Three-dimensional echocardiography in congenital heart disease: an expert consensus document from the European Association of Cardiovascular Imaging and the American Society of Echocardiography

Abstract

Three-dimensional echocardiography (3DE) has become important in the management of patients with congenital heart disease (CHD), particularly with pre-surgical planning, guidance of catheter intervention, and functional assessment of the heart. 3DE is increasingly used in children because of good acoustic windows and the non-invasive nature of the technique. The aim of this paper is to provide a review of the optimal application of 3DE in CHD including technical considerations, image orientation, application to different lesions, procedural guidance, and functional assessment.

Introduction

Three-dimensional echocardiography (3DE) has become important in the management of patients with congenital heart disease (CHD), particularly with pre-surgical planning, guidance of catheter intervention, and functional assessment of the heart. 3DE is increasingly used in children because of good acoustic windows and the non-invasive nature of the technique. The aim of this paper is to provide a review of the optimal application of 3DE in CHD including technical considerations, image orientation, application to different lesions, procedural guidance, and functional assessment.

Three-dimensional echocardiographic imaging techniques

Transducers

The evolution of 3DE techniques and transducer technology has been well described.^1–3 The development of the matrix array probe with parallel processing has made real-time 3DE possible since the 1990s.^4,5 Later generations of transducers have become smaller with footprints similar to two-dimensional echocardiography (2DE) transducers. The development of a small high-frequency paediatric 3DE transducer (2–7 MHz) has enhanced spatial and temporal resolution, especially pertinent for small children with high heart rates.^6,7 Similarly, miniaturization has enabled the development of adult-size 3D transoesophageal echocardiography (TOE) probes.⁸

Workflow

Ideally, 3DE transducers should be capable of producing 2D images which are at least equivalent to 2DE transducers. Some 3D transoesophageal transducers achieve this, but transthoracic 3DE probes still do not generally match the image quality of a dedicated 2D transducer. The difference remains most marked for high-frequency paediatric 3DE probes compared with the 2DE equivalent transducer. Consequently, the use of the combined 2DE–3DE transducer is not routine in smaller patients. Manufacturer recommendations suggest that current 3D TOE probes are used for patients >30 kg. Some paediatric cardiologists will extend use of such probes to smaller patients. Those undertaking such procedures should be aware of the specific manufacturer recommendation for the transducer. In any patient, the risk of complications such as damage to the oropharynx and oesophagus caused by an oversized probe needs to be balanced against the additional value of 3DE. For patients who are currently too small to accommodate 3D TOE transducers, epicardial 3D imaging with a transthoracic 3DE transducer is a feasible alternative technique during surgery.⁹

Data acquisition modes

Good spatial and temporal resolution in 3DE is a priority for imaging of CHD, particularly valve pathology and complex lesions. The matrix transducer has different modalities of data acquisition whose use is dictated by the clinical question. For example, in the assessment of double outlet RV, the incorporation of the AV valves, ventricular septum, and great arteries is necessary for decision-making, whereas measuring the size of an isolated VSD does not require such an extended field of view. The exact configuration and nomenclature of different modes is vendor specific but with features in common.

2D simultaneous multiplane mode

Current matrix probes allow 360° electronic rotation of the imaging plane as well as simultaneous display of more than one 2D imaging plane that can be electronically steered in the elevation or lateral plane. The crop plane is marked on the projection but with the drawback that temporal resolution is reduced.¹⁰ Applications include assessment of atrial septal defect (ASD) size and rim length,¹¹ size and shape of VSDs, AV valve morphology and regurgitation (Figure 1A and B ), outflow tracts and arterial valves.

Real-time 3DE mode

Real-time 3DE permits a display of an adjustable pyramidal volume, minimizing the issue with poor co-operation in children because there is no potential for ‘stitch’ artefacts between adjacent subvolumes. Increasing the region of interest decreases frame rate, and the limited field of view is a disadvantage for complex CHD where the relationship of different structures to each other is crucial for decision-making. Some manufacturers have a further 3D mode which permits the operator to select an area of interest but with relatively low frame rates particularly if colour flow Doppler is added. This mode is mainly used during catheter intervention, particularly ASDs, VSDs, and the AV valves. Depending on the system, vendor settings may be adjusted to allow the operator to prioritize volume rate at the expense of line density, thus achieving a higher temporal but lower spatial resolution.

ECG-gated multi-beat acquisition

With current imaging technology, ECG-gated multi-beat image acquisition is frequently used for paediatric 3DE because it acquires a large field of view with sufficient temporal resolution. However, the electronic ‘stitching’ of narrow volumes of data over 2–6 heartbeats may produce artefacts related to patient breathing or movement, particularly in young children. This is not an issue in ventilated children under general anaesthesia, because ventilation can be suspended briefly, and is less of a problem during sleep or with sedation. Although ‘single-beat’ volume acquisition has been introduced, the limited temporal resolution is insufficient for high heart rates of infants and children unless the region of interest is small thereby permitting narrowing of the imaging sector, or if the region of interest is relatively static.

3DE colour flow Doppler

3DE colour flow Doppler can be added to any of the above modalities. In common with 2DE, the addition of colour flow Doppler reduces temporal resolution. Depending on the size of the region of interest, the achievable frame rate may be too low for fast moving structures such as AV valves. Some manufacturers may permit the user to prioritize temporal resolution over spatial resolution to maintain an acceptable frame rate. The alternative is to use an ECG-gated multi-beat acquisition to maintain an acceptable frame rate.

Principles of 3DE acquisition

3DE for CHD utilizes the same transducers and ultrasound systems as adults with the addition of high-frequency probes suitable for imaging babies and children. General points are summarized below with emphasis on points relevant to the child or adult with CHD. In all patients, meticulous attention to 2D image quality is necessary to optimize the quality of the 3DE dataset. A high-frequency 3DE transducer should be used where possible, especially in small children where ultrasound penetration is not an issue, and the sector width narrowed to include only the regions of interest. Axial resolution is higher than either lateral or elevational planes which impacts on the optimal transducer position for the lesion being evaluated. For example, the mitral valve may be interrogated either from apical or parasternal view to delineate both leaflets and the subvalvar apparatus.¹² Review of the multiplanar images is particularly important prior to display of rendered images to avoid diagnostic error. Comprehensive reviews of 3DE artefacts have been published for reference.¹³

Imaging of relatively static cardiac structures, such as ASDs and VSDs, may be achieved using live 3DE or 3DE ‘zoom’ modes because temporal resolution is sufficient. Gain settings during acquisition and post-processing are particularly important in smaller patients with thin valves to reduce ‘noise’ that may impede visualization (Figure 2A and B) while avoiding ‘holes’ or other artefacts from inadequate gain. The normal gain setting for a 3DE acquisition is slightly higher than conventional 2DE as gain can be reduced during post-processing to optimize the image, whereas if too little gain is used during acquisition, increasing gain during post-processing does not restore parts of the image which have not been adequately visualized during the acquisition. This affects thin structures such as valve leaflets in particular.

