This study was approved by the Ethics Committee of Shanghai Chest Hospital (ethics number: IS23011), and the requirement for informed consent was waived because of the retrospective nature of the study.

From October 2021 to September 2022, 41 patients with BAV underwent aortic valve repair for AR with or without ascending aortic aneurysms at the Shanghai Chest Hospital. Among them, 20 patients with type 1 R/L who underwent isolated valve repair were retrospectively analyzed and classified into two groups: patients who underwent external suture annuloplasty (10 patients before April 2022) and patients who underwent internal suture annuloplasty (10 patients after April 2022). The remaining 21 patients were excluded based on the following criteria: patients without a complete diagnostic workup with adequate quality pre- and post- operative ECG-gated CT scans; concurrent aortic stenosis with more than moderate severity; patients with aortic root dilatation with a cut-off diameter of 45 mm or type A dissection involving the aortic root, thereby requiring additional root reimplantation; and BAV of subtypes other than type 1 R/L. The baseline characteristics of patients with preoperative echocardiographic data are presented in Table 1.

The choice between the two annuloplasty techniques has evolved with time and surgical access. Before April 2022, full median sternotomy was routinely performed for surgical access in the external suture annuloplasty group. From then on, we turned to partial upper sternotomy for the group with internal suture annuloplasty, with less impact on chest wall integrity and a better cosmetic outcome, which is meaningful for young patients. However, it is more difficult to perform deep root dissection and hemostasis after recovery of heart beating for external annuloplasty. Therefore, we adopted the less commonly used internal suture annuloplasty approach to avoid these shortcomings.

After cardiopulmonary bypass was initiated, the aorta was cross-clamped, the heart was arrested, and the following steps were performed in the group with external suture annuloplasty. The aorta was transected 5 mm above the sinotubular junction. Commissural resuspension sutures are used to expose the aortic valve. Aortic valve leaflets were inspected for tissue quality and quantity by measuring the geometric height of the non-fused cusp and half of the fused R/L leaflet. The decision for aortic valve repair was based on a comprehensive evaluation of the degree of raphe fusion, commissural orientation, and leaflet mobility. The right coronary (RC) and left coronary (LC) origins were isolated but not detached from the sinus through blunt dissection with a right-angle clamp and were secured by a stay suture passing beneath the ostia. Dissection was performed externally along the aortic root between the right and left coronary origins. Due to variations in the level of the VAJ relative to the VBR and the frequent presence of “sinking sinus” in this area [22], we did not aim to reach a deep subvalvular plane to avoid extensive myocardial dissection, but only to create a proper space for passing the needle of suture annuloplasty externally, as suggested in the literature [6]. The root dissection was continued along the left and non-coronary (NC) sinuses posterior to the level of the VBR, which was easily accomplished with routine dissection, as the VAJ is either almost at the same level as the VBR or absent where the curtain is located. Annulus dilatation, defined as a diameter $>$25 mm measured in a preoperative imaging study and confirmed intraoperatively with a Hegar dilator, was treated by external circular suture annuloplasty using a CV-0 polytetrafluoroethylene single-needle suture (W. L. Gore & Associates, Newark, DE, USA) according to Schaf̈ers’ technique [6, 14]. We aimed to place the suture deep while avoiding injury to the surrounding structures.

In the internal suture annuloplasty group, aortic root dissection was not required between the right and left coronary origins. Similar to the technique described by Holst et al. [23], following the VBR plane under direct vision, the suture was started from inside the left ventricular outflow tract and passed outside the area of the NC sinus. Care was taken to elevate the internal suture line along the membranous region to avoid injury to the conduction system.

At all times, the sutures were tightened across the aortic annulus using a Hegar dilator. The sizing strategy with the Hegar dilator was more aggressive, with a smaller diameter in the internal group to increase coaptation length (see Table 1 for Hegar diameter). Thus, we aimed to delay the recurrence of AR in the event of postoperative annular redilatation. This is particularly important considering there is limited evidence of the effectiveness of annular stabilization using the internal suture technique.

