes were then placed closely around the patient’s left nipple (in order to reduce distortion of the electrocardiogram) and connected to a telemetry system. They were then wrapped in a blanket and a firm foam mattress to restrict their movements. All the infants were imaged supine inside a 32 cm proton head coil. Simultaneous sections, 5 mm thick, separated by 0.5 mm and gated to the patients’ electrocardiogram were acquired using a Tl weighted spin echo sequence with echo time of 30 msec. A matrix size of 256 X 256 was used throughout with a field of view of 200 to 250 mm. Reduced acquisition using only 180 signals instead of 256 to construct the final image enabled scanning times to be reduced with only slight reduction in image resolution in the phase encoded direction. With two signal averages being performed for each examination a set of 7 sections would take about 6 minutes. Sections were acquired in a combination of standard orthogonal, and oblique imaging planes. In all cases the selection of planes was made by one of two cardiologists (J.M.P., E.J.B.) present throughout each examination. Oblique imaging
planes were chosen in order to demonstrate specific morphological features relevant to the suspected known cardiac pathology and varied to suit individual patient anatomy. The angles for these planes were determined individually and depended on the anatomical structures displayed from pre-existing images. Frequently used oblique imaging planes included oblique transverse rotated around an anterior-posterior axis (producing an image equivalent to an apical four chamber cross-sectional echocardiographic view) to demonstrate the great arteries in cross section; an oblique sagittal plane through the aortic arch to demonstrate the aorta and an oblique transverse plane rotated around a left-right axis to demonstrate the pulmonary trunk and arteries. Typical scanning times for a complete examination were between 60 and 90 minutes. Analysis of the magnetic resonance images was made by two cardiologists (J.M.P., E.J.B.) both of whom were aware of the patients’ clinical diagnosis and cross-sectional echocardiographic examination. A comparison with subsequent angiographic, surgical and autopsy findings (where available) was also made.
Results Image quality All the magnetic resonance images were of diagnostic quality and enabled assessment of ventriculo-arterial connexions, identification of the relationship between great arteries and a morphological description of the aorta and pulmonary arteries to be made in each infant. Ventriculo-arterial
connexions
In six infants (all with aortic coarctation), intra-cardiac anatomy was not studied and no attempt was made to define the ventriculo-arterial connexions in this group. In the remaining 54 infants ventriculo-arterial connexions were correctly defined in every case (Table 2). In 39 infants, the ventriculo-arterial connexions were concordant (Figs. 1, 2). Five had discordant ventriculo-arterial connexions, 3 with a concordant
atrioventricular connexion (congenitally complete transposition Fig. 3) and 2 with a discordant atrioventricular connexion (congenitally corrected transposition). Six infants had double outlet connexion from a morphologically right ventricle (Fig. 4). Four infants had a single patent great artery arising from the heart. In 3, this was an aorta in association with pulmonary atresia (Figs. 5, 6, 7) while in the other it was a common arterial trunk (Fig. 8). Ventriculo-arterial connexions were best defined using standard transverse imaging planes. A series of transverse sections, encompassing structures from the apex of the heart to the aortic arch, enabled the spatial relationship between the great arteries and the ventricles to be observed. Transverse sections, producing four-chamber views, enabled the trabeculated morphologically right ventricle to be identified and differentiated from the smooth walled morphologically left ventricle in all those that underwent intracardiac imaging. Transverse sections taken at more superior levels then demonstrated the great arteries in cross-sec-
TABLE
2
Vent&do-arterial
(VA) connexions.
VA connexions
Pts.
Concordant Discordant Doublet outlet right ventricle Single arterial connexion No intra-cardiac imaging
39 5 6 4 6
Total
60
tion in every infant. Sections through the aortic arch identified the aorta whilst the pulmonary trunk was recognised by its branching into right and left arteries. Sagittal sections were also helpful in identifying the spatial relationships between great arteries and ventricles. This was especially important in those infants with double outlet right ventricle (Fig. 4). Occasionally it was possible to identify coronary arteries (Fig. 3) and their origins (from both transverse and sagittal sections) and this aided differentiation of the aorta from the pulmonary trunk. Relationships
Fig. 1. Concordant ventriculo-arterial connexion demonstrated in transverse section. The pulmonary trunk (pt) is anterior to the ascending aorta. The proximal left pulmonary artery is shown and the right pulmonary artery can be seen to the level of the first hilar branches (arrow). The descending aorta is also identified.
The relationship between great arteries was defined in every infant. Transverse sections were most helpful although coronal and sagittal planes yielded additional useful information. In 45 infants, the pulmonary trunk was demonstrated anterior to the aorta (Fig. 1). In the infants with complete transposition, the aorta was demonstrated anterior and to the right of the pulmonary trunk in one, anterior and slightly to the left of the pulmonary trunk in another (Fig. 3) and, in the third, the great arteries were seen side-by-side with the aorta to the right of the pulmonary trunk. The pulmonary trunk was demonstrated anterior and to the left of the aorta in both infants with discordant ventriculo-arterial and atrioventricular connexions (congenitally corrected transposition). Of the six infants with double outlet connexions the aorta was anterior and to the left of the pulmonary trunk in one infant, side-by-side and to the left of the pulmonary trunk in another infant and post-
Fig. 2. Concordant
ventriculo-arterial
connexion.
In this transverse section the left pulmonary level of the first hilar branches (arrow).
