Poor survival after radiation treatment for lung cancer patients has recently started to focus on cardiac toxicity and toxicity of the immune system. Several studies have shown that higher doses to the heart (1), as well as higher doses to the immune system (2) are correlated to poorer survival. The exact cause of the poorer survival for patients receiving higher cardiac or immune system doses is unknown, and requires a large dataset to disentangle the two effects. Gathering such a dataset in a single institute is cumbersome and arguably impossible, and privacy sensitive, as it requires anonymization and sending of large quantities of data. These issues can be resolved using federated learning, where statistical models (including complex machine learning and deep learning models) can be trained on oncology data distributed all over the world, without the requirement to take patient data outside of the originating hospital (3). By sharing the model instead of patient data, federated learning allows models to learn from data in a privacy-sensitive method. In this way we expect to study the association of cardiac and immune system toxicity with overall survival in thousands of patients.

Your research will be embedded in the Department of Radiation Oncology of the Netherlands Cancer Institute, in close collaboration with the research group at MAASTRO and you will be part of an international federated learning consortium.

We are looking for a highly motivated and enthusiastic PhD student with a strong affinity for radiotherapy, machine learning and/or (bio)statistical modelling. Moreover, interest in federated learning and the ICT tools required to make the approach work will be required. The candidate should have a master’s degree in physics, biomedical engineering , mathematics, informatics or a related field. Strong communication skills are required in this multidisciplinary project. A background in radiotherapy or modelling as part of the master program is recommended. Experience with programming, preferably in Python, expertise with Docker containerization and federated learning and processing DICOM imaging data is appreciated.

Fixed-term contract: 4 years.

Compensation

Your temporary employment will be for a period of 4 years. The gross salary per month for our PhD will range from €3.355,- to a maximum of €4.073,- according to the standard PhD scales. The terms of employment will be in accordance with the Collective Labor Agreement for Hospitals. In addition you will receive a fixed end-of-year bonus (8,33%) and you will receive 8,33% holiday pay.

For more information in regards to the secondary conditions please visit our website: https://www.nki.nl/careers-study/how-to-apply/.