Our vision is to enable a wider adoption of fully-quantitative molecular image information in the context of personalized medicine.
Our mission is the advancement of basic correction and quality measures to facilitate the extraction of information and knowledge from clinical images and non-imaging information. We engage in translational research through local and international collaborations and source knowledge from multiple disciplines, including medicine, pharmacology and radiochemistry. We believe in a synergistic ‘Physician + artificial intelligence’ approach to further increase our understanding of diseases and help support clinical decision making.
Hybrid Imaging Technology | Data Analytics | AI and quantum computing
This research topic is dedicated to all aspects of hybrid imaging technology. In this field, we engage in performance evaluation and quality control improvements, Monte–Carlo simulations of (pre) clinical PET and SPECT systems, reconstruction algorithms and advanced data correction approaches such as partial volume and positron range corrections. We work in close collaboration with the Division of Nuclear Medicine and have access to clinical and preclinical SPECT/CT, PET/CT and PET/MRI devices including access to the on side radio-pharmaceutical production side equipped with a cyclotron and generators for 68Ge and 99mTc.
In this research topic, we focus on harnessing the true potential of the hybrid imaging systems (PET/CT and PET/ MR) by utilizing the synergistic anatomical and functional information obtained during a single imaging session. In particular we develop methods to enable non-invasive absolute quantification and multi-parametric imaging in clinical routine. To promote transparency and reproducibility, all the source codes developed during our research endeavors are open source, routinely curated in GitHub and are available for free.
This research topic focuses on the combined analysis of imaging and non-imaging medical data with artificial intelligence (AI) approaches to establish highly-performing, repeatable predictive models. The utilized approaches cover the range of quantitative radiomics, deep learning, AI-driven data preparation and pre-processing as well as mixed ensemble learning (a.k.a. super learners). In addition, this topic also includes applied quantum computing research activities with the focus on optimizing modeling, simulation, imaging and AI processes.
Our group brings together physicists, computer scientists and engineers in an effort to establish and validate clinical and research protocols using quantitative hybrid imaging technologies, such as PET/CT, PET/MRI and SPECT/CT.
Hybrid imaging modalities are considered hardware combinations of nuclear medicine and radiology imaging systems aiming at: