• Delegation from Kiel University opens the first express connection for artificial intelligence (AI) between Kiel and San Francisco (UCSF) in San Francisco together with Minister President Daniel Günther.
• AI networks can thus be trained jointly using patient data from the university hospitals in San Francisco and Kiel
• First application: AI should be able to use X-ray images to predict bone fractures or recommend the right treatment at the first signs of a stroke

Looking into the future becomes the present: using simple X-ray images, an artificial intelligence (AI) should be able to recognize whether or not a hip fracture is likely to occur in the next ten years. What recently sounded like science fiction is now reality and only part of the cooperation between Kiel University (CAU) and the University Medical Center Schleswig-Holstein (UKSH) with the University of California, San Francisco (UCSF). On June 7, 2023 at 3:00 p.m. local time (24:00 German time), a CAU delegation will open the new infrastructure for AI in the USA together with Schleswig-Holstein's Minister President Daniel Günther. CAU President Prof. Dr. Simone Fulda will join the event live from Kiel at night and reaffirm CAU's commitment to this cooperation.

Data from both locations converge

The "AI Exchange" project between the "Intelligent Imaging Lab" at the CAU and the "Center for Intelligent Imaging" at UCSF uses "Federated Learning" technology, which protects sensitive data such as X-ray images in particular. "With this technology, all sensitive medical data remains on site, at the UKSH or at UCSF, so that data protection is guaranteed," says Prof. Dr. Claus-Christian Glüer, who heads the project. Together with Prof. Dr. Jan-Bernd Hövener, he set up the Intelligent Imaging Lab and the Biomedical Imaging Section in Radiology. "Instead of sending the data back and forth, we train the networks at each location within the respective firewalls," explains Hövener. "After a certain period of time, we combine the results of the local networks so that we can take advantage of the large data volumes at both sites and protect our data."

The AI networks at both sites are initially trained independently at both sites, "they gain practical experience," says Hövener. These are sent to a central server at the University Medical Center Schleswig-Holstein (UKSH) and merged. This results in a joint new network that is more experienced than the individual networks. This is then sent back to the local training locations again and again until it is optimally trained. "The more data flows into an AI network, the more precisely and accurately it can work later and help people," says Hövener. CAU President Simone Fulda emphasizes: "With this cooperation, the CAU can further strengthen its role as a central and internationally networked scientific player in Schleswig-Holstein. And this project in particular impressively demonstrates the added value of international cooperation for the use of data-driven AI technologies in medicine."

Hit rate to be increased

 One initial application will be the prediction of hip fractures based on X-ray images. It is not possible for people to accurately predict whether a fracture will occur within the next ten years based on simple images. The AI at the CAU currently achieves a prediction accuracy that is better than that of current screening tests. Together with the data from UCSF, the aim is to further improve the accuracy rate. Further projects are also planned in the short and medium term. One example is the automated detection of the cause of strokes in the emergency room: is there a hemorrhage that needs to be stopped or does a blocked artery need to be unblocked? It is vital and urgent for patients to receive the right treatment for these extremely different conditions. AI can help doctors to ensure that nothing is overlooked.