Background
Coronary artery disease (CAD) is the leading cause of death in the world. Physiological lesion assessment with indices such as fractional flow reserve (FFR) is now accepted as the invasive gold-standard for diagnosing the significance of CAD and for guiding treatment. Patients undergoing percutaneous coronary intervention (PCI) guided by FFR have better clinical outcomes than those undergoing standard assessment. Furthermore, FFR-guided PCI is associated with decreased stent implantation and reduced long-term cost. Only a minority of patients undergoing invasive coronary angiography are currently afforded these benefits due to a number of procedure, operator, and economic related factors. There may be additional benefits from combined pressure and flow measurement. There is therefore a need for a technology that delivers the benefits of physiological lesion assessment without the factors which limit use of the invasive technique.
Hypothesis
Computational fluid dynamics (CFD) modelling based upon invasive coronary angiographic images (ICA) can characterise and predict intracoronary physiology.
Aims
(i) To develop a CFD-based model capable of simulating and predicting clinically relevant intracoronary physiology and (ii) validate model performance using clinical data from patients with CAD.
Methods
A workflow, based upon 3-D CFD modelling, capable of predicting intracoronary pressure and ‘virtual’ FFR from ICA, was developed. The model was validated against in vivo clinical measurements in 35 unique arterial datasets. The model predicted physiological lesion significance with 97% overall accuracy. Computation was prolonged (>24hrs). Two novel methods for solving the 3-D CFD were therefore developed. These methods enabled accurate computation of results in clinically tractable timescales (<5mins), at least equivalent to invasive measurement. The critical influence of system boundary conditions was explored, characterised, and quantified. A novel approach to patient-specific tuning of the outlet boundary conditions was developed and evaluated. The workflow was adapted to compute the pressure-flow relationship from measured pressure boundary conditions within a fully patient-specific in silico model. Results were validated within a novel experimental flow circuit incorporating patient-specific 3-D printed coronary arterial phantom models.
Conclusions
It is possible to compute clinically relevant intracoronary physiology (pressure or flow) from ICA. Results can be generated in clinically tractable timescales. The CFD model can be tuned to individual patient characteristics. The developed tools may be commercially desirable. Prior to full clinical translation, the model must be evaluated in a clinical trial