Leaflet repair procedures were performed to achieve valve competency in both groups. The effective height of the cusp was assessed with a caliper, and any cusp prolapse was corrected by free margin plication with 5-0 polypropylene suture to obtain an equivalent free margin of both leaflets with an effective height of 9–10 mm.

Contrast-enhanced ECG-gated CT of the aortic root was performed in both groups 2–3 days before and 5–7 days after surgery. The CT protocol and image reconstruction methods were introduced in our previous study [21]. All CT data were systematically analyzed using Osirix software version 9.5.1 (Pixmeo, Geneva, Switzerland). Multiplane reconstruction was performed to visualize the planes of interest (VBR and VAJ) for BAV repair.

Through dedicated multiplanar reconstruction with the application of a double oblique view, an axial plane perpendicular to the long axis of the aortic root was obtained. The axial image passing through the nadir in the NC sinus and the midportion of the respective half of the fused anterior leaflet (approximate nadir of each fused leaflet) was identified as the VBR in type 1 R/L BAV. In both groups, preoperative and postoperative perimeters and areas were measured during systole and diastole at 20% and 80% of the R-R interval, respectively [24] (Fig. 1).

Fig. 1.

Definition and measurement of VBR by reconstructed ECG-gated computed tomography. The axial image passing through the nadir in the NC sinus and in the midportion of the respective half of the fused anterior leaflet (approximated nadir of each fused leaflet) was identified as the VBR in type 1 R/L BAV. The pictures show the annular area and perimeter before BAV repair in systole (A) and diastole (B) at 20% and 80% of the R-R interval (green circle). VBR, virtual basal ring; ECG, electrocardiogram; BAV, bicuspid aortic valve; NC, non-coronary.

According to the anatomic study, the VAJ was more of a circular structure consisting of different tissue components (interventricular septum, aortomitral curtain, and connective tissue) rather than a planar circle strictly above the VBR, as initially thought. Owing to its three-dimensional curvilinear configuration, the height of the VAJ relative to the VBR varies along the root circumference. We measured these heights in 6 long axis views of aortic root perpendicular to the VBR plane at 20% of the R-R interval preoperatively, with each view corresponding to a specific predefined landmark of the aortic root circumference. These six landmarks were set at the nadir in the NC sinus, RC and LC ostium, right/non (RN) and left/non (LN) commissure, and non-functional commissure adjacent to the raphe of the fused cusp (raphe) (Fig. 2).

Fig. 2.

Identification of VAJ and measurement of heights from VBR at predefined anatomical landmarks. (A) The short axis view of the aortic root. Anatomical landmarks are identified on short axis view of the aortic root by rotation of 2 orthogonally-crossed plane perpendicular to the VBR. Green line: orthogonally-crossed planes in the long axis view of the aortic root. (B) Height of VAJ from VBR in the long axis views of the aortic root along the anterior aortic annulus. (C) Height of VAJ from VBR in the long axis views of the aortic root along the posterior aortic annulus, the white line did not appear in the first and second pictures as it coincides with the yellow line. Yellow line: level of the VBR. White line: level of the anatomical landmarks. Red arrows: position of anatomical landmarks. VAJ, ventricular-aortic junction; VBR, virtual basal ring; RC, right coronary; LC, left coronary; LN, left-non; NC, non-coronary; RN, right-non.

The CV-0 suture was not radiopaque; therefore, the annuloplasty positioning was assessed based on the presence of a characteristic narrowing effect induced by tightening the suture within or outside the aortic root on postoperative CT images. The level of suture positioning, either higher or lower than the VBR plane, produces a notch over the aortic root or LVOT on long-axis views of the aortic root perpendicular to the VBR plane. This narrowing effect is a feature of root distortion, referred to as the waist sign (Fig. 3A). The distance between the waist sign and VBR plane was measured in the same manner as the height of the VAJ using the aforementioned landmarks (Fig. 2). The distance between the annuloplasty suture and the VBR was considered as 0 if no waist signs were identified (Fig. 3B).

