Cancer is a widely known and widely prevalent disease that is a leading cause of death worldwide. In 2022, it accounted for 9.7 million deaths, and there were an estimated 20 million new cancer cases. This disease encompasses more than 100 different types, categorized by the type of cell they start from. Among them, the most common are lung, breast, prostate, colon, and skin cancers.

Depending on the case, the detection method can vary significantly, including X-rays, CT scans, MRI scans, and blood tests. Despite advances in diagnostic technologies, a significant challenge remains the lack of accessible data, crucial for developing machine learning (ML) systems. Traditional data gathering is hindered by privacy concerns, preventing effective data sharing.

To address these issues, Federated Learning (FL) has emerged as a promising solution. It enables collaborative model training without direct data transfer, therefore adhering to privacy regulations and having the potential to become the primary learning paradigm for distributed cancer research. 

While numerous studies are benchmarking FL and developing methods in research environments, the transition to clinical trials has not started yet. Federated Learning’s compatibility with regulatory standards could accelerate and potentially revolutionizing cancer diagnosis and treatment by enabling using AI by clinicians.