Image Data Platform (RACOON)

The RACOON platform represents a significant advancement in the research infrastructure for medical image-based studies. It is meticulously designed as a national research platform that establishes and maintains a comprehensive ecosystem tailored for modern image-based medical research projects. RACOON prioritizes user-friendliness and fosters a collaborative environment by integrating all radiology departments within the German Network of University Medicine as partner sites in strict compliance with data protection regulations. Its extensible, modular system supports multiple individual research projects (subprojects). It preserves the workflows and methodologies they establish, enhancing its functionality for future endeavors.

Goals and Challenges

The central vision of RACOON is to develop a comprehensive, collaborative platform for medical image-based research that facilitates radiologic studies intuitively and user-friendly. By incorporating all radiology departments within the German Network of University Medicine as partner sites, RACOON aims to provide a complete, modular ecosystem for research projects, adhering to stringent data protection regulations.

However, several challenges must be addressed to ensure seamless integration of infrastructure components while simultaneously expanding functionality. Establishing a regulatory framework for legally compliant, transnational research is imperative. Additionally, security vulnerabilities must be identified and mitigated to ensure maximum data protection. Standardized processes must be implemented to maintain high data quality. The functionality of infrastructure modules needs to be adapted to meet the diverse requirements of various research projects. Furthermore, the scalability of research studies, from single-center designs to multi-center projects, must be considered, especially given the typically incompatible data storage systems of different research partners.

Implementation and Overview

RACOON is developing a platform for medical image-based research in collaboration with renowned technology partners, university working groups, cross-site computer scientists, medical professionals, and researchers. The project develops and continuously updates a regulatory framework for RACOON subprojects, promoting nationwide collaboration among partner networks. It is founded on interoperable interface and data standards, facilitating sustainable use of the infrastructure and its components. RACOON relies on machine-readable and interoperable data formats for research datasets.

A cornerstone of RACOON is developing an information model that considers existing data models and involves top-tier experts. Developers are deployed to various sites and connected with the development teams of technology partners and users in cross-site, interdisciplinary expert groups.

RACOON’s modular design allows for the continuous integration of new infrastructure modules, ensuring that the system can map the entire lifecycle of research projects with multimodal datasets. Process definitions, interactive quality management modules, standard operating procedures (SOPs), and quality standards are coordinated and provided through a unified project management environment.

An open-source software environment and a robust server network enable the development and deployment of artificial intelligence (AI) within RACOON. Monitoring modules support the oversight of network functionality and the progress of research projects, which is essential for controlling network operations and tracking project advancements.

To ensure the highest level of security, dedicated efforts are made to establish security processes, including independent security audits and penetration tests.

RACOON has successfully implemented a hardware and software infrastructure with network nodes at all German university hospitals. This architecture combines decentralized components with a secure central environment (RACOON-CENTRAL), forming a resilient overall infrastructure. Research projects benefit from rapid, secure commissioning and seamless scalability across multiple locations. RACOON-CENTRAL also encompasses the functionalities of the decentralized instances, such as structured reporting, image annotation, segmentation, training or inference of AI, and central monitoring of the individual RACOON nodes. These components create a trustworthy, cross-partner research environment, continually enhanced through generic method packages and specific workflows developed within RACOON research applications. Meeting all applicable requirements (medical device class IIb), the platform supports routine clinical use for image diagnosis and analysis, including cross-site consultation support.

NUM Geschäftsstelle TUM Medizin
 
Research Team
Prof. Marcus R. Makowski
Prof. Marcus R. Makowski
Prof. Dr. Rickmer Braren
Prof. Dr. Rickmer Braren