P23 Inter-observer variability of coronary stenosis characterized by coronary angiography: a single-centre retrospective chart review by staff cardiologists

Local practice varies considerably in criteria used to assess critical stenosis in coronary angiography. In this retrospective observational cohort, we examined the inter- and intra-rater reliability of visual interpretation of coronary arteriography. We assessed the impact of this inter-observer variability of coronary lesions on clinical decision-making with independent, blinded review of angiogram clips by three experienced interventional cardiologists including the original operator. A review of 200 angiograms performed at the Toronto General Hospital (TGH) showed a mean agreement between all participating observers (mean ICC) of 77.4 (i.e., an interobserver variability of 22.6%). The interobserver variability in proximal parts of the main coronary vessels was shown to be the lowest (i.e., highest agreement), while the midportion parts was the highest (i.e., lowest agreement) and the intermediate results were achieved in the distal parts. This analysis re-demonstrates the variability in visual interpretation of coronary angiograms and highlights the importance of considering augmenting visual interpretation with more objective measures in cases of clinical uncertainty.