Fig. 3.

Postoperative long axis views of the aortic root after BAV repair. (A) Waist sign caused by the narrowing effect of the external suture annuloplasty. (B) No waist sign was seen by BAV repair with internal suture annuloplasty. Yellow line: level of the virtual basal ring. Green line: orthogonally-crossed planes in the long axis view of the aortic root. The asterisk: right coronary ostium. Black arrow: waist sign. BAV, bicuspid aortic valve.

Measurement data were assessed to compare different patient groups or subgroups using the chi-square test and Fisher’s exact probability test for categorical variables and the two-tailed Student’s t-test for continuous variables. Continuous variables were summarized as mean and standard deviation. Categorical variables were expressed as numbers and percentages. A p-value of 0.05 was considered statistically significant. All analyses were conducted using SPSS software (version 22.0; IBM Corporation, Chicago, IL, USA).

The height of VAJ related to the VBR circumference was uneven with the highest at the raphe and RC ostium (7.9 $\pm$ 1.5 mm, 7.7 $\pm$ 3.3 mm, respectively), and VAJ is at the same level as the VBR at the NC sinus and LN commissure. The height of the RC and LC ostia from the VBR and the distance between them, both of which are related to the complexity of anterior annulus dissection, were also assessed. There were no significant differences between the internal and external groups in any of these parameters (Table 2).

Results from the external group showed that the suture annuloplasty was farthest from the VBR along the anterior annulus at the RC ostium (5.3 $\pm$ 1.1 mm), followed by the raphe (4.8 $\pm$ 1.0 mm). Conversely, the suture was closest to the VBR at the LC ostium (1.9 $\pm$ 1.7 mm), LN commissure (2.2 $\pm$ 1.8 mm), and NC sinus (2.3 $\pm$ 1.5 mm), where VAJ was also closer to the VBR plane. The internal group showed minimal identification of waist sign, except at the RN commissure, where sutures deviated the most from the VBR, as seen in the external group (4.3 $\pm$ 1.4 mm vs. 5.8 $\pm$ 1.5 mm, p = 0.026). This deviation was intended to prevent injury to the conduction system in both groups. However, the internal group had significantly shorter distances between the suture annuloplasty and the VBR than the external group at all landmarks (Table 3).

The preoperative geometrical parameters of VBR were comparable between the two groups during the cardiac cycle. Both groups showed a significant reduction in the systolic and diastolic annular dimensions after annuloplasty (Table 4). To account for the different annular reduction strategies between the groups (smaller Hegar dilator diameter used in the internal group compared to the external group, 22.1 $\pm$ 1.4 mm vs. 24.0 $\pm$ 1.4 mm, p = 0.008), postoperative VBR was not directly analyzed. Instead, the effect of annuloplasty was assessed by comparing annular expansibility (variation in area and perimeter between systolic and diastolic sequences) and the dimensional discrepancy between postoperative VBR and Hegar dilator between the two groups. Postoperative annular area expansibility decreased in the internal group compared to preoperative levels (4.9 $\pm$ 2.3% vs. 8.9 $\pm$ 5.5%, p = 0.038) with a trend of decreased perimeter expansibility (2.3 $\pm$ 2.4% vs. 3.6 $\pm$ 3.9%, p = 0.259), while no significant change in expansibility was found in the external group. The internal group showed less discrepancy between the VBR and Hegar dilator, both in area and perimeter, at systole and diastole compared with the external catheter group (Table 5).

These findings need to be confirmed in a larger number of patients undergoing isolated BAV repair. Annular reduction strategies with the Hegar dilator differed between the two groups, which made a direct comparison of the postoperative annular reduction effect between the two annuloplasty techniques impossible. Our study had a single-center design and only immediate postoperative outcomes; to obtain more definite conclusions on long-term annulus stability after suture annuloplasty, a longer and more complete follow-up is required, incorporating the validation of MS-CT measurements by a core laboratory. A multicenter, prospective, randomized trial is required to minimize this bias.