AI diagnostic support for rare diseases
Diagnosing rare diseases often presents doctors with challenges. Digitalization can help to make a diagnosis - Frankfurt University Hospital is showing how: It is playing a leading role in the research project Smart Physician Portal for People Affected by Rare Diseases (SATURN). Funded by the German Federal Ministry of Health, a platform is being created that uses artificial intelligence (AI) to help with diagnosis. As a software expert, Fraunhofer IESE is investigating in the project how AI can be used to make comprehensible and transparent suspected diagnoses for rare diseases with small amounts of data. A test version of SATURN is now available online - just in time for Rare Disease Day on February 29, 2024.

It is estimated that around four million people in Germany alone live with one of over 6,000 different rare diseases. For doctors, diagnosing those affected poses a major challenge, as rare diseases are a group of very different and complex clinical pictures, meaning that the actual disease is usually not recognizable at first glance. Another challenge is that there are only a few experts in Germany who can treat and research the respective rare diseases: The data situation is correspondingly thin. There are already the first digital solutions, such as diagnostic apps, which provide suspected diagnoses. However, these applications still leave local doctors alone with their patients. There is no support in planning further steps. This is where the SATURN project comes in.
Smart doctor's portal for patients with an unclear diagnosis
The smart doctor's portal for patients with an unclear diagnosis, SATURN for short, is a digital application to support diagnosis. “Our aim is to support doctors in primary care with a powerful tool that facilitates the diagnosis and treatment of rare diseases,” says Dr Michael von Wagner, Head of Medical Information Services and Digitalization at Frankfurt University Hospital. The portal uses AI to create diagnostic suggestions based on both the knowledge of experts and real clinical data. In addition, the Care Atlas for People with Rare Diseases - SE-ATLAS (www.se-atlas.de) is connected as a database of experts and self-help organizations for rare diseases. Scientists at Fraunhofer IESE are using various data sources to develop an AI system based on both expert knowledge and structured and unstructured data that can support the diagnosis of rare diseases.
In practice, SATURN is used as follows: The attending physician enters the patient's symptoms into the SATURN portal. These symptoms are compared with the clinical data stored in the portal. The result is a suggested diagnosis. In the case of common clinical pictures, additional instructions are given on how to proceed. The integrated SE-ATLAS can then be used to find and contact the right center or the right contact person for the respective diagnosis.
To ensure that the smart doctor portal also meets the requirements of everyday practice, various requirements analyses have been carried out with patients and practitioners during development to date. “By actively involving patients and doctors in the development process, we are ensuring that SATURN meets the real needs of the healthcare sector,” explains Dr. Jannik Schaaf from the Institute of Medical Informatics at Goethe University.
Funded by the Federal Ministry of Health
SATURN has received funding from the Federal Ministry of Health since January 1, 2022, which runs until December 31, 2024. The project partners are the Medical Information Systems and Digitization Unit at Frankfurt University Hospital, the Institute of Medical Informatics and the Institute of General Medicine at Goethe University Frankfurt, the Institute of Medical Informatics and Biometry at Dresden University of Technology and the Fraunhofer Institute for Experimental Software Engineering IESE in Kaiserslautern. The project milestones to date and the current test version of SATURN can be found at this link: www.saturn-projekt.de
For further information:
Dr. Michael von Wagner
Medical Director of the Medical Information Systems and Digitization department
Universitätsklinikum Frankfurt
E-Mail: michael.wagner@ukffm.de
Dr. Andreas Jedlitschka
Department Head Data Science
Fraunhofer IESE
E-Mail: andreas.jedlitschka@iese.fraunhofer.de
Last modified: