Fractional flow reserve from coronary CT angiography compared with quantitative flow ratio in complex CAD

Diagnostic concordance among Fractional Flow Reserve derived from computed tomography (FFR_CT), and the Quantitative flow ratio (QFR) and Murray's Law-based QFR (μFR) derived from invasive coronary angiography (ICA) is implicitly assumed.

Coronary CT angiography (CCTA) and ICA were analyzed in a central imaging core lab in this post-hoc imaging sub-study of the FASTTRACK CABG trial that enrolled 114 patients with de-novo three-vessel and/or left-main coronary artery disease. FFR_CT, QFR, and μFR were analyzed at corresponding bifurcation points on CCTA and ICA, and virtual pullback pressure gradient index (PPGi) and FFR derivatives (dFFR/ds) were assessed to patho-physiologically categorize the lesion phenotype into diffuse or focal.

In 199 vessels, mean distal estimates of FFR_CT (0.70), QFR (0.71), and μFR (0.69) were similar (p ​= ​0.127). QFR was significantly higher than FFR_CT (p ​< ​0.01) and μFR (p ​< ​0.01) in the main branches of the two most proximal bifurcations. Concordance between FFR_CT and QFR, and FFR_CT and μFR was 76.3 ​% (kappa ​= ​0.451) and 80.3 ​% (kappa ​= ​0.544), respectively, when using a cut-off of ≤0.80. Concordance in the pathophysiological lesion phenotype (diffuse or focal) as derived from virtual PPGi was poor between FFR_CT vs QFR (k ​= ​0.04) and FFR_CT vs μFR (k ​= ​0.16). QFR (20.9 ​%) tended to identify focal lesions more frequently than FFR_CT (13.4 ​%) and μFR (7.5 ​%).

In the two most proximal bifurcations, QFR values were higher than FFR_CT and μFR, resulting in lesion severity being underestimated, which may impact revascularization decisions. The pathophysiological phenotype classification was poorly correlated among FFR_CT, QFR, and μFR.

NCT04142021.