AI revolution in oncology: Dresden researchers develop autonomous system

AI revolution in oncology: Dresden researchers develop autonomous system

Technische Universität Dresden, 01062 Dresden, Deutschland - A research team at the Else Kröner Fresenius Center (EKFZ) for digital health at the Technical University of Dresden (TUD) has developed a groundbreaking autonomous AI agent who is intended to support clinical decision-making in oncology. This innovative approach could revolutionize the way in which medical specialists analyze complex data and creates personalized treatment plans. The results of this study were published in the renowned journal Nature Cancer , which underlines the importance of this development. With their system, the researchers would like to reduce the challenges associated with the evaluation of diverse data sources, including medical imaging and genetic information that is crucial for the treatment of cancer patients.

The AI ​​agent is based on the powerful language model GPT-4 and is equipped with various digital tools. This includes the creation of radiologists from MRI and CT scans, medical image analysis and the prediction of genetic changes from histopathological tissue. In addition, the agent has access to around 6800 documents from official oncological guidelines to ensure that the clinical decisions are based on a solid knowledge basis. In test runs on 20 simulated patient cases, the system achieved correct clinical conclusions in 91 percent of cases and correctly cited relevant oncological guidelines in more than 75 percent of the cases. Despite these encouraging results, restrictions were also found, since the system was only tested on a limited number of cases, which requires further validation.

integration and challenges

In the discussion about the future integration of the AI ​​system, it was noted that the focus should be on the development of conversation skills and data protection-compliant applications. The challenges of interoperability with existing systems and the observance of data protection requirements as well as admission procedures must also be prioritized. Professor Jakob N. Kather from the TU Dresden emphasizes the need to use AI responsibly and to define ethical and safety -relevant framework conditions. In the long term, the use of AI agents could also be expanded in other medical fields, provided that they are equipped with suitable tools and comprehensive data.

The research team emphasizes how crucial the further training of doctors is in order to effectively work with these advanced technologies. It must be ensured that the medical staff retains the decision to make decisions, even if AI agents are increasingly integrated into decision-making processes. These developments could not only increase the speed and efficiency of clinical decision -making, but also transform the research landscape in cancer research. The integration of autonomous AI models could help to accelerate the entire development process in cancer research, including literature research, project planning, medication modeling and design clinical studies.

The way to tomorrow's cancer research

Progress in biomedical research and the use of AI go hand in hand. While traditional AI applications are often limited to specific tasks, newly developed AI agents enable automation of multi-stage work processes, which promotes more efficient cooperation between researchers and AI systems. Previous AI systems were able to read hundreds of publications and process complex data in a very short time. These technologies could give scientists: more valuable time in order to turn more creative and strategy aspects of research.