erior and to the right of the pulmonary trunk in the remaining 4 infants. Central intrapericardial pulmonary arteries were not identified in 3 of the 4 infants with pulmonary atresia. Of this group, the aorta arose posteriorly from the left ventricle in 2 and, in the other infant, the aorta arose anteriorly from a hypoplastic morphologically right ventricle (Fig. 7). In one infant, in whom central intrapericardial pulmonary arteries were present, the aorta arose from the left ventricle. In the infant with a common arterial trunk, a common pulmonary confluence was identified arising anteriorly and slightly to the left from the main trunk (Fig. 8). Pulmonary arteries Images from standard transverse planes provided good visualisation of the pulmonary trunk and right pulmonary artery to the level of its first
artery
is well demonstrated
out to the
hilar branches (Fig. 1). The proximal portion of the left pulmonary artery was also demonstrated from this view but, because of its oblique course, additional images from an oblique transverse plane rotated in an anterior-posterior axis were occasionally necessary to demonstrate the distal portions, again out to the level of the first hilar branches (Fig. 2). Large dilated pulmonary arteries and dilatation of the pulmonary trunk was demonstrated in an infant with a perimembranous ventricular septal defect and significant left-to-right shunt. In an infant who had undergone banding, an adequate band and normal proximal pulmonary arteries were clearly identified (Fig. 9). Hypoplasia of the pulmonary arteries and trunk was demonstrated in two infants, one with tetralogy of Fallot and another with atrioventricular septal defect and double outlet right ventricle. In the three infants with pulmonary atresia, in which magnetic reso-
78
Fig. 3. Transverse
section in a patient with discordant ventriculo-arterial connexion. The aorta (Ao) is anterior trunk (pt). In this section a left coronary artery can be seen arising from the aorta (arrow).
nance imaging had correctly predicted absence of the central intrapericardial arteries, it was also possible to identify aorto-pulmonary collateral arteries, all confirmed by angiography. In one infant with pulmonary atresia, magnetic resonance imaging had identified central intrapericardial pulmonary arteries measuring 3 mm in diameter (Fig. 5). An oblique transverse plane rotated approximately 45 o in a left-right axis proved to be extremely helpful in differentiating central intrapericardial pulmonary arteries from aortopulmonary collateral arteries and the main bronchi. The origins and proximal portions of the aorto-pulmonary collateral arteries were best demonstrated using oblique transverse and oblique sagittal imaging planes (Figs. 6, 7). Two of the infants had undergone previous palliative operations with modified Blalock-Taus-
to the pulmc mary
sig shunts. Excellent views of the shunt were demonstrated in both from images in a sagittal plane. The entire length of the shunt, including the distal anastomosis with the left pulmonary artery, was displayed in a single image. The proximal anastomotic site was not clearly seen, although narrowing of the proximal portion of the shunt was identified in one infant (Fig. 4) and confirmed angiographically. Aorta Aortic coarctation was clinically suspected in 23 infants. Magnetic resonance imaging positively identified coarctation in 16 of these. The remaining 7 infants had equivocal echocardiographic findings and magnetic resonance imaging was able to identify complete hypoplasia of the thoracic
19
Fig. 4. Sagittal section from a patient with double outlet connexion from a morphologically right ventricle (rv). The ventricular septal defect (VSD) is closely related to the pulmonary trunk (pt) which is posterior to the aorta (Ao). Subvaivar pulmonary stenosis is demonstrated (arrows) and a modified Blalock Taussig shunt can be identified. The distal anastomosis with the left pulmonary artery appears patent, there is narrowing of the proximal portion of the shunt (single arrow).
aorta in one infant and normal aortic anatomy in 6. Angiography has been performed in 4 of these 7 infants, confirming the diagnosis made with magnetic resonance imaging. The remaining 3 patients have all had a subsequently normal clinical course for a mean period of 22.6 months. Multiple transverse sections produced a series of images which displayed the ascending and descending aorta in cross section. Identification of locahsed narrowing in the descending aorta was an accurate predictor of aortic coarctation. Imaging along the plane of the arch (oblique sagittal) produced a sagittal view of the aorta (Fig. 10). This permitted an accurate view of the site of coarctation and provided information about its
relationships to the brachiocephalic vessels and also demonstrated the aortic isthmus. An oblique coronal plane through the descending aorta provided a further coronal view of the aorta in which the site of coarctation site was identified. An anomalous right subclavian artery was identified from this view in one infant. Two infants were also studied who had undergone previous repair of coarctation using a subclavian flap. In one, narrowing at the site of repair, indicating recoarctation, was identified from transverse and oblique sagittal planes. No evidence of recoarctation was seen in the other infant. A double aortic arch was identified in one infant, seen most clearly from the transverse plane.
80
Fig. 5. Transverse section from a patient with pulmonary atresia. Small confluent central intrapericardial pulmonary arteries (single arrow) can be seen immediately posterior to the ascending aorta (Ao). Bronchi are demonstrated posterior to these (arrows).
Fig. 7. Oblique sagittal section from a patient with pulmonary atresia. The aorta arises anteriorly from the right ventricle (TV). The ascending aorta, aortic arch and descending aorta are clearly demonstrated and an aorto-pulmonary collateral vessel (arrow) can be seen branching off the descending aorta.
In occasional infants, the origins and proximal portions of left and right coronary arteries have been identified from sections in standard transverse (Fig. 3) and sagittal planes. Comparison of magnetic other findings
Fig. 6. Oblique sagittal section from a patient with pulmonary atresia. The aortic arch is demonstrated with an aortopulmonary collateral vessel (arrow) branching inferiorly.
resonance
images
with
Accuracy of the findings with magnetic resonance imaging was confirmed by angiography, performed in 33 infants, surgery in 39 infants and by autopsy findings in three infants. Seven infants, all with suspected aortic coarctation, had equivocal echocardiographic examinations because of difficulty in imaging the distal portion of the descending aorta. In all of these cases, the magnetic resonance images were better than echocardiography. Magnetic resonance images were also consistently better than echocardiography at demonstrating abnormalities of the pulmonary arteries. In the infants with pulmonary atresia, echocardiography was not able to differentiate confidently aorto-pulmonary col-
81
Fig. 8a. Oblique sagittal section demonstrating a common arterial trunk with its relationship to the underlying outlet ventricular septal defect (vsd). A single pulmonary confluence (PC) can be seen branching to the left,. cfb body; tv = tricuspid valve; ra = right atrium. b. Sagittal section from the same patient. A large perimembranous septal defect is demonstrated (vsd) and its relationship with the persistent common arterial trunk can be seen. The confluence is seen branching off anteriorly (arrow).
lateral vessels from central intrapericardial pulmonary arteries whereas magnetic resonance imaging was correct in each case.
