Contents

1 In Vitro Validation
 1.1 Compliant Aneurysm
 1.2 Compliant Thoracic Model
2 In Vivo
 2.1 Distal complication
 2.2 Post traumatic aortic dissection
 2.3 Cardiac

1 In Vitro Validation

A discontinuous pump adapted from a Ventricular Assist Device (Thoratec) is used to feed the phantom with pulsatile fluid flow near to physiological arterial pressure level (100mmHg). Inlet / outlet ’hemodynamic’ conditions (time-dependent functions) reconstructed from the 3D velocity mapping were synchronized with the moving mesh and subsequently used as boundary conditions as appropriate.

1.1 Compliant Aneurysm

The blood is replaced by water, a non physiologic newtonian fluid with viscosity μ = 1.10-3Pa.s and density ρ = 1000kg∕m3. Note that the problem is thus simplified since the difficulties related to the complex blood rheology do not need to be accounted for in what follows. We want to demonstrate that we can numericaly reproduce CFD what is observed on MRI, for the moving geometry and the velocity profiles.


PIC PIC

Figure 1: Compliant Aneurysm, MRI experimental device


The computational domain is depicted in Figure 1 where the unstructured mesh (738043 tetrahedra) is also shown. The flow rate obtained from 3D velocity aquisition is imposed at the inlet (left boundary) while an essentially non-reflecting boundary condition [12] is prescribed at the outlet (right boundary). At the largest cross section area, the aneurysm diameter varies in the range 47-56 mm over the cardiac cycle. Thus its variation due to the wall compliance is 19 %.


PIC


Figure 2: Velocity elements in order to correlate MRI measurements and CFD results


1.2 Compliant Thoracic Model

Here, the configuration corresponds to the phantom of a human aortic cross whose compliance is typical of actual values. The blood is replaced by a non physiologic newtonian fluid with high shear viscosity and density relevant to blood, viz. μ 4 × 10-3 Pa.s and ρ 1056 kg/m3 respectively. Hemodynamic conditions (time-dependent functions) from the MRI-PC sequences were synchronized with the wall motion and subsequently used as boundary conditions for the ascending aorta (inlet) and the supra-aortic vessels (exit). An essentially non-reflecting boundary condition [12] is prescribed at the outlet of the descending aorta. The computational domain is depicted in figure 3 where the position of the cross section displayed in figure 4 is also shown. Note that the extra pipe section at the inlet of the computational domain is part of the in vitro model. This one was designed is such a way to promote the swirl motion of the fluid flow at the inlet of the physiological part of the phantom in order to reproduce the hemodynamic in vivo conditions more closely.


PIC

Figure 3: Computational domain for the phantom of the aorta cross. The bold line shows the position of the cross section displayed in figure 4



PIC

Figure 4: Velocity vectors obtained from CFD (white background) and MRI (gray background) at systole (left) and diastole (right) phase. Plane intersects the upper arch.


The simulations began from an initially quiescent flow state and continued for a number of full cardiac cycles in order to allow the development of a fully periodic flow, representative of a regular heartbeat. It was found that the main features of the vascular flow field became periodic within four cycles.


PIC


Figure 5: Sagittal velocity field in order to correlate MRI measurements (right) and CFD results (left)


2 In Vivo

A discontinuous pump adapted from a Ventricular Assist Device (Thoratec) is used to feed the phantom with pulsatile fluid flow near to physiological arterial pressure level (100mmHg). Inlet / outlet ’hemodynamic’ conditions (time-dependent functions) reconstructed from the 3D velocity mapping were synchronized with the moving mesh and subsequently used as boundary conditions as appropriate.

2.1 Distal complication


PIC PIC PIC PIC

Figure 6: Stent : Distal complication, correction


2.2 Post traumatic aortic dissection


PIC


Figure 7: Post traumatic aortic dissection, morphologic and functional imaging


2.3 Cardiac


PIC


Figure 8: OCFIA applied for ventricular flow problem


References

[1]   Nicoud F. Defining wave amplitude in characteristic boundary conditions. Journal of Computational Physics, 149(2), 418–422, 1999.

[2]   Nicoud F. Integral boundary conditions for unsteady biomedical cfd applications. International Journal for numerical methods in fluids, 40, 457–465, 2002.