17.1 Introduction

CMR is an established non-invasive imaging modality for morphologic and hemodynamic evaluation of congenital heart defect (CHD) in children and adults. Performing CMR in infants and neonates is different from adults in some imaging aspects as well as regarding specific referral indications. The overall examination setting requires adjustments of the imaging sequences due to the small patient size and fast heart rates. In young children CMR is mainly used for definitive diagnosis complementary to echocardiography. It provides useful information which can help to define treatment strategies and is valuable in planning interventions, such as cardiac surgery or catheter-guided treatment (Fig. 17.1). Subsequently, CMR can be used to assess the results of the performed intervention and for documenting adequate development and growth of cardiovascular structures.

Fig. 17.1
figure 00171

Newborn patient with pulmonary atresia and ventricular septal defect. Pulmonary perfusion is warranted by a patent ductus arteriosus. (a) Surface rendered CE-MRA depicting a complex anatomical situation: ductus arteriosus is tortuous and perfuses the left pulmonary artery; both pulmonary arteries are connected by a severely hypoplastic and stenotic confluent segment (**). (b) On the base of this information further treatment was planned, and the anatomy of pulmonary bifurcation was reconstructed by catheter-guided stenting (arrows) of the hypoplastic confluent segment. Another stent was placed in the ductus arteriosus to keep its patency (arrow). AAOascending aorta, LPAleft pulmonary artery, PDApatent ductus arteriosus, RPAright pulmonary artery

These particular aspects of paediatric cardiac imaging will be further discussed in this chapter.

17.2 Fundamentals of Paediatric Imaging

17.2.1 Settings

Paediatric patients with complex cardiac anatomy that can not be adequately assessed by echocardiography require advanced imaging with CMR. At the time of first diagnosis, particularly in complex CHD, understanding all clinical questions that need to be answered by the examination and the ability to interpret each image as it is acquired, are essential factors to direct the scan appropriately. As children may not be compliant for a long examination time and the youngest ones require general anaesthesia, the CMR scan should be kept as short as possible (shorter than 60 min) and tailored to the specific questions that need to be answered. Ideally, CMR in young children should be performed by a cardiac imager together with or supervised by an experienced paediatric cardiologist trained in cross-sectional imaging.

17.2.2 Sequences

CMR assessment of CHD in children can be performed with the same basic sequences applied in adults (Chap. 1). However high heart rate and small patient size require some adjustment of the imaging parameters, in order to improve both temporal and spatial resolutions of the images. Table 17.1shows the acquisition parameters of the main sequences, tailored for examining neonates and infants using a 1.5 T Signa HDxT MRI twinspeed scanner (General Electric Medical System, Milwaukee, Wisconsin, USA). These parameters may need some changes to adapt to a different type of scanner.

Table 17.1 Acquisition parameters (specific for a 1.5 T GE Signa HDxT scanner)

High spatial resolution(in-plane resolution ∼1 mm2) can be achieved by using a small field of view, a sufficiently large matrix, and thin slices. In order to decrease the acquisition time one can use rectangular FOV (reducing the number of phases) and whenever appropriate parallel imaging. Appropriate coil selection is crucial for obtaining sufficient signal from the resulting small voxels. If dedicated cardiac coils (phased-array, multi-channel) are not available for neonate examinations, the smallest coil covering the entire chest of the baby and providing the highest signal-to-noise ratio should be chosen. For this purpose an adult phased-array shoulder coil, a multi-channel knee coil or a head coil (Fig. 17.2) are possible solutions [1].

Fig. 17.2
figure 00172

3 days old boy examined in general anaesthesia in a head coil. The patient is covered with preheated hotpacks and with blankets, in order to prevent hypothermia

High temporal resolution(20–40 ms) is particularly important for functional measurements such as ventricular function using steady-state free precession (SSFP) and blood flow quantification with velocity encoded phase contrast cine (PC). For SSFP the true temporal resolution is calculated as the number of views per segment (or lines of the k-space) that can be obtained by one repetition time: TR × VPS. For cine PC the time is multiplied by a factor of two since the sequence is repeated twice for obtaining an anatomical image corresponding to each phase contrast image: 2 × TR × VPS. Temporal resolution can therefore be optimized by adjusting the number of VPS (between 2 and 8) to heart rate and keeping TR as low as possible.