Presentation of 3DE images of CHD merits particular consideration, especially for valves, and individualized tilt or rotation to display the region of interest can help (Figure 3A and B). The 3D display modes for CHD are analogous to those used for adults and include (i) volume rendering, (ii) surface rendering, and (iii) multiplanar reformatted (MPR) image display.^14,15 Volume-rendered datasets can be electronically segmented, allowing the operator to ‘crop’ the heart in multiple sections to expose the cardiac structure of interest from any desired angle. This is especially useful for CHD prior to surgery or intervention. Surface rendering presents the surfaces of structures or organs with a solid appearance and is mainly used to visualize size, shape, and function of cardiac structures.¹⁶ Analysis software for this mode includes semi and fully automated endocardial border detection for quantification of left ventricular (LV) and RV function^17,18 as well as semi-automated mitral leaflet motion detection for quantification of valve function, both of which may be limited by the abnormal ventricular shape or valve morphology found in CHD.

MPR allows the 3DE dataset to be viewed on a quad screen with the 3D dataset cut into three planes (sagittal, coronal, and transverse) which are configurable by the user. Adjustment of the MPR planes is illustrated in see Supplementary data online, Powerpoint 1. This display of planes not available by 2DE can aid with understanding complex anatomies and facilitate measurement of many CHDs such as ASD and VSD size. Valve areas, effective orifice areas, and vena contracta areas of regurgitant jets can all be measured by manipulating cutting planes to avoid foreshortening or oblique measurements.^19–21

Future directions

Improved temporal resolution of single-beat acquisitions or post-processing software to deal with stitch artefacts would enhance 3DE in younger patients. Software packages that can accommodate analysis of valves and chambers of abnormal morphology would also benefit the patient with CHD.

Recommendations

The 3DE approach should be tailored according to the patient. Small footprint and high-frequency 3DE transducers should be used in infants and young children and for epicardial 3DE.

3D TOE should be considered when patient size permits if 3D TTE provides insufficient imaging to plan therapy.

Image display and orientation

Published standards have been produced for the echocardiographic assessment of patients with paediatric and CHD, including TTE,²² TOE,²³ and quantification.²⁴ Recent published work has defined standards for adult practice using 3DE,¹⁴ but the latter specifically excluded CHD. In the patient with CHD, cardiac position, situs, connections, and alignments may be abnormal which presents a major challenge compared with acquired heart disease. 3DE facilitates en face projections of cardiac septums and AV valves,²⁵ which can be rotated in the z-plane to any desired orientation. If the analogy of a clock face is used, the use of z-plane rotation turns the image in a clockwise or counter clockwise direction to achieve an anatomically correct orientation (see Supplementary data online, Presentation 1). The retention of adjacent anatomic landmarks is desirable, where possible, to assist orientation. Echocardiographic evaluation of the patient with CHD is often complemented by other imaging modalities including magnetic resonance imaging (MRI), computed tomography (CT), and angiography. To gain maximum value, the orientation of 3DE images should be both consistent and intuitive, as exemplified by the ‘anatomic’ approach to image display,²⁶ projecting the heart in the same orientation as a person standing in an upright position. With this approach, superiorly positioned structures will be displayed uppermost on the image. This anatomically correct approach is consistent with the projection of MRI and CT images.^26,27 The application of an anatomic orientation can be illustrated with specific examples.

The atrial septum

The atrial septum can be visualized from the right or left atrial aspect. A projection from the right atrium permits visualization of important landmarks such as the superior vena cava, inferior vena cava, ascending aorta, tricuspid valve, oval fossa, and os of the coronary sinus. The preferred anatomic image orientation has the superior vena cava uppermost and tricuspid valve seen to the right of the atrial septum (Figure 4).

The ventricular septum

The ventricular septum can be visualized from either the RV or the LV aspect. By convention, the components of the ventricular septum are named as if the septum is viewed from the RV side with the free wall of the RV removed. With the anatomic approach, the diaphragmatic border of the heart is lowermost, the RV apex is seen to the right and the RV outflow tract viewed uppermost on the projected image so that landmarks such as the tricuspid valve, moderator band, and septomarginal trabeculation are seen in anatomically appropriate position (Figure 5A). Similarly, the LV aspect of the ventricular septum can be viewed in an anatomic projection (Figure 5B), to include both the septum and LV outflow tract.

AV valves

In CHD, the left-sided AV valve may not be a bileaflet valve of ‘mitral’ type and the right-sided AV valve may not be the tricuspid valve. Regardless of the morphology of the valve, or whether an en face view is projected from the ventricular or atrial aspect, the 3D rendered image is rotated so that the diaphragmatic surface of the heart is shown lowermost (Figure 6A and B). This means, for example, that the superior bridging leaflet in an AV septal defect will be shown uppermost on the image and the inferior bridging leaflet lowermost (Figure 7).

Aortic and pulmonary valves

The morphology, position and patency of the aortic, and pulmonary valves cannot be assumed in CHD. These valves are therefore projected in an anatomic format using the standard conventions for nomenclature of the valve leaflets. For example, the aortic valve may be projected as if seen from the ascending aorta or from the LV outflow tract. The conventional nomenclature of left, right, and noncoronary leaflets and sinuses is used in exactly the same fashion as in 2DE. An example of the preferred orientation of the aortic valve is shown (Figure 8) and a similar approach is used for the pulmonary valve.

Complex abnormalities of cardiac connections

When the main cardiac connections are abnormal, an anatomic presentation of images is particularly important so that the abnormal anatomy is displayed in a manner as close as possible to the actual spatial locations. An accurate understanding of the relationship of intracardiac structures has a direct impact on the surgical approach.

