Medical implants are highly developed in terms of the compatibility of the materials. Nevertheless, the failure rate is still very high compared to technical systems. Especially the biophysical integration is often not well understood. Here, modelling and simulation techniques can help to gain a better understanding. Although powerful computer models are available, patient-specific modelling is still very complex. As a consequence these models are not used in clinical practice until today. To advance these models for clinical application, newer developments in computer mechanics, namely model reduction techniques (surrogate models), should pave the way.
The aim of subproject A07, therefore, is to develop surrogate models for patient-specific prognosis in treatment with hip joint and cochlear implants to support the assessment of implant status.
Based on established high-fidelity finite element models, the parameter spaces are defined in close cooperation with physicians and the high-fidelity models have to be parameterised for fully automated runs. Efficient techniques for reducing the model order have to be tested for these systems, especially taking into account the dimensionality of the problems. In a second step, the reduced models should additionally be tested with clinical data. In a clinical study with 30 CI users we will collect impedance and trans-impedance measurements, evoked composite action potentials with different stimulation parameters, medical imaging data to estimate the electrode position and magnetic resonance imaging of the auditory nerve before implantation.