Computed tomography−based prediction of commissural positions facilitates valve-sparing aortic root replacement

Figures

• 

  Predicting commissural positions by computed tomography facilitates David reimplantation.

• 

  Figure 1. Study eligibility criteria. CT, Computed tomography.

• 

  Figure 2. Preoperative aortic root/valve measurement parameters. Schematic drawings of the aortic root structures are shown as a 3-cusp view (A); a head-on view (B); and a cusp overlap view (C). VBR areas are depicted in pink. The ICDs are shown as bidirectional green solid lines (a-c). Angles α, β, and γ were approximated from the ICDs a-c. The predicted CH (f) (bidirectional blue solid arrows) was calculated from dimensions d and e (bidirectional blue dotted arrows) using the Pythagorean theorem. For the CT analyses, the aortic root structure is visualized in 2-dimensional images with measurement parameters (D-G) and 3-dimensional images by the algorithm (H-K), showing the VBR (D and H), CCA (E and I), ICD (F and J), and CH (G and K) in blue, red, and orange lines. For the TEE analysis, the VBR diameters in 3 directions between their nadirs and the opposite commissure sides are shown as bidirectional yellow, red, and blue arrows (L-O). The formulas used to calculate each parameter are shown in the lower panel. VBR, Virtual basal ring; ICD, intercommissural distances; CH, commissural heights; CT, computed tomography; CCA, commissure circle area; TEE, transesophageal echocardiography.

• 

  Figure 3. Intraoperative adjustment of the commissural positions and cusp repair during David reimplantation. The number of patients (A) and commissures (B) requiring CP adjustments based on inspection of the aortic valve and the water test, as well as those requiring cusp repair, are shown. CP, Commissural position.

• 

  Figure E1. Bland-Altman plots of the manual and automatic measurement values of the patients who underwent David reimplantation. A, Estimated VBR diameter; B, commissure circle area; C, intercommissural distance; and D, commissural height. The solid red lines represent the means of the differences between the algorithm measurement and those of expert 1. The dashed red lines depict the limits of agreements (±1.96 SD). Cases that were outside the limits of agreements are marked in blue, green, or red on the basis of the reason for exceeding the agreement limits. Blue indicates poor data conditions (poor contrast or blurred images). Green indicates anatomical reasons such as irregular VBR shapes (CT #7), disproportionate SOVs (CT #7), and chronic type-A dissection (CT #20 and Control #8). Red marker indicates discrepancies between the measurement criteria of the algorithm and the expert. In the cases of CT #13 and control #1, the shadow of the cusp was fused with the thickened aortic wall on imaging, preventing accurate commissure detection. VBR, Virtual basal ring; SD, standard deviation; CT, computed tomography; SOV, sinus of Valsalva; Suppl, supplementary case.

• 

  Figure E2. Kaplan-Meier curves for recurrent AR above the mild level after David reimplantation. The graph compares the Kaplan-Meier curve of patients whose David reimplantation planning was based on CT with that of a control group in whom conventional planning methods were used. Dotted lines show 95% confidence limits. The rates of recurrent AR above the mild level 1 year after David reimplantation were 11.1% in the CT group and 20.0% in the control group (Log-rank test, P = .21). AR, Aortic regurgitation; CT, computed tomography.

• 

  Figure E3. Representative cases of outliers from the Bland-Altman plots of the measurement values of patients who underwent David reimplantation. A-C, The algorithm's automated measurements underestimated irregular VBR in case CT #7. D–F, In the same patient (CT #7), the algorithm wrongly detecting the commissure positions in disproportionately dilated SOVs. G-I, In a patient with chronic type A dissection (CT #20), the algorithm identified the commissural positions for the ICD measurement on the intimal flap, whereas expert 1 set them at their original position, assuming that the aortic wall was not dissected. CT #7 and CT #20 are marked in green on the Bland-Altman plots (Figure E1), indicating anatomical reasons for the measurement discrepancies. J-L, In CT #13 the cusps' shadow was fused with the thickened aortic wall around the commissure on imaging, preventing the accurate detection of commissures. This case is marked in red on the Bland-Altman plots, indicating differences in the measurements criteria of the algorithm and the expert. The manual measurements (A,D, G, J), automated measurements from 2-dimensional images (B, E, H, K), and automated measurements from 3-dimensional images…