Sameer Ahmed1, Stefan L Zimmerman2, Pamela T Johnson1, Hong Lai1, Satomi Kawamoto1, Karen M Horton1, Elliot K Fishman1. 1. The Russell H. Morgan Department of Radiology and Radiological Sciences, Johns Hopkins University School of Medicine, 601 N. Caroline Street, JHOC 3142, Box 0818, Baltimore, MD 21287, USA. 2. The Russell H. Morgan Department of Radiology and Radiological Sciences, Johns Hopkins University School of Medicine, 601 N. Caroline Street, JHOC 3142, Box 0818, Baltimore, MD 21287, USA. Electronic address: szimme11@jhmi.edu.
Abstract
BACKGROUND: Treatment decisions for ascending aortic aneurysms are guided by measurements from CT angiograms. OBJECTIVE: The aim was to evaluate the reproducibility of these measurements by using manual techniques and advanced imaging software. METHODS: Two radiologists measured maximal ascending aorta diameter on CT angiograms in 30 subjects at 4 separate reading sessions-2 with manual techniques and 2 with semiautomated software analysis. Interobserver and intraobserver variabilities were assessed with Bland-Altman plots and Spearman correlation coefficients. RESULTS: Interobserver variability was smaller for the software-assisted method. Limits of agreement for the manual method were -4.2 mm and 9.2 mm compared with -4.0 mm and 4.6 mm for the software-assisted method; coefficients of repeatability were 6.8 mm and 4.3 mm. Intraobserver variability was inconsistent between readers. Strong correlation was found between observers who used both methods (R(2) = 0.8078-0.9881; P < .05 for all). CONCLUSION: The use of an advanced imaging software for measurement of ascending aortic aneurysm reduces interobserver variability.
BACKGROUND: Treatment decisions for ascending aortic aneurysms are guided by measurements from CT angiograms. OBJECTIVE: The aim was to evaluate the reproducibility of these measurements by using manual techniques and advanced imaging software. METHODS: Two radiologists measured maximal ascending aorta diameter on CT angiograms in 30 subjects at 4 separate reading sessions-2 with manual techniques and 2 with semiautomated software analysis. Interobserver and intraobserver variabilities were assessed with Bland-Altman plots and Spearman correlation coefficients. RESULTS: Interobserver variability was smaller for the software-assisted method. Limits of agreement for the manual method were -4.2 mm and 9.2 mm compared with -4.0 mm and 4.6 mm for the software-assisted method; coefficients of repeatability were 6.8 mm and 4.3 mm. Intraobserver variability was inconsistent between readers. Strong correlation was found between observers who used both methods (R(2) = 0.8078-0.9881; P < .05 for all). CONCLUSION: The use of an advanced imaging software for measurement of ascending aortic aneurysm reduces interobserver variability.
Authors: Prachi P Agarwal; Peter S Liu; Peter Hagan; Anna M Booher; Kuanwong Watcharotone; Leslie E Quint Journal: Clin Imaging Date: 2015-10-27 Impact factor: 1.605