In children with CHD cine imaging with SSFP and contrast-enhanced MR angiography (CEMRA) are the sequences of choice for anatomical assessment. In contrast to in adults, the use of T1 weighted black-blood images remains limited to selected indications, such as tumours (Fig. 17.3and Movie 17.1) that necessitate advanced tissue characterization and the central anatomy of the airways.

Fig. 17.3
figure 00173

Double Inversion recovery Black Blood image in a short axis view showing a large fibroma (**) infiltrating the inferior wall of the left ventricle. LVleft ventricle, RVright ventricle

Moreover similarly as in adults SSFP is the sequence of choice for measuring ventricular volume and assessing ­ejection fraction; this method has been shown to have an excellent reproducibility for both the left- and the right ventricles not only in adults but also in children. Specific paediatric normal values have been recently published [2].

CEMRA is considered to be the most striking and revolutionary sequence for vascular imaging in CHD. In our paediatric CMR program, CEMRA is being performed in >90 % of all examinations. CEMRA acquires a stack of contiguous thin slices covering the entire chest, which corresponds to a 3D data set. In order to optimally visualize all intrathoracic vessels, CEMRA is usually performed in a coronal plane including the chest wall anteriorly and spine posteriorly. The image acquisition is repeated three subsequent times, which provides imaging of the arterial and the venous phase of contrast medium passage. The excellent SNR obtained after contrast medium injection allows the use of a small field of view (18–20 cm) and very thin slices (down to 1.4 mm). In small children a double dose of Gadolinium (0.2 mmol/kg) is usually injected [3]. If the amount of contrast medium required is small (1–2 ml), we dilute it into saline solution up to a total volume of 3–5 ml and inject it manually through a peripheral intravenous line. Volumes larger than 5 ml can be injected by using an automated injection pump, after ensuring full patency of the i.v. line. In uncorrected CHD with right to left shunt, particular care needs to be taken for avoiding injection of air bubbles, since contrast medium can not be injected through an air filter which would dampen the bolus effect.

Correct timing of image acquisition synchronized to the maximal signal intensity from the contrast in the vessel of interest is crucial for achieving a good quality angiography data. The injection rate should be set depending on the amount of contrast medium and targeting a total injection time of 10 s. In modern scanners automated bolus detection or MR fluoroscopic triggering can be used for exact timing of image acquisition. We use a sequence with elliptic-centric k-space filling and a delay time of 5 s between bolus detection and image acquisition to start a breath-holding. If automated contrast medium detection is not available, we recommend using outcentric linear k-space filling and to start image acquisition simultaneously with contrast injection.

From the acquired 3D dataset it is possible to delineate a specific vessel in any oblique plane following sub-volume maximum intensity projection (MIP) reconstructions. Accurate measurements of the vessel diameters even in small children can be achieved, which is crucial for proper planning of further surgical or catheter-guided intervention [4].

Time-resolved 3D CEMRA is a new technique that overcomes the need for bolus timing and has been successfully performed in free-breathing children with CHD [5]. However, the acquisition with an improved temporal resolution implies a significant decrease of both the in-plane and through-plane spatial resolution as a trade off. This may result in insufficient accuracy when evaluating small vascular structures (Fig. 17.4). In contrast, with conventional CEMRA by using a slice thickness of 1.2–2 mm, FOV 180–200 mm, matrix 256 × 160 and zero interpolation a sub-millimeter spatial resolution can be obtained while maintaining the same acquisition time and signal intensity; a reconstructed voxel size of 0.45 × 0.55 × 0.3–0.5 mm3corresponds to a true spatial resolution of 0.9 × 1.1 × 1.2–2 mm3[1].