Discussion Magnetic resonance imaging has several advantages over existing imaging techniques which make it suited for examination of cardiac anatomy. First, it is non-invasive and, furthermore, is not affected by the depth of structures within the thoracic cavity. Second, contrast agents are not required to opacify vessels, as the presence of flowing blood (which does not produce an image signal) acts as natural contrast. By acquiring a series of sections in any desired plane, the spatial relationships between structures can be appreciterms. This is imated in “three-dimensional” portant in the analysis of complex cardiac defects.
perimembranous = central fibrous outlet ventricular single pulmonary
It is, therefore, particularly suitable for examination of the great arteries in patients with congenital heart malformations [lO,ll]. Earlier reported studies were limited by use of only standard orthogonal imaging planes and many were performed on a relatively older age group of patients [2-61. Since the majority of patients with congenital heart disease present under one year of age [12], it is essential that any new imaging technique be evaluated in this age group. Although cross-sectional echocardiography at this age does not have to deal with poor echo windows, chest wall deformities and intra-thoracic air, problems which are all associated with imaging older patients, inaccuracies can still occur in imaging the aorta and pulmonary arteries [7-91. We have, therefore, chosen to study the great arteries in a group of infants using a system employing a strong field of 1.5 Tesla. This has an
Fig. 9. Oblique coronal section rotated around a left-right axis and a superior-inferior axis demonstrating a short axis view in a artery band. The right ventricular outflow tract is clearly demonstrated (rvot). The pulmonary trunk is seen patic :nt with a pulmonary with the band in position (arrow) and the proximal pulmonary arteries, which are not being distorted by the band, are also identified. tv = tricuspid valve; ra = right atrium; la = left atrium.
increased signal-to-noise ratio and allows thin sections of high resolution to be acquired in shorter scanning times. This makes it suitable for use in smaller sized children and infants [13]. We also used oblique imaging planes, which we feel is particularly important in the study of congenital heart disease. Technical problems with electrocardiographic gating were not encountered, and images of acceptable diagnostic quality were produced in all infants. Imaging as thus performed was able accurately to define the ventriculo-arterial connexions and demonstrate the relationships between the great arteries and intra-cardiac anatomy in all patients. Such knowledge is particularly important for
managing patients with complex heart defects such as double outlet right ventricle [14] and cannot always be straightforward to obtain with echocardiography and angiography. This is one area where magnetic resonance imaging can provide useful complementary information to existing imaging techniques. The ability to interpret clearly spatial relationships of the great arteries with intra-cardiac structures from simultaneous sections taken in any imaging plane is an important aspect of magnetic resonance imaging that enables a greater appreciation of the cardiac anatomy to be made. Examination of the aorta from sections in the oblique sag&al plane displayed the entire thoracic
83
Fig. 10. Oblique sagittal section demonstrating aortic coarctation with a discrete shelf (arrow).
aorta in long axis. This permitted excellent visualisation of the sites of coarctation, the aortic isthmus and the head and neck vessels, all usually within one section. Consequently, magnetic resonance imaging provided better resolution images and more detailed information than could be obtained with echocardiography, as we have reported and other studies have confirmed elsewhere [15,16].
These findings were initially confirmed from subsequent angiograms. In light of the quality of the magnetic resonance images, however, we now consider that angiography is no longer necessary in infants with coarctation diagnosed by this method. Identification of the origins and proximal portions of coronary arteries with magnetic resonance imaging was not consistent and could have been
84
improved by use of more specific oblique imaging planes. Accurate visualisation of the pulmonary trunk and pulmonary arteries is often essential in order to make appropriate surgical decisions. Magnetic resonance imaging was able accurately to demonstrate both normal sized and hypoplastic pulmonary trunk and pulmonary arteries to the level of their first branches, confirmed in all cases by angiography. Echocardiography, however, was rarely able to demonstrate little more than the proximal portions of the pulmonary arteries. This is another area where magnetic resonance imaging has advantages over echocardiography and, in some cases, may avoid the necessity for additional angiography. In patients with pulmonary atresia, identification of the presence of central intrapericardial pulmonary arteries is an important part of their management [17]. It may be difficult to differentiate these from aorta-pulmonary collateral vessels using echocardiography [8], as was the case in our group. Magnetic resonance imaging, however, was able correctly to demonstrate central intrapericardial pulmonary arteries when present and to differentiate them from aorto-pulmonary collateral vessels. Sections from oblique transverse imaging planes were most helpful in this respect with oblique sag.ittal planes being useful at demonstrating the proximal portions of the aorto-pulmonary collateral vessels. As indicated by -Rees and his colleagues, this is undoubtedly an area where magnetic resonance imaging has an important future role [II]. Magnetic resonance imaging is currently the only non-invasive method of demonstrating Blalock-Taussig shunts. Absence of signal within the shunt has been used as a method of indicating flowing blood and confirming shunt patency [18]. In one of our patients we were able to identify an area of stenosis. The ability to differentiate stenosed shunts from occluded shunts and to demonstrate the site of stenosis would be helpful in selecting those patients suitable for balloon dilation [19]. Crucial to every examination was the use of oblique imaging planes. In all cases, these were selected by a cardiologisp who was already familiar
with the echocardiographic findings and used this knowledge to direct the sequence of imaging planes used. We do not consider that predetermined sequences of imaging planes are appropriate for the study of congenital heart disease. Our findings have shown that magnetic resonance imaging is able to provide a detailed morphological evaluation of the great arteries in infants with congenital heart disease. It has identified important areas where this information can supplement echocardiographic and angiographic findings and, in some cases, where it may prove to be a better alternative non-invasive imaging technique.
