Physics-Informed Neural Networks (PINNs) represent a novel class of machine learning models that integrate physical laws into neural networks to solve complex scientific problems. This work explores the application of PINNs to immunotherapy models in cancer, particularly focusing on solving the inverse problem using available data. By leveraging the principles of immunology and oncology, we utilize PINNs to estimate parameters within cancer-immune interaction models. This research demonstrates that PINNs can effectively predict treatment outcomes and provide insights into the optimization of immunotherapy strategies.