Bicuspid aortic valve morphology and aortic valvular outflow jets: an experimental analysis using an MRI-compatible pulsatile flow circulation system

The characteristics of aortic valvular outflow jet affect aortopathy in the bicuspid aortic valve (BAV). This study aimed to elucidate the effects of BAV morphology on the aortic valvular outflow jets. Morphotype-specific valve-devising apparatuses were developed to create aortic valve models. A magnetic resonance imaging-compatible pulsatile flow circulation system was developed to quantify the outflow jet. The eccentricity and circulation values of the peak systolic jet were compared among tricuspid aortic valve (TAV), three asymmetric BAVs, and two symmetric BAVs. The results showed mean aortic flow and leakage did not differ among the five BAVs (six samples, each). Asymmetric BAVs demonstrated the eccentric outflow jets directed to the aortic wall facing the smaller leaflets. In the asymmetric BAV with the smaller leaflet facing the right-anterior, left-posterior, and left-anterior quadrants of the aorta, the outflow jets exclusively impinged on the outer curvature of the ascending aorta, proximal arch, and the supra-valvular aortic wall, respectively. Symmetric BAVs demonstrated mildly eccentric outflow jets that did not impinge on the aortic wall. The circulation values at peak systole increased in asymmetric BAVs. The bicuspid symmetry and the position of smaller leaflet were determinant factors of the characteristics of aortic valvular outflow jet.

Bicuspid aortic valve (BAV) is the most prevalent congenital heart malformation; it occurs in 0.5-2% of the population 1 . Up to 80% of patients with BAV develop ascending aortic dilatation, called aortopathy, with serious clinical sequelae, such as 6-9 times increased risk of aortic dissection [1][2][3][4] . The natural history of aortopathy in BAV is quite different from that in tricuspid aortic valve (TAV). Progression of aneurysmal diameter is significantly faster with BAV than with TAV, even after aortic valve replacement 5,6 , possibly due to underlying intrinsic medial fragility combined with long-term exposure to BAV-related hemodynamic alterations. Therefore, the current guidelines on aortic diseases generally recommend early surgical intervention in a significantly dilated ascending aorta with BAV-defined as maximum diameter > 45 mm-to prevent sudden death from aortic dissection or rupture [7][8][9] . However, the treatment of a mildly dilated ascending aorta (40-45 mm) at the time of aortic valve surgery remains controversial due to the markedly heterogeneous characteristics of aortopathy in BAV 10,11 .
With the recent improvements in pre-procedural imaging and devices, transcatheter aortic valve implantation (TAVI) is considered a safe and feasible treatment option for aortic valve stenosis with BAV 12 . Furthermore, it has been reported that TAVI could be performed safely in patients with a mildly dilated ascending aorta and low intraprocedural risk 13 . Given the expanding indications for TAVI in patients with BAV, individualized strategies to manage BAV based on a reliable index for the prediction of the risk of aortopathy are required. www.nature.com/scientificreports/ It is being increasingly recognized that aortopathy in BAV is affected by valve-related hemodynamic alterations. Recent studies based on time-resolved three-dimensional (3D) phase contrast magnetic resonance imaging (4D-flow MRI) have demonstrated that BAV morphology affects the eccentricity of the outflow jet and helical systolic flow, which are related to aortic wall remodeling 14,15 . However, due to several cofactors that affect the aortic hemodynamics, such as age, blood pressure, valvular function, and dimensions of the aorta, the relationships between BAV morphology and hemodynamics are not well understood clinically. This study aimed to elucidate the effects of BAV morphology on the characteristics of aortic valvular outflow jet using an MRI-compatible pulsatile flow circulation system to identify BAV phenotypes at high risk of aortopathy.