‘Surgical’ views of the heart

The term ‘surgical’ view has been used to describe 3D projections that are most akin to the surgeon's view during an operation. There are specific considerations for this term, particularly in contrast to the ‘anatomic’ view. The anatomic view is projected as if the person is standing upright, whereas a surgical view is projected as if the patient is lying supine with the lead surgeon operating from the right side of the patient. The effect of this is that an anatomic en face view of the right side of the atrial or ventricular septum would be rotated counter clockwise 90° when projecting a ‘surgical’ view (Figure 9). Views of the atrial aspect of the AV valves are often referred to as ‘surgical’ even though the surgeon may take a different access route to repair the valve in question. For example, the truly surgical view of the mitral valve would be presented with the patient supine with the left atrium accessed through the atrial septum. This contrasts with the usual projected 3DE view obtained by cropping away the posterior aspect of the atriums and rotating the AV valves en face with rotation of the whole dataset into an anatomic position. In practice, our preference is an anatomic orientation, which maintains consistency with projections of MRI and CT scans, knowing that ‘surgical’ visualization of the structures may be different. A visual demonstration of the common manipulations of 3D datasets accompanies this document (see Supplementary data online, Presentation 1).

Future directions

The availability of orientation markers, already available for CT angiography and MRI, is needed for 3DE to mark, for example, the true left/right or superior/inferior orientation of an image. This should be coupled with the ability to ‘landmark’ important anatomical structures to enhance display of complex CHD as the data is manipulated. Fusion imaging, where 3DE datasets can be co-registered with datasets from other modalities (fluoroscopy, CT angiography, and MRI), will allow for both automated anatomic orientation of 3DE images and visualization of regions which may be sonographically inaccessible. Widespread implementation would require co-registration across vendors, modalities, and platforms.

Recommendations

An ‘anatomic’ approach to image display is recommended as it reflects the real position of structures in space and is consistent with other modalities such as MRI and CT.

En face views of septums and AV valves should retain important landmarks and be rotated into a correct anatomic orientation.

The term ‘surgical’ view should be used only for projections that show anatomy as the surgeon would visualize the region of interest.

Optimal sonographic projections for different congenital heart lesions

Standard imaging planes in 2DE have been developed from a combination of anatomical constraints and accessible sonographic windows, forming the basis for published standards.²² 3DE is less constrained since post-processing will allow interrogation of structures contained within the acquired volume. A corollary is that a structure of interest can be displayed similarly after post-processing despite being acquired from a range of different echocardiographic windows. However, the principles of imaging physics apply as much to 3DE as they do to 2DE. For this reason, there are optimal approaches for data acquisition that should allow for optimal demonstration of the structure of interest. General recommendations can be made about the optimal acoustic windows to show different regions of interest. These include the following: For example, acquiring a 3D dataset to display an ASD is best achieved in a subcostal window because insonation from this view is orthogonal to the plane of the atrial septum. With TOE there is less flexibility to adjust the sonographic window; however, this is offset by higher image quality than TTE.²⁸

• clear visualization of the region of interest on 2DE

• insonation orthogonal to the plane of the structure of interest where possible

• inclusion of clinically relevant adjacent structures

• optimization of volume width and depth

Although image quality is central to interrogation of regions of interest and good 3D reconstruction, there are important differences from 2DE. For example, parasternal long-axis (PLAX) and parasternal short-axis (PSAX) views are used in 2DE to evaluate a perimembranous VSD (pmVSD) because the defect is in the near field, and good alignment to Doppler jets can be achieved. However, 3DE is especially useful for its ability to display VSDs en face and to delineate adjacent structures. PLAX and PSAX views are limited because of the small size of the ultrasound sector in the near field. Thus, subcostal and modified apical views may better define adjacent structures because the VSD is in the centre of the imaging field with a wider sector. This tailored approach is also employed for more complex lesions such as double outlet RV where AV valves, ventricular septum, and outlets all have to be incorporated into the 3DE volume.²⁹Table 1 summarizes optimal TTE sonographic views and the usefulness of 3D TOE for common forms of CHD.

Technical limitations may, however, persist irrespective of the imaging plane. Both the aortic and pulmonary valves are thin, rapidly moving structures. These are often imaged by 3D TTE using a PLAX or PSAX view where the plane of insonation does not lend itself to good quality rendering of the entirety of individual valve leaflets, particularly the body of the leaflet. This issue can be overcome by TOE in patients large enough to accommodate the 3D TOE probe.

Recommendations

Optimal sonographic projections for different congenital heart lesions

The angle of insonation should be tailored to the region of interest and ideally should be orthogonal to the relevant structure.

The size of 3DE region of interest should be adjusted to optimize temporal and spatial resolution.

Added value of 3DE for different congenital heart lesions

The literature on application of 3DE to CHD covers a wide variety of lesions including the AV valves, atrial septum, ventricular septum, and the outflow tracts. The use of 3DE techniques has increased as technology has improved, but there is wide institutional variability in the adoption of the technique. A central point in this regard is the evidence of additional diagnostic information compared with either 2DE or other imaging modalities. There have been no randomized trials relating to procedural success, morbidity or mortality related to the application of 3DE. Rather, 3DE has been adopted into practice on the basis of a clinical need to provide additional diagnostic information. Tables 2 and 3 present our consensus view of the added value of 3DE to assess some major groups of lesions. A selection of the key references relating to each of the different lesions is also included within the tables as well as a summary of the additional information provided. The type of lesions for which 3DE has a major role is heavily weighted towards valvar lesions and defects in both the atrial and ventricular septum (see Supplementary data online, Appendix 1, Presentation 1). A good example of the application of 3DE to more complex CHD is decision-making in patients with double outlet RV where particular considerations include the size and location of the VSD, and the relative position of the great arteries (Figure 10A–D, Supplementary data online, Video 1A–D). The depth of field enhances visualization of the position and size of the VSD relative to the great arteries, and projections are achieved from the RV aspect and apex which cannot be achieved by 2DE.

Recommendations

Added value of 3DE for different congenital heart lesions

3DE is recommended for the assessment of valvar lesions, septal defects, and complex abnormalities of the cardiac connections.

3DE should be regarded as a technique that complements rather than replaces 2DE for assessment of CHD.

