Cyclic changes in area- and perimeter-derived effective dimensions of the aortic annulus measured with multislice...

Multislice computed tomography (MSCT) is recommended for annular sizing prior to transcatheter aortic valve implantation (TAVI), but it remains unclear whether systolic or diastolic reconstructions should be used and whether the effective annular diameter should be derived by area or perimeter. In this study these different approaches were compared with intraoperative sizing.

In 52 patients who were evaluated but deemed unsuitable for TAVI, the annulus was measured during conventional surgery using metric sizers (AnnOp) and compared with MSCT measurements (cross-sectional diameter derived by area [AnnAsys, AnnAdia; AnnAmean = (AnnAsys + AnnAdia)/2] and perimeter (AnnPsys, AnnPdia) in systole and diastole). Furthermore, TAVI was simulated based on AnnOp and the impact of the various MSCT approaches on sizing strategy was determined.

The best agreement with AnnOp [mean difference (limits of agreement)] was shown for AnnAmean [0.03 mm (-1.9 to 1.96)], whereas the strongest deviation was noted for AnnPsys [-1.08 mm (-3.01 to 0.86)]. Mean differences between systole and diastole were significant but small: 0.82 mm (3.5 %) for area- and 0.81 mm (3.3 %) for perimeter-derived measurements. Simulation of TAVI revealed the least change of strategy for AnnAmean (76.9 %) as compared with AnnPsys (53.8 %); between AnnAsys and AnnAdia sizing would have been deviant in 17.3 % due to relatively large intraindividual cyclic differences.

AnnAmean demonstrated the best agreement with AnnOp, whereas perimeter-derived measurements were somewhat overestimated. Despite a negligible average difference between systolic and diastolic annular values, in a subset of patients the intraindividual cyclic variability was relatively large and potentially of clinical impact.