Yiannis M. Chatzizisis MD, Abstract Co-Author: Nothing to Disclose

Elizabeth George MD, Abstract Co-Author: Nothing to Disclose

Tianrun Cai MD, Abstract Co-Author: Nothing to Disclose

Richard Thomas Mather PhD, Abstract Co-Author: Employee, Toshiba Corporation

Frank John Rybicki MD, PhD, Abstract Co-Author: Research Grant, Toshiba Corporation

Dimitris Mitsouras PhD, Presenter: Nothing to Disclose

Urvi Pravin Fulwadhva MD, Abstract Co-Author: Nothing to Disclose

Kanako Kunishima Kumamaru MD, PhD, Abstract Co-Author: Nothing to Disclose

Kurt Schultz RT, Abstract Co-Author: Employee, Toshiba Corporation

Yasuko Fujisawa MS, Abstract Co-Author: Employee, Toshiba Corporation

Michael Lally Steigner MD, Abstract Co-Author: Speaker, Toshiba Corporation

Ron Blankstein MD, Abstract Co-Author: Nothing to Disclose

Carlos Rassi, Abstract Co-Author: Nothing to Disclose

TAG from 320-detector row CTA offers incremental value for the prediction of functionally significant coronary artery disease (CAD). However, TAG measurements are labor-intensive. We thoroughly assessed TAG accuracy and reproducibility using new automated software compared to a manual method previously validated against invasive fractional flow reserve (FFR).

540 TAG measurements were performed in 20 patients that underwent clinically indicated CCTA on a 320-row detector CT (Toshiba AquilionONE). Two readers measured TAG manually by placing 1mm2 luminal ROIs in cross-sections orthogonal to the vessel centerline every 5 mm; two independent readers used an automated software (Cardiac Analysis Package, Toshiba). Accuracy of automated vs manual TAG and inter- and intra-observer reproducibility of both methods were assessed via linear regression and Bland-Altman analyses. Subanalyses were performed with respect to factors that can affect accuracy or reproducibility, namely a) presence/severity of CAD by CCTA or invasive angiography and SPECT (n=14 patients), b) plaque composition (calcified vs noncalcified), c) CT reconstruction method (iterative vs FBP), and d) physiologic parameters (BMI, left vs right coronaries).  

Correlation of automated and manual measurement was excellent (y=1.0x+2.0, r=0.92, p<0.001), with low mean difference (1 HU/cm) and narrower inter- and intra-observer limits of agreement for automated vs manual (25% and 36% average reduction, respectively, Figure). The median TAG in vessels with angiography/SPECT-confirmed obstructive CAD was -23.8 (IQR: -26.6, -19.3), similar to that previously reported for vessels with FFR ≤0.8 (median -21, IQR: -27, -16). Presence and severity of CAD, plaque composition, reconstruction method, and physiologic factors did not significantly affect accuracy or reproducibility. Analysis time was reduced by 68% per patient (13.9±3.3 min automated vs 39.8±8.7 min manual TAG). 

Automated assessment of TAG is fast, accurate, and enhances reproducibility. 

Automated TAG measurement alleviates labor-intensive calculations and can thus enable studies over larger populations to establish its clinical significance for detecting hemodynamically significant CAD.

Chatzizisis, Y, George, E, Cai, T, Mather, R, Rybicki, F, Mitsouras, D, Fulwadhva, U, Kumamaru, K, Schultz, K, Fujisawa, Y, Steigner, M, Blankstein, R, Rassi, C, Accuracy and Reproducibility of Automated, Standardized Transluminal Attenuation Gradient (TAG) Measurements in Coronary Computed Tomography Angiography (CCTA).  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14006973.html