Coronal reformations of the chest on 64-row multi-detector row CT: Evaluation of image quality in comparison with 16-, 8-and 4-row multi-detector row CT

Purpose: To evaluate image quality of coronal reformations of chest performed on 64-row MDCT in comparison with 16-, 8- and 4-row MDCT. Materials and methods: Consecutive patients who underwent pulmonary CT angiography using four different MDCT scanners were retrospectively studied with IRB approval: (1) n = 30,64-row MDCT; (2) n = 30,16-row MDCT; (3) n = 30, 8-row MDCT; (4) n = 30,4-row MDCT. Coronal reformatted images (2 mm thickness and 2 mm intervals for 64-row MDCT; 5 mm thickness and 5 mm intervals for 16-, 8- and 4-row MDCT) were evaluated by consensus reading of two board-certified radiologists who were blinded to scanner type. The image quality of overall chest appearance and individual thoracic structures including heart, aorta and pulmonary arteries was graded using five-point scale. Grades from four different scanners were compared using Kruskal-Wallis test. A second evaluation was performed in 48 randomly selected patients (12 patients for each scanner). Reproducibility was assessed using weighted-kappa analysis. Result: Significant reproducibility was observed between the first and second evaluations in 48 patients both for image quality of overall chest (weighted kappa = 0.826) and each thoracic structure (mean weighted kappa = 0.803; range, 0.729-0.858). Image quality of overall chest and individual thoracic structures differed significantly among four different MDCT groups, with 64-row MDCT having the highest grades, followed by 16-, 8- and 4-row MDCT (mean grades for overall chest in each scanner: 3.9, 3.0, 2.4 and 1.9, respectively) (P < 0.0001 for overall chest and each thoracic structure). Conclusion: When comparing coronal reformations of chest using four different MDCT scanners, the 64-row MDCT had the highest image quality for overall chest appearance and individual thoracic structures, followed by 16-, 8- and finally 4-row MDCT. (c) 2006 Elsevier Ireland Ltd. All rights reserved.