Fig. 17.4
figure 00174

Comparison of image quality in the same 1 month-old patient, between time-resolved CEMRA and conventional CEMRA. (a) Time-resolved CEMRA was performed with parallel imaging using a cardiac coil and a field of view of 32 cm; image resolution was 1.5 × 1.5 × 2–3 mm3and acquisition time 4 s. (b) “Conventional” CEMRA was performed by using a receiver-transmit head coil and a field of view of 20 cm; image resolution was 0.45 × 0.55 × 0.6 mm3, acquisition time 20 s

The recent concerns regarding gadolinium administration in neonates and infants and the potential risk for developing nephrogenic systemic fibrosis in patients with renal failure led to a search for a 3D sequence without the use of gadolinium based contrast medium. 3D SSFP allows whole-chest imaging with 3D reconstruction without administration of contrast-medium. This sequence applies a spectrally selective fat-saturation pulse to reduce signal from fat, followed by a T2 preparation pulse to reduce signal from blood which results in an enhancement of the contrast between blood, fat and myocardium. The application of a navigator allows image acquisition in free-breathing, since image acquisition is long lasting up to 15 min. 3D SSFP presents with an increased spatial resolution compared to anisotropic 2D and contrast-enhanced 3D imaging; however in small children small voxel size and fast heart rate result in decreased signal-to-noise ratio and contrast-to-noise ratio as well as blurred borders of anatomic structures. Recently imaging acquisition parameters could be tailored for imaging in small patients, and high image quality with good image contrast and sharp borders between blood-pool and myocardium have been obtained (Fig. 17.5).

Fig. 17.5
figure 00175

3D SSFP image in a patient with double outlet right ventricle, depicting the intracardiac anatomy, with large VSD, transposed position of the ascending aorta, and overriding pulmonary artery with subvalvar narrowing (conus) (Courtesy of Dr. Greil London). AAOascending aorta, PApulmonary artery, VSDventricular septal defect

3D cine SSFP images, a further development of 3D SSFP, have been successfully acquired with good spatial and temporal resolution in free-breathing infants [6]. 3D cine SSFP provides data on ventricular volumes and function, without the need for repetitive breath-holding and cumbersome planning procedures.

17.3 Anaesthesia and Sedation Considerations for CMR in Infants and Young Children

Younger children, mainly before school-age (7–8 years), require either sedation or general anaesthesia for a successful CMR examination. The preference regarding these procedures may differ between centres.

If a deep sedationis performed, the patient can breathe spontaneously and the images are acquired during shallow free-breathing. This approach is primarily less invasive, but the airways remain unprotected with a certain risk of aspiration, airway obstruction and hypoventilation. Thus sedation in the CMR unit should always be performed and monitored by an anaesthesiology team. From an imaging point of view, respiratory motion generates artefacts that deteriorate image quality and the negative diagnostic impact of such artefacts is more pronounced in younger children, in whom exact delineation of smaller vascular structures requires sharp images.

General anaesthesiawith endotracheal intubation allows an adequate degree of sedation, in which the airways are protected and ventilation can be controlled. In addition, potential hemodynamic changes associated with continuous intravenous sedation can be avoided [7]. Thus, even experienced anaesthesiologist may prefer handling the patient under complete controlled ventilatory and hemodynamic conditions, particularly children with CHD, who are considered high risk patients for sedation and anaesthesia [8].

Working in the CMR unit may present some more challenges for the anaesthesiology team. The procedure is performed in a noisy and unfamiliar environment, the ambient temperature is cool (20–22 °C) with risk of patient hypothermia, the access to the patient and the equipment may be limited, and reliable monitoring of the patient may not be warranted during the whole MRI scan.

MR compatible anaesthesia equipment, without mutual interference between scanner and monitoring system, is required. The respirator in the CMR unit should provide complete ventilation and respiratory gas monitoring, including pletysmographic and end-tidal carbon dioxide analysis. Hemodynamic monitoring should include blood pressure, and continuous ECG and pulse oximetry (Fig. 17.6). In neonates hypothermia can be avoided and body temperature is maintained constant by placing commercially available hotpacks, preheated to body temperature, around the head, the pelvis and the legs of the patients, and by covering the baby with a blanket (Fig. 17.2).