References 1 Parsons JM, Ladusans EJ, Baker EJ, Ayton V, Tynan M. Magnetic resonance imaging of the great arteries in infants. Eur Heart J 1988;9:(abstr suppl 1)284. 2 Higgins CB, Byrd BF, Farmer DW, Osaki L, Silverman NH, Cheitlin MD. Magnetic resonance imaging in patients with congenital heart disease. Circulation 1984;70:851-860. 3 Fletcher BD, Jacobstein MD, Nelson AD, Riemenschneider TA, Alfidi RJ. Gated magnetic resonance imaging of congenital cardiac malformations. Radiology 1984;150:137140. 4 Jacobstein MD, Fletcher BD, Nelson A, Goldstein S, Alfidi J, Riemenschneider TA. ECG-gated nuclear magnetic resonance imaging: appearance of the congenitally malformed heart. Am Heart J 1984;107:1014-1020. 5 Higgins CB, Stark D, McNamara MT. Magnetic resonance imaging of the heart: a review of the experience in 172 subjects. Radiology 1985;155:671-679. 6 Didier D, Higgins CB, Fischer MR, Osaki L, Silverman NH, Cheitlin MD. Congenital heart disease: gated MR imaging in 72 patients. Radiology 1986;158:227-235. 7 Huhta C, Gutgesell HP, Latson LA, Huffines FD. Two-dimensional echocardiographic assessment of the aorta in infants and children with congenital heart disease. Circulation 1984;70:417-424. 8 Huhta JC, Piehler JM, Tajik AJ, et al. Two-dimensional echocardiographic detection and measurement of the right pulmonary artery in pulmonary atresia-ventricular septal defect: angiographic and surgical correlation. Am J Cardiol 1982;49:1235-1240. 9 Smallhorn JF, Huhta JC, Adams PA, Anderson RH, Wilkinson JL, Macartney FJ. Cross-sectional echocardiographic assessment of coarctation in the sick neonate and infant. Br Heart J 1983;50:349-361. 10 Fletcher BD, Jacobstein MD. MRI of congenital abnormalities of the great arteries. Am J Roentgen01 1986;146:941948.
85 Rees RSO, Somerville J, Underwood SR, et al. Magnetic resonance imaging of the pulmonary arteries and their systemic connections in pulmonary atresia; comparison with angiographic and surgical findings. Br Heart J 1987;58:621-626. Hoffman JIE, Christianson R. Congenital heart disease in a cohort of 19,502 births with long term follow up. Am J Cardiol 1978;42:641-647. Smith MA, Baker EJ, Ayton V, Parsons JM, Ladusans ET, Maisey MN. Magnetic resonance imaging of the infant heart at 1.5 T. Br J Radio1 1989;62:367-370. 14 Macartney FJ, Rigby ML, Anderson RH, Stark J, Silverman NH. Double outlet right ventricle cross-sectional echocardiographic findings. their anatomical explanation, and surgical relevance. Br Heart J 1984;52:164-177. 15 Baker ET, Ayton V, Smith MA, et al. Magnetic resonance
16
17
18
19
imaging of coarctation of the aorta in infants: use of a high field strength. Br Heart J 1989;62:97-101. Von Shulthess GK, Higashino SM. Higgins SS. Didier D. Fischer MR, Higgins CB. Coarctation of the aorta: MR imaging. Radiology 1986;158:289-296. Anderson RH, Macartney FJ. Shinebourne EA. Tynan M. Paediatric cardiology. Edinburgh: Churchill Livingstone. 1987;823-827. Jacobstein MD, Fletcher BD, Nelson AD, Clampitt M, Alfidi RJ. Riemenschneider TA. Magnetic resonance imaging: evaluation of palliative systemic-pulmonary artery shunts. Circulation 1984;70:650-656. Parsons JM, Ladusans E.I. Qureshi SA. Balloon dilatation of a stenosed modified (polytetrafluoroethylene) BlalockTaussig shunt. Br Heart J 1989;62:228-229.
CARD10
Journal of Cardiology,
13
28 (1990) 73-85
01087
Magnetic resonance imaging of the great arteries in infants J.M. Parsons Departments
I, E.J. Baker ‘, A. Hayes ‘, E.J. Ladusans ‘, S.A. Qureshi M.N. Maisey * and M. Tynan *
of ’ Paediatric
‘, R.H. Anderson
Cardiology and 2 Radiological Sciences, Guy’s Hospital, London, U.K.; ’ Department National Heart and Lung Institute, London, U.K. (Received
20 September
1989; revision
accepted
Parsons JM, Baker EJ, Hayes A, Ladusans EJ, Qureshi Magnetic resonance imaging of the great arteries in infants.
2 February
SA, Anderson Int J Cardiol
3,
of Paediatrics,
1990)
RH, Maisey 1990;28:73-85.
MN,
Tynan
M.
Sixty infants, aged 1-48 (median 8) weeks, with suspected congenital heart disease underwent a morphological evaluation of the great arteries using magnetic resonance imaging at 1.5 Tesla. Cross-sectional echocardiography was performed in all infants, angiography in 33 and surgery in 44. Multiple sections, 5 mm thick and gated to the patients’ electrocardiogram were acquired in standard and oblique imaging planes. Ventriculo-arterial connexions were correctly identified in 54 infants (6 did not have intracardiac imaging performed) and an accurate description of the relationships of the great arteries was made in all. Magnetic resonance imaging clearly demonstrated normal and hypoplastic pulmonary arteries to the level of the first hilar branches and was better than echocardiography at confirming the presence or absence of central intrapericardial pulmonary arteries in 4 infants with pulmonary atresia. All parts of the thoracic aorta were accurately demonstrated and, in 23 infants with clinical suspicion of aortic coarctation, magnetic resonance images provided more information than echocardiography. Magnetic resonance imaging accurately demonstrates great arteries non-invasively supplementing echocardiographic and angiographic findings. In many cases, it replaces the need for invasive investigations. Key words:
Magnetic
resonance
imaging;
Great
arteries
Introduction Correspondence IO: Dr. E.J. Baker, Dept. of Paediatric Cardiology. Guy’s Hospital, St Thomas Street, London SE1 9RT. U.K. This research was supported by the British Heart Foundation. We would also like to acknowledge Sir Philip and Lady Harris, the Special Trustees of Guy’s Hospital and Philips Medical Systems who all helped in the funding of this project. Presented at the 10th Congress of the European Society of Cardiology, September 1988, Vienna.
0167-5273/90/$03.50
0 1990 Elsevier Science Publishers
Magnetic resonance imaging is a non-invasive technique which has been used to demonstrate congenital cardiac abnormalities [2-61. The majority of these studies have been performed in older children and detailed examination of infants has not been reported. In this age group, accurate diagnosis can usually be achieved non-invasively
B.V. (Biomedical
Division)
74
Patients and Methods
using cross-sectional echocardiography although abnormalities of the aorta and pulmonary arteries may not always be adequately defined [7-91. In this investigation, we have assessed the ability of magnetic resonance imaging to provide an accurate morphological description of great arteries in infants with suspected congenital heart disease. We have then attempted to define possible areas where this technique may be of benefit in their management.
