EUREKA ITEA ASSIST -Automation, Surgery Support and Intuitive 3D visualization to optimize workflow in IGT SysTems

Purpose and goal:

The main goal of the ASSIST project is to help doctors to work more efficiently, by taking advantage of recent developments in deep learning. Radiation therapy is an efficient way to treat a tumor, but the planning takes time as it is necessary to collect medical images, segment tumor and risk organs, and generate a treatment plan that will kill the tumor cells while not damaging other organs.

Deep learning can be used in all these steps, to shorten the time for planning, which will lead to a shorter waiting time for the patients. Expected results and effects: The expected results of the project are - faster and more accurate segmentation of tumors and risk organs, through deep learning

- a framework for training deep networks through so called federated learning, where medical images do not leave the hospital

- to be able to plan tumor treatment with only MR images by synthesizing CT images, this will save time and reduce the radiation for the patient Approach and implementation:

The Swedish part of the project will be carried out by Linköping University in collaboration with the companies RaySearch, Spectronic, Inovia and Eigenvision. The Swedish parties provide different kinds of expertise; RaySearch and Spectronic are experts in radiation therapy, Eigenvision are experts in medical imaging, Inovia are experts in machine learning and infrastructure for this, and LiU are experts in medical imaging and machine learning.

In the project, we will also collaborate with Quantib and Leiden university medical center (Netherlands), among others.