Experimental Medical Physics

Microbeam Radiation Therapy

Microbeam Radiation Therapy (MRT) is an innovative, but still preclinical concept in radiation therapy. In MRT radiation fields are subdivided into arrays of micrometre sized beamlets. Preclinical investigations showed that this treatment concepts stands out due to substantially reduced side effects at equivalent tumour control rate. We investigate the physical and radiobiological principles of MRT.

Microbeam Radiation Therapy (MRT) was developed at large 3rd generation synchrotrons such as the European Synchrotron (ESRF) in Grenoble, France. It is based on the observation that normal tissue can more easily repair damage confined to small regions than spatially extended lesions. In MRT the radiation field is spatially modulated on a micrometre scale. Micrometre sized radiation fields create extremely high peak doses of several hundred Gray, while doses between the peaks remain below the tissue tolerance level. Abundant preclinical evidence demonstrates that MRT spares normal tissue more effectively than conventional radiation fields at equal tumour control rates. While conventional radiation therapy assumes that only a high homogeneous dosage of the tumour target prevents a tumour regrowth, MRT is able to control tumours despite remaining low dose areas. In our team we investigate MRT focusing on biological, physical and technical research questions that need to be solved in order to translate this promising new approach into first clinical applications.

Radiation biology

Although the mechanisms behind MRT are not yet understood, there are several hypotheses. Most are related to differences in the microenvironment between tumour and healthy tissue:

1. Higher sensitivity of immature tumour vasculature to microbeams than mature normal tissue vasculature.
2. Intercellular signalling and bystander effects influence the cell survival.
3. Triggering of immune and inflammatory responses in the tumour due to high peak doses and an immunogenic cell death.

We investigate the mechanisms of MRT in-vitro and in-vivo. We have set-up a compact MRT source on a small animal irradiator system that allows in-vivo and in-vitro experiments independent of the availability of synchrotron radiation. Using various imaging techniques such as MR imaging, CT or phase contrast imaging we investigate the effects of MRT treatment after radiation exposure. We probe for intercellular and bystander signals that influence cell survival using 2D and 3D cell culture techniques.

Generation of microbeams with compact radiation sources

One of the major obstacles for a clinical application of MRT is the availability of appropriate radiation sources. Currently only large synchrotrons provide the necessary dose rates and radiation quality required for clinical treatments. In our team we develop novel concepts of microbeam sources and investigate new collimator designs for preclinical and clinical applications. As part of these efforts a compact preclinical microbeam irradiator was developed allowing for the first time in-vivo microbeam irradiations with a conventional x-ray tube.

Part of this work is:

1. Construction of a prototype for microbeam therapy (Emmy Noether project).
2. Dose calculation and treatment planning for experimental radiation sources.
3. Planning studies for future clinical applications.
4. Mathematical modelling of radiobiological effects.

Radio-oncomics: “Big data in radiotherapy”

The ubiquitous availability of large amounts of data is transforming our modern society and evaluation of large data sets has led to substantial improvements in economy, society and security. Also in the medical area patient data could be exploited to increase the effectivity of treatments, be it by stratification of patients, analysis of risk factors or personalized treatments. We investigate how medical image data can be used for refinement of prognoses, the prediction of treatment outcomes and individually tailored treatment protocols in radiation oncology.

In close collaboration with the University Hospital of TUM (Klinikum rechts der Isar) we mine data that is accumulated in radiotherapy treatment. We combine dose distributions, segmentations and medical imaging such as MR or CT with the aim to predict treatment outcomes. Of particular interest are time series of images acquired in the course of the treatment, documenting temporal changes in the tissue.

Machine learning methods, more specifically deep neuronal networks learn the relationship between image features, dose distribution and treatment outcome. Treatment outcome can either relate to normal tissue side effects or tumour growth/shrinkage.

Vision of our research ambition is a data based treatment planning system that acts as decision support in radiation oncology. Furthermore we aim to validate findings in preclinical radiobiology with clinical data, e.g. partial volume or bystander effects.

Particular research focusses are:

1. Combination of data of different imaging devices.
2. Correlation of radiomics information with 3D dose distributions.
3. Inclusion of prior bio-medical knowledge to overcome limited sample sizes.