Three-dimensional echocardiography in adults with congenital heart disease: a scientific statement of the European Society of Cardiology Working Group on Adult Congenital Heart Disease and the European Association of Cardiovascular Imaging of the European Society of Cardiology

Abstract

Congenital heart disease (CHD) is the most common inborn defect affecting up to 1% of newborns. The prevalence of CHD has shifted from childhood to adulthood, and the number of adult patients living with CHD continues to increase. This patient population presents unique challenges in diagnostic imaging and management due to complex underlying cardiac morphology, previous operations and interventions, and haemodynamic conditions. 3D echocardiography (3DE) has significantly improved our understanding of complex anatomic and haemodynamic substrates and emerged as a clinically useful tool that provides incremental information and complements the routine echocardiographic examination. The advantages of 3DE, including more accurate visualization of anatomic structures, absence of geometrical assumptions regarding shape of cardiac structures, and ability to obtain a complete view of the structures of interest from multiple perspectives in a beating heart, are especially relevant for diagnosis and follow-up of CHD in adult population, as well as interventional and surgical planning and guidance. In this scientific statement, we provide detailed and simple-to-follow descriptions of the added value of 3DE in evaluation of specific cardiac structures encountered in CHD, its role in diagnosis and follow-up, and training requirements for proficiency in 3DE in adult CHD.

Introduction

Congenital heart disease (CHD) is the most common inborn defect affecting up to 1% of newborns. The prevalence of CHD has now shifted from infancy/childhood to adulthood, and the number of adult patients living with CHD continues to increase due to significant advances in early diagnosis and management.¹ Residual or post-surgical lesions and haemodynamic abnormalities are common in this patient population, and adults with CHD (ACHD) present unique challenges in diagnostic imaging and management due to complex cardiac anatomy and haemodynamics.

The underlying congenital defect and the history of surgical and/or interventional procedures require multimodality imaging techniques for accurate diagnosis, follow-up, and optimal timely management, as symptoms may appear late.² Although 2D echocardiography (2DE) remains the core imaging modality for CHD, 3D echocardiography (3DE) has significantly improved our understanding of complex anatomic and haemodynamic substrates. 3DE has emerged as a clinically useful tool that provides incremental information and complements the routine 2DE examination. Unique advantages of 3DE, such as more accurate visualization of anatomic structures that is obtained without inherit limitations of 1D or 2D techniques, freedom from geometrical assumptions, and ability to provide complete view of the heart from multiple perspectives, are especially relevant in ACHD diagnosis, risk stratification, interventional and surgical planning, and guidance. We aim to provide herewith the scientific statement on the applications of 3DE in ACHD, the added value of 3DE in evaluation of specific cardiac structures encountered in ACHD, its role in patients’ management, and training requirements for proficiency in 3DE in ACHD.

3DE imaging in ACHD

The crucial milestones in the history of 3DE have been the development of a fully sampled matrix array transducer for 3D transthoracic echocardiography (TTE) and later—for transoesophageal echocardiography (TOE) to enable real-time 3D imaging.³ Continuous developments of 3D TTE transducer technology over the years included increased ultrasound penetration and resolution, lighter probes for daily routine scanning with a smaller footprint and better access through narrow rib spaces, higher frequency (up to 2.5/5 MHz) harmonics, improved 3D image quality, and higher frame rates with 3D TTE and 3D colour imaging. Suboptimal acoustic window (for example, due to body habitus or lung disease), however, remains an important limitation of 3D TTE. The trade-off between spatial resolution, temporal resolution, and volume size has been optimized to allow the acquisition of larger data sets with sufficient volume rate in real time. With both 2DE and 3DE modalities being available in the same probe, 3D data set acquisition can now be easily implemented into a standard TTE or TOE examination of ACHD patients with challenging underlying anatomy and haemodynamics (Table 1). These recent developments underscore the significant advantages of incorporating 3DE into the clinical management of patients with CHD, leading to improved diagnostic accuracy, better surgical planning, and enhanced guidance during interventional procedures,^4–9 and address many limitations of conventional 2DE.

Figure 1 illustrates the proposed algorithm for optimal use of 3DE in ACHD.

Added value of 3DE for assessment of specific cardiac structures in ACHD

Ventricles

3DE assessment of the LV

Morphology

Assessment of ventricular morphology is an essential component of echocardiographic examination in ACHD. Real-time or full-volume rendering 3DE can provide detailed assessment of ventricular anatomy and help to identify key morphological features of the left ventricle (LV), such as fine apical trabeculations, presence of two papillary muscles, and specific morphological characteristics of the mitral valve included into the inlet portion of the LV.^10,11 (Supplementary data online, Video S1). Additionally, it is a useful tool to visualize pathological features in the LV (e.g. outflow tract lesions, crypts/diverticula, abnormal position or number of the papillary muscles, and mitral arcade).