Methods
Tissue-based aortic valve models. Three types of aortic valve models with annulus diameter of 24 mm, including TAV, asymmetric BAV (ABAV), and symmetric BAV models, were prepared using bovine pericardium and a short segment of bovine thoracic descending aorta. The tissues were obtained from a local abattoir. The followings were three steps in devising the aortic valve model: (1) preparation of the bovine pericardium and a short segment of the aorta; (2) creating pericardial valve leaflets; and (3) suturing the pericardial leaflets to a tubular aorta. Preparation of the bovine pericardium and aorta began with manual removal of fat on their outer surfaces. The pericardium was rinsed using physiological saline solution and not treated with glutaraldehyde. The bovine aorta was excised to prepare a short segment of the tubular aorta with an inner diameter of 24 mm and length of 10 cm. The intercostal arteries branching from it were ligated with 5-0 monofilament.
Morphotype-specific valve-devising apparatuses, complementary pair of a leaflet-trimming template, and leaflet-suturing guide were developed to prepare the aortic valve models. Valve-devising apparatuses were designed using computer-aided design (CAD) software (Solidworks 2019, Dassault Systemes Solidworks Corporation, Paris, France) using the following data regarding the geometry of the aortic cusp: annulus diameter, 24 mm; coaptation height, 6.2 mm; commissural length, 5.2 mm; leaflet length, 20 mm; and depth of the central portion of the leaflets from the commissures, 2.7 mm (Fig. 1). The coaptation height of the aortic valve was based on a previous study in healthy humans 16 . The values of the commissural length and the depth of the central portion of the leaflets from the commissures were determined based on the average values in the porcine aortic root. Seven porcine aortic roots with an annulus diameter of approximately 24 mm (mean annulus diameter: 24 ± 0.4 mm, n = 7) were obtained from the local abattoir. Using an X-ray micro-computed tomography (CT) (TDM1300-IS, Yamato Scientific, Tokyo, Japan), the geometric valves were measured under the condition that the aortic valves were closed under the static pressure of 30 mmHg. The cusp angles were defined as 120°-120°-120°, 240°-120°, and 180°-180° in TAV, asymmetric BAV, and symmetric BAV models, respectively (Fig. 1a). CAD data were exported to a 3D printer (Eden260vs, Stratasys, Valencia, CA, USA) and modeled using acrylic resin (FullCure720, Altech, Tokyo, Japan).
In the valve-devising procedure, the leaflets were first cut from the bovine pericardium using a morphotypespecific leaflet-trimming 3D template (Fig. 1a). The bovine pericardium was pressed on the template and cut along the edge with 1-mm seam allowance (Fig. 1b). Next, a complementary morphotype-specific leaflet-suturing guide ( Fig. 1c) was inserted into the short segment of the tubular aorta. The annular margins of the leaflets were sutured with 5-0 monofilament continuous sutures to the aortic wall along the annular curve of the leafletsuturing guide. The commissural margins of adjacent leaflets were sutured together with interrupted sutures and attached to the aortic wall (Fig. 1d).
Classification scheme of BAV morphology. The classification of BAV morphology was based on two characteristics: the symmetry and position of the leaflets (Fig. 2). The spatial position of the leaflets was defined based on the aortic quadrant faced by the leaflets. ABAV models included three morphologies (ABAV-1, ABAV-2, and ABAV-3); they included the smaller leaflet facing the right anterior (RA), left posterior (LP), and left anterior (LA) quadrants of the ascending aorta, respectively. Symmetric BAV models included two morphologies (SBAV-1 and SBAV-2) with symmetric leaflets facing RA-LP and LA-right posterior (RP) quadrants of the ascending aorta, respectively. MRI-compatible pulsatile flow circulation system. The MRI-compatible pulsatile flow circulation system, which duplicated the thoracic aortic circulation, was composed of an elastic left ventricle (LV) model, aortic valve model, aortic arch model, aortic compliance chamber, resistive unit, left atrium pressure chamber, and polymeric valve as an alternative to the mitral valve (Fig. 3a,b). The elastic LV model was operated using a pneumatic control system (VCT-50 (χ), Nippro, Osaka, Japan) to simulate periodically contracting and relaxing myocardium. The aortic arch model with a tunnel-shape ( Fig. 3c) was created using silicone (Shin-Etsu Silicone [KE-1603-A: KE-1603-B: KF-96-50CS; 10:10:1], Shin-Etsu Chemical, Tokyo, Japan) to provide a static signal for background phase