Toward pulmonary digital twins: recent efforts in motion tracking using mechanical regularization, lung poromechanical modeling and estimation, model reduction using finite element neural networks
Martin Genet (Ecole Polytechnique)
Thursday 27th March 14:00-15:00 Maths 311B
Abstract
In this presentation, I will go through several research topics related to our endeavour toward building pulmonary digital twins from clinical data to improve diagnosis, prognosis and treatment of various pulmonary diseases.
The first topic is motion tracking from clinical images, especially mechanical regularization. I will first describe the general motion tracking problem and provide some literature review on mechanical regularization. I will then introduce a recent generalization of the equilibrium gap principle to the nonlinear finite strain framework, and the associated consistent finite element discretization. Finally, I will illustrate the tracking performance of the method, and provide various elements of validation based on synthetic and real images. [Genet et al., 2018] [Genet, 2023]
The second topic is lung poromechanical modeling and estimation. I will first present recent and ongoing works on the modeling of the lungs at both alveolar, tissue, and organ scales. More specifically, starting with tissue scale modeling efforts, I will describe constitutive choices suitable for the lungs [Patte et al., 2022]. Then, at the organ scale, I will introduce various sets of boundary conditions that we proposed for the lungs, describing the organ environment with various levels of details [Patte et al., 2022; Peyraut & Genet, 2024]. I will finish the modeling part of the talk with current work at the alveolar scale, including a consistent formulation of the micro-poro-mechanics problem as well as a bridge between micro- and macro- poro-mechanics [Álvarez-Barrientos et al., 2021; Manoochehrtayebi et al., In rev.]. I will then present the personalization pipeline we designed alongside the model, in order to personalize parts of the model geometry, behavior and boundary conditions based on clinical data [Patte et al., 2022; Laville et al., 2023; Peyraut & Genet, 2025]. As an example, I will discuss early results obtained on Idiopathic Pulmonary Fibrosis, a progressive form of interstitial lung disease that has a strong impact on the tissue mechanical properties and could be strongly influence by the tissue stresses and/or strains, but that remains poorly understood, poorly diagnosed, and poorly treated [Gonsard et al., 2024; Brillet et al., 2025].
The third topic is model reduction using finite element neural networks. Our work extends the HiDeNN framework, a sparse and interpretable neural network architecture that integrates the finite element method into traditional neural networks, for surrogate modeling. We introduced various improvements at the interpolation and training levels through reference element, adaptive meshing and multigrid strategies. Moreover, by combining HiDeNN with Proper Generalized Decomposition (PGD), we addressed high-dimensional parametric problems using tensor decomposition and dynamic architecture optimization. [Škardová et al., In rev.; Daby-Seesaram et al., In rev.]
References:
[Álvarez-Barrientos, F., Hurtado, D. E. & Genet, M. (2021). Pressure-driven micro-poro-mechanics: A variational framework for modeling the response of porous materials. International Journal of Engineering Science.](https://doi.org/10.1016/j.ijengsci.2021.103586)
[Brillet, P.-Y., Peyraut, A., Bernaudin, J.-F., Fetita, C., Nunes, H. & Genet, M. (2025). What is personalized lung poromechanical modeling and how can it improve the understanding and management of fibrotic interstitial lung diseases? Expert Review of Respiratory Medicine.](https://doi.org/10.1080/17476348.2025.2464886)
[Daby-Seesaram, A., Škardová, K. & Genet, M. Finite Element Neural Network Interpolation. Part II: Hybridisation with the Proper Generalised Decomposition for non-linear surrogate modelling. In Revision.](https://doi.org/10.48550/ARXIV.2412.05714)
[Genet, M. (2023). Finite strain formulation of the discrete equilibrium gap principle: Application to mechanically consistent regularization for large motion tracking. Comptes Rendus. Mécanique.](https://doi.org/10.5802/crmeca.228)
[Genet, M., Stoeck, C. T., von Deuster, C., Lee, L. C. & Kozerke, S. (2018). Equilibrated Warping: Finite Element Image Registration with Finite Strain Equilibrium Gap Regularization. Medical Image Analysis.](https://doi.org/10.1016/j.media.2018.07.007)
[Gonsard, A., Genet, M. & Drummond, D. (2024). Digital twins for chronic lung diseases. European Respiratory Review.](https://doi.org/10.1183/16000617.0159-2024)
[Laville, C., Fetita, C., Gille, T., Brillet, P.-Y., Nunes, H., Bernaudin, J.-F. & Genet, M. (2023). Comparison of optimization parametrizations for regional lung compliance estimation using personalized pulmonary poromechanical modeling. Biomechanics and Modeling in Mechanobiology.](https://doi.org/10.1007/s10237-023-01691-9)
[Manoochehrtayebi, M., Bel-Brunon, A. & Genet, M. Finite strain micro-poro-mechanics: Formulation and compared analysis with macro-poro-mechanics. In Revision.]
[Patte, C., Brillet, P.-Y., Fetita, C., Gille, T., Bernaudin, J.-F., Nunes, H., Chapelle, D. & Genet, M. (2022). Estimation of regional pulmonary compliance in idiopathic pulmonary fibrosis based on personalized lung poromechanical modeling. Journal of Biomechanical Engineering.](https://doi.org/10.1115/1.4054106)
[Patte, C., Genet, M. & Chapelle, D. (2022). A quasi-static poromechanical model of the lungs. Biomechanics and Modeling in Mechanobiology.](https://doi.org/10.1007/s10237-021-01547-0)
[Peyraut, A. & Genet, M. (2024). A model of mechanical loading of the lungs including gravity and a balancing heterogeneous pleural pressure. Biomechanics and Modeling in Mechanobiology.](https://doi.org/10.1007/s10237-024-01876-w)
[Peyraut, A. & Genet, M. (2025). Finite strain formulation of the discrete equilibrium gap principle: Application to direct parameter estimation from large full-fields measurements. Comptes Rendus. Mécanique.](https://doi.org/10.5802/crmeca.279)
[Škardová, K., Daby-Seesaram, A. & Genet, M. Finite Element Neural Network Interpolation. Part I: Interpretable and Adaptive Discretization for Solving PDEs. In Revision.](https://doi.org/10.48550/ARXIV.2412.05719)
Add to your calendar
Download event information as iCalendar file (only this event)