Modeling and Simulation of Tumor Growth by Finite Element and Machine Learning Method
Open AccessTumor has been an object of computational model studies for decades. Most mechanobiology phenomena commonly involve biological growth and deformation. Statistical and machine learning methods have greatly benefited scientific discovery and understanding through the analysis of biological data. It is essential to develop and apply machine learning models that can incorporate knowledge of mathematical models simulating tumor growth and predicting tumor growth via a data-driven method. There are mainly two methods for simulation of tumor growth. The first one is using the continuum model. We propose an innovative cancerous growth model that describes the tumor as a poroelastic medium consisting of solid and fluid components. In our biologically informed mechanical description of tumor growth dynamics, we derive the governing equations of the tumor’s growth and incorporate them with large deformation and materially nonlinear constitutive equations to improve our simulation's accuracy and efficiency.Meanwhile, the dynamic finite element equations (DFE) for coupled displacement and pressure fields are formulated and solved. The 3-dimensional porous model is introduced. Numerical results are presented and discussed. The FEA-based ML approach for modeling the biomechanical behavior of the tumor growth is proposed. We introduced five regression models and tuning them to perform the prediction results. Nine breast tumor models are built using SolidWorks. The models were imported into COMSOL then performed the simulations with a developed nonlinear FEM solver. Dataset generated via COMSOL are then put into the ML models training and model validation. Regression model evaluations are discussed. XGBoost model is suitable for predicting biology growth.
- All rights reserved
Notice to Authors
If you are the author of this work and you have any questions about the information on this page, please use the Contact form to get in touch with us.