PurposeTo assess inter-observer variability of the Coronary Artery Disease – Reporting and Data System (CAD-RADS) for classifying the degree of coronary artery stenosis in patients with stable chest pain.Material and methodsA prospective study was conducted upon 96 patients with coronary artery disease, who underwent coronary computed tomography angiography (CTA). The images were classified using the CAD-RAD system according to the degree of stenosis, the presence of a modifier: graft (G), stent (S), vulnerable plaque (V), or non-diagnostic (n) and the associated coronary anomalies, and non-coronary cardiac and extra-cardiac findings. Image analysis was performed by two reviewers. Inter-observer agreement was assessed.ResultsThere was excellent inter-observer agreement for CAD-RADS (k = 0.862), at 88.5%. There was excellent agreement for CAD-RADS 0 (k = 1.0), CAD-RADS 1 (k = 0.92), CAD-RADS 3 (k = 0.808), CAD-RADS 4 (k = 0.826), and CAD-RADS 5 (k = 0.833) and good agreement for CAD-RADS 2 (k = 0.76). There was excellent agreement for modifier G (k = 1.0) and modifier S (k = 1.0), good agreement for modifier N (k = 0.79), and moderate agreement for modifier V (k = 0.59). There was excellent agreement for associated coronary artery anomalies (k = 0.845), non-coronary cardiac findings (k = 0.857), and extra-cardiac findings (k = 0.81).ConclusionsThere is inter-observer agreement of CAD-RADS in categorising the degree of coronary arteries stenosis, and the modifier of the system and associated cardiac and extra-cardiac findings.