International consensus statement on nomenclature and classification of the congenital bicuspid aortic valve and its aortopathy, for clinical, surgical, interventional and research purposes

Abstract

This International Consensus Classification and Nomenclature for the congenital bicuspid aortic valve condition recognizes 3 types of bicuspid valves: 1. The fused type (right-left cusp fusion, right-non-coronary cusp fusion and left-non-coronary cusp fusion phenotypes); 2. The 2-sinus type (latero-lateral and antero-posterior phenotypes); and 3. The partial-fusion (forme fruste) type. The presence of raphe and the symmetry of the fused type phenotypes are critical aspects to describe. The International Consensus also recognizes 3 types of bicuspid valve-associated aortopathy: 1. The ascending phenotype; 2. The root phenotype; and 3. Extended phenotypes.

INTENDED AUDIENCE AND PURPOSE

This international evidence-based nomenclature and classification consensus on the congenital bicuspid aortic valve (BAV) is intended to be universally used by clinicians (both paediatric and adult), echocardiography sonographers and physicians, cardiovascular advanced-imaging specialists, interventional cardiologists, cardiovascular surgeons, pathologists, geneticists and researchers encompassing these clinical and basic research areas. In addition, if and when new landmark research is available, this international consensus may be subject to change in accordance with evidence-based data.

GENERAL NOSOLOGY OF THE CONGENITAL BICUSPID AORTIC VALVE CONDITION

The congenital BAV condition is fundamentally a valvulo-aortopathy characterized by significant heterogeneity of its valvular and aortic phenotypic expressions, of its associated disorders, of its complications and its prognosis [1–5]. From the nosology perspective, and in order to reconcile this clinical and prognostic heterogeneity, the BAV condition is broadly categorized into 3 clinical-prognostic (Fig. 1) subgroups: (i) complex valvulo-aortopathy [5, 6], where concomitant or associated disorders may be clinically and prognostically worse than the BAV condition per se (i.e. Turner syndrome, Loeys-Dietz syndrome, Shone complex, severe aortic coarctation) and/or there is early/accelerated valve dysfunction and/or aortopathy, more commonly diagnosed earlier in the paediatric, adolescent and young adult population [7, 8]. This presentation frequently requires early surgical/invasive treatment and close surveillance. (ii) Typical valvulo-aortopathy [2, 6], the most common group, with progressive BAV dysfunction and/or aorta dilatation without major associated or concomitant disorders, more commonly diagnosed in the young adult and adult, requires long-term surveillance and usually necessitates subsequent surgical/invasive treatment. Patients with complex-presentation and those with typical-presentation valvulo-aortopathies are at risk of developing infective endocarditis and aortic dissection (Fig. 1), although aortic dissection is extremely rare in young children with BAV and rare in adults without aortic dilatation [2, 9]. Importantly, complex-presentation valvulo-aortopathies may also occur in adults and typical-presentation valvulo-aortopathies may occur in children. (iii) Undiagnosed or uncomplicated BAV, a subgroup [2], is a lifelong silent condition with mild or non-progressing valvulo-aortopathy that does not manifest clinically but may come to light at autopsy or incidentally by imaging (Fig. 1); therefore, it represents a retrospective definition, yet it requires surveillance if incidentally diagnosed. Some of these cases will never be diagnosed which hampers the assessment of the true incidence and prevalence of BAV complications due to a smaller denominator of diagnosed cases.

A critical difference between the typical and complex valvulo-aortopathies is the preserved long-term overall life expectancy, which is similar to that of the age- and sex-matched general population with typical valvulo-aortopathy [11], whereas life expectancy may be reduced in those with the complex valvulo-aortopathy. For example, long-term survival in patients with severe aortic coarctation requiring surgery is significantly inferior to that in the general population [12]. Similarly, long-term survival in patients with Turner syndrome is also significantly compromised compared to the general population [13].

FUNDAMENTALS OF IMAGING ASSESSMENT OF THE CONGENITAL BICUSPID AORTIC VALVE CONDITION

At the centre of the BAV condition is echocardiography, which serves as the first-line imaging modality in 6 major capacities [6]: (i) BAV diagnosis, (ii) valvular phenotyping, (iii) assessment of valvular function [6], (iv) measurement of the thoracic aorta (the expression of BAV aortopathy is dilatation of the thoracic aorta), (v) exclusion of aortic coarctation and other associated congenital lesions [2, 7] and (vi) assessment of uncommon but serious complications such as infective endocarditis [14] and aortic dissection [9]. Transthoracic echocardiography (TTE) is the first-line BAV diagnostic and phenotyping modality, the best modality for haemodynamic assessment of valvular dysfunction, and the initial modality for assessment of thoracic aorta size, presence of aortic coarctation and other congenital lesions. Transoesophageal echocardiography may aid in the diagnosis and phenotyping of BAV that is not well visualized by TTE, has excellent accuracy for the diagnosis of aortic dissection [15] and is mandatory in the assessment of infective endocarditis [16], whether it is native or prosthetic.

Also at the centre of the BAV condition are advanced imaging modalities: electrocardiographic (ECG)-gated cardiac computed tomography (CCT) and ECG-gated cardiac magnetic resonance (CMR). These imaging techniques improve diagnostic accuracy and phenotyping of BAV [17, 18] and represent the gold standard for measuring the thoracic aorta because they accurately assess aortic diameters that are truly perpendicular to the longitudinal axis of the aorta by use of the double-oblique technique. In addition, interval measurements can be performed at the same exact anatomical locations for comparison. After initial TTE imaging, if any aortic segment cannot be visualized or coarctation cannot be ruled out or any thoracic segment measures ≥45 mm by TTE, then ECG-gated computed tomography (CT) angiography or magnetic resonance angiography is recommended [19], with magnetic resonance angiography preferred for younger patients (i.e. <50 years old) to avoid repeated radiation exposure at follow-up examinations. Further recommendations on echocardiographic and CCT/CMR assessment of congenital BAV and aortopathy have been recently published [6, 19], including echocardiographic assessment of BAV function [6].

SYNOPSIS OF THE CLINICAL HISTORY OF THE CONGENITAL BICUSPID AORTIC VALVE CONDITION

The most common complication of the BAV condition in adults is valve dysfunction that necessitates surgical aortic valve replacement (AVR) or repair, and it is strongly determined by the development of aortic stenosis (AS) [2, 20]. The community risk of AVR 25 years after BAV diagnosis is greater than 50% [2]. Surgical AVR is the gold standard for treating BAV-related AS. Nonetheless, with the latest generation of transcatheter aortic valve replacement (TAVR) devices, guided by careful preprocedural ECG-gated CCT analysis [21, 22], the technical success of TAVR has improved significantly, and TAVR may be an alternative to AVR for patients with BAV with AS and a high surgical risk (see Section Interventional cardiology considerations); indeed, up to 20% of patients ≥80 years old undergoing AVR have a congenital BAV [23]. Significant aortic regurgitation (AR) in BAV is considerably less common than AS (30% vs 70%) and is more frequent in men [3]. Surgical AVR remains the gold standard for treatment of BAV-related AR; nonetheless, surgical repair is an option, and echocardiography plays a critical role in determining reparability of the regurgitant BAV [6, 24], which is successful more frequently in BAV than in tricuspid aortic valves, with a low cumulative reoperation incidence of 20% at 15 years when combined with root remodelling [25].

The next most common complication of the BAV condition is aortopathy [19], which manifests clinically as dilatation of the thoracic aorta. The prevalence of any aortic dilatation in patients with BAV is reported to be from 40% to 70% depending on the population studied and the definition of dilatation [2]. The population incidence of aortic dilatation ≥45 mm is greater than 25% at 25 years of follow-up, with more than 20% undergoing surgery for aorta repair [9]. Coarctation of the aorta is present in 7–10% of adults with BAV [26], whereas BAV is present in 50–60% of patients with coarctation [27]. Concomitant coarctation is associated with a higher risk of aortic complications [27]. Mitral valve prolapse affects 2–3% of patients with BAV; this value is not different from that of the general population, but isolated anterior prolapse including ‘giant’ anterior leaflet prolapse is 2 times more frequent in patients with BAV and may hamper successful mitral repair [28]. The least frequent yet most deadly complications are infective endocarditis and aortic dissection. The incidence of BAV endocarditis [native and prosthetic (aortic position)] has been reported at 2% in most contemporary cohorts with BAV [2, 29]; the population incidence of approximately 14 cases per 10 000 patient-years is 11 times that of the general population [14]. Among patients with BAV, the overall community incidence of aortic dissection is approximately 3 cases per 10 000 patient-years, which is 8 times that of the general population, increasing to 0.5% in patients with aortic diameters ≥45 mm [9] but generally <1% [29].

WHY A STANDARD NOMENCLATURE AND CLASSIFICATION CONSENSUS FOR THE CONGENITAL BICUSPID AORTIC VALVE CONDITION?

Nomenclature refers to the choice of ‘name’ that is given to a particular structure, abnormality or phenotype, whereas classification refers to the process of ‘arranging or categorizing’ something according to shared features. The clinician evaluating the patient with BAV must be able to communicate in a common language all specific morphological, functional and prognostic aspects of the BAV condition to patients, other clinicians, surgeons, interventionalists and researchers [6, 10]. In addition, there are multiple gaps in the knowledge and understanding of the BAV condition [2]. In order to advance the clinical, biological and genetic understanding of the BAV condition, a common language must be articulated among researchers in all clinical and laboratory research disciplines. There are multiple nomenclatures and classifications for the BAV condition, and they are as heterogeneous or more so than the BAV condition itself (Table 1) [4, 30–38]. For example, the Sievers and Schmidtke [34] and Schaefer et al. [33] classifications use multiple numbers and letters for the BAV and aorta phenotypes, with Sievers including an incomplete definition of unicuspid aortic valves within the BAV classification (Table 2). Although the morphological spectrum of human congenital aortic valve abnormalities includes unicuspid, bicuspid and quadricuspid aortic valves, their genetic and embryological origin may not necessarily be closely linked [39, 40], and their prevalence, age at presentation, prognosis and associated conditions are not equivalent [6, 41, 42], with BAV being much more prevalent and heterogeneous. In addition, the surgical Sievers classification does not incorporate the evaluation of the symmetry of the BAV, a critical surgical-repair feature in current times [25, 43] (Table 2). Other BAV classifications are extremely succinct-dichotomous, as proposed by Sun et al. [36], or extremely complex as proposed by Kang et al. [30], with 5 numerical types of BAV phenotypes and 4 numerical types of aortic phenotypes (Table 1). Others have used a combination of previous classifications and added new observations: For example, Murphy et al. [38] proposed the clock-face orientation combined with the Sievers classification, adding partial cusp fusion and leaflet asymmetry by CMR (Table 1). Additionally, the use of one or another classification system for research varies by author and institution. A consistent description of the subtle variations in valve morphology, as well as newly developed in vivo metrics of haemodynamic changes associated with differing aortic valve morphologies, highlights the need for a universal, uniform classification scheme [44]. Finally, there are specific nomenclatures that lead to confusion such as the ‘true’ BAV: Does it mean that the others are not really BAV? And, as mentioned, Sievers’ type 2 BAV is actually not bicuspid; it is unicuspid (Table 2). These numerous and heterogeneous classifications cause confusion in clinical practice, failure to identify phenotypes that may predict outcomes, inability to analyse clinical outcomes data in registries, systematic review and meta-analysis formats, failure to capture anatomical information critical for surgical aortic valve repair and TAVR and hamper identification of phenotypic-genetic associations. Herein, we present an imaging-based, descriptive, simple-but-comprehensive nomenclature and classification system that is based on the English language and not on numbers or letters and is based on important and available anatomical, clinical, surgical and pathological scientific data [10]. This new nomenclature/classification system represents the combined efforts of international BAV experts including clinicians (both adult and paediatric), surgeons, interventionalists, pathologists, geneticists and imagers (echocardiography, CT and magnetic resonance experts).

DEFINITION of CONGENITAL BICUSPID AORTIC VALVE AND AORTIC ROOT COMPLEX

Congenital bicuspid aortic valve

The aortic valve includes the cusps and the annulus. The congenital BAV is most commonly diagnosed by base-of-the-heart short-axis aortic valve imaging with TTE, ECG-gated CCT or CMR, demonstrating the existence of only 2 commissures delimiting only 2 valve cusps [2, 45] (Fig. 2; Video 1). On echocardiographic long-axis imaging, systolic doming of the conjoined cusp may be appreciated particularly for right–left coronary cusp fusion (Fig. 2; Video 2), but it is less reliable for other BAV phenotypes. The diagnosis can also be made by direct surgical observation [31, 43] and pathological examination [32]. It is important to recognize that a tricuspid aortic valve that is fibrotic and calcified or rheumatic may present a pattern of acquired (non-congenital) fusion of 2 cusps that may be difficult to differentiate from congenital BAV. In these cases, surgical inspection and/or pathological examination may identify whether the fusion is congenital or not. In the operating theatre, although it is not always possible, the surgeon can define the congenital bicuspid nature by observing the height of the ‘pseudocommissure’ (the attachment of the raphe at the aortic wall), which is lower within the root compared to the height of the true commissures, whose attachment is higher (Fig. 3). Additional gross features can be used on surgical or pathological inspection, such as the angle formed between the fused cusps (obtuse: congenital fusion; acute: acquired fusion) and the cleavage plane on the ventricular aspect of the fused cusps (absent: congenital; present: acquired) (Fig. 2). It is critical to utilize the information provided by the surgeon and especially by the pathologist [46] to determine the presence of a congenital BAV in cases of severely calcified AS.

