Fig 1.
Theoretical expressions of FFR and CFR.
FFR can be expressed as a function of a dimensionless parameter Π, which relates aortic pressure (Pa), pressure loss coefficient (ζL) and hyperemic microvascular resistance (HMR). CFR and FFR are interrelated through the basal-to-hyperemic microvascular resistance ratio.
Fig 2.
Hemodynamic parameters of the 1st dataset (n = 199). Pa.
= aortic pressure; BMR = basal coronary microvascular resistance; HMR = hyperemic coronary microvascular resistance; QCA = diameter reduction (%) by quantitative coronary analysis; ζL = pressure loss coefficient; FFR = fractional flow reserve. The table reports median ± robust standard deviations. The dendrogram represents the average-link similarities returned by an agglomerative hierarchical clustering. Modulus of the Spearman’s rank correlation coefficient (ρ) was used as a distance metric; it reflects the proximity between two objects by measuring at what point they are similar. Reported values = median (interquartile range).
Fig 3.
Left panel: estimated vs. actual FFR variations around the 0.8-threshold value (Eq 5). Right panel: respective impacts of hyperemic microvascular resistance, stenosis severity and aortic pressure on FFR variation.
Fig 4.
Relationship between HMR, ζL and FFR.
Left panel: The dot colors represent the FFR measured by the pressure guide wire. The colored background is the theoretical FFR (Eq 4) assuming a proximal pressure of 95 mmHg (red: FFR<0.75; grey: 0.75<FFR<0.8; green: FFR>0.8). The dashed lines identify the modes of ζL and HMR distributions. Right panel: Independent effects of ζL and HMR on FFR around their respective modes. HMR and log10(ζL) were fixed at 1.9 ± 0.19 mmHg/cm/s and 1.4 ± 0.14, respectively. The blue curves are theoretical (Eq 4).
Fig 5.
CFR was predicted from Eq (6). The green (red) dots correspond to lesions with FFR greater (less) than 0.8. The inset represents the corresponding median CFR values ± robust standard deviation.
Fig 6.
Relationship between FFR, CFR and (BMR/HMR).
CFR and FFR are related through the ratio of basal to hyperemic vascular resistances (BMR / HMR); see Eq (6). The dot colors represent the measured BMR-over-HMR ratios. The colored background illustrates the theoretical ratio (red: ratio<2; grey: 2<ratio<2.5; green: ratio>2.5).
Fig 7.
Hemodynamic parameters of the 2st dataset (n = 75).
Same acronyms as in Fig 2. Subscripts “pre” and “post” refer to pre- and post-revascularization, respectively. Reported values = median (interquartile range).
Fig 8.
Accuracy of the CFR-to-FFR ratio to discriminate post-PCI CFR>2 from post-PCI CFR<2. AUC = area under the ROC curve. The colored disks represent the Cohen’s kappa statistic. The ROC curves of the FFR and CFR classifiers are also represented for comparison.
Fig 9.
Measured CFR before vs. after PCI.
The white dots in the boxplots represent the median values. Non-italicized p values (1st rows) refer to the one-sample right-tailed t-tests with the alternative hypothesis that post-PCI-CFR mean was greater than 2. Italicized p values (top) refer to the one-sample right-tailed t-tests with the alternative hypothesis that post-PCI-CFR mean was greater than pre-PCI-CFR mean.
Fig 10.
Likelihood of ischemia relief after PCI.
This flowchart speculates on ischemia relief by revascularization based on our theoretical and experimental findings. The numbers represent the occurrences among the 299 lesions.