CHAIMELEON is a project aimed at establishing an EU-wide repository for health imaging data to enhance AI-driven cancer management research. This repository will adhere to the legal and ethical standards of the involved countries and build on existing data infrastructures such as the PRIMAGE repository for pediatric cancer, the Euro-BioImaging node, the Radiomics Imaging Archive by Maastricht University, DRIM AI France, and the Oncology imaging biobank by Pisa University. The repository will gather multimodality imaging and associated clinical data for lung, prostate, colon, and rectal cancers from both historical and newly diagnosed patients. An analytical engine will facilitate the extraction and interpretation of this data, supporting AI pipelines that automate data de-identification, curation, annotation, and harmonization—critical for ensuring reproducibility across large, diverse datasets. The repository’s effectiveness as a tool for AI experimentation will be validated through open challenges to the AI community and in-silico validation in observational studies across European hospitals, focusing on key clinical decisions related to major cancer types.