Aortic root and root complex

Understanding the topographical anatomy of the proximal aorta is critical because it is an integral part of the aortic valve function, akin to the annulus and subvalvular apparatus for the mitral valve. Although ‘ascending aorta’ and ‘aortic root’ are sometimes used interchangeably to indicate the entire vascular segment from the aortic valve to the brachiocephalic artery take-off (beginning of the arch), the term aortic root refers only to the most proximal part of the ascending thoracic aorta, from the distal end of the left ventricular outflow tract to the sinotubular junction (STJ), formed by the sinuses of Valsalva and containing the aortic valve [47] (Fig. 3). The anatomy and physiology of the aortic root complex and its interaction with the valve have been thoroughly investigated as contemporary techniques for aortic valve repair have been introduced and more widely adopted [48, 49]. Functionally, and particularly in relation to the competency of the BAV and surgical repair of the regurgitant BAV, 3 elements form the aortic root complex and cooperate in determining physiological valve dynamics [50]: (i) the STJ, (ii) the aortic sinuses with the crown-like attachment line of the aortic valve cusps to the aortic wall at the aortic sinuses which, as mentioned, assumes a peculiar form in the fused BAV, with 1 of the 3 ‘crown tips’ corresponding to the under-the-raphe pseudocommissure, reaching a lower height than the other 2, i.e. not reaching the STJ (Fig. 3) and (iii) the aortic annulus, which is a virtual circular line inside the left ventricular outflow tract, running through the nadir of the aortic cusps and the respective bases of the inter-cusp triangles (Fig. 3). The aortic annulus is a virtual surrogate for the ventriculo-aortic junction, which is the real boundary of the aortic root complex identified anatomically as the transition from the ventricular muscle to the aortic media. It is located circumferentially slightly above the nadir of the aortic cusps, crossing the semilunar lines of each cusp’s attachment (Fig. 3). In both surgery and imaging, however, the surrogate of the ventriculo-aortic junction (aortic annulus) is the practical and clinically used anatomical landmark that constitutes the third component of the root complex, as described above. It has been reported that the distance between the ventriculo-aortic junction and the virtual annulus levels is variable and usually greater in BAV than in the normal aortic valve, particularly in the right coronary sinus [48]. The aortic root complex, particularly the size of the aortic annulus and the STJ, is indispensable in the maintenance of sufficient diastolic cusp coaptation area to prevent the progression of AR [51] and its recurrence after surgery [52]. Therefore, the aortic root complex is the anatomical scaffold that maintains BAV competency, with the BAV cusps acting as a stentless valve and the root complex as its native stent [50].

The tract of the proximal aorta spanning from the STJ to the brachiocephalic artery take-off should be referred to as the ‘tubular ascending aorta’ or the ascending aorta. The subsequent tract, from the brachiocephalic artery to the isthmus (the physiological narrowing just distal to the left subclavian artery origin), is called the aortic arch.

CONSENSUS ON BICUSPID AORTIC VALVE NOMENCLATURE AND CLASSIFICATION FOR CLINICAL, SURGICAL, INTERVENTIONAL AND RESEARCH PURPOSES

Bicuspid types and specific phenotypes

There are 3 BAV types: the fused BAV, the 2-sinus BAV and the partial-fusion BAV, each with specific phenotypes [10] (Fig. 4).

The fused bicuspid aortic valve type

The fused BAV is the most common type (Figs 5 and 6), accounting for approximately 90–95% of cases [2, 32]. The fused BAV is characterized by 2 of the 3 cusps appearing fused or joined within 3 distinguishable aortic sinuses, resulting in 2 functional cusps (1 fused or conjoined and the other non-fused) that are usually different in size and shape, with non-fused cusp commissural angles of varying degrees (Figs 6–8). Commonly, both adult and paediatric patients with BAV demonstrate eccentric dominance of the non-fused aortic sinus and its cusp (compared to the other 2 sinuses and 2 fused cusps), irrespective of age [53] (Figs 6 and 7). Frequently (approximately 70%), but not always, there is a congenital fibrous ridge between the fused cusps, termed raphe [32, 54]. The presence of a raphe has been associated with the progression of valvular dysfunction (particularly AS) and future valvular surgery [45, 54, 55]. A raphe may be present but not initially visible by echocardiography and may become visible years later [56]. Significant calcification of a raphe can be identified by echocardiography (highly echogenic, casting a shadow) but less-severe calcification versus raphe-fibrosis cannot be easily discerned. Conversely, raphe calcification can be readily identified by the specific attenuation pattern on CCT (highly dense, usually more than 130 HU).

There are 3 specific BAV phenotypes within the fused type: right–left cusp fusion, right non-(non-coronary) cusp fusion and left non (non-coronary) cusp fusion (Figs 6 and 7; Videos 1–4). The right–left cusp fusion phenotype is the most common (70–80%) across American, European and Asian populations [2, 32, 57]. The right–left cusp fusion phenotype is also the most common across all phenotypic variations of the aorta (normal aorta, dilated ascending aorta, dilated arch or dilated root) and across valve dysfunction (regurgitation or stenosis). Although this right–left fusion phenotype statistically develops more AS [2], it has been associated in some patients, both children/adolescents [58] and adults [59, 60], with aortic root dilatation, AR and male preponderance (these associations have been termed the ‘root phenotype’). The right–left cusp fusion is also strongly associated with aortic coarctation in children [61].

The right non-cusp fusion phenotype is the next most common (20–30%); it is associated with a higher prevalence of AS in adults [55] and independently predicts AR progression in adults [51]. Similarly, the right non-cusp fusion phenotype is associated with a more rapid progression of AS and regurgitation in children and adolescents [61, 62]. The right non-cusp fusion phenotype is also more prevalent in Asian populations, as is the left non-cusp fusion phenotype [57, 63], which is the least common phenotype (3–6%) across studies. Interestingly, African American patients are reported to have a lower prevalence of BAV and aortopathy altogether [64].