TABLE
Patients Between April 1987 and October 1988, 60 infants with clinically suspected congenital heart disease were studied (Table 1). The median age of this group was 8 weeks (range l-48 weeks). Cross-sectional echocardiography had been performed in all of the infants prior to magnetic
1
Patient details. No.
Age
Diagnosis
No.
Age (wk)
Diagnosis
coarctation coarctation coarctation normal anomalous LCA coarctation, anomalous RSA multiple VSDs DORV, BTS pul. atresia, AP ~011. double aortic arch TAPVC normal VSD, PA band DORV ASD coarctation, VSD coarctation pul. atresia, VSD, AP ~011. hypoplastic aorta VSD normal VSD, coarctation pul. atresia, DILV, AP ~011. coarctation coarctation, DORV TGA, VSD Tetralogy of Fallot common arterial trunk coarctation, VSD CCTGA
31 32 33 34 35 36 3-l 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
1 8 2 20 4 7 36 3 29 44 2 1 4 9 8 2 2 48 44 16 12 1 4 12 8 2 4 40 16 16
normal coarctation normal TGA CCTGA, coarctation repair coarctation VSD VSD, recoarctation normal tricuspid atresia, BTS RAI, TAPVC, pul. atresia, AP ~011. normal AVSD coarctation HOCM, PS coarctation AVSD, TOF rhabdomyoma AVSD AVSD TGA, VSD TAPVC DORV coarctation DORV RAI, AVSD, DORV TAPVC coarctation coarctation
fwk) 1 2 3 4 5 6 7 a 9 10 11 12 13 14 1.5 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
3 3 1 2 8 1 12 8 3 12 8 1 16 12 26 6 40 16 12 6 1 16 16 12 4 20 4 3 8 48
wk = weeks; LCA = left coronary artery; RSA = right subclavian artery; VSD = ventricular septal defect; DORV = double outlet right ventricle; BTS = Blalock-Taussig shunt; pul. atresia = pulmonary atresia; AP ~011.= aorta-pulmonary collateral vessel; DILV = double inlet left ventricle; TGA = transposition of the great arteries; CCTGA = congenitally corrected transposition of the great arteries; RAI = right atrial isomerism; TAPVC = total anomalous pulmonary venous circulation; AVSD = atrioventricular septal defect; HOCM = hypertrophic obstructive cardiomyopathy; PS = pulmonary stenosis; TOF = tetralogy of Fallot.
resonance imaging. Angiography was subsequently performed in 33 infants and 44 infants have undergone surgery (including 4 infants who had operations before the study). Three infants have since died and autopsies have been performed. Six infants with suspected coarctation and one infant with suspected double aortic arch had equivocal cross-sectional echocardiographic examinations and were diagnosed as having normal hearts by magnetic resonance imaging. The diagnosis has since been confirmed with angiography in 3 infants and by a normal clinical course in the other four (mean follow-up 21.2 months). Methods Full informed consent was obtained from parents or guardians before starting the study. All the infants were examined with a 1.5 Tesla whole body imaging system (Gyroscan) manufactured by Philips Medical Systems. They were all sedated at the start of the procedure using chloral hydrate, 50-75 mg/kg. Standard electrocardiogram electrodes were then placed closely around the patient’s left nipple (in order to reduce distortion of the electrocardiogram) and connected to a telemetry system. They were then wrapped in a blanket and a firm foam mattress to restrict their movements. All the infants were imaged supine inside a 32 cm proton head coil. Simultaneous sections, 5 mm thick, separated by 0.5 mm and gated to the patients’ electrocardiogram were acquired using a Tl weighted spin echo sequence with echo time of 30 msec. A matrix size of 256 X 256 was used throughout with a field of view of 200 to 250 mm. Reduced acquisition using only 180 signals instead of 256 to construct the final image enabled scanning times to be reduced with only slight reduction in image resolution in the phase encoded direction. With two signal averages being performed for each examination a set of 7 sections would take about 6 minutes. Sections were acquired in a combination of standard orthogonal, and oblique imaging planes. In all cases the selection of planes was made by one of two cardiologists (J.M.P., E.J.B.) present throughout each examination. Oblique imaging
planes were chosen in order to demonstrate specific morphological features relevant to the suspected known cardiac pathology and varied to suit individual patient anatomy. The angles for these planes were determined individually and depended on the anatomical structures displayed from pre-existing images. Frequently used oblique imaging planes included oblique transverse rotated around an anterior-posterior axis (producing an image equivalent to an apical four chamber cross-sectional echocardiographic view) to demonstrate the great arteries in cross section; an oblique sagittal plane through the aortic arch to demonstrate the aorta and an oblique transverse plane rotated around a left-right axis to demonstrate the pulmonary trunk and arteries. Typical scanning times for a complete examination were between 60 and 90 minutes. Analysis of the magnetic resonance images was made by two cardiologists (J.M.P., E.J.B.) both of whom were aware of the patients’ clinical diagnosis and cross-sectional echocardiographic examination. A comparison with subsequent angiographic, surgical and autopsy findings (where available) was also made.
Results Image quality All the magnetic resonance images were of diagnostic quality and enabled assessment of ventriculo-arterial connexions, identification of the relationship between great arteries and a morphological description of the aorta and pulmonary arteries to be made in each infant. Ventriculo-arterial
connexions
In six infants (all with aortic coarctation), intra-cardiac anatomy was not studied and no attempt was made to define the ventriculo-arterial connexions in this group. In the remaining 54 infants ventriculo-arterial connexions were correctly defined in every case (Table 2). In 39 infants, the ventriculo-arterial connexions were concordant (Figs. 1, 2). Five had discordant ventriculo-arterial connexions, 3 with a concordant
atrioventricular connexion (congenitally complete transposition Fig. 3) and 2 with a discordant atrioventricular connexion (congenitally corrected transposition). Six infants had double outlet connexion from a morphologically right ventricle (Fig. 4). Four infants had a single patent great artery arising from the heart. In 3, this was an aorta in association with pulmonary atresia (Figs. 5, 6, 7) while in the other it was a common arterial trunk (Fig. 8). Ventriculo-arterial connexions were best defined using standard transverse imaging planes. A series of transverse sections, encompassing structures from the apex of the heart to the aortic arch, enabled the spatial relationship between the great arteries and the ventricles to be observed. Transverse sections, producing four-chamber views, enabled the trabeculated morphologically right ventricle to be identified and differentiated from the smooth walled morphologically left ventricle in all those that underwent intracardiac imaging. Transverse sections taken at more superior levels then demonstrated the great arteries in cross-sec-
TABLE
2
Vent&do-arterial
(VA) connexions.
