About us

At CCIBonn.ai (Divison for Computational Radiology & Clinical AI, University Hospital Bonn / Faculty of Medicine at the University of Bonn, Germany), we focus on the development and application of artificial intelligence in medical imaging to enhance diagnostic accuracy, optimize clinical workflows, and enable large-scale data-driven research. Led by Prof. Dr. Philipp Vollmuth, Else Kröner CS Professor for AI in Medical Imaging (Faculty of Medicine at the University of Bonn), our multidisciplinary team combines expertise from AI researchers, data scientists and clinicians to bridge the gap between computational research and clinical application. Our research includes foundational AI models for radiology, privacy-preserving federated learning, the use of synthetic data to overcome limitations in medical datasets, and the integration of large language models into healthcare workflows.

Infrastructure

We operate state-of-the-art infrastructure, including multiple NVIDIA DGX H200 servers, develop and operate unique radiology research infrastructure like ADIT and RADIS that enables large-scale data mining and analysis, and have through our affiliation with the Clinic for Neuroradiology access to cutting-edge imaging systems such as 0.064T, 3T and 7T MRI scanners. We are key contributors to the Human Radiome Project, the MICCAI Brain Tumor Segmentation (BraTS) challenge and the Federated Tumor Segmentation (FeTS) initiative. Moreover, collaborations with leading institutions including the German Cancer Research Center (DFKZ), the German Center for Neurodegenerative Diseases (DZNE), and the European Organization for Research and Treatment of Cancer (EORTC) ensure the clinical and methodological rigor of our work.

Funding

We have secured >4.5 million EUR in funding and we are currently supported by the European Research Council (ERC Consolidator Grant for "AI-Next"), the German Research Foundation (DFG - Priority Program 2177: Radiomics - Next Generation of Biomedical Imaging), BONFOR, and the Else Kröner Fresenius Foundation among others.

Publications

With publications in journals such as Lancet Oncology, Nature Communications, and Lancet Digital Health, we are committed to translating computational research into clinically impactful solutions for precision medicine and radiology. Full list of publications is available here.

Open Science

We prioritize open science and share our research through our GitHub organizations CCIBonn and OpenRadX. Notable repositories include HD-BET and HD-GLIO, which have become widely used tools in medical image processing. Additionally, our team is developing ADIT (Automated DICOM Transfer) and RADIS (Radiology Report Archive and Discovery System), which provide scalable infrastructure for accessing and mining previously siloed radiological data:

• ADIT streamlines DICOM data management with features such as secure transfers, on-the-fly pseudonymization, trial name tagging, and batch processing. It allows users to schedule transfers, manage permissions, and operate through an intuitive web interface or REST API, ensuring efficient, secure, and compliant workflows for both clinical and research applications.
• RADIS is an open-source web application for archiving, querying, and discovering radiology reports. It integrates advanced semantic and keyword search capabilities, enhanced by large language models (LLMs), enabling complex queries, report classification, and organization. This significantly improves search accuracy and facilitates nuanced question answering. Combining LLM capabilities with secure, local deployment, RADIS enables large-scale data mining for both clinical and research applications.