Fig. 17.6
figure 00176

CMR compatible anaesthesia equipment available in the examination room; the ventilatory system is integrated with a full monitoring system. The anaesthesiologist has full control on patient’s conditions and can stop ventilation manually during image acquisition

Repetitive safety instruction is mandatory for all additional medical personal entering the CMR unit, since rigorous behaviour is crucial for avoiding any incident related to ferromagnetic medical equipment and objects in the CMR environment. An additional crucial aspect for a successful and safe CMR examination consists of a close collaboration between the imaging team (cardiologist and/or radiologist) and the anaesthesiology team, with a joint meticulous planning and tailoring of the examination to the patient’s hemodynamic and respiratory status [7]. Of particular interest are patients with pulmonary hypertension on whom the frequency of repeated breath-holds need to be adjusted to the O2 saturation and CO2 levels, since longer and frequent breath-holds may induce life-threatening pulmonary hypertensive crisis. Even though it is recognized, that patients under anaesthesia, infants under 1 year, inpatients and ICU patients present higher risk for adverse events than general outpatient population, some reports demonstrate that CMR can be performed safely in intensive care infants, without significant intraprocedural complications or hemodynamic consequences [9, 10]. Nevertheless, even though CMR can be considered a safe diagnostic modality, careful analysis of the risks and benefits of a CMR examination has to be carried out for children with CHD and in the high risk group. Moreover in critically ill infants, anaesthesia may be performed directly by the neonatal intensive care team, who better knows the clinical status of the patient.

In a well organized paediatric infrastructure, with an anaesthesiology premedication clinic and dedicated beds in a day-clinic ward, CMR can be usually performed in an outpatients setting, and the patients discharged the same day of the examination. Indications for hospitalization are recommended for patients under 3 months of age, patients with a single ventricle physiology, and children with pulmonary hypertension or with unrepaired cyanotic CHD [8].

17.4 Specific Clinical Applications in Children

In children and infants, CMR is usually performed for ­complementary diagnosis, planning interventions or assessing the results of interventions. The main indications for performing CMR in the paediatric age group in our institution are shown in Fig. 17.7. In older children CMR is mainly performed during follow-up after surgical repair of CHD for detecting or ruling out significant findings and for planning a reintervention, such as pulmonary valve replacement in TOF. Thus, the majority of the patients are imaged for the assessment of right ventricle, right ventricular outflow tract and for quantification of pulmonary regurgitation. Postoperative assessment of TOF has become one of the most common clinical indications for CMR not only in adults but also in children.

Fig. 17.7
figure 00177

Graphic showing the most frequent indications for performing CMR in paediatric patients at the University Children’s Hospital Zurich

Main indications for performing CMR in neonates and infants are summarized in Table 17.2.

Table 17.2 Common clinical indications for CMR in neonates and infants with CHD

In newborns and infants the main referral reason for CMR is anatomical delineation of the intrathoracic vascular anatomy, including the aorta, the pulmonary arteries and the pulmonary veins, mainly in the context of complex CHD (Fig. 17.8). The information obtained by CMR is often of crucial significance for planning further treatment steps, and particularly invasive interventions. A retrospective analysis of 101 CMR examinations performed at our institution in infants and newborns (unpublished data) showed that in 70 % of the cases the CMR finding were relevant for planning cardiac surgery (47 %) or a catheter guided intervention (23 %). In 20 % of the cases findings requiring further intervention were ruled out, and in 5 % compassionate care was decided, as CMR confirmed a dismal diagnosis and prognosis.

Fig. 17.8
figure 00178

Newborn with heterotaxy syndrome, double-outlet right ventricle, transposition of the great arteries, subaortic conus and hypoplastic aortic arch with coarctation (arrow). A patent arterial ductus (**) provides continuity to the proximal descending aorta. AOascending aorta, LVleft ventricle, MPAmain pulmonary artery, RVright ventricle

Since conotruncal anomalies, aortic arch anomalies, the pulmonary arteries and the pulmonary veins are being discussed in different chapters of this book, the following paragraphs will focus on more specific paediatric applications and their preoperative assessment.

17.4.1 Heterotaxy Syndrome

Heterotaxy syndrome is frequently associated with complex CHD and is characterized by inconsistency of the situs of the thoracic and abdominal viscera (situs ambiguous), with similar shape of the atrial appendages (atrial isomerism), bronchial symmetry and splenic anomalies, including asplenia or polysplenia. The most common characteristics of cardiac defects occurring in heterotaxy syndrome are summarized in Table 17.3. Hearts in left atrial isomerism tend to be biventricular and those with right atrial isomerism univentricular.

Table 17.3 Characteristics of complex CHD associated with heterotaxy syndrome

As the spectrum of anomalies is exceedingly broad, a comprehensive and accurate delineation of cardiovascular anatomy and hemodynamic is essential for clinical decision-making and for planning the optimal surgical approach for each individual patient.

Thus a well-structured and segmental approach is recommended for planning and interpreting preoperative diagnostic testing in these patients. This should include assessment of the abdominal situs, spleen – asplenia, polysplenia, thoracic situs – bronchi, morphology of the atrial appendages, cardiac position, cardiac segments, intracardiac anatomy (atrial defects, atrioventricular defects, ventricular defects, coronary sinus), ventriculoarterial alignment, systemic venous anatomy, pulmonary venous anatomy and site of connection, pulmonary arteries, and aortic arch.

CMR may represent the ideal imaging modality to depict all these characteristics in one single and non-invasive examination. The thoracic and abdominal status can be defined using either black-blood fast spin echo sequence or SSFP. Fast spin echo may be particular advantageous for demonstrating the anatomy of the main stem bronchi and their relations to the branch pulmonary arteries (Fig. 17.9). In situs solitus, the course of right bronchus is superior to the right pulmonary artery (epiarterial bronchus), whereas the left bronchus runs inferiorly to the left pulmonary artery (hyparterial bronchus). Symmetrical length of the bronchi or symmetrical superior-inferior relation between the main bronchi and the proximal pulmonary arteries are indicative for a situs ambiguous.

Fig. 17.9
figure 00179

Black Blood Double Inversion Recovery coronal image depicting symmetrical bronchial tree consistent with a bilateral right bronchial system in a newborn patient with right atrial isomerism

Good diagnostic performance of CMR, in some cases superior to other imaging modalities, has been reported solely by using spin echo [11]. More recently used sequences, such as 2D SSFP, 3D SSFP and CEMRA may contribute significantly to improve diagnostic accuracy in these patients. SSFP can be particularly useful in defining intracardiac anatomy and the morphology of the atrial appendages. The morphologically right atrial appendage has an external triangular shape with a broad base and the apex pointed upward; internally the prominent crista terminalis and the pectinate muscles can be identified (Fig. 17.10). The left atrial appendage has a tubular finger-like shape with a narrow base, pointed anteriorly and downward. Delineation of additional features determining situs, such as the splenic status is possible by covering the upper abdomen when acquiring axial SSFP images. In presence of unbalanced atrioventricular septal defect, CMR can provide crucial information about the size of the ventricle, allowing clinical decision for performing biventricular repair or univentricular palliation (Movie 17.2).

Fig. 17.10
figure 001710

CEMRA reconstructed MIP images in an axial plane (a) and a coronal plane (b) at the level of the atrial appendages (raaand arrows) depict as symmetrical broad-based triangular shaped ­appendages, ­indicating the presence of right atrial isomerism. AOascending aorta, MPAmain pulmonary artery, raaright atrial appendage

Anomalies of the pulmonary venous connection occur very frequently in heterotaxy syndrome and are major determinants of outcome. Pulmonary venous anomalies can be accurately depicted by CEMRA (Fig. 17.11).

Fig. 17.11
figure 001711

Pulmonary venous anomalies in heterotaxy syndrome. (a) Newborn with right atrial isomerism. Echocardiography suspected stenosis of the pulmonary veins, due to turbulent flow. CEMRA (in a view from posterior) demonstrates that all pulmonary veins connect to the left-sided atrium without anatomical narrowing. There are three segmental pulmonary veins draining the left lung (*), and two right-sided pulmonary veins (**). LAleft atrium, LPAleft pulmonary artery, RPAright pulmonary artery. (b) Newborn with left atrial isomerism. CEMRA show that the left sided pulmonary veins connect to the left-sided atrium (*); the right-sided pulmonary veins (**) connect to the right-sided atrium. (c) Newborn with right atrial isomerism and total anomalous pulmonary venous connection as demonstrated in a CEMRA view from posterior. Right-sided pulmonary veins (**), left-sided pulmonary veins (*), VVvertical vein LAleft atrium, RAright atrium

In general, involvement of a paediatric cardiologist experienced in imaging complex CHD is strongly recommended for successful examination of children with heterotaxy syndrome.

