Discrepancy in the assessment of jailed side branch lesions by visual estimation and quantitative coronary angiographic analysis: comparison with fractional flow reserve

OBJECTIVE: We sought to evaluate the variability in the assessment of jailed side branch (SB) lesions by visual estimation and quantitative coronary angiography (QCA) and to compare those results with fractional flow reserve (FFR). METHODS: Twenty jailed SB lesions with available FFR (median 0.76; range, 0.39-0.94) were selected from the PRESSURE trial. Lesions were assessed by three independent QCA core laboratories with different QCA systems and by three different cardiologist groups (five European bifurcation club members, five Korean experts, and five trainees). Agreements of the continuous measurements were expressed as the intraclass correlation coefficient (ICC) and average coefficient of variance (CV), and those of the categorical values as kappa. RESULTS: Mean minimum lumen diameter (MLD) and % diameter stenosis differed among the three QCA systems up to 0.30 mm and 9.65%, respectively (P < 0.001). Three QCA systems showed fair agreement for the measurements of reference diameter, % diameter stenosis, MLD, and lesion length (ICC 0.346-0.686, CV 8.7-29.5%), and a poor agreement on stenosis of 75% or more (Fleiss κ 0.14 and mean κ 0.18). Agreements of visual estimation among the three groups were poor to fair (Fleiss κ 0.167-0.367). Sensitivity and specificity for predicting ischemia-inducible lesion (FFR < 0.75) were 64.7% and 48.0% for visual estimation and 56.6% and 56.6% by QCA, respectively. Visual estimation overestimated the % diameter stenosis and functional significance of the lesions compared with QCA (P < 0.001) and FFR (P = 0.036). CONCLUSIONS: Angiographic assessment of jailed SB lesions by both QCA and visual estimation showed variability. Visual estimation tended to overestimate the severity of jailed SB lesions compared to FFR and QCA.ope