VA connexions
Pts.
Concordant Discordant Doublet outlet right ventricle Single arterial connexion No intra-cardiac imaging
39 5 6 4 6
Total
60
tion in every infant. Sections through the aortic arch identified the aorta whilst the pulmonary trunk was recognised by its branching into right and left arteries. Sagittal sections were also helpful in identifying the spatial relationships between great arteries and ventricles. This was especially important in those infants with double outlet right ventricle (Fig. 4). Occasionally it was possible to identify coronary arteries (Fig. 3) and their origins (from both transverse and sagittal sections) and this aided differentiation of the aorta from the pulmonary trunk. Relationships
Fig. 1. Concordant ventriculo-arterial connexion demonstrated in transverse section. The pulmonary trunk (pt) is anterior to the ascending aorta. The proximal left pulmonary artery is shown and the right pulmonary artery can be seen to the level of the first hilar branches (arrow). The descending aorta is also identified.
The relationship between great arteries was defined in every infant. Transverse sections were most helpful although coronal and sagittal planes yielded additional useful information. In 45 infants, the pulmonary trunk was demonstrated anterior to the aorta (Fig. 1). In the infants with complete transposition, the aorta was demonstrated anterior and to the right of the pulmonary trunk in one, anterior and slightly to the left of the pulmonary trunk in another (Fig. 3) and, in the third, the great arteries were seen side-by-side with the aorta to the right of the pulmonary trunk. The pulmonary trunk was demonstrated anterior and to the left of the aorta in both infants with discordant ventriculo-arterial and atrioventricular connexions (congenitally corrected transposition). Of the six infants with double outlet connexions the aorta was anterior and to the left of the pulmonary trunk in one infant, side-by-side and to the left of the pulmonary trunk in another infant and post-
Fig. 2. Concordant
ventriculo-arterial
connexion.
In this transverse section the left pulmonary level of the first hilar branches (arrow).
erior and to the right of the pulmonary trunk in the remaining 4 infants. Central intrapericardial pulmonary arteries were not identified in 3 of the 4 infants with pulmonary atresia. Of this group, the aorta arose posteriorly from the left ventricle in 2 and, in the other infant, the aorta arose anteriorly from a hypoplastic morphologically right ventricle (Fig. 7). In one infant, in whom central intrapericardial pulmonary arteries were present, the aorta arose from the left ventricle. In the infant with a common arterial trunk, a common pulmonary confluence was identified arising anteriorly and slightly to the left from the main trunk (Fig. 8). Pulmonary arteries Images from standard transverse planes provided good visualisation of the pulmonary trunk and right pulmonary artery to the level of its first
artery
is well demonstrated
out to the
hilar branches (Fig. 1). The proximal portion of the left pulmonary artery was also demonstrated from this view but, because of its oblique course, additional images from an oblique transverse plane rotated in an anterior-posterior axis were occasionally necessary to demonstrate the distal portions, again out to the level of the first hilar branches (Fig. 2). Large dilated pulmonary arteries and dilatation of the pulmonary trunk was demonstrated in an infant with a perimembranous ventricular septal defect and significant left-to-right shunt. In an infant who had undergone banding, an adequate band and normal proximal pulmonary arteries were clearly identified (Fig. 9). Hypoplasia of the pulmonary arteries and trunk was demonstrated in two infants, one with tetralogy of Fallot and another with atrioventricular septal defect and double outlet right ventricle. In the three infants with pulmonary atresia, in which magnetic reso-
78
Fig. 3. Transverse
section in a patient with discordant ventriculo-arterial connexion. The aorta (Ao) is anterior trunk (pt). In this section a left coronary artery can be seen arising from the aorta (arrow).
nance imaging had correctly predicted absence of the central intrapericardial arteries, it was also possible to identify aorto-pulmonary collateral arteries, all confirmed by angiography. In one infant with pulmonary atresia, magnetic resonance imaging had identified central intrapericardial pulmonary arteries measuring 3 mm in diameter (Fig. 5). An oblique transverse plane rotated approximately 45 o in a left-right axis proved to be extremely helpful in differentiating central intrapericardial pulmonary arteries from aortopulmonary collateral arteries and the main bronchi. The origins and proximal portions of the aorto-pulmonary collateral arteries were best demonstrated using oblique transverse and oblique sagittal imaging planes (Figs. 6, 7). Two of the infants had undergone previous palliative operations with modified Blalock-Taus-
to the pulmc mary
sig shunts. Excellent views of the shunt were demonstrated in both from images in a sagittal plane. The entire length of the shunt, including the distal anastomosis with the left pulmonary artery, was displayed in a single image. The proximal anastomotic site was not clearly seen, although narrowing of the proximal portion of the shunt was identified in one infant (Fig. 4) and confirmed angiographically. Aorta Aortic coarctation was clinically suspected in 23 infants. Magnetic resonance imaging positively identified coarctation in 16 of these. The remaining 7 infants had equivocal echocardiographic findings and magnetic resonance imaging was able to identify complete hypoplasia of the thoracic
19
Fig. 4. Sagittal section from a patient with double outlet connexion from a morphologically right ventricle (rv). The ventricular septal defect (VSD) is closely related to the pulmonary trunk (pt) which is posterior to the aorta (Ao). Subvaivar pulmonary stenosis is demonstrated (arrows) and a modified Blalock Taussig shunt can be identified. The distal anastomosis with the left pulmonary artery appears patent, there is narrowing of the proximal portion of the shunt (single arrow).
aorta in one infant and normal aortic anatomy in 6. Angiography has been performed in 4 of these 7 infants, confirming the diagnosis made with magnetic resonance imaging. The remaining 3 patients have all had a subsequently normal clinical course for a mean period of 22.6 months. Multiple transverse sections produced a series of images which displayed the ascending and descending aorta in cross section. Identification of locahsed narrowing in the descending aorta was an accurate predictor of aortic coarctation. Imaging along the plane of the arch (oblique sagittal) produced a sagittal view of the aorta (Fig. 10). This permitted an accurate view of the site of coarctation and provided information about its
relationships to the brachiocephalic vessels and also demonstrated the aortic isthmus. An oblique coronal plane through the descending aorta provided a further coronal view of the aorta in which the site of coarctation site was identified. An anomalous right subclavian artery was identified from this view in one infant. Two infants were also studied who had undergone previous repair of coarctation using a subclavian flap. In one, narrowing at the site of repair, indicating recoarctation, was identified from transverse and oblique sagittal planes. No evidence of recoarctation was seen in the other infant. A double aortic arch was identified in one infant, seen most clearly from the transverse plane.
