Physics-Informed Neural ODEs for Tumor-Immune Dynamics Modeling
Anish Sarkar, Anindya Sarkar
We present a physics-informed neural ODE approach for modeling tumor-immune dynamics using Julia. Dual neural networks learn immune suppression and corrections within Gompertz growth framework. Model achieves R² > 0.95 (approx. 99.7%) accuracy and captures 30.25% volume reduction effects. Implementation uses DifferentialEquations.jl, Flux.jl, and SciMLSensitivity.jl, training in <1hr on Atsou tumor datasets.
The JuliaHealth Mini-Symposium
Lawrence Room 104 - Function Room