ABSTRACT
We performed a meta-analysis of 31 studies comparing the results of fractional flow reserve (FFR) against quantitative coronary angiography (QCA) and/or noninvasive imaging of the same lesions. Studies were retrieved from PubMed (last search February 2006). Across 18 studies (1,522 lesions), QCA had a random effects sensitivity of 78% (95% confidence interval [CI] 67 to 86) and specificity of 51% (95% CI 40 to 61) against FFR (0.75 cutoff). Overall concordances were 61% for lesions with diameter stenosis 30% to 70%, 67% for stenoses >70%, and 95% for stenoses <30%. Compared with noninvasive imaging (21 studies, 1,249 lesions), FFR had a sensitivity of 76% (95% CI 69 to 82) and specificity of 76% (95% CI 71 to 81) by random effects. Summary receiver-operator characteristic estimates were similar. Most data addressed comparisons with perfusion scintigraphy (976 lesions, sensitivity 75%, specificity 77%), and some data were also available for dobutamine stress echocardiography (273 lesions, sensitivity 82%, specificity 74%). In conclusion, QCA does not predict the functional significance of coronary lesions. FFR shows modest concordance with noninvasive imaging tests. The prognostic implications of discordant FFR and imaging results need further study.