80
Fig. 5. Transverse section from a patient with pulmonary atresia. Small confluent central intrapericardial pulmonary arteries (single arrow) can be seen immediately posterior to the ascending aorta (Ao). Bronchi are demonstrated posterior to these (arrows).
Fig. 7. Oblique sagittal section from a patient with pulmonary atresia. The aorta arises anteriorly from the right ventricle (TV). The ascending aorta, aortic arch and descending aorta are clearly demonstrated and an aorto-pulmonary collateral vessel (arrow) can be seen branching off the descending aorta.
In occasional infants, the origins and proximal portions of left and right coronary arteries have been identified from sections in standard transverse (Fig. 3) and sagittal planes. Comparison of magnetic other findings
Fig. 6. Oblique sagittal section from a patient with pulmonary atresia. The aortic arch is demonstrated with an aortopulmonary collateral vessel (arrow) branching inferiorly.
resonance
images
with
Accuracy of the findings with magnetic resonance imaging was confirmed by angiography, performed in 33 infants, surgery in 39 infants and by autopsy findings in three infants. Seven infants, all with suspected aortic coarctation, had equivocal echocardiographic examinations because of difficulty in imaging the distal portion of the descending aorta. In all of these cases, the magnetic resonance images were better than echocardiography. Magnetic resonance images were also consistently better than echocardiography at demonstrating abnormalities of the pulmonary arteries. In the infants with pulmonary atresia, echocardiography was not able to differentiate confidently aorto-pulmonary col-
81
Fig. 8a. Oblique sagittal section demonstrating a common arterial trunk with its relationship to the underlying outlet ventricular septal defect (vsd). A single pulmonary confluence (PC) can be seen branching to the left,. cfb body; tv = tricuspid valve; ra = right atrium. b. Sagittal section from the same patient. A large perimembranous septal defect is demonstrated (vsd) and its relationship with the persistent common arterial trunk can be seen. The confluence is seen branching off anteriorly (arrow).
lateral vessels from central intrapericardial pulmonary arteries whereas magnetic resonance imaging was correct in each case.
Discussion Magnetic resonance imaging has several advantages over existing imaging techniques which make it suited for examination of cardiac anatomy. First, it is non-invasive and, furthermore, is not affected by the depth of structures within the thoracic cavity. Second, contrast agents are not required to opacify vessels, as the presence of flowing blood (which does not produce an image signal) acts as natural contrast. By acquiring a series of sections in any desired plane, the spatial relationships between structures can be appreciterms. This is imated in “three-dimensional” portant in the analysis of complex cardiac defects.
perimembranous = central fibrous outlet ventricular single pulmonary
It is, therefore, particularly suitable for examination of the great arteries in patients with congenital heart malformations [lO,ll]. Earlier reported studies were limited by use of only standard orthogonal imaging planes and many were performed on a relatively older age group of patients [2-61. Since the majority of patients with congenital heart disease present under one year of age [12], it is essential that any new imaging technique be evaluated in this age group. Although cross-sectional echocardiography at this age does not have to deal with poor echo windows, chest wall deformities and intra-thoracic air, problems which are all associated with imaging older patients, inaccuracies can still occur in imaging the aorta and pulmonary arteries [7-91. We have, therefore, chosen to study the great arteries in a group of infants using a system employing a strong field of 1.5 Tesla. This has an
Fig. 9. Oblique coronal section rotated around a left-right axis and a superior-inferior axis demonstrating a short axis view in a artery band. The right ventricular outflow tract is clearly demonstrated (rvot). The pulmonary trunk is seen patic :nt with a pulmonary with the band in position (arrow) and the proximal pulmonary arteries, which are not being distorted by the band, are also identified. tv = tricuspid valve; ra = right atrium; la = left atrium.
increased signal-to-noise ratio and allows thin sections of high resolution to be acquired in shorter scanning times. This makes it suitable for use in smaller sized children and infants [13]. We also used oblique imaging planes, which we feel is particularly important in the study of congenital heart disease. Technical problems with electrocardiographic gating were not encountered, and images of acceptable diagnostic quality were produced in all infants. Imaging as thus performed was able accurately to define the ventriculo-arterial connexions and demonstrate the relationships between the great arteries and intra-cardiac anatomy in all patients. Such knowledge is particularly important for
managing patients with complex heart defects such as double outlet right ventricle [14] and cannot always be straightforward to obtain with echocardiography and angiography. This is one area where magnetic resonance imaging can provide useful complementary information to existing imaging techniques. The ability to interpret clearly spatial relationships of the great arteries with intra-cardiac structures from simultaneous sections taken in any imaging plane is an important aspect of magnetic resonance imaging that enables a greater appreciation of the cardiac anatomy to be made. Examination of the aorta from sections in the oblique sag&al plane displayed the entire thoracic
83
Fig. 10. Oblique sagittal section demonstrating aortic coarctation with a discrete shelf (arrow).
aorta in long axis. This permitted excellent visualisation of the sites of coarctation, the aortic isthmus and the head and neck vessels, all usually within one section. Consequently, magnetic resonance imaging provided better resolution images and more detailed information than could be obtained with echocardiography, as we have reported and other studies have confirmed elsewhere [15,16].
