IAEA Launches New Research Project on AI in Radiotherapy Contouring

The International Atomic Energy Agency (IAEA) announced the initiation of a new two-year Coordinated Research Project (CRP) aimed at exploring the integration of artificial intelligence (AI) in the contouring process essential to radiotherapy for head and neck cancer. This project was officially launched on June 13, 2025, and aims to refine the accuracy and efficacy of contouring, which involves delineating tumors and surrounding tissues to optimize radiation therapy delivery.

In recent decades, advancements in radiation oncology have significantly evolved, driven by innovative treatment technologies and enhanced imaging modalities. However, discrepancies in contouring practices remain prevalent, often leading to suboptimal treatment outcomes. "Missteps during contouring can significantly impact treatment outcomes—be it from missing the tumors or irradiating normal tissues unnecessarily," stated Dr. Tomoaki Tamaki, Head of Applied Radiation Biology and Radiotherapy in the IAEA's Division of Human Health. He emphasized the importance of accurately identifying tumors amid the growing global cancer burden, limitations in diagnostic imaging, and the inherent challenges that clinicians face during contouring.

The CRP will build upon a preceding project that concluded in 2024, which assessed the utility of AI tools in identifying organs-at-risk, crucial for protecting healthy tissues during treatment. The new initiative focuses on enhancing contouring precision through deep learning algorithms and explainable AI (XAI), which aims to provide clarity on AI decision-making processes to healthcare providers. Dr. Tamaki highlighted that previous studies have demonstrated a reduction in contouring time and variability when AI tools are utilized effectively.

The overarching objective of this CRP is to improve the quality of radiotherapy contouring specifically for head and neck cancer by evaluating the effects of e-learning and deep learning-based auto-segmentation on inter-observer variation and bias in contouring gross tumor volumes. The study will be executed in two distinct phases. The initial phase will compare AI-assisted contouring outcomes—with and without explainability—against traditional manual contouring methods. Subsequently, the second phase will investigate the risk of bias associated with AI tools, again considering the presence or absence of explainability. An optional third phase may explore optimal visualization methods for presenting AI-generated explanations.

The IAEA invites participation from all member states, asking that eligible radiotherapy centers treat a minimum of 20 head and neck cancer patients annually and possess advanced 3D radiotherapy capabilities. The agency encourages research institutions to incorporate women and young researchers into their project proposals, reflecting a commitment to inclusivity within the scientific community. Proposals must be submitted by August 30, 2025, to the IAEA’s Research Contracts Administration Section.

This innovative CRP comes at a crucial time as global cancer rates rise, underscoring the need for advanced techniques that can enhance the precision of cancer care. The project not only aims to refine the contouring process but also to address potential biases introduced by AI technologies, ensuring that the integration of AI enhances patient outcomes while maintaining ethical standards in medical practice. As the project unfolds, its implications could reshape radiotherapy practices worldwide, facilitating improved treatment for millions of patients facing head and neck cancers.

In conclusion, the IAEA's commitment to leveraging AI in radiotherapy reflects an ongoing effort to modernize cancer treatment methodologies, ensuring that advancements in technology translate into better clinical outcomes for patients around the globe.