Use of 3DE to guide catheter intervention

3D TOE is a rapid and useful imaging technique for the assessment of CHD during catheter-based interventions, including device closure of ASDs^{11,28,30–33,64–68} and VSDs.^{43–45,68,69,70} 3D TOE complements rather than replaces 2D TOE, and both modalities are used to assess defects and adjacent margins, rims, and structures. 3DE is particularly helpful for irregularly or asymmetrically shaped defects where 2D assessment of size by rotation of the TOE probe is insufficient. En face views of defects permit more precise appreciation of adjacent structures than 2D TOE alone, particularly for more complex lesions. 3D TOE of younger patients is typically done under general anaesthesia and is only feasible in patients large enough to accommodate the 3D TOE probe. Current manufacturers' recommendations propose a minimum patient weight of 30 kg. Time is limited during the procedure; therefore, rapid, simple, and minimal post-processing real-time 3D acquisition modes are often most effective. A 3D full-volume acquisition is favoured at some centres to produce a high-resolution view of the entire region of interest at the start of the procedure, followed by more focused imaging using live techniques. Lesion-specific targeted views have to be obtained, as discussed in subsequent sections. En face views are consistently used to demonstrate the relevant anatomy and interventional hardware. In addition to rendered 3D views, MPR imaging is useful for quantitation and image display during interventions¹⁹ (Figure 10).

ASD device closure

Transcatheter device closure of a secundum ASD has become the preferred method of treatment when the anatomy is favourable.^71,72 Accurate assessment of ASD type, size, position, number of orifices, shape, and rim sizes (Figure 11) is essential for correct patient selection, device selection, and deployment. Detailed analysis of device position, configuration, anchorage, residual shunt as well as the relationship of the device to the aorta, mitral valve, tricuspid valve, superior and inferior vena cava, and the pulmonary veins is necessary (Figure 11). Demonstration of these features is usually enhanced and at times only possible with 3DE.^{30,31,64–67,73}

The 3D TOE right atrial en face views from the mid-oesophagus selecting the region of interest to produce real-time 3D images can often demonstrate the key features during secundum ASD intervention. 3D full-volume acquisition may be needed if the volume rate using focused live 3D acquisition is too low. The deep trans-gastric sagittal bicaval view best demonstrates the inferior rim and septal length and provides a good view from which to monitor device deployment. Precise measurement of the ASD is best performed using the MPR four panel format, but measurement of the rendered image is an alternative (Figure 10). Colour Doppler flow analysis of the size, position, and mechanism of residual shunts is best performed with live 3D colour or biplane imaging. There are reports of using transthoracic 3DE to guide ASD closure from a subcostal view.^73,74

VSD device closure

Transcatheter closure of VSDs has developed as an alternative treatment to surgical closure of muscular VSDs (mVSDs) and pmVSDs.^75–81 Specific advantages of 3DE over 2DE are improved visualization of the VSD shape, size, and location as well as characterization of the tricuspid pouch and surrounding structures.^{43,44,46,68–70}En face presentation of the ventricular septum from both RV and LV aspects can be accomplished most expeditiously from a four-chamber view using live 3D (Figure 12) or ECG-gated full-volume acquisition. Monitoring of interventional device hardware and device deployment is seen in a frontal four-chamber view (Figure 12). Following device deployment, live 3DE, cross plane, or MPR imaging with colour flow Doppler is optimal for assessing the interventional result.

Additional applications of 3DE guidance: 3D TOE has been used during catheter-based closure of Fontan fenestrations,⁸² ruptured sinus of Valsalva aneurysms,⁸³ coronary artery fistulas,^84,85 prosthetic valve para-valvar leaks,^86–88 atrial switch baffle leaks or obstruction,^89,90 atrial septum trans–septal puncture,⁹¹ and biventricular pacemaker synchrony assessment and lead placement.⁹² Recently, intracardiac echocardiography transducers have been developed with 3D capability in a segment of ∼60 × 15° using a 10-French probe. Early work has demonstrated visualization of the atrial and ventricular septums, aortic valve, mitral valve, and atrial appendages for guidance of intervention.^93–95

Future directions

Miniaturization of 3D TOE probes for use in smaller patients and enhanced automation are likely future developments. The introduction of 3D intracardiac probes will provide an alternative to the TOE approach. The application of fusion imaging of 3D TEE with fluoroscopy, 3D rotational angiography, cardiac MRI, and CT angiography is also likely to expand as interventions become more complex.

Recommendations

Use of 3DE to guide catheter-based interventions

3DE is recommended to assist interventional closure of selected ASDs and VSDs, particularly multiple, irregularly shaped, or residual defects.

Real-time 3D imaging is recommended for visualization of catheters, delivery systems, and devices during catheter intervention in CHD.

Use of 3DE to measure defects (either MPR or rendered 3D images) is recommended to assist interventional catheter procedures involving CHD.

Three-dimensional echocardiographic assessment of ventricular volumes and ventricular function

There are many challenges with assessment of the RV and LV in patients with CHD, including abnormal cardiac position, abnormal connections, septal defects, non-contractile patch material, and abnormal loading conditions. These difficulties are additive to generic considerations such as adequate temporal resolution to determine the chamber volume throughout the cardiac cycle with accuracy. Above all, however, in CHD ventricular geometry may be far removed from the normal geometry around which software packages have been designed, so that analysis algorithms may not be valid. Nonetheless, 3DE has been applied in CHD because assessment of cardiac volumes and function are being increasingly used to plan patient management,⁵⁴ and subjective assessment is unreliable.⁹⁶

3DE assessment of the right ventricle

Assessment of RV size and function is important in clinical practice, particularly in patients with repaired tetralogy of Fallot (TOF) and other surgical repairs utilizing an RV to pulmonary artery conduit where RV measurements impact on the timing of pulmonary valve replacement. The position of the RV immediately behind the sternum complicates imaging windows, and subcostal imaging may not be adequate beyond the early childhood years. Prominent trabeculations complicate endocardial border delineation, and normal RV geometry is complex with a triangular appearance in the sagittal plane and a crescent shape in the coronal plane. RV inflow and outflow tracts are also located in different imaging planes. All these hinder capturing the RV in its entirety by ultrasound and preclude the use of a simple geometrical formula to calculate volumes and ejection fraction (EF). 3DE has potential advantages for assessment of RV volume because it makes few assumptions about ventricular shape. Three 3DE techniques have been applied to measure RV volumes and EF, namely summation of discs, semi-automated border detection, and knowledge-based reconstruction.