LV volumes and EF

Accurate assessment of LV volumes and function is crucial for clinical diagnosis, decision-making, risk stratification, and predicting outcomes in ACHD. LV ejection fraction (EF) calculated by 2DE has significant limitations, because it relies on assumptions regarding LV shape, which are frequently invalid in the ACHD population. Due to frequently present right ventricular (RV) dilatation or hypertrophy, the LV does not always have a ‘bullet’ shape in patients with CHD.^10,12 3DE is free of geometric assumptions, less affected by LV foreshortening due to the possibility to re-align the planes and adjust the LV size to its maximum longitudinal axis, less labour-intensive, more reproducible, and accurate than conventional 2DE.^13,14 It can contribute significantly to the assessment of LV volumes, function, mass, and its routine incorporation into clinical practice is advised.¹⁵

Data acquisition and analysis

The 3DE full-volume data set is usually obtained from apical four-chamber view; however, modified transducer position is also acceptable. Care should be taken to include the entire LV in the 3DE volume. Commercially available LV endocardial tracking algorithms should be used offline during post-processing for volumetric analysis. Semi-automated tracking algorithms involve user definition of key reference points in the LV followed by semi-automated tracking of the endocardium.¹⁶ Fully automated method applies artificial intelligence in the process of LV chamber segmentation and placing endocardial boundaries in the LV for volumetric analysis.¹⁷ Both methods provide opportunity to manually adjust the endocardial border, if necessary. In the context of CHD, if the LV is in systemic position and there is no significant distortion of its shape, the fully automated method can provide fast and accurate quantification; however, in the presence of unusual LV shape (e.g. compressed systemic LV due to RV volume or pressure overload or in patients with the LV in sub-pulmonary position; Figure 2; Supplementary data online, Video S2), manual editing of the LV endocardial contours is essential to ensure accurate evaluation. Published studies showed good correlation of 3DE-derived LV volumes and EF with cardiac magnetic resonance (CMR) measurements in patients with CHD.^16,18,19 Thus, in a population of 32 CHD patients who had same-day evaluation by 3DE with LV manual border detection and CMR, a correlation between 3DE and CMR for LV end-diastolic volume (EDV) was r = 0.97, for LV end-systolic volume (ESV) r = 0.98, and for LV EF r = 0.94. Feasibility of 3DE data acquisition was 91%.¹⁶ A good correlation between LV stroke volume measured by 3DE and by 2D phase-contrast CMR has been described recently in a cohort of 83 adult patients with atrial septal defects (ASDs) (r = 0.73).²⁰

Published meta-analyses demonstrated that 3DE-derived LV volumes were underestimated compared with CMR, whereas LV EF revealed excellent accuracy.^13,14 Pooled analysis of 23 studies (including 1638 echocardiograms of patients with acquired and CHD) demonstrated pooled biases for 3DE- vs. CMR-derived LV EDV −19.1 ± 34.2 mL, LV ESV −10.1 ± 29.7 mL, and LV EF −0.6 ± 11.8%. Importantly, when the 2DE-derived LV volumes and EF were compared with CMR, they demonstrated significantly higher biases. Authors concluded that compared with traditional 2D methods, 3DE is more accurate for LV volumetric assessment and more precise in all three measurements.¹⁴ This, however, highlights the importance of using method-specific reference values for the LV volumes (Table 2).¹⁵

A recent study in a small cohort of patients with CHD with various forms of LV (mean age 18.7 years) compared two different 3DE software packages for LV mass quantification vs. CMR. Despite overestimation of 3DE-assessed LV mass by 5–10%, results from both 3DE software demonstrated high correlation with the gold standard technique (all r > 0.97).²¹ Similarly, a very good correlation was observed between 3DE and CMR for LV mass in a study of 20 adult patients with CHD and at least moderate 3DE image quality (r = 0.98).²²

When comparing volumetric assessment, it is essential to acknowledge that 3DE and CMR are fundamentally different imaging modalities, each based on distinct physical principles. CMR is considered the ‘gold standard’ for volume quantifications, and direct comparison of absolute volume values between 3DE and CMR cannot be performed. The measurements obtained from each modality are method specific, meaning that reference values and normal ranges should be interpreted within the context of the imaging technique used and they are not interchangeable.

3DE assessment of LV intraventricular dyssynchrony

Ventricular dyssynchrony is an important form of ventricular dysfunction and is very common in ACHD due to previous surgical repair, conduction disorder, and/or ventricular–ventricular interaction. The capability of 3DE to capture the entire LV volume offers the opportunity to assess global LV dyssynchrony, which is expressed as the standard deviation of the times taken by LV segments to reach their minimal systolic volume, indexed to the cardiac cycle length, also called ‘Systolic Dyssynchrony Index’.²³ Similar to 3D speckle tracking, 3DE assessment of the LV dyssynchrony currently remains a research tool, and more prospective studies are needed to establish its diagnostic and prognostic role in ACHD population.

3DE assessment of the RV

Morphology

3DE full-volume data sets incorporate all three components of the RV (inflow, apical portion, and outflow) and allow accurate morphological assessment of the RV and right-sided valves and detailed analysis of RV size, shape, wall motion, function, and contraction patterns (Figure 3; Supplementary data online, Videos S3 and S4). This approach also provides 3D morphological assessment in the beating heart, which is unique compared with other imaging modalities and useful for assessment of dynamic changes in the RV, such as dynamic RV outflow tract (RVOT) obstruction or double-chamber RV (Figure 4; Supplementary data online, Videos S4 and S5).

