Scott K. Nagle MD, PhD, Presenter: Stockholder, General Electric Company Research Consultant, Vertex Pharmaceuticals Incorporated

Sarah Sweetman, Abstract Co-Author: Nothing to Disclose

Carrie Bartels, Abstract Co-Author: Nothing to Disclose

Giorgio Gimelli MD, Abstract Co-Author: Nothing to Disclose

Amish N. Raval MD, FRCPC, Abstract Co-Author: Nothing to Disclose

Christopher Jean-Pierre Francois MD, Abstract Co-Author: Research support, General Electric Company

Alejandro Munoz Del Rio PhD, Abstract Co-Author: Research Consultant, Cellectar Biosciences, Inc Reviewer, Wolters Kluwer nv

To determine the precision of CTA aortic annulus area measurements and the impact on TAVR device selection.

This retrospective study included 86 consecutive clinical TAVR screening CTAs performed on a 64-slice scanner (LightSpeed VCT, GE Healthcare) using retrospective ECG gating. A 1st year medical student (R1, after training on 10 separate CTAs), a 3D lab technologist (R2, 3 yrs experience), and a cardiothoracic radiologist (R3, 6 yrs experience) independently measured the aortic annulus in systole in a random, blinded fashion. The annular plane, containing the hinge points of all 3 valve cusps, was located using multiplanar reformats (Vitrea, Vital Images). The annular area was measured using a freely drawn contour. All measurements were repeated >2 weeks later to avoid recall bias. Bland-Altman analysis was used to assess each reader’s repeatability. The difference between the 95% limits of agreement and the bias was used to estimate the measurement precision. To assess differences between readers, variance ratios (VR) were calculated along with their 95% confidence intervals and compared with an F test. The impact on device sizing was evaluated using the Edwards SAPIEN valve as an example. Annular size was grouped into 5 categories, based on the recommended device: too small, 23mm, either, 26 mm, or too large. Percent agreement between the measurements was calculated for each reader.

Bias between measurements was 6 [-1,13] (R1), -3 [-11,5] (R2), and 1 [-5,7] (R3) mm2. Precision was ±64 [52,76] (R1), ±70 [57,83] (R2), and ±55 [44,66] (R3) mm2. The difference in precision between R2 and R3 was statistically significant (VR: 1.60 [1.04,2.46], p=0.03). Device size recommendations from the 2 measurements differed in 23% (R1), 29% (R2), and 22% (R3) of the cases and differed by more than 1 category in 2% (R1), 4% (R2), and 1% (R3) of the cases.

Within reader annular area measurement imprecision results in different TAVR device size recommendations ~25% of the time, even for an experienced cardiovascular CTA reader. Reports should include estimated measurement precision to aid in the interpretation of the results.

Knowing the precision of CTA-based aortic annulus area measurements is very important for multidisciplinary TAVR treatment planning. A single point estimate of the annular area may not be sufficient.

Nagle, S, Sweetman, S, Bartels, C, Gimelli, G, Raval, A, Francois, C, Munoz Del Rio, A, Precision of CTA-based Aortic Annulus Area Measurements for Transcatheter Aortic Valve Replacement (TAVR): Effects of Reader Experience and Implications for Appropriate Device Sizing.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14007262.html