In complex-presentation forms like BAV associated with genetic syndromes, right non-cusp fusion is more common in patients with Down syndrome, and right–left cusp fusion is more common in patients with Turner’s syndrome and Shone complex, suggesting different abnormalities in developmental pathways [8]. Based on the results from animal experiments, it can be assumed that the embryological background of the fused types is that of abnormal remodelling/maturation (excavation) of the valve cushions (the 3 fused types may be explained by defective excavation) or a mild defect during outflow tract septation for fused right–left phenotypes and during endocardial cushion formation/positioning for the fused right non- and left non-phenotypes [65–69].

Referring to the fused phenotypes as BAV with right–left cusp fusion, right non-cusp fusion or left non-cusp fusion is appropriate. Occasionally, it is possible to recognize a BAV with 3 aortic sinuses but not be able to discern the fusion phenotype, in which case BAV with indeterminate cusp fusion is appropriate (Fig. 4). It is important to recognize that some fused BAVs may not have a congenital raphe [32] or have a raphe that is not visible by imaging [56], yet they have 3 distinguishable aortic sinuses and the 2 fused cusps can be identified (Fig. 8; Video 5).

Symmetry of the fused bicuspid aortic valve types

Evaluation of BAV symmetry for the fused BAV type is defined by the angle between the commissures of the non-fused cusp and has recently become a critical aspect in the planning and performance of BAV repair for pure AR [10, 43, 70]. From a regurgitation-treatment perspective, the BAV concept offers a simple, single-line coaptation surface [a tricuspid aortic valve has 3 coaptation lines (Fig. 5, left)]; as long as that single coaptation line is straight or almost straight (Figs 8 and 9, symmetrical), the repair of the regurgitant BAV is reproducible (see Section Surgical considerations). As the angle between the commissures of the non-fused cusp decreases to <160° [70], the BAV becomes less symmetrical, more closely resembling a tricuspid (especially <140°) valve (Fig. 9, very asymmetrical), which becomes technically more challenging for the surgeon to ‘bicuspidize’ during the repair yet remains repairable in experienced hands. Asymmetrical valves may exhibit retraction of the free edge of the fused cusp at the raphe level, which is best appreciated by direct surgical visualization (Figs 2 and 9) or gross pathological inspection, and not reliably by imaging. This retraction may contribute to valve regurgitation. Figure 8 shows a fused BAV with right–left cusp fusion with a 180° non-fused cusp commissural angle (symmetrical), although the 2 cusps are not the same size/shape. Measuring the non-fused cusp commissural angle on precardiopulmonary bypass transoesophageal echocardiography aids the surgeon in planning the repair (Fig. 10; Video 6). Therefore, the symmetry of a fused-type BAV is defined by the angle between the commissures of the non-fused cusp.

The 2-sinus bicuspid aortic valve type

The 2-sinus BAV is uncommon, accounting for approximately 5–7% of cases [2, 10, 32]. In contrast to that of the fused type, the appearance of the 2-sinus BAV does not suggest that 2 of the 3 cusps have fused; instead, it suggests that 2 cusps, roughly equal in size and shape, each cusp occupying 180° of the annular circumference, were ‘formed’ within only 2 aortic sinuses, resulting in a 2-sinus/2-cusp valve (Figs 11–13; Videos 7–10) without raphe and with 180° commissural angles. It is often difficult to determine which 2 cusps could have coalesced to form a 2-sinus BAV, but it is usually evident whether the cusps are laterolateral (side-to-side) or anteroposterior (front-and-back) within the short-axis base of the heart plane (Figs 11–13; Videos 7–10); thus, these are the 2 specific phenotypes of the 2-sinus BAV category. The 2-sinus laterolateral BAV has 1 coronary artery arising from each cusp, whereas the anteroposterior BAV may have 1 coronary artery arising from each cusp or both coronary arteries arising from the anterior cusp (Figs 11 and 13). Based on results from animal experiments, it can be assumed that the embryological background of the 2-sinus BAV is that of abnormal endocardial cushion formation/positioning for the laterolateral and abnormal outflow tract septation for the anteroposterior. The 2-sinus BAV likely represents a more severe expression of the embryological mechanisms leading to the fused BAV. Referring to these phenotypes as 2-sinus laterolateral BAV and 2-sinus anteroposterior BAV is appropriate. Occasionally, despite suspicion, it may be difficult to be certain whether there are only 2 sinuses, in which case, terms such as possible or probable 2-sinus BAV may be used. There is a lack of scientific data on the clinical/prognostic associations of the 2-sinus BAV, which represents a ‘morphologically severe’ form of BAV. Therefore, we hope that through this nomenclature/classification, the research community directs more attention towards this type of BAV.

The partial-fusion bicuspid aortic valve (or forme fruste bicuspid aortic valve) type

The partial-fusion BAV (or forme fruste BAV) type has recently been recognized; its prevalence is unknown [71] (Fig. 14). The appearance of the partial-fusion BAV [72] is that of a typical tricuspid aortic valve with 3 symmetrical cusps with a systolic triangular opening and commissural angles of 120°, yet on surgical inspection or high-resolution imaging, cusp fusion of less than 50% is noted at the base of a commissure, forming a small ‘mini-raphe’ [10, 71, 73, 74]. It is important to recognize and further study the partial-fusion BAV, which has been described mostly in the operating room in patients undergoing surgery for aorta dilatation [71] (Fig. 15; Videos 11 and 12) [74]. This forme fruste BAV results in alteration of the aortic flow patterns, consisting of increased flow eccentricity and increased vortexes [73], perhaps partially explaining the apparent high prevalence of aorta dilatation in these patients. Referring to this phenotype as partial-fusion BAV or forme fruste BAV is appropriate. Based on results from animal experiments, it can be assumed that the embryological background of the partial-fusion BAV is that of a mild defect during outflow tract septation or during remodelling/maturation (excavation) of the valve cushions [65, 66, 69, 75, 76].

The bicuspid aortic valve anatomical spectrum.

The BAV phenotypic expression represents an anatomical continuum that is likely related to the severity of its embryological mechanisms [10]. Therefore, we propose a general BAV anatomical spectrum (Fig. 16) of BAV phenotypes in order of ‘bicuspidity’, defined as the resemblance to a 2-sinus BAV. This spectrum represents a continuum of increasing non-fused cusp commissural angles and increasing similarity of cusp size and shape. The spectrum begins with the partial-fusion BAV, which most closely resembles a tricuspid aortic valve and represents the mildest embryological defects, on to asymmetrical fused phenotypes, to symmetrical fused phenotypes with and without a raphe, ending with the 2-sinus BAV, which represents the most severe embryological defects and is anatomically close to perfect ‘bicuspidity’. This BAV anatomical spectrum can be demonstrated surgically and pathologically (Fig. 17). Virtually the same spectrum has been described in animal models, in which the anatomical variation depends on the severity of the embryonic defect [66, 67, 69, 76].

