Optimization of Radiation Applications in Medicine

Our mission is improving the efficiency of medical applications of ionising radiation. The research focus of our working group is to make personalised applications of ionising radiation in medicine in an efficient way with regard to medical diagnostic and therapeutic procedures. Scientific strengths of our working group are the expertise in biokinetic modelling, imaging geometries and tomographic reconstruction methods as well as uncertainty and sensitivity analysis.

Tomografic application and visualisation

In tomographic imaging processes, the reconstruction algorithm plays one of the most important roles. In the ideal case, the geometry of the tomographic scanner used for data acquisition is optimized for the data geometry that is required by the reconstruction algorithm applied. For the new reconstruction algorithm OPED, two scanning geometries have been designed: WATCH (Well Advanced Technique for Computed tomography with High resolution, Fig.1) and CT D'OR (Computed Tomography with Double Optimal Reading, Fig. 2).

Fig. 1 WATCH (Well Advanced Technique for Computed tomography with High resolution)

The data collected by both geometries can be used by the OPED reconstruction algorithm directly without any further adaptations. Besides, OPED can also cope with data from conventional CT systems without noticeable loss of image quality.

Fig. 2 CT D'OR (Computed Tomography with Double Optimal Reading).

Biokinetic investigations in nuclear medicine

In nuclear medicine, radioactively labeled molecules (radiopharmaceuticals) are administered to patients for diagnostic or therapeutic purposes.

We study and model the biokinetics of these radiopharmaceuticals used in diagnostic as well as in therapeutic applications in cooperation with Nuclear Medicine Departments. In diagnostic applications, e.g. using PET or SPECT, the knowledge of the biokinetics in the organs of interest as well as in the surrounding soft tissues enables us to optimize the time scheme for image acquisitions. The biokinetic information of the radiopharmaceuticals can be used to estimate the uncertainty of the radiation dose to the patients.

Fig. 1 Reconstructed PET images at different times after injection of 18F-Choline and measured activity in e.g. the liver


In the European project MADEIRA, which we coordinated, this optimization is coupled with improved methodologies for image reconstruction and noise reduction in order to enable good quality diagnostic studies while administering less radioactivity to the patients. In therapeutic applications, individualized biokinetics studies are performed to obtain patient-specific dose estimates that can be correctly correlated to the success of the therapy and used to improve the therapy protocol. 

Fig.2 Development of the biokinetic model of 18F-Choline which matches the experimental activity data

Dynamic processes in molecular imaging


We develop and validate new methods using targeting molecules and small particles in order to improve radiation therapy and diagnostics.

Radiation Therapy

Aiming for the use of interface effects in radiation therapy we investigate dose enhancement close to small particles. According to the idea of accumulating small particles at the target area of radiation therapy we examine small particles in dosimetric experiments. Here we use polymer-gel dosimetry as it offers an experimental setup to investigate dose deposition in tissue like material with high spatial resolution.

Fig.1 Gold particles at a cell membrane

The experimental results are applied to cell experiments in order to connect dose deposition to biological effects. In parallel we validate targeting molecules in order to aim these effects to highly specific sites e.g. tumors.

Fig.2 Illustration of Localized Radiation Interaction around Small Particles Diagnostics


The knowledge about the localization and the dynamic behavior is an important basis for the experiments with particles. Especially changes in their biokinetic behavior connected to different surface treatment with targeting molecules and functionalization are to be explored.

Fig.3 Gold Particles can be connected to targeting molecules in order to serve as contrast agent

The particles themselves can serve as contrast agent in radiography and computed tomography or magnetic resonance imaging. The development of low noise, high resolution and fast imaging techniques helps to better understanding of targeting mechanisms of the functional molecules in use.



Energy-sensitive imaging

Although the efficiency and life-time of x-ray tubes could continuously be increased, the fundamental method of creating x-rays has not changed since their discovery by Wilhelm Conrad Röntgen 1895. Electrons accelerated in a high potential electric field are hitting a target where they lose almost all their energy in form of heat and only a small fraction is converted (mainly by Bremsstrahlung) into x-rays with a spectrum of energies. Although for the imaging process, the low-energetic x-rays are filtered, the resulting x-ray spectrum still covers a large energy range. Furthermore, all common x-ray imaging modalities are based on recording all photons transmitted through the patients without resolving their energy. Since the absorption of x-rays in the body is energy dependent, different tissue compositions can lead to the same recorded signal, and thus, such tissues cannot be discriminated in the final image. This problem can be overcome by either using mono-energetic x-ray sources or by resolving the x-ray energy in the imaging plate (detector).

Currently, mono-energetic x-rays can only be created in large accelerators, and thus can hardly be used in clinics. However, lately research on a laser-driven x-ray source has been started by the Munich Centre of Advanced Photonics (http://www.munich-photonics.de/research-areas/area-a3/), which has the potential to provide quasi mono-energetic x-rays by a small-scale device. The impact of such a source on mammographic imaging has been explored by simulations and a considerable improvement of the image signal-to-noise ratio compared to conventional mammography could be attested. Thus, less radiation dose would be required with this source.

Fig.1 Simulation of novel CT systems

Fig.1 Simulation of novel CT systems

X-ray fluorescence imaging

X-ray fluorescence imaging The visualisation of physiological processes (e.g., in neurology or cardiology) or detection of diseases (e.g., cancer) is enabled by molecular imaging, where molecules sensitive to processes or cell types are coupled to a signalling substance. In the majority of cases, the signalling substances are radioactive isotopes, thus the technique is called nuclear imaging. The most commonly used methods are positron emission tomography (PET) and single-photon-emission tomography (SPECT), which however have low spatial resolution, such that now-a-days these modalities are often coupled with additional tomographic systems (thus, e.g., PET-CT, SPECT-CT). The drawback of nuclear imaging methods is the radiation burden of the patient due to the radioactive decay of the incorporated radionuclide. Particularly, organs involved in the metabolism and excretion of the tracer are exposed to radiation although usually not being the target of the diagnosis.

An imaging modality where radiation is only present during the procedure and only in the region of interest is thus to be preferred. This can be achieved by exciting target isotopes bound to biological markers by external x-rays and measuring their atomic signature, the fluorescence. For atoms with atomic number greater than about 50, the fluorescence energy is about 30keV, and thus is not completely absorbed in the human body. The challenge in detecting the x-ray fluorescence is to separate its signal from scattered x-rays, which are usually much more abundant. By an innovative design envisaged at our working group this task should become feasible (Figure 1). First simulations have shown that signalling atom concentrations similar to PET or SPECT can be detected with a resolution almost as high as for computed tomography, such that a double imaging-modality device is not required. The high sensitivity and resolution of the systems enables new and better diagnostics, and will also offer new possibilities in deciphering molecular processes in the development of diseases.



Müller,B.; Hoeschen,C; Grüner,F.; Arkadiev, V.A.; Johnson, T.R.C. Molecular imaging based on X-ray fluorescent high-Z tracers. Phys. Med. Biol. 58, 8063-8076 (2013)

Figure 1: Schematic view of x-ray fluorescence imaging setup. An x-ray beam (toward observer) excites fluorescent x-rays. Multiple scattered x-rays are absorbed by the collimator, while single scattered x-rays are filtered by the energy-sensitive analyzer crystal. Thus, only primary fluorescence x-rays are detected.

Figure 1: Schematic view of x-ray fluorescence imaging setup. An x-ray beam (toward observer) excites fluorescent x-rays. Multiple scattered x-rays are absorbed by the collimator, while single scattered x-rays are filtered by the energy-sensitive analyzer crystal. Thus, only primary fluorescence x-rays are detected.