Bridging Accuracy and Efficiency: The Role of PINNs in Immune Response Simulations

Myocarditis, characterized by inflammation of the heart muscle, has seen a notable 62.2% increase in incidence over the past three decades, leading to 324,490 deaths in 2019. Despite advancements in understanding its etiology, several critical questions remain unresolved, highlighting the need for improved patient treatment strategies. Computational methods play a crucial role in unraveling the complex interactions between pathogens and the immune system. This study explores the use of Physics-Informed Neural Networks (PINNs) to accelerate a newly developed model for myocardial edema formation in acute infectious myocarditis. The model describes the concentrations of pathogens and leukocytes using a nonlinear system of ordinary differential equations (ODEs). Our findings demonstrate a mean computational speedup of 5.93, while maintaining a Root Mean Squared Error of 0.0015. These results indicate that PINNs can accurately replicate the ODE system’s solutions, offering substantial computational efficiency for this application.