Definition of aorta dilatation and bicuspid aortic valve aortopathy

Definition of aorta dilatation

The clinical expression of the BAV-related aortopathy is dilatation of the thoracic aorta. The definition of aortic aneurysm [77] is rarely applied in clinical practice, and the term aneurysm carries a somber or dismal connotation for patients. Therefore, we propose a simple and universal term: aortic dilatation. Qualitative-descriptive terms such as saccular or fusiform dilatation or STJ effacement may be important for aorta specialists and surgeons. Echocardiographic studies in populations of apparently normal individuals have shown that the diameters of the root and ascending aorta are proportionally related to body size (most commonly expressed as body surface area), age (increasing by 0.1 mm/year in ‘healthy’ adults) and male sex in adults [78–80]. These studies and normative data in children [78, 81] allow identification of aortic root and/or ascending aorta dilatation by echocardiography when the aortic diameter is above the upper 95% confidence limit of ‘normal’ values (Fig. 18) or the calculated z-score exceeds +2.0. However, data on ‘normal’ aortic diameters are limited, with continued publications reporting varying ‘normal’ values depending on different demographics and anthropometrics of the populations observed and on methodological aspects: i.e. diastolic leading-edge to leading-edge (adult echo) versus systolic inner-edge to inner-edge (paediatric echo) measurements, echocardiography(Fig. 18) versus CCT/CMR (inner wall-to-inner wall versus outer wall-to-outer wall). These factors should be also considered when comparing serial imaging results in an individual patient during follow-up: The difference between current and previously reported aortic diameters (at the same level) can be considered a reliable quantifier of the progression of the dilatation only when measured by the same modality and exact anatomical location and method [82–84]. In adults with BAV, TTE systematically underestimates the aortic root measurement (asymmetrical aortic sinuses) compared to CCT, whereas the measurements are generally unbiased between TTE and maximum diastolic inner wall-to-inner wall CCT for the ascending aorta [85]. Therefore, in adults, diastolic leading-edge to leading-edge echocardiography is generally equivalent to diastolic inner wall-to-inner wall CCT/CMR except for the root, where CCT/CMR should be used for accurate measurement when it is enlarged (i.e. >45 mm) or asymmetrical [15, 19].

Due to the tremendous change in body size and cardiac structures that occurs from infancy to adolescence, utilization of z-scores to compare obtained aortic measurements to normative data is essential. This approach allows for easy identification of infants, children and adolescents who have echocardiographic aortic dimensions that fall outside the normal range for their age and body size, typically identified as a z-score that is 2 standard deviations above the mean (97.7th percentile); +2.0 [86]. Alternatively, CMR-derived percentile curves for normal cross-sectional areas of the ascending aorta, arch and descending thoracic aorta in children, adolescents and young adults have been published [87]. However, for clinical care in most settings, categorization of aortic dilatation as mild, moderate or severe for adults with BAV is more practical than referring to z-scores. Because most available data in adults relate the risks of aortic complications to the measured absolute aortic diameter without further indexing for body size, age or sex, it is reasonable at present to ‘initially’ separate these categories by simple aortic diameter partitions. Thus, in general, dilatation of the root or ascending aorta in patients with typical valvulo-aortopathy BAV (Fig. 1) is considered mild if the diameter is between the age-, body size- and sex-specific upper limit of normal (Fig. 18) [78] and 45 mm; moderate for diameters between 46 mm and 50–54 mm; severe for diameters ≥55 mm (elective surgical cut-off) if no associated risk factors are present, and also severe for ≥50 mm (elective surgical cut-off) if there are associated risk factors (any risk factor) [1, 19]. These risk factors that increase the likelihood of aortic complications (i.e. dissection) in patients with BAV with typical-presentation valvulo-aortopathy are the ‘root-phenotype’, severe BAV regurgitation, uncontrolled hypertension, personal history of coarctation, family history of aortic dissection or early unexplained sudden cardiac death or aortic diameter increase >3 mm/year [1, 19]. For patients with complex valvulo-aortopathy (Fig. 1), for example associated with genetic syndromes [5], the severity of aortic dilatation varies according to the specific underlying disease: In Loeys-Dietz syndrome, severe dilatation may be within 40–45 mm [88] depending on sex, and for women >15 years of age with Turner syndrome(short stature and small body size), severe dilatation is considered at 2.5 cm/m² of aortic diameter corrected for body surface area [89]. Indeed, because patients may vary significantly in body size, for patients with typical valvulo-aortopathy, it is important also to report the aortic diameters adjusted for the patient’s size; for example, utilizing the aortic root cross-sectional area-to-height ratio [r² Π (cm²)/height (m)] where values >10 cm²/m are associated with worse aortic outcomes [90, 91]. Alternatively, imagers may choose not to report ‘severity’ but just the measurements in millimetres, and let the clinician/surgeon define the severity according to each patient’s clinical circumstance [5].

Bicuspid aortic valve aortopathy phenotypes

The importance of recognizing BAV aortopathy phenotypes is that their presence and association with specific valvular phenotypes and patterns of valvular dysfunction may imply different clinical histories for the BAV patient [92]. There are 2 major forms of aortic dilatation BAV phenotypes: the ascending phenotype (dilatation preferentially located at the tubular ascending tract beyond the STJ) (Fig. 19), which accounts for approximately 70% of BAV aortopathy cases; and the root phenotype [dilatation preferentially located at the root (sinuses of Valsalva), possibly involving also the ventriculo-aortic junction/annulus], which accounts for approximately 20% of BAV aortopathy cases (Fig. 19) [10, 59, 60, 93]. Importantly, the root phenotype may have mild ascending dilation but significantly prevails at the root, and the ascending phenotype may have mild root dilatation but significantly prevails at the ascending portion. In addition, these 2 categories often correspond to 2 clearly distinct overall patient phenotypes: roughly, the older patient with BAV, either male or female, presenting more often with aortic valve sclerosis/stenosis (ascending phenotype); and the younger BAV patient, usually male, presenting with mild to severe AR (root phenotype) [59, 94, 95]. The greater prevalence of the ascending phenotype in BAV is consistent with the tubular ascending tract being the site of maximal growth rate of the BAV aorta in multiple studies [60, 93, 96–98], the growth rate ranging from 0.2 to 2.3 mm per year, usually 0.4 to 0.6 mm per year. A small percentage of patients demonstrate more rapid growth rates [93, 97]. Besides age, baseline aortic diameter and family history of aorta disease, the associated valve dysfunction (regurgitation vs stenosis) and the location of the dilatation (ascending versus root) impact the rate of growth [93, 96–98].