Assessment of RV volumes and EF

3DE is the only echocardiographic technique providing direct measurements of RV volumes and EF.¹⁵ 3DE-derived RV volumes and EF closely correlate with CMR-measured parameters in both children and adults with different cardiac conditions.^24–28 Similarly to LV, 3DE underestimates RV volumes compared with CMR²⁸; hence, method-specific reference values should be used when quantifying the RV by 3DE (Table 2), and CMR and 3DE reference values are not interchangeable.^29,30 In a meta-analysis aimed to assess agreement of different imaging modalities for evaluation of RV EF compared with the gold standard technique CMR, 3DE has overestimated RV EF only by 1.16% vs. CMR.³¹

The most recent systematic review and meta-analysis aiming to validate real-time 3DE against CMR based on pooled data from 518 patients with various cardiac conditions also provided promising results: the pooled mean differences for RV EDV and ESV and RV EF between 3DE and CMR were −5.48 mL [95% confidence interval (CI) −9.33, −1.62 mL], − 0.78 mL (95% CI −5.63, 4.07 mL), and −0.20% (95% CI −1.22, 0.82%), respectively. The differences between 3DE and CMR for ESVs and EF were non-significant.³²

In a small study of patients with repaired tetralogy of Fallot (TOF), 3DE-derived RV EF demonstrated significantly better agreement with CMR-derived EF compared with conventional echocardiographic parameters of RV systolic function such as tricuspid annular plane systolic velocity (TAPSE), myocardial performance index, or S’.³³

Current guidelines for chamber qualification recommend 3DE-derived RV EF > 45% as the lower limits of normal values and specify 3DE as a method of choice for assessment of RV volumes and systolic function.¹⁵

Several studies specifically focused on patients with CHD. In patients with repaired TOF and pulmonary stenosis post-valvotomy, a good correlation has been demonstrated between 3DE and CMR for measuring RV volumes.^33–35 Although 3DE tended to underestimate the RV volumes, when method-specific reference values were used, all patients with dilated RV according to CMR were identified correctly by 3DE.³⁴ A study including unselected patients with complex CHD demonstrated similar results with 3DE underestimating RV EDV and ESV by 34 ± 65 and 11 ± 55 mL, respectively, compared with CMR.³⁶

Good correlation has been also demonstrated for RV EF in ACHD population, such as in patients with repaired TOF³⁵ or in patients with dilated RV due to ASD.³⁷  Table 3 provides recently available data on estimation of RV volumes and EF by 3DE in CHD and correlation with CMR (while studies published before 2016 have been summarized in the expert consensus document by Simpson et al.¹⁰).

Training and experience are essential components for correct data acquisition and post-processing to ensure optimal results of RV quantification by 3DE. Several steps during data analysis (such as proper alignment of the data set, scrolling and manual checking the borders between blood and myocardium, counting only the compact portion as myocardium and the rest as the blood pool, and necessity of manually correcting the 3D model created by the software to obtain optimal results) should be followed.^39,41 This strategy leads to high accuracy of 3DE in the quantification of RV volumes compared with CMR.^26,39

3DE-derived RV EF is significantly associated with cardiac and all-cause mortality and major adverse cardiac events in patients with various cardiovascular conditions.^27,42–45 A recent meta-analysis demonstrated that reduction in 3DE-derived RV EF showed stronger association with adverse clinical outcomes than conventional RV functional indices and might further refine the risk stratification of patients with cardiac diseases.⁴⁶ In CHD, several 3DE-derived parameters demonstrated strong predictive value in identification of patients with repaired TOF who have impaired exercise capacity as assessed by cardiopulmonary exercise test.⁴⁷ More data, however, need to be generated on prognostic role of 3DE in ACHD.

Limitations of 3DE in ACHD population include potential incomplete visualization of the whole chamber in case of severe RV dilation and challenges in endocardial delineation in case of hypertrophied/heavily trabeculated RV walls (e.g. systemic RV). Nevertheless, several studies including unselected patients with complex CHD demonstrated good feasibility and shorter data acquisition and analysis time compared with CMR³⁶ (Table 3). In case of challenging apical windows, full-volume 3DE data set for the RV can be acquired from subcostal view (Figure 5; Supplementary data online, Videos S6 and S7). A study in CHD patients has demonstrated that for apical and subcostal views, 3DE-derived ventricular volumes agree well with CMR.⁴⁰

Assessment of the relative contribution of different components of the RV systolic function

Further post-processing of 3DE RV data set enables quantification of the relative contribution of longitudinal, radial, and anteroposterior motion components of the RV to the global RV systolic function and provides detailed characterization of RV contraction pattern. The relative reduction in radial component (inward motion of the RV free wall) can be an early sign of RV pressure overload.⁴⁸ In contrast, in RV volume overload, the longitudinal RV component is more impaired than either the radial or the anteroposterior components.⁴⁹ In patients after cardiac surgery, there is a temporal reduction in RV longitudinal shortening; however, the RV EF usually remains maintained due to compensatory increase in the radial motion. Data on the prognostic role of different components of RV systolic function in congenital population are scarce.

