Refrigid AI project decodes cancer metastasis in Mainz!

Refrigid AI project decodes cancer metastasis in Mainz!

Mainz, Deutschland - A newly launched research project called Decipher-M aims to decrypt the mechanisms of cancer metastasis using artificial intelligence. On March 1, 2025, the project started working to gain fundamental insights into the creation and spread of metastases that represent a main cause of death for cancer patients. The project is made by an interdisciplinary team under the direction of Prof. Dr. med. Jakob N. Kather at Else Kröner Fresenius Center for Digital Health.

Decipher-M combines research institutions from various German cities, including the University Hospital RWTH Aachen, the University Hospital Carl Gustav Carus Dresden, the University Hospital Essen, the German Cancer Research Center Heidelberg, the University Medical Center Mainz and the Helmholtz Zentrum Munich and the Technical University of Munich. Overall, funding amounts of around 9 million euros are available for the five -year term from 2025 to 2030, of which around 5.5 million euros are provided during the first three years by the Federal Ministry of Education and Research.

goals of the project

The project has set itself the goal of using multimodal foundation models of artificial intelligence to analyze various data types such as radiological images, pathological reports and genetic information. A central aspect of research is to understand mechanisms of the occurrence of metastases, to predict potential development locations and opportunities and to evaluate the effectiveness of treatments for different patients.

The identification of characteristics and signatures associated with metastasis and forecast are to be developed. These tools will serve to adapt the screening and treatment procedures in the event of a high metastasis risk and to predict the most effective therapeutic approaches for patients with metastatic disease.

technological approaches

The use of modern AI technologies enables researchers to identify complex patterns in clinical routine data. An important part of the project are approaches that link radiology, pathology, text reports and electronic health records. The development of a so -called "Cancer Foundation Model", led by Julia Schnabel, director of the Institute of Machine Learning in Biomedical Imaging at Helmholtz Munich, plays a key role. This model combines Vision Transformers and Large Language Models for analyzing image and text data.

The Helmholtz Zentrum Munich receives around 544,000 euros for the first funding period to further promote research efforts in the field of cancer metastasis. The DECIPHER-M project is supplemented by the experts in Mainz by analyzing university and multimodal models in order to ensure comprehensive validation of the developed risk models.

For more information about_decipher-m_ Visit the websites of Health Research BMBF and Helmholtz Munich