correction during MRI. The 3D aortic arch model with the geometries based on the literature values of young human aortic arch were devised ( Fig. 4) 17,18 . The aortic compliance chamber was installed behind the aortic arch model to simulate arterial pulse pressure. The working fluid, which was physiological saline solution, was recirculated to the left atrium chamber via the peripheral resistive unit. Pneumatic positive pressures in the LV model were adjusted to drive pulsatile flow through the flow circuit at systolic LV pressure of 150 mmHg. A pulse rate of 70 beats per minute (bpm) and systolic fraction of 35% were set in the pneumatic control system. Valvular hemodynamics in aortic valve models. Valvular hemodynamics of the aortic valve models, including the mean aortic flow, forward flow, regurgitation (regurgitant volume related to the dynamics of the valve closure), and leakage were measured at a pulse rate of 70 bpm, systolic fraction of 35%, and target aortic  14) and an oblique-sagittal slab encompassing the aortic arch model. A simulated electrocardiographic signal with an RR interval of 857 ms, corresponding to a pulse rate of 70 bpm with mean flow of 4.9 ± 0.10 L/min, was used for gating. Scan times ranged from 10 to 15 min (mean, 12 min). Before analyses, the images were corrected for Maxwell phase effects, encoding errors due to gradient field distortions, and eddy current-related phase offsets.
Qualitative evaluation of eccentric outflow jets at peak systole. Corrected velocity data were imported into 3D visualization software (iTFlow; Cardio Flow Design, Tokyo, Japan) to analyze the fluid flow in the aortic arch model using velocity vector fields. Streamlines, which are instantaneous paths tangential to The degree of outflow jet eccentricity, defined as the flow displacement of the high-velocity outflow jet from the center of the vessel, was evaluated in three cross-sections of supra-valve (S 0 ), at the level corresponding to pulmonary bifurcation (S 1 ), and distal edge of the brachiocephalic artery (S 2 ), as described previously 15 . It was graded as follows: no eccentric jet = centrally focused high-velocity systolic flow occupying a majority of the vessel lumen; mild eccentric jet = high-velocity systolic flow occupying between 1/3 and 2/3 of the vessel lumen; and marked eccentric jet = high-velocity systolic flow occupying ≤ 1/3 of the vessel lumen.
Quantitative analysis of helical flow components at peak systole. The circulation values (m 2 /s) that corresponded to the surface integral values of two-dimensional (2D) vorticity (/s) were calculated at four cross-sections (level corresponding to the pulmonary bifurcation, S 1 ; distal edge of the brachiocephalic artery, S 2 ; distal edge of the left subclavian artery, S 4 ; and fourth thoracic vertebral body upper edge, S 5 ) to quantify the magnitude of helical flow components at the four aortic levels, and to investigate the level of the aorta at which helical flow developed, maximized, and faded out. For each cross-section, a 2D vorticity (ω) map of timeresolved through-plane vorticity was computed to calculate the circulation value (Γ) according to the following formula:

Results
Valvular hemodynamics in aortic valve models. The aortic flow rate waveform in the TAV is shown as a representative in Fig. 5a. There were no differences in the aortic flow rate curves among the 6 valve morphologies. The peak flow rate was slightly higher in the TAV in comparison with those of the BAVs. Among the five BAVs, no differences were observed in the mean aortic flow, forward flow, regurgitation, and leakage (Fig. 5b). The data of BAVs were comparable to those of the TAV; however, the forward flow in TAV was significantly greater than those in BAVs (see Supplementary Table S1 online).
Qualitative evaluation of eccentric outflow jets at peak systole. The BAV models demonstrated morphotype-dependent aortic valvular outflow jets at peak systole (Figs. 6, 7). In the ABAVs, markedly eccentric outflow jets exclusively directed to the aortic wall facing the smaller leaflet were observed. In the ABAV-1, an RA-directed high-velocity outflow jet impinged on the outer curvature of the ascending aorta. In the ABAV-2, an LP-directed high-velocity outflow jet shifting to the proximal transverse arch was observed, and the longtraveling eccentric outflow jet impinged on the outer curvature of the proximal transverse arch. In the ABAV-3, a short-traveling, markedly eccentric outflow jet impinged on the LA quadrant of the ascending aorta at the supra-valve level. In the symmetric BAVs, mildly eccentric outflow jets were present in the ascending aorta.
