• Muhammad Aamir Latif, Frank W Sanchez, Karl Sayegh, Emir Veledar, Muhammad Aziz, Rehan Malik, Imran Haider, Arthur S Agatston, Juan C Batlle, Warren Janowitz, Constantino Peña, Jack A Ziffer, Khurram Nasir, and Ricardo C Cury.
• From the *Miami Cardiac and Vascular Institute, Baptist Health South Florida, Miami; †Center for Healthcare Advancement and Outcomes, Baptist Health Medical Group, Miami Beach; and ‡West Kendall Baptist Hospital, Baptist Health South Florida, Miami, FL; §Department of Radiology, Emory University Hospitals and Clinics, Atlanta, GA; ∥Department of Cardiology, University of Texas Medical Branch, Galveston, TX; ¶The Johns Hopkins Ciccarone Center for the Prevention of Heart Disease, Baltimore, MD.
• J Comput Assist Tomogr. 2016 Sep 1; 40 (5): 763-72.

BackgroundCardiac computed tomography (CT) image quality (IQ) is very important for accurate diagnosis. We propose to evaluate IQ expressed as Likert scale, signal-to-noise ratio (SNR), and contrast-to-noise ratio (CNR) from coronary CT angiography images acquired with a new volumetric single-beat CT scanner on consecutive patients and assess the IQ dependence on heart rate (HR) and body mass index (BMI).MethodsWe retrospectively analyzed the data of the first 439 consecutive patients (mean age, 55.13 [SD, 12.1] years; 51.47% male), who underwent noninvasive coronary CT angiography in a new single-beat volumetric CT scanner (Revolution CT) to evaluate chest pain at West Kendall Baptist Hospital. Based on patient BMI (mean, 29.43 [SD, 5.81] kg/m), the kVp (kilovolt potential) value and tube current were adjusted within a range of 80 to 140 kVp and 122 to 720 mA, respectively. Each scan was performed in a single-beat acquisition within 1 cardiac cycle, regardless of the HR. Motion correction software (SnapShot Freeze) was used for correcting motion artifacts in patients with higher HRs. Autogating was used to automatically acquire systolic and diastolic phases for higher HRs with electrocardiographic milliampere dose modulation. Image quality was assessed qualitatively by Likert scale and quantitatively by SNR and CNR for the 4 major vessels right coronary, left main, left anterior descending, and left circumflex arteries on axial and multiplanar reformatted images. Values for Likert scale were as follows: 1, nondiagnostic; 2, poor; 3, good; 4, very good; and 5, excellent. Signal-to-noise ratio and CNR were calculated from the average 2 CT attenuation values within regions of interest placed in the proximal left main and proximal right coronary artery. For contrast comparison, a region of interest was selected from left ventricular wall at midcavity level using a dedicated workstation. We divided patients in 2 groups related to the HR: less than or equal to 70 beats/min (bpm) and greater than 70 bpm and also analyzed them in 2 BMI groupings: BMI less than or equal to 30 kg/m and BMI greater than 30 kg/m.ResultsMean SNR was 8.7 (SD, 3.1) (n = 349) for group with HR 70 bpm or less and 7.7 (SD, 2.4) (n = 78) for group with HR greater than 70 bpm (P = 0.008). Mean CNR was 6.9 (SD, 2.7) (n = 349) for group with HR 70 bpm or less and 5.9 (SD, 2.2) (n = 78) for group with HR 70 bpm or greater (P = 0.002). Mean SNR was 8.8 (SD, 3.2) (n = 249) for group with BMI 30 kg/m or less and 8.1 (SD, 2.6) (n = 176) for group with BMI greater than 30 kg/m (P = 0.008). Mean CNR was 7.0 (SD, 2.8) (n = 249) for group with BMI 30 kg/m or less and 6.4 (SD, 2.4) (n = 176) for group with BMI greater than 30 kg/m (P = 0.002). The results for mean Likert scale values were statistically different, reflecting difference in IQ between people with HR 70 bpm or less and greater than 70 bpm, BMI 30 kg/m or less, and BMI greater than 30 kg/m.

25 keywords...

hide…