Medical imaging based in silico head model for ischaemic stroke simulation
Yun Bing1, Daniel Garcia-Gonzalez2, Natalie Voets3
1Department of Engineering Science, University of Oxford, Parks Road, Oxford, OX1 3PJ, UK.
Journal of the Mechanical Behavior of Biomedical Materials
|October 12, 2019
Summary
Stroke simulations reveal midline shift (MLS) may not accurately predict severity or tissue damage. Stroke location significantly impacts MLS evolution and potential axonal injury, suggesting a need for location-specific analysis in clinical decision-making.
Area of Science:
- Computational neuroscience and biomechanics
- Medical imaging and simulation
- Neurological injury modeling
Background:
- Stroke is a leading cause of death and disability globally, often leading to brain edema and midline shift (MLS).
- Current clinical use of MLS to assess stroke severity and guide surgical decisions has known limitations.
- In silico experiments offer a promising approach to understand stroke's impact on brain tissue, aiding clinical decision support.
Purpose of the Study:
- To develop and utilize biologically-informed finite element head models for simulating stroke-induced brain changes.
- To investigate the relationship between stroke location, brain edema, midline shift (MLS), and axonal injury.
- To propose an in silico methodology for predicting stroke evolution and aiding clinical decision-making.
Main Methods:
- Construction of human and rat finite element head models incorporating MRI-derived tissues (grey matter, white matter, CSF, skull) and vasculature.
- Development of constitutive models to represent tissue mechanical behavior during edema development.
- Calibration of swelling parameters using the rat model, followed by simulation of human stroke at three common locations (basal ganglia, fronto-opercular/anterior insula, temporo-parietal).
Main Results:
- Simulations showed a quadratic MLS evolution over time for all simulated stroke locations.
- The basal ganglia stroke location exhibited the largest MLS, while the temporo-parietal location showed the smallest.
- A proposed axonal tract injury criterion was higher in the temporo-parietal stroke simulation, indicating potential location-dependent damage.
Conclusions:
- Stroke location is crucial when interpreting MLS as an indicator of stroke severity.
- MLS values may not reliably correlate with the extent of underlying tissue damage.
- The proposed in silico methodology shows potential for predicting stroke progression using MLS and location data.
