Patient-specific coronary stenoses can be modeled using a combination of OCT and flow velocities to accurately predict hyperemic pressure gradients.

Computational fluid dynamics (CFD) is increasingly being developed for the diagnostics of arterial diseases. Imaging methods such as computed tomography (CT) and angiography are commonly used. However, these have limited spatial resolution and are subject to movement artifact. This study developed a new approach to generate CFD models by combining high-fidelity, patient-specific coronary anatomy models derived from optical coherence tomography (OCT) imaging with patient-specific pressure and velocity phasic data. Additionally, we used a new technique which does not require the catheter to be used to determine the centerline of the vessel. The CFD data were then compared with invasively measured pressure and velocity. Angiography imaging data of 21 vessels collected from 19 patients were fused with OCT visualizations of the same vessels using an algorithm that produces reconstructions inheriting the in-plane (10 µm) and longitudinal (0.2 mm) resolution of OCT. Proximal pressure and distal velocity waveforms ensemble averaged from invasively measured data were used as inlet and outlet boundary conditions, respectively, in CFD simulations. The resulting distal pressure waveform was compared against the measured waveform to test the model. The results followed the shape of the measured waveforms closely (cross-correlation coefficient = 0.898 ± 0.005, ), indicating realistic modeling of flow resistance, the mean of differences between measured and simulated results was -3. 5 mmHg, standard deviation of differences (SDD) = 8.2 mmHg over the cycle and -9.8 mmHg, SDD = 16.4 mmHg at peak flow. Models incorporating phasic velocity in patient-specific models of coronary anatomy derived from high-resolution OCT images show a good correlation with the measured pressure waveforms in all cases, indicating that the model results may be an accurate representation of the measured flow conditions.

A numerical study of aortic flow stability and comparison with in vivo flow measurements.

The development of an engineering transitional turbulence model and its subsequent evaluation and validation for some diseased cardiovascular flows have been suggestive of its likely utility in normal aortas. The existence of experimental data from human aortas, acquired in the early 1970s with catheter-mounted hot film velocimeters, provided the opportunity to compare the performance of the model on such flows. A generic human aorta, derived from magnetic resonance anatomical and velocity images of a young volunteer, was used as the basis for varying both Reynolds number (Re) and Womersley parameter (α) to match four experimental data points from human ascending aortas, comprising two with disturbed flow and two with apparently undisturbed flow. Trials were made with three different levels of inflow turbulence intensity (Tu) to find if a single level could represent the four different cases with 4000 < Re < 10,000 and 17 < α < 26. A necessary boundary condition includes the inflow "turbulence" level, and convincing results were obtained for all four cases with inflow Tu = 1.0%, providing additional confidence in the application of the transitional model in flows in larger arteries. The Reynolds-averaged Navier-Stokes (RANS)-based shear stress transport (SST) transitional model is capable of capturing the correct flow state in the human aorta when low inflow turbulence intensity (1.0%) is specified.