RV shape analysis

3DE also allows quantification of global and regional RV shape indices based on analysis of the RV curvature.^50,51 In patients with pulmonary hypertension and RV pressure overload, the curvature of the RV inflow tract was a more robust predictor of mortality than 3DE-derived RV EF, volumes, or other regional curvature indices.⁵⁰ The reference values of 3DE regional curvature indices are currently available.⁵²

3DE assessment of the systemic RV

3DE is a useful tool for the assessment of the size and systolic function of the systemic RV^53,54 (Figure 6). Although data acquisition and analysis may be challenging due to systemic RV dilatation and heavy trabeculation, feasibility may reach 70% in experienced centres.^55,56 It has been demonstrated that 3DE-derived systemic RV EF strongly correlated with BNP level and this correlation was higher compared with other conventional echocardiographic parameters including fractional area change, TAPSE, and S’.⁵⁵ Data on the correlation of 3DE- and CMR-derived volumes and EF of the systemic RV are limited; however, small studies of ACHD patients demonstrated acceptable agreement between the two modalities.^56,57

Assessment of the contraction patterns of the systemic RV using 3DE demonstrated significant differences in RV wall motion components in patients with transposition of the great arteries (TGA) post-atrial switch repair and with congenitally corrected TGA. The anteroposterior component was dominant in TGA post-atrial switch repair, providing compensation for impaired longitudinal and radial components, while in congenitally corrected TGA, all three components contributed equally to the total EF.⁵⁵

3DE assessment of the functionally single ventricle

Morphology

There are many anatomical variations and ventricular morphologies in this group of patients. The majority of adult patients with single ventricular physiology will have Fontan type of operation during childhood; some may only have palliation with arterial shunts. In very rare cases, patients may survive into adulthood with native anatomy and balanced circulation with either pulmonary hypertension or pulmonary stenosis. In experienced centres, 3DE can be helpful in the identification of complex and various morphologies of the univentricular hearts (Figure 7; Supplementary data online, Videos S8 and S9).

Assessment of ventricular volume and function

Accurate assessment and identification of subtle changes in ventricular volumes and function are crucial in the clinical management. Even minor degree of ventricular dysfunction may have a profound effect on the fragile circulation of these patients. Due to variable ventricular shape, semi-automatic tracing method is a preferable solution for assessment of ventricular volumes and EF. Examples of 3DE data acquisition and volumetric analysis of different morphologies of functionally single ventricles are presented in Figure 8. Studies assessing the agreement of 3DE-derived volumes and EF with CMR have limited sample size and mainly focus on paediatric population, demonstrating good correlation of 3DE- and CMR-derived parameters, but underestimation of ventricular volumes and EF when assessed by 3DE, both for morphologically left and right single ventricles.^58–61 3DE-derived EF and volumes in patients with single ventricle demonstrate good reproducibility and correlate with main prognostic cardiopulmonary exercise test parameters, such as VO₂max and VE/VCO₂ slope; however, feasibility remains limited.⁶² Owing to a small patient population and lack of normal values, the 3DE-derived ventricular parameters are the most useful for self-comparison during clinical follow-up.⁶³

Summary

3DE aids in identifying complex ventricular morphologies in CHD and facilitates the assessment of changes in ventricular volumes and function.

3DE is the only echocardiographic technique that enables measurements of RV volumes and EF and shows superior diagnostic and prognostic value compared with conventional 2DE.

3DE should be attempted for ventricular volumes and EF assessment in experienced centres and in patients with reasonable image quality.

Method-specific reference values for LV and RV volumes should be used.

Valves

Mitral valve/left AV valve

Morphology

AV valves are complex units composed of annulus, leaflets, tendinous chords, and papillary muscles. Consider each of these components when assessing valve morphology and function. 3DE provides a depth of field and allows to obtain ‘en-face’ views of the valve and, therefore, enables recognition of crucial morphological features, such as saddle-shaped annulus, curved coaptation line, leaflets scallops, fibrous continuity in between anterior mitral leaflet and aortic non-coronary leaflet, insertion of primary and secondary order chords, and respective position of papillary muscles in between each other (‘chordal projection’),⁶⁴ thus, ensuring a better description for the surgeon and overcoming issues related to geometrical assumptions and interpretation of images (Figure 9; Supplementary data online, Video S10).⁶⁵

Quantification

Assessing the severity of stenosis in congenital mitral valve disease might be challenging due to multiple levels of obstruction and eccentric orifices. 3DE helps to visualize the level of obstruction (supravalvular, valvular, or subvalvular) by measuring area at different levels and evaluating dysplasia of mitral valve components (leaflet thickening, shortening and fusion of chordae, papillary muscle hypoplasia). It can also add quantitative data regarding mitral valve planimetry and the extent of tethering (Figure 10). 3D planimetry has been reported as the most accurate method for mitral valve area evaluation^66–68 and seems to be more reproducible than the 2D area planimetry.⁶⁹

