Enabling AI-Based Prediction of Neoadjuvant Treatment Response: A FAIR Multimodal Dataset Within EUCAIM.

This work presents a FAIR-compliant, multimodal PET-CT and clinical dataset developed within the EUCAIM framework to support future AI-based prediction of response to NST in LABC patients. A real-world dataset was curated, harmonized, and integrated into the EUCAIM CDM within a federated infrastructure. While predictive modeling is ongoing, this study focuses on data preparation and infrastructure, key bottlenecks for AI development, demonstrating the feasibility of integrating local hospital data into a European federated ecosystem to enable future multicentric AI applications.