Seen here, the collimator developed by the research team for the project. Image source: Stefan Bartzsch

Helmholtz AI project call showcase: AI to apply precise Radiation Therapy

Insights into projects funded by the Helmholtz AI project calls. Starting with a project that's a collaboration between Helmholtz Munich and Helmholtz-Zentrum Dresden-Rossendorf (HZDR)


The Helmholtz AI project calls so far funded 36 high-risk high-gain AI and ML projects proposed by Helmholtz researchers. In the coming weeks, we will shine the spotlight on the researchers working on these projects and give insights into the topic itself. We begin the Helmholtz AI project calls showcase series with a look at the project headed by Stefan Bartzsch at Helmholtz Munich. This project is a result of a collaboration between Helmholtz Munich and Helmholtz-Zentrum Dresden-Rossendorf (HZDR). Here's what Stefan has to say about the project.

Seen below, a figure showing the dose distribution of minibeams. Image copyright: Stefan Bartzsch

Could you introduce yourself, giving your affiliation, area of work, and of course, the project title?

I am leading a research team at Helmholtz Centre Munich and at the Klinikum r. d. Isar of the TU Munich. I am investigating novel strategies in radiation oncology that are less destructive to healthy tissue and organs. One research focus of my team is spatially fractionated radiation therapy, where we expose the tumour with razor blade like high dose beams. The width of these beams is just a few micrometre. In the Helmholtz AI project “AI to apply precise Radiation Therapy” we investigate the efficacy of these beams for the treatment of brain tumours.

In simple words, what specifically is your project about? And, how and why do you think it is a high risk, high gain endeavour?

Spatially fractionated radiation therapy is effectively destroying tumour tissue, but does less harm to healthy tissue than conventional radiation therapy. Therefore, the technique may strongly enhance the success of radiation therapy treatments in the future. However, we need a better understanding of the mechanisms before we can use it for patient treatments. We think that the razor blade-like beams in spatially fractionated radiation therapy damage specifically the chaotically organized tumour vasculature. To investigate this hypothesis we treat mouse tumours with minibeams at the proton facility OncoRay in Dresden. Afterwards we use tissue clearing methods and light sheet microscopy to visualize the vasculature and changes to it. The huge amount of resulting 3D imaging data can only be analysed using artificial intelligence. The results will help us to understand mechanistic differences in the radiobiology of vascular damage and repair after treatment with minibeams compared to conventional treatments. This is an important step towards clinical translation of minibeam therapy.

How important has the Helmholtz AI funding and platform been to carry out this project?

Helmholtz AI funding initiated a really fruitful cooperation between Dresden and Munich! We are financing a PhD student from the grant and we have been able to get the hardware required to produce minibeams in Dresden.