Abnormalities of mitral valve leaflets, chordal apparatus, and papillary muscles can be better visualized by 3DE demonstrating a parachute mitral valve or arcade. In mitral regurgitation, 3DE accurately displays isolated clefts and differentiates them from the zone of apposition within the atrioventricular (AV) septal defect (AVSD) complex. The exact aetiology can be clearly defined by 3DE that has shown greater efficiency than 2DE in detecting defects such as isolated cleft or double orifices mitral valve^70–72 (Figures 10 and 11; Supplementary data online, Videos S11 and S12). Additionally, 3DE provides quantitative evaluation of mitral regurgitation severity, showing superiority in defining eccentric jets and the dimensions of vena contracta. The 3D vena contracta area correlates with the regurgitant volume.⁷³ The 3DE Proximal Isovelocity Surface Area (PISA) technique allows measurement of the 3D surface of the proximal flow convergence region without shape assumptions and, therefore, increases the accuracy of regurgitant orifice area measurement, which correlates well with CMR-derived regurgitant volume.⁷⁴

3D TOE provides higher resolution and superior image quality and can aid left AV valve evaluation, especially in case of suboptimal transthoracic acoustic window and if transcatheter interventional procedures are considered. Finally, 3DE is a promising tool for reconstruction of patient-specific 3D models of the valves, which can be helpful in planning surgical or transcatheter interventions. In Table 4, we provide the main advantages and limitations of 3DE in assessing mitral valve/left AV valve.

Tricuspid valve/right AV valve

For the detailed assessment of the tricuspid valve, the same factors described for the mitral valve should be considered: leaflets, papillary muscles, chordal attachments, and annulus. Tricuspid valve structure, however, is physiologically less constant than its left counterpart; leaflets, as well as papillary muscles, can vary in shape and in number.⁷⁵ Due to the complexity of tricuspid valve geometry and its anterior position in the mediastinum, 3DE is the preferred imaging technique for assessing the anatomy and the pathophysiological mechanisms of valve dysfunction (Figure 12; Supplementary data online, Videos S13 and S14).^76,77 It has been demonstrated that 3D TTE is feasible for the assessment of the tricuspid valve in complex CHD, such as TOF or hypoplastic left heart syndrome, and identifies unique differences in valve morphology.⁷⁸ Future studies are needed to clarify the clinical significance of tricuspid valve morphology in these patient populations.

Simultaneous visualization of all tricuspid valve leaflets is highly feasible using 3DE, unlike 2DE. Due to unlimited rotation of the cropping plane, the ‘en-face’ view of the valve can be obtained in experienced centres even in such complex morphologies, such as Ebstein anomaly. ‘En-face’ views from the RV and atrial perspectives (surgical view) are essential to define the spatial relationship of the tricuspid valve leaflets with the surrounding structures (Figure 13; Supplementary data online, Video S15).⁷⁹ In particular, the ventricular view allows evaluating commissures and coaptation defect.

During semi-quantitative assessment of tricuspid valve function, it is important to keep in mind that lower flow velocities and pressures in the RV can have an impact on proximal convergence and regurgitant jet analysis. Once colour Doppler scale has been optimized and the image and the cropping plane are orthogonally oriented, 3D vena contracta area can be measured with good correlation with EROA.⁸⁰ In this way, multiple jets may be measured and summed together. Moreover, 3DE allows to quantify RV volumes and EF, to provide a better description of haemodynamic impact of AV valve dysfunction.⁸¹

However, there are also limitations to be considered. First, tricuspid leaflets are thinner than mitral ones and obliquely oriented, which may lead to signal dropouts. Second, the fibrous body of the heart as well as any prosthetic material may cause acoustic shadowing during TOE, needing a deeper insertion of the probe, with the drawback of originating ‘non-isotropic voxels’ and thus limiting spatial resolution of structures parallel to the probe.⁷⁶

Common AV valve

A common AV valve is the key feature of AVSDs. This is a broad spectrum of lesions, with significant variability of sizes and relationship of the bridging leaflets with the atrial and ventricular septum. 3DE is the only imaging modality able to provide real-time 3D visualization of leaflets and chordae and their relations with surrounding structures, such as LV outflow tract (LVOT) or ventricular septal defect (VSD) (Figure 14; Supplementary data online, Video S16). Postoperative left AV valve morphology is different from a normal mitral valve: there is an additional third mural leaflet, a clockwise rotation of papillary muscles and LV inlet-to-outlet disproportion.⁶⁴ 3DE can help at different steps in AVSD management: pre-surgical assessment, intraoperative monitoring, early follow-up, and long-term follow-up when it can support the identification of the mechanisms of the eventual failure of the left AV valve.⁶⁵ Moreover, 3DE evaluation of regurgitation is free from geometrical assumptions and therefore can be applied also to these valves. It seems that AVSD without ventricular component (primum ASD) is more prone to develop left AV valve failure after correction; other, uncommon, risk factors include single papillary muscle or closely spaced papillary muscles, short mural leaflet, short tendinous chordae, extremely dysplastic valve or double orifice valve, leaflet prolapse, and anterolateral papillary muscle attachment on the anterior bridging leaflet.^82,83