17.4.2 The Borderline Ventricle – the Single Ventricle

In patients with a borderline left ventricle, usually in the context of a hypoplastic left heart complex (Movie 17.3), or with a borderline right ventricle, such as in unbalanced atrioventricular septal defect, decision for biventricular repair versus univentricular palliation has to be taken during the first months of life. Such decision can be challenging if we rely only on echocardiographic images. In these cases CMR is effective in quantifying the correct ventricle volume [12]. Using CMR Grosse-Wortman et al. demonstrated that patients for biventricular repair could be correctly selected in a group of 20 newborns with hypoplastic left heart complex and 80 % underwent successful biventricular repair. Moreover in patients with hypoplastic left heart complex late enhancement imaging may help to demonstrate the presence of endocardial fibroelastosis, which is thought to be a prognostic negative factor for a biventricular repair.

More recently further clinical applications of CMR in small children with CHD have been validated. So, the use of routine diagnostic cardiac catheterization before performing cavopulmonary anastomosis in children with a single ventricle physiology is being increasingly questioned. By combining CMR findings with clinical evaluation and echocardiography, most necessary preoperative information can be achieved and an increasing number of institutions opt for this non-invasive and non-ionizing radiation approach [3, 12].

17.5 CMR in Relation to Other Non-invasive Modalities

In the modern era of pediatric cardiology the different imaging modalities should not be considered as competitive techniques, but as diagnostic tools complementing each other. It is therefore the responsibility of every imaging expert to know strengths and weaknesses of each modality and to use and combine them appropriately.

17.5.1 Echocardiography

Echocardiography remains the first line bedside imaging modality with the capability to accurately describe intracardiac anatomy, particularly in newborns and small children with an intact acoustic window. Particularly in small patients with high heart rate, echocardiography provides superior spatial and temporal resolution than CMR imaging.

However echocardiography is highly dependent on operator skills, regarding the ability to visualize remote and unusual anatomical structures. Complex vascular anatomycan be difficult to assess and interpret in three dimensions, and the topographic relationships within the chest may be challenging to demonstrate. The course of the vessels beyond the central mediastinum and within the lung remains elusive for echocardiographers. Thus in presence of complex ­vascular anatomy, a second line advanced imaging modality is frequently required.

The major strength of CMR compared to echocardiography is quantitative functional assessment, such as measurement of ventricular volumes and function, as well as quantification of blood flow. Echocardiography assessment of left ventricular volumes and function is based on geometric algorithms, assuming the ventricle to be shaped like an ellipsoid. Obviously these models can not be applied for quantification of the abnormal left ventricle, the right ventricle or a single ventricle, each of these structures with a totally different shape. In contrast CMR, with the acquisition of real parallel adjacent images of the ventricles and with the calculation based on the summation-disc method, provides accurate volume measurements, reflecting the actual shape of every ventricle. New echocardiographic techniques, such as 3D echocardiography are designed to overcome these limitations, and are currently being validated for young children.

Similarly, in contrast to Doppler echocardiography, which can only estimate semi-quantitatively blood flow, PC cine imaging is able to provide accurate and reproducible blood flow quantification.

17.5.2 CT

Multislice CT angiography shares several advantages with CMR, particularly regarding the outstanding imaging quality. CT is superior to CMR for visualization of the airways and their relationship to vascular structures, as well as for assessment of parenchymal lung disease. Particularly in small children the superior spatial resolution of CT can be of great advantage for imaging small vascular structures. Ferromagnetic artefacts caused by clips, coils or stents are limited on CT compared to CMR, so that CT examination is still of diagnostic value. The readily availability of CT in urgent situations and the very short scanning time, which may overcome the need for general anaesthesia are further strength of this modality. In contrast functional measurements by CT are limited, and require ECG-gated image acquisition, which raises the amount of radiation used for the examination. In fact major concerns exist regarding the stochastic effects of exposure to ionizing radiation in young children. Therefore in the paediatric population the use of CT should be limited to selected indications, such as patients with contraindications for CMR (ferromagnetic devices, implants), assessment of the airways or of the lung tissue, or when sedation or anaesthesia can’t be used. If a CT is required, the scanning parameters need to be adapted to the size of the patient, in order to keep radiation as low as possible.