It is possible that the 2 aortic phenotypes may have different genetic bases [99, 100] explaining their occurrence, but the influence of different 4-dimensional (4D) CMR aortic flow patterns has also been suggested (see Section Cardiac magnetic resonance considerations), mostly based on the fact that BAV stenosis and the right non-cusp fusion valvular phenotype are infrequently associated with the root phenotype and frequently associated with dilatation at the level of the ascending aorta and arch [101]. Conversely, the right–left cusp fusion exerts greater wall shear stress (WSS) on the root/proximal aorta and is frequently associated with the root phenotype [55, 102, 103]. However, those associations are not unequivocal, and the right–left cusp fusion BAV can be associated with either aortic phenotype [95]. In addition, the presence of concomitant BAV stenosis can complicate the pattern of WSS expression independently of the cusp fusion phenotype [104]; therefore, the severity of the AS must be considered in the investigation of the valve-mediated aortopathy.

Notably, in some cases, the dilation of the aorta does not significantly prevail at 1 segment. In a proportion of patients, a localized dilatation at first observation can evolve during the follow-up period, with possible dilatation of previously normal adjacent segments of the aorta. In this scenario, the ascending phenotype can present, especially if a right non-cusp fusion valve is present [30, 33, 94, 105], with associated dilatation of the aortic arch; it is appropriate to refer to this condition as ascending phenotype extended. Similarly, the root phenotype has been demonstrated to be independently associated with faster growth of the ascending tubular tract, so that cases of ‘cross-over’ from an initial root phenotype configuration to significant dilatation of both tracts (and even extension into the proximal arch) have been observed [93, 105] (Fig. 19): Root phenotype extended would be an appropriate definition of this form. In the context of a root phenotype, the presence and progression of effacement of the STJ may be an initial sign of this kind of evolution.

The root phenotype has been associated with greater rates of acute aortic dissection in the postoperative follow-up of patients with BAV who had undergone simple AVR compared to the ascending phenotype [106]. The root phenotype may represent the expression of a bicuspid form of aortopathy fundamentally driven by some still unknown genetically determined connective tissue disorder, and it represents a risk factor for aortic complications within BAV aortopathy [1, 19], as previously mentioned. Conversely, for the ascending phenotype, the inherently altered flow patterns of the bicuspid valve may mainly drive the disease, which is suggested not only by the previously mentioned associations between WSS patterns and the location of the dilatation (more proximal with the right–left cusp fusion, more distal with right non-cusp fusion) but also by the typical asymmetrical dilatation of the ascending tract, i.e. with dominant involvement of the greater curvature, that is, where the greatest WSS nearly invariably occurs [107, 108] (see Section Cardiac magnetic resonance considerations).

Summary

Based on the new nomenclature and classification consensus, Fig. 20 presents a simple algorithm of the critical imaging evaluation for the BAV valvulo-aortopathy. Three critical anatomical aspects must be described in all patients with BAV.

(i) The type and specific phenotype of the BAV and the valve function; (ii) the presence and characteristics of the raphe and the cusp size/shape and symmetry of the BAV; and (iii) the presence and phenotype of aortopathy (aortic dilatation) and whether or not coarctation is present.

SURGICAL CONSIDERATIONS

The current consensus nomenclature/classification proves critical for surgical practice and surgical research. The recent American Association for Thoracic Surgery consensus document [19] recommended surgery for aortic dilatation (root or ascending) exceeding 55 mm in the general population of patients with BAV and 50 mm in patients with BAV with further risk factors for dissection, including significant AR and/or root phenotype (Fig. 19) (see Section Definition of aorta dilatation). The knowledge about the segmental nature of the majority of aortopathy cases with non-syndromic BAV indicates that liberal extension of resection to adjacent non-dilated segments (i.e. extending ascending aorta repair to the root, especially with stenotic and/or right non-fused BAV, or to the arch) is not justified at the time of tubular ascending replacement [19]. Therefore, if the patient with BAV exhibits the most common aortopathy phenotype (ascending dilatation) with a normal or only mildly dilated root/arch, replacement of the tubular portion alone will suffice. Earlier diameter indication (i.e. 50 mm) for root replacement in the root phenotype with severe AR, especially in younger patients, emphasizes the need for valve repair rather than replacement, in centres with extensive experience. Repair of the BAV has become an accepted alternative to replacement in patients with BAV regurgitation [25, 109]. Typically, the main mechanism leading to BAV regurgitation is the prolapse of the fused cusp (for fused BAV types) (Figs 6 and 7) and prolapse of 1 of the symmetrical cusps in the 2-sinus type (Figs 11–13). Other concomitant mechanisms include prolapse of the non-fused cusp and cusp retraction (Fig. 9). In addition, the aortic annulus is often dilated (i.e. >25 mm) [51], the sinuses may be enlarged and there may be STJ dilatation, all of which contribute to AR (root complex) (see Section Aortic root and root complex) (Fig. 3). Therefore, in general, the BAV repair comprises the plication of the free margins of the prolapsing cusps to correct the prolapse (Fig. 21) plus an annuloplasty suture or ring [110] to correct annular dilatation and stabilize the repair (Fig. 21). Additionally, stabilization of the STJ may require placement of a ring or ascending aorta replacement [110]. Alternatively, root replacement via a reimplantation technique will also stabilize the root at multiple levels. A critical discovery has been the importance of valve symmetry (see Section Symmetry of the fused bicuspid aortic valve types) (Figs 8 and 9), which can be measured preoperatively (Fig. 10). The closer the BAV phenotype is to a 2-sinus type with a symmetrical non-fused cusp commissural angle, the more feasible the repair will be [70] (Fig. 22). Otherwise, the surgeon uses techniques directed at ‘bicuspidizing’ the valve more (Section The bicuspid aortic valve anatomical spectrum) (Fig. 21). If the BAV is very asymmetrical, the surgeon will treat it as a tricuspid valve instead [43, 110].

