CHAIMELEON Project Wraps with Major Breakthroughs in Artificial Intelligence for Cancer Imaging

After four years of intensive collaboration, the EU-funded CHAIMELEON project concludes with transformative achievements that are set to reshape how artificial intelligence (AI) is applied in precise cancer diagnosis and treatment. The project successfully established a secure, interoperable health imaging repository that supports the development, validation, and benchmarking of clinical AI tools across Europe.

From large-scale data harmonisation to reproducible AI development pipelines, CHAIMELEON sets a new standard for collaborative, clinically relevant innovation in the digital health domain.

Key Achievements

  • Pan-European cancer imaging repository: CHAIMELEON compiled and harmonised imaging data from over 40,000 cancer cases (lung, breast, colorectal, and prostate), representing more than 20 million anonymised images collected from eight leading clinical centres across Europe.
  • Fully operational AI development platform: A cloud-based infrastructure was created for secure ingestion, annotation, harmonisation and management of imaging datasets, enabling scalable training and evaluation of AI models.
  • Strong clinical and policy engagement: Project outcomes were showcased at major events including the European Congress of Radiology 2025, with outreach to healthcare professionals, policymakers, and researchers.
  • Privacy and interoperability by design: The repository adheres to FAIR principles and demonstrates robust data harmonisation across institutions, scanners, and modalities, while ensuring pseudonymisation and GDPR compliance.

Benchmarking Outcomes

  • Robust validation environment: Internal testing and open benchmarking challenges were used to evaluate AI tools, providing real-world performance insights.
  • Scientific and academic output: Included a total of 54 journal articles; 16 direct and 38 indirect outcomes, with thousands of views and downloads
  • Effective dissemination strategy: Project newsletters achieved open rates of 30%, while coordinated communications activities reached wide audiences through social media, webinars and international conferences.
  • Key performance indicators met or exceeded: Clinical data and images targets, stakeholder engagement goals, and platform readiness milestones were all successfully achieved.

Clinical Relevance

The CHAIMELEON repository is designed to support the development of AI tools for cancer characterisation, staging, treatment planning and prognosis. By enabling multi-centre, real-world validation of algorithms, the project enhances the reliability and clinical trustworthiness of AI in oncology. Its observational study design ensures tools are tested in realistic clinical settings, helping pave the way for future adoption and regulatory approval.

Sustainability and Next Steps

The CHAIMELEON platform is built for long-term use beyond the project’s conclusion:

  • The repository and infrastructure remain available for external users, including clinical researchers, data scientists and industry partners.
  • Sustained alignment with EU policy priorities through participation in EUCAIM, AI4HI, and other European health data initiatives.
  • Plans to expand the repository to cover additional cancer types and data modalities, ensuring the platform continues to grow in relevance and utility.

Knowledge and Lessons Learned

CHAIMELEON leaves behind a legacy of technical excellence and strategic insight:

  • Quality over quantity: Emphasis on data standardisation, harmonisation and annotation quality ensured AI tools could be meaningfully evaluated.
  • Inclusive communication matters: Visual resources, lay-friendly materials, and multilingual outreach increased visibility across diverse audiences.
  • Collaboration is key: Strong engagement through open AI challenges, policy dialogue, and shared infrastructure fostered trust and uptake among stakeholders.

A Lasting Impact on European Health AI

CHAIMELEON has demonstrated that responsible, large-scale secure data sharing can support safe and effective AI solutions for cancer care. Its platform and processes are already serving as a model for future European Health Data Space initiatives and cross-border research infrastructure.