17.6 CMR and Cardiac Catheterization

Cardiac catheterization with conventional cine angiography has traditionally been the second-line modality of choice after echocardiography, as it can provide hemodynamic information and show the extracardiac vessels. Recently a European multi-center survey performed in 70 tertiary pediatric cardiology centres (data in publication), demonstrated that in patients with CHD cardiac catheterisation is still the most frequently used imaging modality after transthoracic echocardiography. Interestingly CMR represents the second more commonly used advanced technique. Thus, in the last decade a clear trend towards a constant decrease of the number of diagnostic catheterisations has been observed, in favour of an increasing number of CMR examinations and interventional catheterisations [13].

The incidence of complications associated with cardiac catheterization in children has been recently reported to be 8 % within the first 24 h after the procedure. The majority of these complications are related to the vascular access, and consist mainly of venous and arterial thrombosis. Infant’s age below 6 months, has been found to be an independent risk factor for vascular complications [14]. With the increasing number of infants with complex CHD, who can be treated nowadays, it is crucial to maintain vascular access patent for future therapeutic interventions and to minimize the number of invasive diagnostic procedures.

Both conventional cine angiography and CEMRA are able to provide excellent visualisation of the extracardiac vasculature. However CEMRA images can be reconstructed in three-dimension and each vascular structure of interest can be shown in every desired plane using multiplanar reformatting projection technique. These characteristics associated with the accuracy of size measurement in CEMRA images make it an ideal tool for planning catheter-guided or surgical interventions [4]. By applying CEMRA in complex CHD, in presence of multicentric lung perfusion and for evaluation of the pulmonary veins, all vascular structures can be accurately visualised by one single peripheral contrast-medium injection. There is no need for multiple selective cannulation, multiple injection of contrast medium for each vascular structure separately, and repetitive use of radiation as for conventional cine angiography

In children conventional cine angiography is still considered the gold-standard for assessment of the coronary arteries, although some preliminary experience has been reported about the use of CMR

CMR is also superior to cardiac catheterisation in quantitative functional assessment of most hemodynamic parameters. Ventricular volumes and function can be determined accurately and with high reproducibility. For quantitative assessment of the right ventricle CMR is notably considered as the reference standard examination. Moreover the capability of accurately measuring the flow enables quantification of valvular regurgitation and differential lung perfusion which cannot be provided by cardiac catheterisation. However cardiac catheterisation is still superior to other non-invasive diagnostic modalities regarding measurement of the pulmonary vascular pressure and resistance, such as in children undergoing the Fontan procedure or in patients with pulmonary hypertension.

In summary comparison of CMR with cardiac catheterisation reveal a similar to slightly superior visualisation of cardiovascular anatomy and higher accuracy for hemodynamic measurements, including ventricular volumes, intracardiac shunts, valvular regurgitation and differential lung perfusion. In addition the more obvious advantages of CMR are its non-invasiveness, the lack of ionising radiation and of iodinated contrast medium.

Assessment of pulmonary vascular pressure and resistance, as well as of complex pressure gradients are still invasively determined by cardiac catheterization.

17.7 Lifelong Follow-Up

Patients with severe CHD tend to present early in life, mainly during neonatal age. At time of initial diagnosis, exact description of the anatomy and hemodynamic assessment is crucial. In some selected cases, such as pulmonary atresia with multicentric lung perfusion, initial presentation and anatomical features can be predictive for later outcome. Nevertheless in congenital cardiology the evidence of the impact of imaging on outcome has not been analysed yet.

Later during life, even after total surgical repair of CHD, new findings may occur and/or residual findings become relevant and new interventions are necessary; therefore advanced imaging assessment may need to be repeated several times during life, particularly before and after interventions. During follow-up the choice of the appropriate imaging modality is determined by two critical aspects: firstly the risk/benefit for the patient, and secondly the reproducibility of the data obtained. In this context CMR presents striking advantages: (1) It is harmless for the body if performed following the standard safety recommendations; (2) the examination can be repeated unlimited time, without the risks associated with cumulative radiation or the complications of cardiac catheterisation discussed above, that may preclude adequate vascular access for future interventions. Congenital cardiologists need to keep in mind that young children present higher radiosensitivity compared to adults because the relative fraction of the body irradiated during catheterization is larger, and that these patients may need repeated examinations during their life.

