Researchers from the Faculty of Medicine at the University of Duisburg-Essen (UDE), Ludwig-Maximilians-Universität München (LMU), and the Berlin Institute for the Foundations of Learning and Data (BIFOLD) have developed an innovative approach to personalized medicine using artificial intelligence (AI). This method integrates diverse data types, including medical history, laboratory values, imaging, and genetic analyses, to enhance clinical decision-making.
Current oncological practices rely on rigid assessment systems, which often overlook individual patient differences such as sex, nutritional status, and comorbidities. Prof. Frederick Klauschen, Director of the Institute of Pathology at LMU, emphasizes that modern AI technologies, particularly explainable artificial intelligence (xAI), can unravel complex interrelationships to personalize cancer treatment more effectively.
The study, published in Nature Cancer, utilized data from over 15,000 patients with 38 different solid tumors. Researchers examined the interaction of 350 parameters, including clinical data and genetic tumor profiles. Dr. Julius Keyl, Clinician Scientist at IKIM, noted that they identified key factors influencing decision-making processes within the neural network.
The AI model was validated using data from over 3,000 lung cancer patients, providing transparent prognoses by illustrating how each parameter influenced outcomes. Dr. Philipp Keyl remarked on the potential of AI to analyze clinical data in context, thereby facilitating personalized, data-driven cancer therapy.
The research team aims to explore complex interrelationships across various cancers, a task challenging for conventional statistical methods. Prof. Martin Schuler, Managing Director of the National Center for Tumor Diseases (NCT), highlighted the potential for clinical trials to demonstrate the real patient benefits of their technology.