Author + information
CT-based measurement of the aortic annulus is often used for prosthesis sizing in TAVR. Since the annulus is elliptical and becomes circular after implantation, various measurement methods exist: Annulus dimensions can be derived by measuring the minimum and maximum diameters before implantation and calculating their average (Dmean), by measuring the area and calculating the diameter under the assumption that the area will become a circle (Darea), or by measuring the circumference and calculating the diameter assuming that this circumference will become that of a circle (Dcirc). It is unknown which parameter is best suited to size the prosthetic valve and what their variability is when measured by CT.
In 65 patients scheduled for TAVR, contrast Dual Source CT imaging was performed in diastole. Dmean, Darea and Dcirc were determined by two independent observers in double-oblique multiplanar reconstructions aligned with the aortic annulus. Differences between the measurements as well as interobserver variability were determined.
Mean age was 83±4years, 29 pts. were male, mean logistic Euroscore was 33±14%. Mean aortic annulus size obtained by averaging the minimum and maximum diameters (Dmean) was 24.4±2.5mm, the mean area-derived diameter (Darea) was 25.0±2.6mm and the mean circumference-derived diameter (Dcirc) was 25.5±2.6mm. Correlations between Dcirc, Darea, and Dcirc were very close (r=0.98 for all), but Darea was significantly larger than Dmean (mean difference 0.6±0.6mm) and Dcirc was significantly larger than Darea (mean difference 0.6±0.5mm) and Dmean (mean difference 1.1±0.6mm, p < 0.001 for all comparisons). Interobserver variability was small (mean absolute difference 0.8±0.6mm for Dmean; 0.7±0.6mm for Darea and 0.7±0.6mm for Dcirc) and not significantly different between the three measurements.
In CT, circumference- and area-derived determination of the aortic annulus size yields significantly larger dimensions than averaging the minimum and maximum diameter, so that different thresholds must be developed and used for TAVR prosthesis sizing. However, correlations are very close and there is no difference in interobserver variability.
Poster Sessions, Expo North
Sunday, March 10, 2013, 3:45 p.m.-4:30 p.m.
Session Title: Multimodality Imaging in Valvular Heart Disease
Abstract Category: 20. Imaging: CT/Multimodality
Presentation Number: 1270-349
- 2013 American College of Cardiology Foundation