Model Order Reduction (MOR) of Function‐Perfusion‐Growth Simulation in the Human Fatty Liver via Artificial Neural Network (ANN)

Abstract

Numerical modeling of biological systems has become an important assistance for understanding and predicting hepatic diseases like non‐alcoholic fatty liver disease (NAFLD) or the detoxification of drugs and toxines by the liver. We developed a model for the simulation of hepatic function‐perfusion processes using a multiscale and multiphase approach. Here, the liver lobules are described using a homogenization approach with a coupled set of partial differential equations (PDE) based on the Theory of Porous Media (TPM) to describe the coupled blood transport and tissue deformation. For the description of metabolic processes on cellular scale ordinary differential equations (ODE) are used. For many practical and clinical applications, e.g. optimization procedures or uncertainty quantification, a fast but reliable computation is required. Thus, we use a non‐linear model order reduction (MOR) based on an artificial neural network (ANN) for the prediction of simulation results. The practicability of this approach is shown in a comparison between the high fidelity numerical simulation of a NAFLD and the predicted results by the ANN.

Quantifying Fat Zonation in Liver Lobules: An IntegratedMultiscale In-silico Model Combining DisturbedMicroperfusion and Fat Metabolism via aContinuum-Biomechanical Bi-scale, Tri-phasic Approach

“…While simulating liver diseases to demonstrate the growth capability of the liver during fat accumulation leading to MASLD, several models deal with the tissue remodeling and alterations in the elasticity of the liver lobule. Waschinsky et al (2016), Lambers et al (2021Lambers et al ( , 2019 or Lambers (2023) focus on the effects of growth due to a high-fat diet, as well as other models using a one-phasic, cf. Menzel and Kuhl (2012), or a bi-phasic, cf.…”

Quantifying fat zonation in liver lobules: an integrated multiscale in silico model combining disturbed microperfusion and fat metabolism via a continuum biomechanical bi-scale, tri-phasic approach

“…While simulating liver diseases to demonstrate the growth capability of the liver during fat accumulation leading to MASLD, several models deal with the tissue remodeling and alterations in the elasticity of the liver lobule. Waschinsky et al ( 2016 ), Lambers et al ( 2021 ), Lambers et al ( 2019 ) or Lambers ( 2023 ) focus on the effects of growth due to a high-fat diet, as well as other models using a one-phasic, cf. Menzel and Kuhl ( 2012 ), or a bi-phasic, cf.…”

Simulation of zonation-function relationships in the liver using coupled multiscale models: Application to drug-induced liver injury

“…Therefore, multiscale models spanning different size scales of the liver have been presented to simulate the detoxification of substances in the liver at subtissue, tissue, and whole body scales [23, 24]. By combining a porous media approach based on the theory of porous media (TPM) at the lobular scale and systems biology models at the cellular scale, multiscale models are able to describe the function-perfusion relationship with respect to glucose metabolism [25], lipid metabolism [26, 20, 27] or acetaminophen (APAP) detoxification [28, 29].…”