Variability of the aortic annulus measured in different planes with multislice computed tomography: Implications on clinical strategy

Introduction: Precise imaging of the aortic annulus is crucial for proper patient selection and safe performance of transcatheter aortic valve implantation (TAVI). However, accurate and reproducible 3-dimensional measurement is not trivial and prone to certain sources of error. Aim of our study was to determine the variation of annular dimensions acquired with multislice computed tomography (MSCT) in different planes parallel to the true annular plane and to verify whether there is any effect on clinical strategy. Methods: 146 patients with severe aortic stenosis who underwent TAVI between January 2012 to September 2012 and had pre-procedural MSCT were included in this retrospective study. Using a customized software (3 mensio Medical Imaging B.V., BS Bilthoven, The Netherlands) the effective annulus diameter was derived by planimetry of the annular area (pr 2 ). By shifting the original annular plane in a stretched axial view towards aortic and left ventricular outflow tract (LVOT) direction taking 1 mm and 2 mm steps, in total 4 parallel images (towards aorta with 1 mm (AnnAo1) and 2 mm (AnnAo2) distance, towards LVOT with 1 mm (AnnLV1) and 2 mm (AnnLV2) distance from the original annular plane) were obtained. We examined the deviation of the annulus size in all different planes and analyzed the impact on the choice of valve prosthesis or TAVI strategy. Results: Mean age was 82 ± 5 years and logistic Euroscore 25.8 ± 13.6%. Effective annulus diameter was 22.6 ± 2.7 mm. The paired T-Test revealed a significant difference of annulus size when acquired in AnnAo1 (22.2 ± 2.1 mm; p < 0.001) and AnnAo2 (21.4 ± 2.4 mm; p < 0.001), while values in direction to the LVOT were almost identical (AnnLV1 22.8 ± 2.1 mm; AnnLV2 22.7 ± 2.1 mm; ns respectively). With respect to choice of valve prosthesis size, measurements in parallel planes AnnAo1, AnnAo2, AnnLV1 and AnnLV2 would have implied change of strategy in 26, 63, 15 and 17 patients with smaller valve size in 22 (15.1%), 63 (43.2%), 8 (5.5%) and 11 (7.5%) and larger valve size in 4 (2.7%), 0, 7 (4.8%) and 6 (4.1%) patients, respectively. Only measurements in AnnAo2 would have resulted in omittance of TAVI in 9 cases given too small annulus sizes. Conclusion: Variability of the aortic annulus in different axial planes increases towards aortic direction and shows no relevant difference when measured in orientation to LVOT.