Real-world laboratory for data intelligence

As part of the Fraunhofer CIMD, two real-world laboratories are to be set up in coordination with each other, one for the “4D Clinic” and the present real-world laboratory “Data Intelligence”. The two real-world laboratories will be developed separately but in close coordination with each other. The real-world laboratory “Data Intelligence” focuses in particular on the structure and functionality of a data platform with a self-service focus for the needs of the participating Fraunhofer CIMD institutes. In the future, the platform is to be used and expanded in various innovative cooperation projects.

A wide range of systems are used in diagnostics. The necessary knowledge for using them often has to be acquired autodidactically. Centralized data storage is often not available, meaning that data has to be transferred manually. Simple tools such as Excel are often used first to analyze (digital) documents. Python and R are usually used for statistical tests. The real-world laboratory for data intelligence aims to put this on a new and integrated footing.

After an intensive requirements analysis with the core institutes, a future-proof data management and a cloud-capable and secure data storage, which is still hosted locally in the first step, will be designed and set up. This will be followed by the development of a visual-interactive analysis platform with subject-specific visualization and analysis methods.

The data analysis platform supports the generation of hypotheses and the selection of cohorts for clinical studies

The overarching goal is to develop a data management and an extensible data analysis platform that can be used independently by medical professionals. The Fraunhofer Institute for Computer Graphics Research IGD and the Fraunhofer Institute for Intelligent Analysis and Information Systems IAIS are providing support and assistance. At the end of the project, Fraunhofer CIMD will have a cloud- and AI-enabled platform that is also available for use and expansion in various innovative collaborative projects.  

The platform enables the discussion and adaptation of new hypotheses in direct dialog between analysts and physicians. It also offers the possibility to analyze the distribution of cohorts and to assess whether the database is sufficient for statistically sound statements in clinical studies.

As a technical partner institute, the Fraunhofer Institute for Computer Graphics Research IGD provides its expertise in the sections of visual analytics, dashboarding in the medical field, and user-centered design. These are complemented by the know-how of the Fraunhofer Institute for Intelligent Analysis and Information Systems IAIS in the field of cloud-based techniques and machine learning.

Outlook

A first usable version of the analysis platform for daily, self-sufficient work with study data should be available by 2024. The platform will then be successively expanded and extended.

Further research is needed with regard to the flexible connection of data sources, secure and usage-related access concepts for AI and cloud readiness of data management, as well as image-based analysis methods.