GENETIC CONSIDERATIONS

Patients with a transforming growth factor beta (TGF-ß) ligand and receptor mutations that cause Loeys-Dietz syndrome (TGFBR1, TGFBR2, TGFB2, TGFB3) and ACTA2 mutations that cause heritable thoracic aortic aneurysms and dissections (HTAD) have a higher prevalence of BAV (4–15%) than the general population (1%), along with rapidly progressive aortic root dilation [5], a highly penetrant risk for aortic dissection, a variety of other congenital heart defects and, in some cases, a recognizable appearance with Marfanoid body features [88, 111]. Mutations of other HTAD genes that are not known to cause BAV, including FBN1, were identified in some patients with BAV with aortic root dilation who lack syndromic features, leading to speculation that 2 different genetic mutations may cause BAV and root phenotype aortopathy in rare individuals [99, 100, 112, 113]. In these cases, recommendations about medical therapies or the timing of interventions may be based on the specific HTAD gene [114]. However, more than 95% of BAV cases are sporadic, lack recognizable syndromic features and are not caused by mutations in known HTAD genes. Instead, rare or unique sequence or copy number variants in dozens of cardiac developmental genes have been identified in BAV [115]. Because any single gene may contribute to fewer than 1% of BAV cases, it is not possible to correlate mutated genes with specific valvular or aortic structural features before the results of large-scale sequencing studies involving thousands of patients with BAV with a common nomenclature and classification are available. Until then, clinical genetic testing should be reserved for the minority of patients with BAV with suspected HTAD gene mutations due to syndromic features, early onset or severe vascular disease or a family history of aortic dissection. This group includes a substantial proportion of individuals with TGFBR1 pathogenic variants and BAV, who do not have recognizable features of Loeys-Dietz syndrome but who may present with rapidly progressive aortic root dilation [5].

CARDIAC MAGNETIC RESONANCE CONSIDERATIONS

Compared to echocardiography, CMR offers additional functional, anatomical, perfusion and myocardial viability information. It also allows for tissue characterization and myocardial fibrosis imaging and quantification (delayed gadolinium enhancement, T1-mapping). In addition, CMR has greater spatial resolution than echocardiography and is an ionizing radiation-free technique that is preferred over CT angiography (CTA) imaging when possible in younger patients and those who will likely have multiple interval imaging studies over their lifetime. Contrast-enhanced (gadolinium-based) CMR or cine CMR (without contrast media) is indicated in patients with BAV in the following situations: (i) when morphology and/or diameter of the aortic sinuses, STJ, ascending aorta or arch cannot be assessed accurately or fully by echocardiography; (ii) in the serial evaluation of size and morphology of the aorta; at least yearly in patients with BAV with >45-mm diameters or with a family history of aortic dissection; (iii) when echocardiography-derived aortic diameters are discrepant with those obtained using CMR, CMR should be the modality of choice for interval aortic imaging.

In patients with aortic valve stenosis, cellular hypertrophy and diffuse fibrosis progress in a rapid and balanced manner but are reversible after AVR. Mid-wall late gadolinium enhancement may allow for improved clinical outcomes by prompting timely AVR in patients with BAV and AS with fibrosis [116].

Scientific evidence for new CMR applications in BAV research and its associated complications is emerging at a fast pace. For example, 4D-flow CMR has shown potential value in the clinical setting when examining traditional risk factors for maximal aortic diameter (age, gender, body surface area, peak valve velocity and valve morphology), and concepts related to flow displacement or eccentric blood flow have shown encouraging correlations with aortic dilatation [117]. 4D flow is an ECG-gated 3-dimensional (3D) phase contrast-CMR velocity encoding technique that allows the visualization of global and local 3D blood flow characteristics in the heart and large vessels. It also allows for the measurement of different components of vascular mechanics, such as the WSS, which is the viscous shear force that blood flow exerts tangentially to the vessel wall, a known haemodynamic measure implicated in vascular remodelling. As mentioned previously (see Section Bicuspid aortic valve aortopathy phenotypes), 4D flow has allowed the study of 3D aortic blood-flow dynamics and its dependence on BAV phenotypes. In right–left cusp fusion BAV, the flow impinges on the outer curvature of the proximal ascending aorta, whereas right non-cusp fusion displays a posteriorly directed flow jet directed towards the proximal ascending aorta and the outer wall of the distal ascending aorta (Fig. 23; Videos 13–15). Therefore, BAV phenotype-dependent flow abnormalities can cause increased aortic wall segmental stress, which partially explains BAV aortopathy phenotypic associations (i.e. right–left cusp fusion associated with root dilatation, right non-cusp fusion with ascending/arch dilatation). 4D flow CMR has the potential of becoming an imaging biomarker for risk stratification of BAV aortopathy.

CARDIAC COMPUTED TOMOGRAPHY CONSIDERATIONS

Cardiac CT, in particular CTA, owing to its superior spatial resolution and 4D display, provides unparalleled visualization of the aortic valve, the aortic root complex and the ascending aorta and serves as an important complement to echocardiography and other techniques in the evaluation of the BAV. Unlike echocardiography and CMR, CTA permits 4D isovolumetric imaging, which allows precise post hoc selection of imaging planes. A proper protocol is critical to an optimal quality CCT study, as has been described [118].

Appropriate evaluation of the aortic valve requires systolic phase imaging that is best achieved using retrospective ECG synchronized imaging. Whereas a full multiphase CCT data set allows for comprehensive imaging of the aortic valve (systole and diastole), coronary CTA is often performed during diastole. Given the high resolution of CCT, it is useful always to evaluate the aortic valve on all studies to determine, if possible, whether it is bicuspid or tricuspid. If imaging is done only during diastole, as is usually the case for routine coronary CTA, it may lead to overlooking the partial or complete fusion of the cusps and to mistaking the valve as tricuspid. One is unlikely to make this mistake if the tricuspid valve is symmetrical and has no leaflet/cusp thickening or asymmetrical calcifications. Although reconstructions to assess BAV can be obtained using preselected R-R intervals, it is advisable to identify the absolute delay after the R peak, usually specified in milliseconds, for best results. Tube modulation should be turned off during systole to reduce image noise during the critical phase of imaging. Intravenous contrast of 50–100 ml is administered with flow rates of 4–6 ml/s. Multiphasic data sets should be acquired and reconstructed with thin slices (<1 mm). Reformatting can be performed manually or with semiautomated software [118]. The annulus, sinuses and STJ levels can be defined using double-oblique views that permit measurement of the in-plane and through-plane aorta. For these reasons, CCT has been critical in surgical planning for conventional surgical AVR and has become the gold standard for pre-TAVR BAV evaluation [119]. In addition, because of high spatial resolution, ease of reformatting, significantly reduced radiation doses with newer scanners and the ability to simultaneously ‘clear’ the coronary arteries and avoid the need for coronary angiography in these younger patients with BAV, CCT/CTA is also the gold standard for preoperative surgical evaluation of BAV valvulo-aortopathy. A technique based on 3D multiplanar images has recently been described that optimizes visualization of the hinge points of the valve leaflets, allowing a distinction between commissures and raphes and thus may prove valuable in accurate characterization of the anatomy of the BAV [120]. More specifically, CTA can evaluate the specific BAV phenotypes including the presence of fusion and the orientation of the cusps in the 2-sinus BAV type, the extent of cusp symmetry, the degree of raphe calcification and the size of the cusps. The dimensions and morphology of the root, ascending aorta and arch can be optimally assessed on CTA to determine the presence of dilatation, its phenotype with respect to the aortic root or more distal aorta and the presence of aortic coarctation.

