Periodic Reporting for period 3 - ImageToSim (Multiscale Imaging-through-analysis Methods for Autonomous Patient-specific Simulation Workflows)
Reporting period: 2021-10-01 to 2023-03-31
Related to osteoporotic bone fracture:
1. Thermodynamically consistent multiscale formulations of elastoplastic multiphase hierarchical materials, with application to bone. The resulting elastoplastic continuum micromechanics model constitutes a cornerstone of the methodology for automated patient-specific and simulation-based predicition of osteoporosis.
2. Development of the methodological foundations of a fully automated finite element analysis (FEA) framework for computing a fracture risk index from CT scans, including successful small-sample-size validation studies for femur and vertebra bones.
3. Creation of a first software protoype that can independently transform clinical CT-scans into mechanical solution fields without the interaction of an expert computational analyst.
Related to perfusion simulaton in the liver:
4. A DEIM driven reduced basis method for the diffuse Stokes/Darcy model coupled at parametric phase-field interfaces, with application to modeling and efficient simulation of liver lobules.
5. Macro- and mesoscale perfusion modeling in the liver based on synthetic hepatic vascular tree generation methods that incorporate rigorous mathematical optimization techniques
6. Modeling of liver growth using cut finite element methods as a viable alternative for problems with mesh-related obstacles, in particular when large growth simulations on complex patient-specific geometries are of primary interest.
(a) Further sophistication of synthetic hepatic vascular tree generation methods, in particular to synthetically generate several trees (supplying trees - hepatic portal veins, heaptic arteries; draining tree - hepatic veins) at the same time that respect local and global optimality criteria, including validation against imaging data of real-life liver trees
(b) In-depth study on specific properties of synthetic vascular trees (e.g. behavior branching patterns, vessel diameter, distribution of Strahler order, etc.) that can be obtained under different
optimality principles (that is, tree models)
(c) Adjustment of established image processing algorithms for automated image-based spatial segmentation of the liver, including partitioning into meso-scale lobules and macorscale vessel segments
(d) Coupling of hierarchical trees to continuum poro-hyperelastic models of the meso-scale lobule structure
(e) Efficient ultra-fast computing methods based on model order reduction of the lobule structure
(f) Extension of poro-hyperelastic models to include finite growth to include regrowth phenomena in the liver in the perfusion model
At this stage of the project, we have already started to work on all of the items above, some of them being close to being published (items (a), (b), (d)) and some of them still in the implementation stage (items (c), (e) and (f)).