Aortic valve and left outflow tract lesions

The 3DE provides an accurate assessment of the aortic valve morphology and eliminates any geometric assumptions of the LVOT, aortic annulus, and root measurements.^84–87 Using 3D TTE, the aortic valve can be imaged from both parasternal and apical views. For anatomic assessment of the aortic valve, live 3D or 3D-zoom acquisition modes are the preferred imaging modes.⁸⁸ 3DE, measuring the anatomic areas at different levels, adds particular value in CHD where the multiple lesions can be frequently present (such as parachute mitral valve, subaortic stenosis, supravalvular stenosis, and VSD) and can affect the accuracy of flow-dependent Doppler parameters of the aortic valve, for example, velocity or gradient. Adding 3D colour Doppler is used to assess for aortic regurgitation and cusp integrity; multiplanar reconstructions parallel to the axis of the regurgitant flow enable direct planimetry of 3D vena contracta area and provide added value especially in the presence of non-circular or multiple jets.^89,90 Suboptimal 3D TTE images are reasonable alternative in patients with poor acoustic window or heavily calcified aortic valve. 3D TOE can provide better image quality with superior spatial resolution and accurate planimetry of the aortic valve orifice area,⁹¹ displayed from both aortic and ventricular perspectives, which further improves the diagnostic accuracy and helps guide interventional procedures (Figure 15; Supplementary data online, Videos S17–19).⁹² The multiplanar reconstruction mode is useful for accurate measurements of the aortic valve area, aortic annulus dimensions, and aortic root size when planning for interventional procedures.^93,94

There are several congenital malformations of the aortic valve, which include unicuspid, bicuspid, and quadricuspid valves. Although standard 2DE can define the aortic valve morphology in most cases, the diagnosis can be difficult in calcified aortic valves and in cases of bicuspid aortic valve when a raphe mimics a commissure. The ability of 3DE to visualize the aortic valve ‘en face’ from the aorta provides direct visualization of the aortic valve cusp margins and commissures.⁹⁵ The diagnostic accuracy of 3D TOE for tricuspid and bicuspid aortic valves has been validated against contrast-enhanced multi-detector computed tomography.⁹⁶ 3D TOE is the best method to visualize subaortic membrane and unmask attachment to the mitral valve anterior leaflet, which may complicate the repair. The main advantages and limitations of 3DE assessment of the aortic valve and LVOT are summarized in Table 5.

Pulmonary valve and right outflow tract lesions

Optimal imaging of the pulmonary valve is essential in order to clearly define valve and RVOT morphology in CHD diagnosis and management and to aid percutaneous pulmonary valve interventions.^97–100 Despite advances in 3DE imaging, challenges to obtain optimal pulmonary valve images are still present primarily due to signal dropout from the thin valve leaflets and due to its position right behind the sternum, particularly difficult to visualize from the oesophagus.¹⁰¹ In many cases, combination of image acquisition techniques can help to overcome current limitations. The advantage of 3D TTE (because of the anterior position of the PV, thus far away from the TOE probe) is that it allows visualization of the pulmonary valve and RVOT from multiple perspectives, e.g. ‘en-face’ view of the valve from the superior (pulmonary artery) side, the inferior (RV) side, as well laterally in a longitudinal view (Figure 16; Supplementary data online, Videos S17, S20, and S21).¹⁰² 2DE images are very unlikely to visualize all three valve leaflets in cross-sectional view, whereas with the 3DE volume data sets, this is feasible in most cases. The 3D TTE short-axis view can provide detailed information on all three cusps of the pulmonary valve including their thickness, mobility, commissures, and opening area (Figures 16C and 17). Previous studies showed 60% success rate of pulmonary annulus measurement in patients with valve stenosis.^103,104 In addition, 3DE can evaluate malcoaptation of pulmonary valve leaflets, and in combination with colour Doppler, it can be planimetered accurately and used for quantification of the severity of pulmonary regurgitation.^90,105 For assessment of pulmonary stenosis, planimetry of the stenosis area is possible with 3D TTE and might be more accurate, as transvalvular gradient can be affected by multiple factors. Further improvement in 3DE TOE imaging techniques will improve its role in percutaneous pulmonary valve intervention guidance.

Summary

3DE in assessment of the valves should be used for the following:

• Detailed assessment of the valve morphology (anatomy of the leaflets and annulus, tenting, straddling, cleft, prolapse, flail, etc.);

• Display of the valves from any prospective;

• Evaluation of the position of the intracardiac lead(s) and interaction with the valve leaflets;

• Assessment of valve stenosis by 3DE planimetry (including multiple levels of stenotic lesions);

• Assessment of origin, number and significance of regurgitant jets; and

• Periprocedural guidance.

Pre-tricuspid shunts

ASDs represent the most common CHD diagnosed in adulthood.¹⁰⁶ 3DE approach to image interatrial septum (IAS) has been shown to have incremental value over the 2DE^107,108 and is advised for better evaluation of the IAS defects.¹⁰⁹

3D TTE assessment of IAS defects is feasible and less invasive compared with TOE; however, it is prone to dropout artefacts, especially when utilizing the apical four-chamber view with ultrasound parallel rather than perpendicular to the IAS, and hence, the best approach is to utilize the subcostal view or (recognizing the limitation of this view in adult patient population) off-axis views where the ultrasound is more perpendicular to the IAS (Figure 18). 3D TTE is emerging as an accurate technique to assess Qp/Qs when compared with CMR, with a very good feasibility and reproducibility even in dilated RV.²⁰

TOE provides superior image quality in the evaluation of IAS. An example of TOE 3D imaging protocol for assessment of IAS is provided in Figure 19.