Akin to CMR, when the morphology and/or diameter of the aortic sinuses, the STJ or the ascending aorta cannot be assessed accurately or fully by echocardiography, CCT should be used; when echocardiography-derived aortic diameters are discrepant with those obtained with CTA, CTA should be the modality of choice for interval aortic imaging.

INTERVENTIONAL CARDIOLOGY CONSIDERATIONS

Based on numerous large randomized clinical trials, TAVR has emerged as an alternative to surgery for patients with severe aortic valve stenosis [121–123]. Within these studies, however, patients with BAV anatomy were excluded, due in part to concerns that TAVR in bicuspid valves may have suboptimal outcomes and/or increased complications. Indeed, with early generation transcatheter valves and limited use of CCT, TAVR in bicuspid anatomy was associated with lower device success rates and an increased incidence of significant paravalvular leak (PVL) [20, 124]. However, more recently, with careful CCT analysis as the standard for procedural planning and using current generation transcatheter valves designed to minimize PVL, non-randomized registry reports have suggested that TAVR in patients with BAV stenosis shows improving results [125, 126]. Yet, the impact of different bicuspid anatomies on TAVR outcomes remains an area of ongoing research and controversy. Although the classification system outlined here will help interventional cardiologists to categorize patients with bicuspid valves, to date there have been limited studies that have looked at TAVR outcomes stratified by bicuspid anatomy subtype (phenotype). This situation has been compounded by the fact that the Society of Thoracic Surgeons (STS)/American College of Cardiology Transcatheter Valve Therapy Registry, which serves as an archive for all patients undergoing TAVR in the USA, does not collect information on the type of BAV. In contrast, the STS Surgical Database Form began collecting information on the Sievers classification in 2017 for patients with bicuspid valve disease undergoing surgical AVR. Much of the limited data on TAVR outcomes based on different bicuspid anatomical forms comes from Jilaihawi et al. [35], who in 2016 proposed a TAVR BAV classification whereby they characterized patients with a bicuspid anatomy into 3 categories: ‘tricommissural’, ‘bicommissural raphe-type’ and ‘bicommissural non-raphe-type’ (Table 1). Using primarily early generation TAVR devices, they found that for patients with a 2-sinus BAV (Figs 11–13), increased intracommissural distance was associated with increased PVL. There was also a trend towards an increased incidence of new pacemakers in patients with fused BAV with left–right cusp fusion (Figs 6 and 7). Results from the STS/ACC/TVTRegistry compared the outcomes of new-generation, balloon-expandable TAVR devices for bicuspid versus tricuspid AS in 2,691 propensity score matched pairs of bicuspid and tricuspid patients [126]. There were no differences in mortality, symptom improvement, PVL and valve haemodynamics, but there was an increase in 30-day strokes and periprocedural complications requiring surgery in the BAV cohort. A recent study reported on 929 propensity matched pairs (bicuspid versus tricuspid) with self-expandable TAVR devices; the researchers found no difference in 30-day or 1-year all-cause death or stroke; however, patients with a bicuspid valve undergoing TAVR were more likely to require aortic valve reintervention at both 30 days and 1 year compared to patients with tricuspid valve undergoing TAVR [127]. Finally, the Bicuspid AS TAVR Registry, which included 1,034 patients with analyses of CCT images [128] showed that patients with a calcified raphe or excess leaflet calcification had increased early mortality and higher rates of periprocedural complications including aortic root injury and moderate or severe PVL (Fig. 24). Therefore, universal equipoise between TAVR and surgical AVR for BAV AS has not been attained, and it will be critical to better understand the relationship between bicuspid anatomy, calcification patterns and TAVR outcomes, in particular, whether there are specific bicuspid phenotypes that are less conducive to TAVR.

ACKNOWLEDGEMENTS

The authors wish to thank Carl Clingman, Senior Medical Illustrator, Biomedical & Scientific Visualization, Mayo Clinic and Pavel Zacek, Department of Cardiac Surgery, Charles University Hospital, Hradec Kralove, Czech Republic, for their illustrations.

Conflict of interest: Victoria Delgado discloses a financial relationship with Abbott Vascular, Edwards Lifesciences, GE Healthcare, MSD, Medtronic, and Novartis. Emmanuel Lansac: patent Extra Aortic ring annuloplasty device with Coroneo Inc. Phillippe Pibarot: funding from Edwards Lifesciences and Medtronic for echocardiography core laboratory services with no personal compensation. Michael A. Borger: discloses a financial relationship with Edwards Lifesciences, Medtronic, Abbott, and CryoLife. John K. Forrest: grant support and consultant to: Edwards Lifesciences, Medtronic Inc. John Webb: consultant to: Edwards Lifesciences, Abbott, Boston Scientific. Martin B. Leon: institutional clinical research grants from Abbott, BSC, Edwards and Medtronic. Michael Markl: research support—Siemens Healthineers, Research Grant—Circle Cardiovascular Imaging, Consulting—Circle Cardiovascular Imaging, Research Grant—Cryolife Inc. Victor A. Ferrari: Senior Advisory Board, Journal of Cardiovascular Magnetic Resonance. Philipp Blanke: consultant for Edwards Lifesciences and Circle Cardiovascular Imaging; and provides CT core lab services for Edwards Lifesciences, Medtronic, Neovasc, and Tendyne Holdings, for which he receives no direct compensation. Ruggero De Paulis: patent on aortic root graft with Terumo Aortic. Consultant for Edwards Lifesciences, Medtronic, and Terumo Aortic. Maurice Enriquez-Sarano: consulting fees from Edwards LLC, Cryolife, and ChemImage, Inc. The Department of Cardiology of the LUMC received unrestricted research grants from Abbott Vascular, Bayer, Biotronik, Bioventrix, Boston Scientific, Edwards Lifesciences, GE Healthcare, and Medtronic. The other authors report no conflict of interest.

Reviewer information

European Journal of Cardio-Thoracic Surgery thanks Martin Misfeld, Patrick Myers, Ulrich Otto von Oppell Richard Weisel and the other, anonymous reviewer(s) for their contribution to the peer review process of this article.

REFERENCES