Peter Andersson: Simulations and AI may improve radiation treatment
By combining simulations and AI, Peter Andersson has developed new ways to analyze radiation doses during cancer treatment. His research may help improve the quality of radiotherapy.
PETER ANDERSSON
Dissertation defense: 14 May 2025 (click for details)
Doctoral thesis:
Research area: Medical Radiation 91探花s
Sahlgrenska Academy, The Institute of Clinical 91探花s
External beam radiotherapy is commonly used to treat cancer. To increase effectiveness and minimize damage to healthy tissue, more advanced techniques are being used where the radiation is adjusted dynamically throughout the treatment. But this also raises the risk that small deviations go unnoticed.
Digital image of the radiation dose
鈥淭he research shows that you can combine simulations with AI to create a kind of digital image of the radiation after it has interacted with the patient,鈥 says Peter Andersson, researcher at RISE Research Institutes of Sweden and doctoral student at the Institute of Clinical 91探花s.
With the help of so-called in vivo dosimetry鈥攁 method that measures the radiation dose directly in or near the body during treatment鈥攊t is possible to ensure that a patient has received the prescribed dose in the planned location.
Detecting treatment deviations
Peter Andersson鈥檚 research shows how Monte Carlo simulations鈥攁 calculation method based on running millions of random events to simulate complex physical processes鈥攃an be combined with deep learning to generate and analyze image data.
鈥淭his can help healthcare professionals detect treatment deviations that are otherwise difficult to see, which ultimately supports safer treatment,鈥 says Peter Andersson.
Merging tech and care
What has been rewarding and enjoyable in your doctoral work, and what has been challenging?
鈥淏eing able to combine physics, programming, and clinical applications has been both fun and meaningful. The hard part has been making the systems work automatically and reliably, since both the clinical data and the techniques involved are complex,鈥 says Peter Andersson.
By highlighting both the possibilities and limitations of current quality assurance methods, the thesis points to the need for better ways to detect deviations in radiation therapy. The proposed strategy is based on combining simulations with image analysis. It is an approach that could help improve patient safety and make treatments more precise.
Text: Jakob Lundberg