The current version of our clinical software tool MEDVIS 3D supports:
- DICOM series import and navigation (MRT, CT, DSA)
- False-Color encoding of DICOM images
- 3D volume reconstruction from DICOM series
- DICOM header extraction
- Direct volume rendering with shading (ARB, CG, GLSL)
- Interactive 3D segmentation (Rubber, Freehand, Region Grow, Sphere, Interactive Region Grow)
- Volume compression
- Calculation of 3D volume and diameters
- 3D measurement tool
- 2D and 3D contrast windowing
- Automated 3D triangle surface mesh generation (Advanced Skeleton Climbing, Marching Cube)
- Automated 3D tetrahedral volume mesh generation (based on NETGEN and TETGEN)
- Volume segmentation caching
- Interactive transfer function adaptation
- Vessel skeleton extraction
- Automated aneurysm detection
- Interactive regular vessel removal
- Various CFD and CSD simulations
- Steady blood flow
- Steady vessel deformation
- Transient blood flow
- Transient blood flow and vessel deformation (FSI)
- Interactive 3D/4D visualization of simulation results
Automated vessel segment removal
The MEDVIS 3D software is now able to recognize and automatically crop vessels, based on a starting- and end-point. This feature enables the physician to properly segment aneurysms from surrounding vessel structures.
Based on a skeleton extraction of the vessel structures, the software calculates the shortest path between the start- and the end-point and interpolates a virtual vessel into the volume. A number of smoothing and correction operations guarantee the correct structure and dimension of the virtual vessel.
This feature increases the accuracy of the aneurysm segmentation while it decreases the time and effort it takes for a proper segmentation.
Blood Flow Simulation
Together with our research partners we are developing a simulation system that is able to simulate the blood flow through intracranial aneurysms on the basis of the reconstructed 3D data from our MEDVIS 3D tool. Our simulation code calculates numerical solutions for the coupled transient system consisting of blood lumen described by the flow equation (Navier-Stokes) and the vessel wall described by a linear elasticity equation (Navier-Cauchy) on the tetrahedral mesh obtained for the patient-specific geometry. On inflow and outflow areas, boundary conditions modeling the variation of blood pressure and velocity during a pulse cycle have to be defined. We make use of the finite element method (FEM) in order to replace the governing partial differential equations with large systems of linear equations for the nodal values of the physical quantities on the mesh. These systems can be solved by iterative methods, in our case the conjugate gradient (CG) algorithm combined with an algebraic multigrid (AMG) preconditioner. The coupling of fluid and structure equations is also done iteratively: The traction exerted by the blood flow on the surrounding vessel wall is applied as a Neumann boundary condition to the structure equation, which in turn yields a new displacement field and therefore a new shape of the blood domain. At each time step, this procedure is repeated until mechanical equilibrium is reached. As a consequence, the FEM mesh, on which the fluid equations are solved, changes in time, and the blood velocity at the fluid-structure interface has to match the velocity of the vessel wall. This is achieved by using an ALE (arbitrary Lagrangian Eulerian) formulation of the Navier-Stokes equations.
Once all the fluid and structure fields (pressure, velocity, displacement) are known, derived quantities such as wall shear stress (WSS), von Mises stress, oscillatory shear index (OSI) and wall shear stress gradient (WSSG) can also be obtained.
All calculations can be carried out on a standard desktop PC, however, for increased efficiency they can also be run on a high performance computing cluster in the Austrian Grid. The solver code is parallelized and scales almost linearly up to 16 CPUs.