For assessment of ventricular size and function, particularly for the right ventricle, CMR is currently considered the reference standard, to be compared with new non-invasive techniques, such as 3D-echocardiography. The 2–3 fold higher reproducibility of CMR functional measurements compared to echocardiography allow for detection of smaller changes during follow-up. On the basis of these considerations, we conclude that CMR may represent the ideal imaging modality for life-long follow-up in many patients. CMR does not present any significant immediate and later examination related risks and provides high sensitivity for small changes, allowing timely reintervention, if required.

17.8 CMR Limitations and Common Pitfalls

The significant resources needed to run a paediatric CMR examination are the most important limitation of CMR. Patient cooperation remains a great challenge in children and anaesthesia is required in children younger than school age. This requires an infrastructural investment, with purchase of fully equipped CMR-compatible ventilation and monitoring equipment, as well as an interdisciplinary agreement involving four different teams, cardiac imaging, cardiology, and anaesthesiology and imaging technicians.

CMR is not a modality available at the bed of the patient, therefore, for critically ill infants transport and monitoring to the CMR unit requires a significant effort from the caring team. Finally CMR may not be available to clinicians working in small medical centres. CMR is expensive and depending on health systems it may represent a significant financial burden for the hospital.

Pitfalls

Even though arrhythmias are less frequent in children than in adults, inappropriate ECG-gating or arrhythmias may deteriorate image quality and affect accuracy of functional measurements. Results from paediatric CMR examinations need to be compared with normal values that have been established for children with the appropriate technique [2, 15].

The accuracy of flow measurements done by PC cine imaging can be affected by offset errors and in many scanners flow correction with static phantom measurements may be necessary. Each institution needs to be aware of this potential source of error and plan enough time after each examination for performing phantom measurements and calibration if necessary. In children with CHD the accuracy of such flow measurements is crucial, as clinical decisions are often taken on the basis of the information provided by CMR. Eventually internal validation of the measurements, by comparing volumetric data of the ventricle and flow data of the great arteries are recommended. In addition when examining small children, one should be aware that accurate measurement of flow requires a minimal cross-sectional area in the region of interest; measurements should not be performed in small vessels containing less than 5 pixels.

Correct reading of CEMRA images requires caution regarding interpretation of narrowing of small vessel. So in presence of a surgical shunt, such as a Blalock-Taussig shunt between the aorta and the pulmonary arteries, the concomitant blood flow coming from the shunt, which may not contain contrast-medium yet, may result in reduce visibility of the vessel and mimic a stenosis. Moreover artefacts caused by surgical clips have to be ruled out, before reading narrowing of a vessel with surrounding clips.

17.9 Conclusions

Neonates, infants and small children are the patients taking most advantage from CMR. Performing CMR in a paediatric populations requires a specific infrastructure, adjustment of the acquisition parameters for achieving optimal images with adequate spatial and temporal resolution, expertise in complex CHD, and an excellent interdisciplinary collaboration among all teams involved. Once these requirements are achieved, CMR can be performed safely even in critically ill children with CHD.

In children with CHD, CMR is the non-invasive imaging modality of choice after transthoracic echocardiography, complementing its morphological and hemodynamic information before and after surgical or catheter-guided interventions. Thus CMR have a significant impact on establishing non-invasive assessment of CHD and in saving invasive procedures for treatment purposes.

In the near future continuous improvement in terms of software (sequences) and hardware (gradients, coils) may overcome the existing limitations of performing CMR in small children.

17.10 Practical Pearls

  • Performing pediatric CMR with anaesthesia requires an experienced anaesthesiology team and a close collaboration between the imaging team (paediatric cardiologist and/or paediatric radiologist) and anaesthesiology team.

  • Performing CMR in anaesthesia requires a CMR compatible ventilatory and hemodynamic monitoring equipment.

  • High spatial resolution can be achieved by using a small field of view, a sufficiently large matrix, and thin slices.

  • Temporal resolution can be optimized by adjusting the number of views per segment to heart rate and keeping TR as low as possible.

  • Administration of contrast medium can be avoided by applying the 3D SSFP sequence.

  • CMR quantitative data acquired in children need to be compared to normal values established in the paediatric population and obtained with the same technique.