These findings were initially confirmed from subsequent angiograms. In light of the quality of the magnetic resonance images, however, we now consider that angiography is no longer necessary in infants with coarctation diagnosed by this method. Identification of the origins and proximal portions of coronary arteries with magnetic resonance imaging was not consistent and could have been
84
improved by use of more specific oblique imaging planes. Accurate visualisation of the pulmonary trunk and pulmonary arteries is often essential in order to make appropriate surgical decisions. Magnetic resonance imaging was able accurately to demonstrate both normal sized and hypoplastic pulmonary trunk and pulmonary arteries to the level of their first branches, confirmed in all cases by angiography. Echocardiography, however, was rarely able to demonstrate little more than the proximal portions of the pulmonary arteries. This is another area where magnetic resonance imaging has advantages over echocardiography and, in some cases, may avoid the necessity for additional angiography. In patients with pulmonary atresia, identification of the presence of central intrapericardial pulmonary arteries is an important part of their management [17]. It may be difficult to differentiate these from aorta-pulmonary collateral vessels using echocardiography [8], as was the case in our group. Magnetic resonance imaging, however, was able correctly to demonstrate central intrapericardial pulmonary arteries when present and to differentiate them from aorto-pulmonary collateral vessels. Sections from oblique transverse imaging planes were most helpful in this respect with oblique sag.ittal planes being useful at demonstrating the proximal portions of the aorto-pulmonary collateral vessels. As indicated by -Rees and his colleagues, this is undoubtedly an area where magnetic resonance imaging has an important future role [II]. Magnetic resonance imaging is currently the only non-invasive method of demonstrating Blalock-Taussig shunts. Absence of signal within the shunt has been used as a method of indicating flowing blood and confirming shunt patency [18]. In one of our patients we were able to identify an area of stenosis. The ability to differentiate stenosed shunts from occluded shunts and to demonstrate the site of stenosis would be helpful in selecting those patients suitable for balloon dilation [19]. Crucial to every examination was the use of oblique imaging planes. In all cases, these were selected by a cardiologisp who was already familiar
with the echocardiographic findings and used this knowledge to direct the sequence of imaging planes used. We do not consider that predetermined sequences of imaging planes are appropriate for the study of congenital heart disease. Our findings have shown that magnetic resonance imaging is able to provide a detailed morphological evaluation of the great arteries in infants with congenital heart disease. It has identified important areas where this information can supplement echocardiographic and angiographic findings and, in some cases, where it may prove to be a better alternative non-invasive imaging technique.
References 1 Parsons JM, Ladusans EJ, Baker EJ, Ayton V, Tynan M. Magnetic resonance imaging of the great arteries in infants. Eur Heart J 1988;9:(abstr suppl 1)284. 2 Higgins CB, Byrd BF, Farmer DW, Osaki L, Silverman NH, Cheitlin MD. Magnetic resonance imaging in patients with congenital heart disease. Circulation 1984;70:851-860. 3 Fletcher BD, Jacobstein MD, Nelson AD, Riemenschneider TA, Alfidi RJ. Gated magnetic resonance imaging of congenital cardiac malformations. Radiology 1984;150:137140. 4 Jacobstein MD, Fletcher BD, Nelson A, Goldstein S, Alfidi J, Riemenschneider TA. ECG-gated nuclear magnetic resonance imaging: appearance of the congenitally malformed heart. Am Heart J 1984;107:1014-1020. 5 Higgins CB, Stark D, McNamara MT. Magnetic resonance imaging of the heart: a review of the experience in 172 subjects. Radiology 1985;155:671-679. 6 Didier D, Higgins CB, Fischer MR, Osaki L, Silverman NH, Cheitlin MD. Congenital heart disease: gated MR imaging in 72 patients. Radiology 1986;158:227-235. 7 Huhta C, Gutgesell HP, Latson LA, Huffines FD. Two-dimensional echocardiographic assessment of the aorta in infants and children with congenital heart disease. Circulation 1984;70:417-424. 8 Huhta JC, Piehler JM, Tajik AJ, et al. Two-dimensional echocardiographic detection and measurement of the right pulmonary artery in pulmonary atresia-ventricular septal defect: angiographic and surgical correlation. Am J Cardiol 1982;49:1235-1240. 9 Smallhorn JF, Huhta JC, Adams PA, Anderson RH, Wilkinson JL, Macartney FJ. Cross-sectional echocardiographic assessment of coarctation in the sick neonate and infant. Br Heart J 1983;50:349-361. 10 Fletcher BD, Jacobstein MD. MRI of congenital abnormalities of the great arteries. Am J Roentgen01 1986;146:941948.
85 Rees RSO, Somerville J, Underwood SR, et al. Magnetic resonance imaging of the pulmonary arteries and their systemic connections in pulmonary atresia; comparison with angiographic and surgical findings. Br Heart J 1987;58:621-626. Hoffman JIE, Christianson R. Congenital heart disease in a cohort of 19,502 births with long term follow up. Am J Cardiol 1978;42:641-647. Smith MA, Baker EJ, Ayton V, Parsons JM, Ladusans ET, Maisey MN. Magnetic resonance imaging of the infant heart at 1.5 T. Br J Radio1 1989;62:367-370. 14 Macartney FJ, Rigby ML, Anderson RH, Stark J, Silverman NH. Double outlet right ventricle cross-sectional echocardiographic findings. their anatomical explanation, and surgical relevance. Br Heart J 1984;52:164-177. 15 Baker ET, Ayton V, Smith MA, et al. Magnetic resonance
16
17
18
19
imaging of coarctation of the aorta in infants: use of a high field strength. Br Heart J 1989;62:97-101. Von Shulthess GK, Higashino SM. Higgins SS. Didier D. Fischer MR, Higgins CB. Coarctation of the aorta: MR imaging. Radiology 1986;158:289-296. Anderson RH, Macartney FJ. Shinebourne EA. Tynan M. Paediatric cardiology. Edinburgh: Churchill Livingstone. 1987;823-827. Jacobstein MD, Fletcher BD, Nelson AD, Clampitt M, Alfidi RJ. Riemenschneider TA. Magnetic resonance imaging: evaluation of palliative systemic-pulmonary artery shunts. Circulation 1984;70:650-656. Parsons JM, Ladusans E.I. Qureshi SA. Balloon dilatation of a stenosed modified (polytetrafluoroethylene) BlalockTaussig shunt. Br Heart J 1989;62:228-229.