Estimating Active Stress in Cardiac Biomechanical Models Based on Physics-Informed Neural Networks
Please login to view abstract download link
Biophysical models and, more broadly, digital twins of the cardiac function are becoming increasingly popular due to their potential to predict patient outcomes and optimise treatment plans. As such, active stress models in cardiac biomechanics, which account for the mechanical deformation caused by muscle activity, provide a link between the electrophysiological and mechanical properties of the tissue. The accurate assessment of active stress parameters is fundamental for a precise understanding of myocardial function but remains difficult to achieve in a clinical setting. In this work, we study the application of physics-informed neural networks [1] with high-resolution three-dimensional nonlinear cardiac biomechanical models to reconstruct displacement fields and estimate patient-specific biophysical properties related to active stress. The physics of the problem is represented by a mathematical model based on partial differential equations. Additionally, the learning algorithm incorporates displacement and strain data that can be routinely acquired in clinical settings. The presentation includes a series of benchmark tests that demonstrate the accuracy, robustness, and promising potential of this method for the precise and efficient determination of patient-specific physical properties in nonlinear biomechanical models. This approach opens a new pathway to enable the detection and characterisation of tissue inhomogeneities, such as fibrotic regions, and could significantly impact the diagnosis, treatment planning, and management of heart conditions associated with cardiac fibrosis.