Mildly RA-and LA-diverted high-velocity outflow jets that did not impinge on the aortic wall were observed in the SBAV-1 and SBAV-2, respectively. In the TAV model, laminar flow was observed at peak systole.
Quantitative analysis of helical flow components at peak systole. The maximum values of circulation (m 2 /s) at peak systole measured at the four cross-sections are presented in Fig. 8. In the asymmetric BAVs, marked increases in the magnitudes of circulations were observed in the ascending aorta, which decreased to

Discussion
The findings of this study revealed that BAV morphology did not affect the mean aortic flow and leakage, but did significantly alter the aortic wall location where the aortic valvular outflow jet impinged along with the magnitude of circulations at peak systole. The present results indicated that the behavior of the aortic valvular outflow jet depended on the bicuspid symmetry and the position of the smaller leaflet. In asymmetric BAVs, the position of the smaller leaflet affected the impinging portions of the outflow jet on the aortic wall along with the markedly eccentric aortic valvular outflow jets at peak systole. The impinging locations were consistent with the reported regions of increased risk of future aortic events 19 . Furthermore, in asymmetric BAVs, the helical  20 . In symmetric BAVs, the abnormal helical flow did not develop at peak systole but developed mildly eccentric aortic valvular outflow jets that did not impinge on the aortic wall. These findings provide novel insights into the prediction of risk of aortopathy in BAV. The ABAV-1 morphology, which resulted in a markedly eccentric outflow jet impinging on the aortic outer curvature, might be a risk factor of asymmetric aneurysm bulging toward the outer curvature of the ascending aorta that is uniquely and frequently observed in patients with BAV 21 . The ABAV-2 morphology, which resulted in a long-traveling outflow jet impinging on the outer curvature of the proximal transverse arch, might be a risk factor of aortic aneurysm involving the transverse arch. The ABAV-3 morphology, which resulted in a markedly eccentric outflow jet impinging on the supra-valvular aortic wall, might be a risk factor of aortic aneurysm involving the proximal ascending aorta. In symmetric BAVs, secondary factors, such as valve stenosis, deformation, and aortic dimensions may affect the characteristics of the aortic valvular outflow jet, which are expected to play roles in  20,22 . The present results also suggest that aortopathy in BAV may not involve the distal transverse arch, irrespective of the valve morphology. It has been recognized that BAV morphology, categorized according to the bicuspid fusion pattern and number of raphes 23 , affects the natural history of aortopathy in BAV 24,25 . Recent advances in 4D-flow MRI have suggested that outflow jet eccentricity is a potential index for the aortopathy phenotype in BAV; it was reported that type 3 aortopathy (involving the aortic arch) was more common in the right and non-coronary cusp fusion phenotype (RN-BAV), whereas type 2 aortopathy (dilatation in the ascending aorta) was more frequent in the right and left coronary cusp fusion phenotype (RL-BAV) 14,26,27 . However, due to the anatomical varieties and hemodynamic differences among patients clinically, the relationship between BAV morphology and characteristics of the aortic valvular outflow jet has not been determined. Furthermore, uncommon BAV phenotypes, such as left and non-coronary cusp fusion phenotype (LN-BAV), symmetric BAV, or unclassifiable BAV have been difficult to study clinically.
