duCAT

Develop and validate a framework for clinical cancer research, incorporating predictive models developed by machine learning based on data from multiple institutions.

The objective of this project was to develop a new framework for clinical cancer research. This revolutionary IT solution more explicitly takes account of the properties of an individual patient in the development and translation to clinical trials on new diagnostic and therapeutic modalities.

In this project we used machine learning to develop new prediction models with the aim to predict survival, radiation induced lung damage and esophagitis. The high accuracy of these prediction models allows it to be used in clinical practice in the participating centres. duCAT was awarded funding by STW in 2011, ran until 2018 and used distributed learning as its core approach. The tools, knowledge and technologies it generated formed the basis of the PHT. The Varian Learning Portal provides a commercial solution.

DuCAT was built upon the CAT database system implemented at MAASTRO and developed by Siemens. The vision for this system is one that can be used by clinical researchers and pharmaceutical companies to advance clinical research in the Netherlands.