3D TOE is feasible, with high diagnostic accuracy, and provides additional detailed and dynamic anatomic data. Speed, resolution, ease of application, and quantitative capabilities continue to improve and make this method preferable for routine clinical use, especially during ASD device closure.¹¹⁰ Such well-known limitations of 3DE as lower temporal resolution compared with 2DE do not produce significant impact on imaging of the IAS, which is, unlike valves, a relatively less mobile structure.

ASDs are classified according to their location.¹¹¹

Secundum ASD and PFO

Secundum ASD, the most common type of all ASDs (80%), is located within the fossa ovalis due to a true deficiency of septum primum tissue. Small secundum ASD needs to be distinguished from a patent foramen ovale (PFO) as the latter is not a true deficiency of atrial septal tissue, but rather a tunnel-like communication between the septum primum and septum secundum located in the anterosuperior portion of the atrial septum.¹¹¹ Because there is no direct defect in PFO unless it is a stretched PFO, 3D en-face view of the IAS may not be so helpful as it will only show the edge of the septum secundum overlapping over the septum primum (Figure 20; Supplementary data online, Video S22). In contrast, the ‘en-face’ view of the IAS is very informative in secundum ASD as it can show the location, shape of the defect, and give a rough idea about the rims even before detailed measurement (Figure 21A and B). High ostium secundum ASD needs to be differentiated from superior vena cava sinus venosus interatrial communication, and the latter is commonly associated with partial anomalous pulmonary venous drainage.¹¹¹ High ostium secundum ASD is located at the superior part of the septum within the vicinity of the atrium rather than the superior vena cava, and that means they have a short superior rim that makes the percutaneous closure challenging especially if there is another deficient rim (Figure 21C and D).

Primum ASD

An ostium primum ASD, also called partial AVSD (accounting for 15% of all ASDs), occurs due to failure of fusion of the endocardial cushions. This anomaly happens as part of the AVSD spectrum pathology. In 3D imaging, it shows as a defect that is close to the AV junction with the left and right AV valves located at the same level (Figures 18C, D, and E and 22).

Sinus venosus ASD

Sinus venosus defects are typically found within the mouth of one of the venae cavae. Although this term is frequently used, they are not true defects of the IAS from a morphological point of view and should be referred to as interatrial communications. More common is the superior sinus venosus defect (5%) resulting from deficiency of the tissue that separates the right upper pulmonary vein from the superior vena cava; pulmonary veins from part of the right lung often drain into the right atrium or connect anomalously to the superior vena cava (Figure 23; Supplementary data online, Video S23). Inferior sinus venosus defect, on the other hand, is a rare defect (<1%) found in the mouth of the inferior vena cava involving the posterior–inferior aspect of the atrium. In 3DE imaging, the defect shows within the vicinity of the corresponding vena cava rather than the atria.

Unroofed coronary sinus

The unroofed coronary sinus is a rare cardiac anomaly in which a part or the whole common wall between the coronary sinus and the left atrium is absent. In many cases, it is associated with persistent left superior vena cava.¹¹² 3DE allows better delineation of the extent of the unroofing compared with the 2DE and helps surgical planning if needed (Figure 24; Supplementary data online, Video S24).

3D quantitative assessment of pre-tricuspid shunts

Direct measurements of the shunt dimension are not advised, as 3D-rendered images and volumes are liable to two main disadvantages: (i) gain artefacts that can cause oversizing or under-sizing of the defect in lower or higher gain settings, respectively; and (ii) projection, as slight tilting of the septum to one side will cause underestimation of the lateral defect dimension, while backward or forward tilting will cause underestimation of the vertical dimension. It is advisable to use the multi-plane reconstruction (MPR) tool to align the cutting planes over the defect edges that will allow getting an ‘en face’, 3D-derived 2D plane of the defect for accurate measurement. Using the MPR will also allow getting all the necessary views to measure the rims of the defect and assess whether the defect is suitable for percutaneous closure (Figure 25).

Percutaneous closure intraprocedural guidance

3D TOE provides reliable and optimal results for sizing the complex defects, while 2D TOE may lead to selecting a smaller-sized device.⁶ 3DE-derived ASD area or circumference can reliably predict device size used for ASD closure through a certain equation 3DE-derived equation.¹¹³ 3D TOE has also shown to significantly decrease radiation exposure through guiding the right heart catheterization in patient undergoing ASD closure.¹¹⁴ Having the ‘en-face’, anatomically oriented view of the IAS allows better understanding of the anatomy and creates a common language between the interventionist and the echocardiologist in the catheterization laboratory that makes catheter navigation over the septal faster compared with 2D TOE combined with fluoroscopy.¹⁰⁸ It allows monitoring of defect crossing, device deployment, and the confirmation of proper seating of the device on both sides of the IAS without compressing neighbouring structures such as the coronary sinus or vena cava (Figure 26; Supplementary data online, Video S25). Percutaneous closure device complications can be visualized using 3DE, including residual shunts or thrombosis. 3DE was also shown to accurately and systematically characterize ASD residual rim in complex ASDs.¹¹⁵

Post-tricuspid shunts

Evaluation of VSDs should include their location, shape, size, and spatial relation with the nearby structures such as valves. In addition to 2DE imaging, 3DE can provide the informative ‘en-face’ view of the ventricular septum and help with assessment of shape, location, relation to nearby structures, and navigation during interventional closure (Figure 27; Supplementary data online, Video S26).^116–120 It also allows accurate sizing of the defect through the MPR tool.

