New state-of-the-art through AI use for tumor delineation in radiotherapy

Purpose and goal

In radiotherapy, radiation beams are directed towards the tumor, while trying to avoid damage to surrounding healthy tissue. As a basis for the treatment, a three-dimensional image is used, in which a physician manually delineate the tumor and surrounding healthy organs. However, manual delineation is a time-consuming process that is characterized by great individual variation and uncertainty regarding the boundaries of the region to include. In this project, an AI based method has been developed for automatic delineation prior to radiation therapy.

Expected results and effects

Within this project, a world-class AI-based method has been developed for automatic delineation of the structures required for radiation therapy. By including feedback from the treatments carried out, the AI can continuously be updated and maintain state-of-the-art performance also in the future. The developed technology can significantly improve the quality of care for patients, by offering more accurate and consistent delineations than is feasible in today´s workflow, while providing effective decision support with significant time gains for physicians.

Planned approach and implementation

The project has been carried out in collaboration between industry and healthcare. The work has been divided into work packages for which one of the parties were responsible. Algorithm and software development has mainly been carried out by the company while clinical data collection has been carried out by the hospital. By combining the technical cutting-edge expertise of the company with the high-level clinical competence and data available at the hospital, the project has reached a highly successful outcome.