Artificial Intelligence and Big Data in Radiation Oncology
We use Artificial Intelligence to develop and deploy novel methods for effective and personalized radiotherapy treatment planning.
Research
Artificial Intelligence in radiotherapy and medical imaging is in tremendous development, with a wide range of potential applications including deep learning based image segmentation, generative image reconstruction and dose prediction, statistical modeling of tumour spread, and radiomics for outcome prediction.
In our research group, we use Artificial Intelligence (AI) to perform advanced analysis of medical images including CT, PET and MR scans, in large patient cohorts, with focus on head and neck cancer and breast cancer. We primarily address the problem that presently gives rise to the largest source of uncertainty in radiotherapy; namely target definition, including its response during treatment.
Target delineation is a time-consuming process, which is also prone to large uncertainties and subjective assessment by the delineating clinician. We develop deep learning models for AI-assisted target segmentation in an interactive user interface, to speed up the delineation process while at the same time reducing the uncertainty related to delineation for a more accurate treatment.
We investigate uncertainty estimation in deep learning predictions and its potential use for increasing transparency and interpretability, thereby increasing trust in the systems and facilitating safe use.
We also investigate risk models for tumour spread, which can be used for novel approaches using probabilistic optimisation of radiation dose within and around the tumour. This approach will give the potential to differentiate does prescription according to risk and reduce high radiation doses to volumes of tissue where the risk of finding cancer cells is low, hence reducing the risk of side effects.
We have a focus on the translation of research results into clinical practice covering all parts of the path from innovation and development to implementation, and we have initiated and lead several clinical trials testing the use of AI-assisted delineation.
We have a wide international network of collaborators, not least including the International Atomic Energy Agency and a consortium of investigators in low- and middle-income countries, with whom we explore the potential of AI to drive automation of radiotherapy to help address the global disparity in access to treatment.
Overall, we aim to decrease the level of uncertainty in the treatment and enhance personalization of radiotherapy through computational methods empowered by artificial intelligence.
The research group is interdisciplinary and spans medical science, medical physics, statistics, biomedical engineering, computer science, health science, and we have a high degree of collaboration with clinical staff.
We swear to a principle of transparency and all our output, including publications, presentations, and code can be accessed via our Gitlab page:
People
Professor of medical physics
Stine Sofia Korreman
stkorr@rm.dk
Aarhus University, PURE, Further information, Orcid
Postdoc
Jintao Ren
JINREN@rm.dk
PURE, Aarhus University. Further information.
Postdoc
Lasse Hindhede Refsgaard
LASREF@rm.dk.
PURE, Aarhus University. Further information.
PhD student
Emma Skarsø Buhl
EMSKAR@rm.dk
Further information.
PhD student
Nadine Vatterodt
nadine.vatterodt@clin.au.dk
PhD student
Kristoffer Moos
KRIMOO@rm.dk
PhD student
Ihsan Bahij
ihsan.bahij@rm.dk
PhD student
Zixiang Wei (co-supervision)
ZIXWEI@rm.dk
PURE, Aarhus University.
PhD student
Maiken Mondrup Hjelt (co-supervision)
maikenhjelt@oncology.au.dk
Collaborating researchers
Professor, MD, Dep of Oncology, Aarhus University Hospital
Jesper Grau Eriksen
jesper@oncology.au.dk
Professor, MD, Dep. of Oncology, Aarhus University Hospital
Birgitte Offersen
Birgitte.Offersen@auh.rm.dk
Associate professor of medical physics
Jasper Albertus Nijkamp
jaspernijkamp@clin.au.dk
Further information.
Associate professor
Jesper Folsted Kallehauge
jespkall@rm.dk
Further information.
Medical Physicist, Dep. of Oncology, Aarhus University Hospital
Anne Ivalu Sander Holm
annivaho@rm.dk
Head of Medical Physics, AUH Oncology Dept.
Ditte Sloth Møller
dittmoel@rm.dk
Further information.
Medical physicist, PhD
Ulrik Vindelev Elstrøm
ulrik.vindelev.elstroem@auh.rm.dk
PURE, Aarhus University.
Consultant, Associate Professor of Clinical Medicine
Kenneth Jensen
kenneth.jensen@auh.rm.dk
Further information.
Research topics
Head and neck cancer:
- Deep learning models for target segmentation
- Uncertainty estimation in deep learning target segmentations
- Interactive tools for AI-assisted delineation
- Computational target definition based on risk modelling
- Data mining for characterization of practices and toxicity modeling
- Adaptive treatment planning for anatomical variations
Breast cancer:
- Data mining for cardiac toxicity risk factors and modelling
- Data mining for mapping of delineation and treatment planning practices
- Automated delineation and treatment planning