How the First Medical Imaging Cancer Atlas EUCAIM Was Populated: The Experience of a Reference Hospital

Area of Science:

• Oncology
• Medical Imaging
• Artificial Intelligence
• Data Science

Background:

• Fragmentation and decentralization of medical data, particularly radiological images, hinder large-scale observational research in Europe.
• Lack of findable, accessible, and interoperable datasets limits the development, validation, and clinical translation of Artificial Intelligence (AI) models for precision medicine.
• The European Federation for Cancer Images (EUCAIM) project was initiated to address these challenges in oncological imaging data.

Purpose of the Study:

• To establish a secure, centralized, and federated infrastructure for the secondary use of large-scale oncological imaging and clinical data.
• To facilitate harmonized data governance and trusted cross-border sharing of fragmented datasets.
• To present the practical experience of integrating imaging and clinical data into the EUCAIM infrastructure, detailing challenges and solutions for regulatory compliance.

Main Methods:

• Consolidation of fragmented datasets into a harmonized infrastructure.
• Implementation of a robust documentation framework for regulatory compliance and data integrity.
• Integration of imaging and clinical data from a reference university hospital into the EUCAIM infrastructure, adhering to privacy, security, and ethical standards.

Main Results:

• Establishment of a secure infrastructure enabling the secondary use of large-scale oncological imaging and clinical data.
• Demonstration of strategies for procedural, ethical, and legal compliance with data protection regulations (GDPR, EHDSR).
• Successful integration of data from a university hospital, showcasing a practical framework for future harmonization and AI development.

Conclusions:

• The EUCAIM project provides a foundational framework for scalable, reproducible, and legally compliant research in AI-driven oncology.
• The project strengthens Europe's capacity for trustworthy AI solutions by addressing data accessibility and interoperability challenges.
• Insights gained offer a practical model for integrating diverse datasets into a federated infrastructure, promoting cross-border collaboration and advancing precision cancer care.