Supercomputer and AI to boost Danish cancer treatment – new project awarded €7 million

Associate Professor Jens Petersen from the Department of Computer Science, University of Copenhagen, is partner in a new project led by Professor Stine Sofia Korreman from the Department of Clinical Medicine, Aarhus University. The project also involves close collaboration with researchers from Odense and New York and all radiotherapy clinics in Denmark.

The aim is to develop an advanced AI model capable of analyzing medical images, such as CT and MRI scans, so that radiation therapy can be optimized and tailored to each individual cancer patient. 

- We will establish a research centre where all researchers collaborate closely on the project both nationally and internationally. The goal is to develop an AI model that functions somewhat like a large language model - but for medical images - which can make treatment more precise and better suited to the individual patient, while also easing the workload for doctors, says Professor Stine Korreman.

The grant paves the way for 15 new research positions and grants the project access to the supercomputer GEFION, which will process 130 million images from Danish radiation therapy clinics, enabling groundbreaking cancer research at an unprecedented pace. This extensive medical image dataset is believed to be the largest of its kind in the world.

The project’s goal is not only to improve treatment for the approximately 14,000 Danish cancer patients who receive radiation therapy each year, but also to develop solutions that can be scaled globally - particularly to benefit low- and middle-income countries.

The role of the Department of Computer Science and Jens Petersen will be to develop algorithms to detect changes that occur during treatment and create individualised models that incorporate patients' medical history.

- Our AI models will enable more precise and continuous adaptation of radiotherapy treatments to improve patients' prognosis. We will develop new methods that can utilise the information that is available from existing treatment scans, but which currently cannot be fully utilised, says Associate Professor Jens Petersen.

In addition to Jens Petersen, Assistant Professor Raghavendra Selvan from the Department of Computer Science is also involved in the project. Raghavendra Selvan's role will be to develop methods for optimising the AI models so that they can be implemented in resource-constrained environments.