Summation of discs

The semi-automated method of disc summation is conceptually most comparable with MRI analysis and has been validated in children by water displacement⁹⁷ and MRI.⁹⁸ Feasibility of this method is good in healthy children (Table 4), although a recent study in older patients found feasibility in <60%.⁹⁹ Although there is excellent correlation with MRI-derived RV volumes, the values from 3DE tend to be lower.⁹⁸ The disc summation method retains landmarks which may be abnormally positioned or absent in CHD.^18,113 Unfortunately, summation of disks methodology has been removed from some software packages, thereby removing the closest correlate with MRI.

Semi-automated border detection

This is the most common 3DE method to assess RV volumes and EF. A full-volume 3D data set is acquired and segmented into four-chamber, sagittal, and coronal views (Figure 13A). Key RV and LV anatomical landmarks are defined, and end-diastolic and end-systolic contours are manually drawn in each view to construct a dynamic polyhedron model of the RV (Figure 13A and B). In healthy adults, this methodology is feasible;^101,102 volumes and EF correlate well with MRI, although 3DE underestimates volumes compared with MRI (Table 4). Acquisition time is significantly shorter for 3DE than for MRI (∼5 vs. 20 min).^17,106 In healthy adults and children, intra- and inter-observer reliability is good^98,101 (Table 4) but agreement with MRI worsens in adults with cardiac dysfunction (Table 4).^103,109

In adults after TOF repair, correlation with MRI is good for end-systolic volume (ESV), end-diastolic volume (EDV),^104–106 and EF; however, 3DE produces lower volumes compared with MRI (Table 4).¹⁰⁵ Individual differences between the techniques can be substantial, and the limits of agreement are wide. The disparity between 3DE and MRI becomes larger in the severely dilated RV where there are particular difficulties in the incorporation of the whole RV, particularly the RV outflow tract in a single volume.^{105,106,108,114} Intraobserver, interobserver, and test–retest reliability varies among studies, but is generally acceptable (Table 4).^104,106,107 In children with hypoplastic left heart syndrome, 3DE has good reproducibility during serial follow-up.¹¹⁵ However, 3DE measurements are a mean 30% lower than MRI with larger differences in smaller patients, so the techniques cannot be used interchangeably.¹¹⁶

Most published RV data utilize full-volume 3DE datasets acquired over several cardiac cycles and normal adult ranges.¹⁰⁰ Recently, 3DE RV volumes and function has been studied using single-beat acquisition in adults.¹¹² Feasibility was very good (96.7%) as were correlation and agreement with MRI; however, the reduced temporal resolution remains a concern for application to younger patients with higher heart rates.

Knowledge-based 3D reconstruction

Knowledge-based 3D reconstruction evaluates 3D RV volumes from a series of 2DE images localized using a magnetic tracking system (Figure 14).¹¹⁷ RV anatomic landmarks are identified on the images, which are processed over the Internet using a reference lesion-specific MRI database. This technique has been validated against MRI in children after TOF repair.^110,118 Bias, intra- and inter-observer reliability were good, and knowledge-based 3D reconstruction was slightly better than semi-automated border detection method in this regard. In adults with a systemic RV^111,119 and pulmonary arterial hypertension, values have shown good agreement with MRI.^120,121 Limitations of the knowledge-based technique include the necessity for a tracked ultrasound transducer and for the patient to remain still throughout the study.

3DE assessment of the LV

Reliable assessment of LV volume and function is important in patients with CHD. 2DE has important limitations in CHD because, in contrast to 3DE, it makes assumptions of LV shape which are frequently invalid in this population. Thus, 3DE can contribute significantly to the assessment of LV volumes, function, and mass.

Analysis methods

The 3DE dataset is obtained from an apical or modified transducer position to include the entire LV volume, which is usually feasible except when the LV is severely dilated. Current software tracks the endocardium of the LV throughout the cardiac cycle and therefore depends on adequate image quality and acoustic windows. Vendors typically display reference planes used to define the endocardial border (Figure 15) as well as the ‘shell’ of the LV itself (Figure 16). Current tracking algorithms involve user definition of key reference points followed by semi-automated tracking of the endocardium, but the operator can manually override the initial automated selection of the endocardial border.¹²²

Despite difficult endocardial delineation and heterogeneous LV shapes, accurate determination of LV volume and function by 3DE has been reported in adults and children with CHD using MRI as the gold standard. Potential errors can still occur in children who have higher heart rates and smaller LV volumes compared with adults, which is offset by superior image quality. 3DE compares favourably with MRI but has a bias to producing lower LV EDV and ESV compared with MRI in patients with CHD in a meta-analysis of 3055 subjects from 95 studies.¹²³ The repeatability of 3DE for estimation of LV volume is good and supports its role for serial follow-up, although the values are not interchangeable with those obtained by MRI in the CHD population. Validation studies in children and adults with CHD are listed in Table 5.^{98,113,125,126,129,130} 3DE is more accurate and reproducible than either M-mode or 2D Simpson's biplane method and is as feasible as 2DE in children.^127,131 These findings have also been validated in neonates and infants where 3DE has compared favourably with MRI even when LV volumes are small and heart rates are high.¹¹³

LV mass

3DE has been used to assess LV mass in children and compared with both 2DE and M-mode.¹³² In younger patients, M-mode has remained the most widespread technique because of the availability of normal paediatric ranges.^133,134 3DE methods are based on calculating the endo- and epicardial volumes, which are subtracted to compute the mass. 3DE values correlate well with MRI, with low inter- and intraobserver variability¹³⁵ but the limits of agreement remain wide.^{98,113,125,129} Therefore, the clinical application of 3DE for LV mass calculation in patients with CHD remains to be established.

3DE assessment of LV intraventricular dyssynchrony

The capability of 3DE to capture the entire LV volume offers the opportunity to assess global and regional LV function (Figure 16). Ventricular dyssynchrony is expressed as the standard deviation of the time taken for segments to reach their minimum systolic volume, indexed to the cardiac cycle length [Systolic Dyssynchrony Index (SDI)].¹³⁶ Normal values of SDI in children and adolescents are lower than the adult population.^137,138 3DE estimates of LV dyssynchrony appear to be most repeatable if a 16-segment model is used as opposed to 12 or 6 segment models.¹³⁹ 3DE has been used to demonstrate increased LV dyssynchrony in children with Kawasaki disease¹⁴⁰ as well as dilated cardiomyopathy where the 16-segment SDI correlated negatively with 3DE EF and 2DE fractional shortening.¹⁴¹ 3DE has also demonstrated LV dyssynchrony in patients with tricuspid atresia after Fontan palliation¹⁴² and a high incidence of LV regional wall motion abnormalities in CHD patients.¹⁴³ Current software packages define abnormal wall motion with respect to the central LV axis which has limitations for some CHD patients with an LV of unusual shape. Further difficulties in measurement of dyssynchrony occur where LV function is poor due to difficulties in accurate determination of the point of minimum systolic volume in segments where such curves are of low amplitude.¹⁴⁴ Caution is required when using 3DE as a modality to quantify electro-mechanical dyssynchrony in CHD in the absence of a significant body of evidence.

