Background Physiology-guided coronary revascularisation is associated with better outcomes however it is unclear if pressure-derived indices, such as FFR, can reliably assess individual lesions in serial disease.

Methods 3D printed serial disease phantoms were assessed in vitro. FFR of a lesion was predicted from the size of step up on pressure-wire pullback in the presence of serial lesions (FFR _app ) and compared to phantoms with no accompanying lesion (FFR _true ). Mathematical models to minimise error in predicting FFR _true were developed in 32 phantoms and validated in 15 patients and another 20 phantoms. In 43 serial disease patients we compared how different indices (FFR, iFR, Pd/Pa) predict true values, after isolation by PCI or disease-free sidebranch. In a subset of 21 we measured flow to calculate Hyperaemic Stenosis Resistance (hSR).

Results FFR _app underestimated FFR _true in 85% of phantoms, with discrepancy proportional to total FFR. 4.5% of lesions were misclassified (FFR <0.8 threshold); with mathematical models this fell to 1.5% (figure 1). Clinically, FFR, iFR and Pd/Pa significantly underestimated true values; hSR is least affected (p<0.01).

• Download figure
• Open in new tab
• Download powerpoint

Abstract 3 Figure 1

Scatterplots of difference between FFRapp of a stenosis in presence of accompanying lesion and FFRtrue, when stenosis present in isolation. Within the validation cohort, discrepancy was 5% (0.036±0.048), which improved to 0.6% (0.005±0.037) and 0.8% (0.006±0.023) using empirical and theoretical solutions respectively (P<0.01). In the clinical cohort, discrepancy was 6.6% (0.041±0.03), which improved to 2.2% (0.019±0.026) and 2.0% (0.016±0.023) (P<0.01)

Conclusion Pressure-derived indices underestimate stenosis significance in serial disease. Pressure pullback data can be mathematically analysed to minimise this. Pressure-flow based resistance indices are less prone to this error.