Mariusz Kruk1, Łukasz Wardziak2, Gary S Mintz3, Stephan Achenbach4, Jerzy Pręgowski2, Witold Rużyłło2, Zofia Dzielińska2, Marcin Demkow2, Adam Witkowski5, Cezary Kępka2. 1. Coronary Disease and Structural Heart Diseases Department, Institute of Cardiology, ul. Alpejska 42, 04-628 Warsaw, Poland; Invasive Cardiology and Angiology Department, Institute of Cardiology, ul. Alpejska 42, 04-628 Warsaw, Poland. Electronic address: mkruk@ikard.pl. 2. Coronary Disease and Structural Heart Diseases Department, Institute of Cardiology, ul. Alpejska 42, 04-628 Warsaw, Poland; Invasive Cardiology and Angiology Department, Institute of Cardiology, ul. Alpejska 42, 04-628 Warsaw, Poland. 3. Cardiovascular Research Foundation, New York, New York. 4. Department of Cardiology, University of Giessen, Klinikstr. 33, 35392 Giessen, Germany; Department of Radiology, University of Giessen, Klinikstr. 33, 35392 Giessen, Germany. 5. Invasive Cardiology and Angiology Department, Institute of Cardiology, ul. Alpejska 42, 04-628 Warsaw, Poland.
Abstract
BACKGROUND: We evaluated the accuracy of commonly used thresholds for vessel area evaluation on coronary CT angiography (CTA) and assessed ability of CTA to image the adventitial border. METHODS: We evaluated 137 paired (coronary CTA and intravascular ultrasound [IVUS]) coronary artery cross-sections in 30 patients. CTA analysis included measurements of external vessel border area defined at Hounsfield unit (HU) thresholds of 0 (presumed adventitia), 50, and 70 (presumed external elastic membrane [EEM]). IVUS analysis included measurements of lumen, EEM, and outer border of the highly echogenic area adjacent to EEM (presumed adventitia area). RESULTS: High correlation was found between CTA and IVUS measurements for EEM areas (R(2) = 0.65, P < .001 and R(2) = 0.60, P < .001 for CTA thresholds of 50 and 70 HU, respectively). CTA and IVUS measurements of adventitia areas were significantly correlated (R(2) = 0.74; P < .001), with no significant difference between the 2 methods (20.2 ± 6.4 mm(2) vs 19.8 ± 6.4 mm(2), respectively; P = .278). Cross-sectional coronary lumen radiodensity on CTA images and plaque burden measured on IVUS significantly affected the accuracy of CTA in assessment of the EEM area but not the presumed adventitial area. CONCLUSIONS: We have demonstrated that use of a 50-HU threshold for vessel area determination by CTA led to its significant overestimation, whereas 70-HU threshold was close to that of EEM on IVUS. CTA may accurately delineate the coronary adventitial border by using a 0-HU threshold.
BACKGROUND: We evaluated the accuracy of commonly used thresholds for vessel area evaluation on coronary CT angiography (CTA) and assessed ability of CTA to image the adventitial border. METHODS: We evaluated 137 paired (coronary CTA and intravascular ultrasound [IVUS]) coronary artery cross-sections in 30 patients. CTA analysis included measurements of external vessel border area defined at Hounsfield unit (HU) thresholds of 0 (presumed adventitia), 50, and 70 (presumed external elastic membrane [EEM]). IVUS analysis included measurements of lumen, EEM, and outer border of the highly echogenic area adjacent to EEM (presumed adventitia area). RESULTS: High correlation was found between CTA and IVUS measurements for EEM areas (R(2) = 0.65, P < .001 and R(2) = 0.60, P < .001 for CTA thresholds of 50 and 70 HU, respectively). CTA and IVUS measurements of adventitia areas were significantly correlated (R(2) = 0.74; P < .001), with no significant difference between the 2 methods (20.2 ± 6.4 mm(2) vs 19.8 ± 6.4 mm(2), respectively; P = .278). Cross-sectional coronary lumen radiodensity on CTA images and plaque burden measured on IVUS significantly affected the accuracy of CTA in assessment of the EEM area but not the presumed adventitial area. CONCLUSIONS: We have demonstrated that use of a 50-HU threshold for vessel area determination by CTA led to its significant overestimation, whereas 70-HU threshold was close to that of EEM on IVUS. CTA may accurately delineate the coronary adventitial border by using a 0-HU threshold.