P3322Optimising physiological assessment of serial coronary artery lesions using an in vitro model of tandem stenoses

Abstract

Background: Fractional Flow Reserve (FFR) is considered the optimal technique to assess the physiological significance of a diseased coronary artery during angiography. However, the ability to use FFR to assess individual lesions is unclear due to the potential haemodynamic interplay between serial stenoses. The most common technique is to use trans­lesional change in FFR (ΔFFR) on manual pullback of the pressure wire to estimate the significance of each lesion Purpose: To evaluate the determinants of haemodynamic interplay between tandem stenoses and incorporate these factors in a model that predicts the true haemodynamic significance of a stenosis, once an accompanying stenosis is removed. Methods: Coronary arteries with tandem stenoses were modelled by creating 3D­printed phantom tubes in various configurations of lesion severity, length and separation. An in vitro model of continuous coronary circulation was created and validated under varying haemodynamic conditions. 34 configurations of tandem stenoses were assessed by FFR (including fixed­speed pullback) and compared with tubes containing isolated stenoses. The functional significance of an isolated stenosis (FFRtrue) was compared to its apparent significance (FFRapp) in the presence of another lesion, based on ΔFFR from fixed­rate pullback in the serially stenosed “vessel” (where FFRapp=1­ΔFFR). A linear regression model was developed to predict the significance of an isolated stenosis (FFRpred) using the first 26 vessels and validated in the next 10 vessels. Results: FFR of the serially diseased vessels was 0.75±0.24. FFRapp underestimated FFRtrue in 25/34 cases, (discrepancy 0.06±0.014, see figure A) regardless of whether the stenosis of interest was proximal or distal to the confounding stenosis (p=0.22). The discrepancy in true and apparent FFR was proportional to the overall severity of disease with no significant difference regardless of whether it was the the proximal or distal lesion lesion removed (p=0.22, see figure A). Two regression models that incorporated ΔFFR and diameter stenosis of the lesions were generated for scenarios when the lesion of interest was proximal or distal. FFRpred was more accurate than ΔFFR at estimating FFRtrue (discrepancy 0.005±0.03 versus 0.04±0.03 respectively, p=0.02): Figure B illustrates this for a assessment of a distal lesion when a tandem proximal stenosis is removed. Conclusion: Estimating the functional significance of a stenosis in a tandemly diseased vessel using ΔFFR alone, leads to significant errors. This can be more accurately predicted by using a mathematical model that incorporates data from a fixed­rate pullback pressure curve as well as lesion geometry. The utility of this model needs to be evaluated in a clinical setting.