3D wall tracking of the left ventricle

Recent advances in 3D wall tracking have allowed for assessment of myocardial deformation in three dimensions from a single volume of the LV (Figure 17). The 3D technique has a potential advantage over 2D strain in that loss of tracking due to through plane motion can be avoided.¹⁴⁵ This permits a semi-automated analysis of longitudinal, radial, circumferential, and 3D strain from a single volume. In addition, LV volume, EF, and LV twist and torsion can be computed from the same volume. The major limitations of this technique remain temporal resolution and feasibility of incorporating the entire LV. Recent publications have reported normal values in children and adolescents,^146,147 but this currently remains a research tool, and the place of such analysis in the management of patients with CHD remains to be established.

Assessment of the functional single ventricle

Ventricular dysfunction is an important long-term complication in patients with functionally single ventricle circulation. 3DE based on disc summation appears feasible and compares favourably with MRI.^18,124 Semi-automated border detection has been applied to the serial assessment of RV volumes and EF in hypoplastic left heart syndrome¹¹⁵ but with systematic underestimation of RV volume compared with MRI.¹¹⁶ The differences may reflect the tendency for semiautomatic methods to skim the surface of trabeculations, geometric limitations, and failure to include the entire ventricle by 3DE.

Future directions

Technical solutions such as the fusion of multiple 3DE volumes may assist in acquiring the entire ventricular cavity for the purposes of RV functional assessment.¹⁴⁸ The availability of the semi-automated method of disks algorithm for RV volumetry is desirable, as is the availability of analysis software that is purpose-designed, including the facility to landmark structures in normal and abnormal hearts. The availability of large studies to define normal RV and LV volumes and mass across a wide range of body size would enable these measurements to replace 2D measurements in clinical practice. The availability of automated, reproducible algorithms that provide for geometry-independent volumetric analysis of congenitally abnormal ventricles would enhance application in serial follow-up.

Recommendations

Assessment of ventricular volumes and function in CHD

Measurement of ventricular volumes and EF by 3DE is highly repeatable.

3DE has a systematic bias to produce lower values of volume than MRI so that measurements should not be used interchangeably.

Current implementation into clinical practice is hampered by a paucity of normal data from infancy through to adulthood.

Software developed for the normal LV or RV should not be applied to patients with congenitally abnormal ventricles without being validated.

The semi-automated summation of disc method is the least constrained by geometry and should be included with post-processing software particularly for ventricles of abnormal morphology.

3D echocardiographic assessment of atrioventricular and arterial valve function in CHD

3DE assessment of AV valves in CHD is one of the most frequently used applications of the technology in a clinical setting.^149,150 3DE provides comprehensive information on AV valve size, morphology, and motion. The application of 3D colour flow Doppler can be used to visualize regions of AV valve regurgitation including the size, shape, and number of regurgitant orifices.

The AV valves

Assessment of AV valve anatomy

3DE assessment of AV valves in CHD is driven by the need for better visualization and understanding of the mechanism of the valve regurgitation during surgical planning. AV valve function is a complex interaction of atrial contraction;¹⁵¹ annular shape, contraction, and motion;^152,153 leaflet size, shape^154,155 and the changes in its tensor strength,^156,157 chordal length and tension as well as papillary muscle contraction and position.^158,159 Ventricular coordination and contraction^160,161 also impact on AV valve function. 3DE assessment focuses on the valve annulus, leaflets, chords and papillary muscles, as these are the AV valve components that can be manipulated or repaired by current surgical techniques. Although quantitative 3DE assessment of the AV valve in CHD has been reported for AV septal defects^159,162 and tricuspid valves in hypoplastic left heart syndrome,^158,163 it is still not part of routine clinical practice. Commercially available quantitation software frequently assumes normal AV valve morphology and acquired disease, making it less valid for use in CHD. Currently, the assessment of AV valve regurgitation by 3DE remains largely qualitative⁴¹ and is tailored to the individual valve anatomy. There are no data on the impact of 3DE on clinical outcomes in CHD but the technique may impact on the surgical approach.¹⁶⁴

Quantitation of AV valve regurgitation

Assessment of AV valve regurgitation severity by 2DE has been demonstrated to guide management in the adult population¹⁶⁵ despite significant interobserver variation.¹⁶⁶ 2DE guidelines have been produced for quantitative assessment of valvular regurgitation in adults^167,168 but not children.²⁴ Poor reproducibility and accuracy result from morphologically abnormal valves, complex regurgitant jet morphology with an elliptical or linear regurgitant orifice and multiple jets, and effective regurgitant orifice area (EROA) calculation has not demonstrated major benefit compared with qualitative assessment.^169–171 EROA reference values relative to patient or valve size are not available, which is a significant limitation in the interpretation of results in growing patients with congenitally abnormal valves. 3DE methods analogous to MRI for estimation of valve regurgitation are encouraging^172–174 but are limited by the requirement of no significant interventricular shunts. 3DE quantification of AV valve regurgitation vena contracta area in CHD is a promising approach, is relatively simple to measure, and is available on most 3DE navigation software. After acquiring a 3DE colour mapping dataset, the clinician can navigate within the MPR mode to directly measure the vena contracta area perpendicular to the regurgitant jet. Multiple validation studies in the adult population have shown strong correlation between 3DE vena contracta area of mitral regurgitation and MRI calculated EROA and regurgitation fraction with high reproducibility.¹⁷⁵ Further validation in paediatric and CHD patients is still required.

