IMAG/MSM WG Multiscale Modeling and Viral Pandemics Zoom @everyone
March 6, 2025 at 3PM (EST)
Chase Cockrell, University of Vermont.
The Wound Environment Agent-Based Model: A Digital Twin for Wound Healing.
The Wound Environment Agent-Based Model is a digital twin representation of volumetric muscle loss injury and its subsequent healing. The model is informed by a range of clinical data, including CT scans, quantitative histology, proteomics, and genomics. To capture the heterogeneity of the clinical population, we employed machine-learning techniques to discover a set of model parameterizations that could not be invalidated by experimental data. We then use deep reinforcement learning on the digital twin, instantiated with a range of parametrizations to determine both the most generalizable therapies to speed healing and reduce scar formation, and individually, to optimize dose strengths and timings for individuals.