Summary

3DE imaging of the pre-tricuspid or/and post-tricuspid defects should be used for the following:

• Accurate assessment of defect morphology, position, size, and rims;

• Evaluation of suitability for percutaneous defect closure; and

• Interventional guidance of defect closure procedures.

In addition, 3DE is emerging tool for accurate assessment of Qp/Qs.

Other lesions

As 3DE can provide ‘en-face’ view of any structure of interest, it can be very useful in delineating intra-atrial obstruction due to the membranous type of structure. In a patient with cor triatriatum, 3DE helps to assess the size and exact location of the defect on the membrane and severity of obstruction (Figure 28; Supplementary data online, Videos S27–S29). Similarly, in patients with subaortic stenosis due to subaortic membrane or ridge, 3DE is a useful tool in demonstrating the size of the exact orifice. In a patient with endocarditis, 3DE is useful in assessing the location, size of the vegetations, and their relation to the adjacent cardiac structure (although visualization of very small and highly mobile vegetations may be challenging due to the lower temporal resolution of 3DE compared with 2DE) and evaluation of the size and location of root or annulus abscess (Figure 29; Supplementary data online, Video S30). Finally, 3DE may aid with morphological assessments of surgically created baffles or conduits in patients with good acoustic window and has a potential to better appreciate complications, such as the presence of thrombi.

Training requirements

To achieve expertise in 3DE in ACHD, a structured and complete training strategy is necessary. Healthcare professionals require substantive knowledge regarding CHD, surgical and percutaneous interventions, together with a strong background in 2DE in CHD and the techniques involved. A conservative estimate for healthcare professionals to attain proficiency in general 3DE is around 6–12 months of dedicated training, contingent on the individual's prior experience, and the intensity of the training programme. While there are no clear recommendations in place for 3DE training in ACHD, we suggest a period of 12 months in a high-volume ACHD centre, with the possibility to compare 3DE images and measurements with data obtained by CMR, computer tomography, or with the surgical findings, and experienced mentorship for comprehensive ACHD echocardiographic training.

Future perspectives

The use of 3DE in ACHD has the potential to bring about breakthroughs that will improve diagnosis accuracy and the quality of care provided to patients. In the context of ACHD, it is anticipated that the spatial and temporal resolution of 3DE imaging will continue to increase as a result of ongoing technological improvements in this field. These include advancements in matrix array transducers and real-time volumetric acquisition. Enhancement in spatial and temporal resolution will facilitate more accurate assessment of complex cardiac anatomy and functional abnormalities, allowing for better management, interventional and surgical planning, and postoperative evaluation. Additionally, the integration of artificial intelligence algorithms for automatic image analysis and interpretation, including flow quantification for valve assessment, and the incorporation of 3D printing technology, virtual reality, or holographic models for personalized reconstruction of beating cardiac models could change preoperative planning and patient education. Emerging research and clinical studies in this field are crucial for evaluation of the role of 3DE in reducing diagnostic errors in CHD, health economics impact of 3DE use, prognostic significance of 3DE-derived parameters in specific CHD, and role of 3DE in improving interventional results (including complication rates, radiation exposure, or material-associated costs). This will facilitate wider inclusion of 3DE into clinical and procedural guidelines and further expand its potential for patients with CHD.

Conclusion

3DE is an easily obtainable, inexpensive, and valuable tool to delineate the individual unique CHD morphology and haemodynamic status. 3DE is nowadays the only imaging modality able to display cardiac structures in 3D in the beating heart (e.g. valve leaflets and chordae) and visualize their relations with surrounding structures (e.g. VSD and LVOT).

The added value of 3DE in ACHD lies in its impact on diagnostic precision particularly relevant to more complex CHD patients, many with multiple previous interventions, interventional and surgical assessment and planning, interventional guidance, postinterventional evaluation, and overall patient care. The ability of 3DE to provide detailed, real-time, and dynamic images of complex cardiac structures contributes to a comprehensive understanding of ACHD anatomy and cardiac function. In addition, 3DE is emerging as an accurate technique to measure Qp/Qs. As technological innovations continue to unfold, with ongoing advancements in spatial and temporal resolution, as well as the integration of artificial intelligence and 3D printing technologies, 3DE stands as a valuable tool in the multidisciplinary approach to managing the complexities of ACHD.

Supplementary data

Supplementary data are available at European Heart Journal - Cardiovascular Imaging online.

Funding

None declared.

Data availability

No new data were generated or analysed in support of this research.

References

Author notes