Quantitation of aortic and pulmonary valve stenosis and regurgitation by 3DE

The aortic valve

The central role of echocardiography in the investigation of aortic valve disease is well established in both paediatric/congenital and adult practice.^24,176,177 As previously discussed, 3DE can define valve morphology and estimate LV volume and mass. Doppler-derived pressure drop across valves is flow dependent, and accurate measurement of effective aortic valve orifice area is clinically important. This is particularly important in children where use of the continuity equation has not been recommended because of the potential for measurement error.²⁴ MPR images can be adjusted to ensure an en face orientation to assist in direct measurement of the effective aortic valve area in children with aortic valve stenosis.^21,178 ‘Cut off’ values of area in children cannot be used because this is impacted by the size of the aortic valve, which is in turn related to the size of the child. Previous published work has addressed the normal aortic valve area across a wide age and size range,¹⁷⁹ but this normative data used a rotational probe rather than current real-time 3DE techniques. Thus, it is feasible to planimeter the effective aortic valve area in the paediatric population, but data on the impact of such measurements on timing of intervention or prognosis is lacking.

Echocardiographic quantification of aortic valve regurgitation has been addressed in several documents,^24,176,177 and standard indices include LV dimensions, regurgitant jet or vena contracta diameter, and diastolic reversal of flow in the aorta. Recent adult work has used 3DE colour estimation of the size of the 3DE-derived vena contracta against 2DE and MRI with encouraging results.^180,181 This technique has not been validated in paediatric or congenital populations. Surgical planning for patients with aortic valve disease involves quantitation of aortic root size, with evidence of improved accuracy using 3DE in children⁵¹ and validation of automated 3DE in adults.¹⁸²

The pulmonary valve

Assessment of pulmonary valve stenosis in children and adults with CHD is normally based on Doppler-derived pressure drop across the valve.²⁴ Obtaining an en face image of the pulmonary valve by 2DE may be challenging, which hampers definitions of stenosis based on effective pulmonary valve orifice area. 3DE may assist in projecting the pulmonary valve en face either by real-time 3DE, by cropping a multi-beat full volume, or by utilization of cross-plane techniques. Pulmonary regurgitation is an important complication of many patients with CHD, particularly those with repaired TOF or with conduits from the RV to the pulmonary artery. Recent publications^54,114,183 emphasize the importance of quantifying the size and function of the RV by 3DE. The estimation of pulmonary regurgitation by 3DE remains semiquantitative; although there are published data using 3DE to measure EROA,¹⁸⁴ it has not been adopted into CHD practice and has not been validated against cardiac MRI.

Recommendations

Atrioventricular valves

The use of 3DE to visualize AV valves, papillary muscles, and chordal support is recommended to assist surgical planning.

3DE assessment of the location, size, shape, and number of regurgitant jets and orifice area is recommended.

Measurement of EROA needs to be validated in children and adolescents and in a wide range of valve pathology before its measurement can be recommended to quantify AV valve regurgitation and impact timing of valve repair or replacement.

Post-processing software for AV anatomy and valve function needs to be used with caution as it may assume normal valve anatomy.

Arterial valves

3DE is recommended to assess the morphology of the aortic and pulmonary valves.

3DE is recommended for the measurement of annulus, root and effective orifice area measurement.

3DE colour flow Doppler measurements remain to be validated as a means of quantifying arterial valve regurgitation.

Training in three-dimensional echocardiography

Requirements and recommendations for training have been an integral part of position statements and guidelines for echocardiography in CHD.^185–187 No such proposals have been produced for 3DE of CHD, and the European certification process for echocardiography of CHD does not currently include 3DE at all.¹⁸⁸ Training and education play a critical role in increasing the utility of 3DE, particularly in terms of its ability to improve outcomes.¹⁸⁹ There is no question that the adoption of 3DE into clinical routine has a learning curve and demands specific training.

Machines and transducers for 3DE are similar to those for 2DE. There are some fundamental differences in the approach to learning 2DE vs. 3DE. In 2DE, the only image that can be interpreted is the acquired 2D image. A central component of training is the appropriate acquisition of images in planes that are largely predefined. In contrast, 3DE involves acquisition of a 3D dataset with extensive potential for post-processing. Training in 3DE must involve methods to develop and improve the ability to acquire optimal 3DE datasets and to post process the sonographic volumetric datasets to show cardiac morphology or quantify ventricular function. Recognition and evaluation of en face views of valves and septums must be learned as well as other details of complex intracardiac anatomy uniquely available through 3DE. While these structures and views are familiar to surgeons, pathologists, and others who have an implicit understanding of cardiac anatomy, they pose a challenge to those who do not have such a frame of reference. One approach to helping bridge this gap has been the development of educational programmes and publications that include side-by-side depiction of pathology specimen photographs and 3DE images for a wide range of CHD.⁶⁰

Training in acquisition has, of necessity, involved live scanning of subjects in a classroom or clinical setting. Although it provides an authentic learning experience, this mode of learning has practical limitations, including class size, availability of cooperative subjects with a wide spectrum of disease states, busy clinical settings, and issues related to privacy. Over the past few years, simulators to teach 2D TOE and TTE have become available but 3DE simulators are lacking.^190,191

Training in post-processing has primarily occurred in the format of workshops involving manipulation of datasets from a wide range of pathologies.¹⁹² This type of training requires high-quality 3DE datasets of a variety of CHD, a large number of portable computers and vendor-specific software. One of the challenges of 3DE training has been the high frequency of modifications of the user interface for both acquisition and post-processing.¹⁸⁹ While this is to be expected in the evolution of the modality, frequent changes of user interface may discourage adoption of 3DE. Stability of the user interface and consistency between manufacturers would help greatly in developing robust tools for training.

A further consideration is the wide variability in the use of 3DE between centres and adoption of 3DE within certification and training programmes. There are no specific guidelines for education related to 3DE in the published North American guidelines for core fellowship training in paediatric cardiology,¹⁸⁵ and the revised guidelines include only vague references to 3DE experience for advanced fellowship training in non-invasive imaging.¹⁹³ 3DE is not a current component of the European certification process in echocardiography of CHD. The absence of specific training standards for 3DE has resulted from the absence of guidelines and standards for the performance and interpretation of 3DE studies. This document should provide a framework for more structured 3DE training in the future.

With respect to what form such training should take, attendance at a formal training course in acquisition and post-processing of 3DE is essential to understand different 3D modalities and post-processing software. Sonographers and cardiologists in training should continue to establish their core skills in 2DE according to established training programmes. For trainees subspecializing in echocardiographic imaging, many should be able to gain the further experience in data acquisition and post-processing within their unit, provided 3DE is being widely used. Assessment should be competence based but acquisition of and post-processing of at least 50 volumes (mixed TTE/TOE) of different forms of CHD would seem reasonable. Competence should include assessment of cardiac morphology as well as quantitation of LV volume and EF. If a centre does not offer exposure to a high volume of 3DE then a period of training of 3–6 months at a high-volume centre would be appropriate. For the advanced practitioner, application of 3DE in the catheterization laboratory and during surgery requires advanced and rapid acquisition and post-processing skills and will generally require more experience and training than that outlined above. Application of 3DE to the RV or abnormally shaped ventricles will also require further experience and training. Trainers in 3D echocardiography will be drawn from cardiologists with many years of experience of acquisition and post-processing of structure and function using 3D echocardiography. It is hoped that the recommendations in this document, for examples those relating to image orientation, may assist in development of structured programs both for trainees and trainers.

