By medical simulation I mean the simulation of bio-mechanical and physiological processes with the goal of supporting medical treatment. Advantages can be improved diagnosis, training using a simulator, simulation and "virtual try-out" of surgical operations, or, in medical research rather than clinical practise, an improved understanding of physiological processes. In some of these applications, individualised medical simulation is important, that is, taking the specific physiology of an individual into account.
Some concrete examples from my work experience are:
- Prediction of facial surgery outcome, for instance to treat cleft lip and palate. This was a topic of the EU-project GEMSS (Grid-Enabled Medical Simulation Services).
- Simulation of blood flow through vessels and aneurysms, to better understand and estimate the risk of life-threatening rupture of such aneurysms, see for instance the EU-project @neurIST, and some of my case studies.
- Coupled simulation of processes leading to blood clotting and thrombosis. This is a multi-scale process from molecular level up to the blood flow through vessels. These questions were addressed in the EU-project Coast.
Still, there are many research questions open. One problem with individualised medical simulation are difficulties to measure the individual physiological properties, the multitude of factors and mutually dependent processes on different time and space scales, whose interaction often is poorly understood.
A burning technical problem is the transformation of measured noisy data (e.g. from medical imaging) into consistent models for a mathematical simulation. It is this field I have intensively worked on in the past years. From my experience, methods are most successful which exploit problem-specific properties to create a basis for the application of powerful combinatorial and geometric algorithms.