The present study elucidated the relationships between BAV morphology and aortic valvular outflow jets based on a new classification of BAV morphology using 4D-flow MRI, MRI-compatible pulsatile flow circuit, and different morphological types of BAV models that were developed utilizing 3D-modeling technologies. The characteristics of aortic valvular outflow jet observed in the ABAV-1 and ABAV-2 morphologies were comparable to previously reported 4D-flow MRI findings for RL-BAV and RN-BAV, respectively, probably due to the similarity of the position of smaller leaflet. Furthermore, this study uncovered the hemodynamic features of LN-BAV (mostly included in the ABAV-3 morphology) and symmetric BAV, which have hardly been observed clinically. The classification scheme for BAV morphology proposed in the present study was based on two characteristics that directly affected the location of the outflow jet impinging on the aortic wall-the symmetry and position of the leaflets. The spatial position of the leaflets was defined based on the aortic quadrant faced by the smaller leaflet independent of the position of the raphe, commissures, and coronary artery orifices. Our classification of BAV morphology may be useful for predicting the risk of aortopathy in BAV. To the best of our knowledge, this is the first report to demonstrate the effects of BAV morphology on aortic hemodynamics using an MRI-compatible pulsatile flow circulation system. Several studies have experimentally investigated aortic hemodynamics using an MRI-compatible pulsatile flow circuit 28,29 . However, none of them www.nature.com/scientificreports/ included MRI-compatible pulsatile flow circulation system that incorporated different morphological types of tissue-based aortic valve models. The methodology of modeling aortic valves using morphotype-specific valve-devising apparatuses enabled the creation of several BAV models with reproducibility. The analysis of hemodynamic features in uncommon BAVs was performed using the valve-devising technique. Furthermore, the morphologically relevant aortic arch model, designed to minimize artifacts during MRI, enabled the visualization and quantification of 4D-flow MRI data with sufficient quality. The MRI-compatible pulsatile flow circulation system developed in the present study may be useful for studying valve-related aortic hemodynamics, especially in uncommon valve phenotypes that are rarely encountered clinically.
The first limitation of the present study is the relatively large regurgitation, which was possibly due to the absence of the sinus of Valsalva and the lack of the closing force of the sinus vortices. However, a tubular aorta with predefined dimensions, not an aortic root, was intentionally used to create aortic valve models to maintain the uniformity and reproducibility of valve modeling. Regurgitation was maintained under 30% in forward flow, which was considered clinically acceptable in each morphology. Second, the 3D helical flow parameters were analyzed at the prescribed cross-sections over the aortic arch model. The major helical flow components, such as rotation around an axis parallel to the main flow, could be identified and quantified in this study. Third, the circulation values (m 2 /s) at peak systole are measured at S 1 , S 2, S 4 , and S 5 (one sample, each). In the three asymmetric BAVs, the magnitudes of circulation values are markedly high in the ascending aorta, which decrease to approximately half and less than one third in the middle and distal parts of the aortic arch, respectively. In the two symmetric BAVs, the magnitudes of circulation values are comparable to that in TAV.

Scientific Reports
| (2021) 11:2066 | https://doi.org/10.1038/s41598-021-81845-w www.nature.com/scientificreports/ physiological saline solution used as the working fluid of the pulsatile flow circulation system did not reflect the viscosity of blood in human body. The use of saline solution might result in a relatively higher Reynolds number. However, the aortic flow belongs to the highly turbulent flow regime, where the effects of the viscosity is negligible 30 . We believe that the relative differences between the present results and in-vivo findings are expected to be small. Finally, because the present study only aimed to demonstrate the effects of BAV morphology on aortic hemodynamics, the effects of other factors, such as valve stenosis, deformation, and aortic arch dimensions were simplified. The effects of these factors on hemodynamics will be studied in the future.
In conclusion, we elucidated the relationships between BAV morphology and characteristics of aortic valvular outflow jet using an MRI-compatible pulsatile flow circulation system for the first time. This study revealed that the bicuspid symmetry and the position of smaller leaflet in BAV were determinant factors of the characteristics of the aortic valvular outflow jet. Asymmetric BAV morphology with a smaller leaflet facing the right anterior quadrant of the ascending aorta may be a risk factor of asymmetric aneurysm bulging toward the outer curvature of the ascending aorta. Asymmetric BAV morphology with the smaller leaflet facing the left posterior quadrant of the ascending aorta may be a risk factor of aortic aneurysm involving the transverse arch. Asymmetric BAV morphology with the smaller leaflet facing the left anterior quadrant of the aorta may be a risk factor of aortic aneurysm involving the proximal ascending aorta. Symmetric BAV morphologies may contribute to the progression of aortopathy along with other factors that affect the aortic valvular outflow jet. Our classification of BAV morphology may be useful in predicting the risk of aortopathy in BAV.