Future directions

Echocardiography simulators should be modified to include training in 3DE acquisition. The development of a comprehensive simulation-based educational resource would require a large library of 3DE datasets encompassing a wide range of pathologic states. Training in post-processing should involve web-based access to 3DE analytic software in conjunction with an online library of 3DE datasets of various CHD lesions. In order to develop such libraries, anonymization tools for Digital Imaging and Communications in Medicine (DICOM) datasets including removal of both DICOM headers and patient data displayed in the image are essential.

Recommendations

Training in 3DE

Training in 3DE should be an integral part of the training of the congenital echocardiographer. This should include the indications for and added value of 3DE in surgical and interventional planning and guidance.

Training should include 3DE dataset acquisition and post-processing to demonstrate competence in assessment of cardiac morphology and quantification of cardiac function.

Operators should learn to recognize and evaluate cardiac views which are uniquely feasible by 3DE including en face views of valves and septums.

A degree of consistency of the user interface and uniformity of terminology between manufacturers would help greatly in developing robust non-proprietary tools for training.

Advances in three-dimensional echocardiography

While the past decade has brought exciting advances to the fore in 3DE, there are many aspects that would benefit from improvements in technology that are tailored to the patient with CHD. We look forward to advances in transducer design and computer post-processing as well as research into novel methodologies, such as frame reordering¹⁹⁴ and image compounding,^148,195,196 which can achieve benefits in terms of field of view, endocardial border definition, enhanced temporal resolution, and intraventricular flow. Advances in 3D printing^197,198 and in holographic displays promise the ability to view, understand, and utilize 3DE data better in clinical practice.

Novel measurements of ventricular volumes, 3D deformation, valvar morphology, valvar function, and flows should be validated, automated, and lead to publication of robust normal ranges across a wide range of ages and body size. These enhancements should not be limited to left heart structures but need to be comprehensive to include the right heart and congenitally abnormal ventricles and valves. All of these measures need to undergo robust assessment of their repeatability and accuracy vs. other imaging techniques.

The assessment of the impact of 3DE techniques on patient outcome remains a challenge. Many current 3D techniques have been adopted in an ad hoc manner to address specific clinical challenges such as surgical repair of CHD or guidance of interventional procedures but without robust analysis of impact on outcome. Quantitation of ventricular volumes, myocardial deformation, and cardiac function by 3DE needs to stimulate research studies that demonstrate whether or not 3DE measures can be used as surrogates for outcome.

The technical advances and increased use of 3DE techniques in clinical practice will place greater emphasis on improved methods for training and education in 3DE. This needs to be matched at the research and development stage by improvements in the software interface for acquisition and analysis of 3DE including, where appropriate, increased automation to improve workflow and reduce observer error.

Reviewers

This document was reviewed by members of the 2014–2016 EACVI Scientific Documents Committee, and external reviewers.

EACVI reviewers included: Prof. Gilbert Habib, Prof. Thor Edvardsen, Dr. Vittoria Delgado, Prof. Erwan Donal, Dr. Denisa Muraru, Associate Professor Kristina Haugaa, Dr. Raphael Rosenhek, Dr. José Ribeiro, Dr. Xavier Iriart.

This document was reviewed by members of the 2015–2016 ASE Guidelines & Standards Committee, the 2015–2016 ASE Board of Directors, and external reviewers.

ASE reviewers included Deborah A. Agler, RCT, RDCS, FASE, Federico M. Asch, MD, FASE, Merri L. Bremer, EdD, RN, EDCS, ACS, FASE, Benjamin Byrd, MD, FASE, Hollie D. Carron, RDCS, FASE, Frederick C. Cobey, MD, FASE, Meryl Cohen, MD, FASE, Patrick Collier, MD, PhD, FASE, Mary Corretti, MD, FASE, Adam Dorfman, MD, FASE, Benjamin Eidem, MD, FASE, Fadia Makarem Ezzeddine, RT, RCS, FASE, Craig Fleishman, MD, FASE, Neal Gerstein, MD, FASE, Yvonne E. Gilliland, MD, FASE, Lanqi Hua, RDCS, FASE, Allan L. Klein, MD, FASE, Joe R. Kreeger, ACS, RCCS, RDCS, FASE, Roberto M. Lang, MD, FASE, Stephen H. Little, MD, FASE, Sunil Mankad, MD, FASE, Rick Meece, RDCS, RCS, RCIS, FASE, Tasneem Naqvi, MD, FASE, Maryellen H. Orsinelli, RN, RDCS, FASE, Andy Pellett, PhD, RCS, RDCS, FASE, Patricia A. Pellikka, MD, FASE, Michael Quartermain, MD, FASE, Vera H. Rigolin, MD, FASE, Brad J. Roberts, ACS, RCS, FASE, Vandana Sachdev, MD, FASE, Anita Sadeghpour, MD, FASE, Elaine Shea, ACS, RCS, RCCS, FASE, Roman M. Sniecinski, MD, FASE, Raymond F. Stainback, MD, FASE, Cynthia Taub, MD, FASE, Neil J. Weissman, MD, FASE, Susan E. Wiegers, MD, FASE.

Notice and disclaimer: This report is made available by EACVI and ASE as a courtesy reference source for members. This report contains recommendations only and should not be used as the sole basis to make medical practice decisions or for disciplinary action against any employee. The statements and recommendations contained in this report are primarily based on the opinions of experts, rather than on scientifically verified data. EACVI and ASE make no express or implied warranties regarding the completeness or accuracy of the information in this report, including the warranty of merchantability or fitness for a particular purpose. In no event shall EACVI or ASE be liable to you, your patients, or any other third parties for any decision made or action taken by you or such other parties in reliance on this information. Nor does your use of this information constitute the offering of medical advice by EACVI/ASE or create any physician–patient relationship between EACVI/ASE and your patients or anyone else.

References