Machine Learning Model Automates Prostate Cancer Radiation Therapy Planning
A new machine learning model developed by researchers at the Vector Institute for Artificial Intelligence and University Health Network can generate radiation therapy treatment plans for prostate cancer patients. The model, detailed in a Nature Medicine study, reduces the planning process from over a day to just a few hours, significantly freeing up hospital resources. In blind tests, the AI-generated plans were deemed as good as or better than human-created plans approximately 90% of the time. Currently deployed at Princess Margaret Cancer Centre in Toronto, the system involves clinicians choosing between AI and human plans, with AI plans selected for treatment 60% of the time due to initial clinical hesitancy. The technology was developed in partnership with RaySearch Laboratories, allowing for global distribution. While initially targeted at prostate cancer, which affects many patients, the model has the potential to expand to other types of cancer radiation therapy. This innovation represents a significant step in integrating artificial intelligence into clinical workflows, aiming to export specialized expertise to help more patients worldwide while maintaining high standards of patient safety and care quality.
Wire timeline
Machine Learning Model Automates Prostate Cancer Radiation Therapy Planning
A new machine learning model developed by researchers at the Vector Institute for Artificial Intelligence and University Health Network can generate radiation therapy treatment plans for prostate cancer patients. The model, detailed in a Nature Medicine study, reduces the planning process from over a day to just a few hours, significantly freeing up hospital resources. In blind tests, the AI-generated plans were deemed as good as or better than human-created plans approximately 90% of the time. Currently deployed at Princess Margaret Cancer Centre in Toronto, the system involves clinicians choosing between AI and human plans, with AI plans selected for treatment 60% of the time due to initial clinical hesitancy. The technology was developed in partnership with RaySearch Laboratories, allowing for global distribution. While initially targeted at prostate cancer, which affects many patients, the model has the potential to expand to other types of cancer radiation therapy. This innovation represents a significant step in integrating artificial intelligence into clinical workflows, aiming to export specialized expertise to help more patients worldwide while maintaining high standards of patient safety and care quality.
Vector Institute for Artificial Intelligence