European Commission launches European Cancer Imaging Initiative to federate cancer imaging data

Brussels, January 23rd 2023

Summary

›The European Commission launched the European Cancer Imaging Initiative on 23 January 2023 under Europe's Beating Cancer Plan.
›The initiative will build a federated digital infrastructure to link cancer imaging datasets across the EU while aiming to respect GDPR and high ethical standards.
›EUCAIM is the flagship project funded by the Digital Europe Programme with €18 million in co-funding and targets more than 100,000 cancer cases and 60 million annotated images by 2025.
›A testing and experimentation facility called TEF-Health will let SMEs test AI tools in real clinical settings and European Digital Innovation Hubs will support rollout.
›Official timelines foresee a platform design by December 2023, a first release end 2024, final release end 2025 and full operation in 2026.
›The initiative faces familiar challenges including data quality, interoperability, regulatory alignment with the European Health Data Space and long term sustainability.

What the European Cancer Imaging Initiative is and why it matters

On 23 January 2023 the European Commission announced the launch of the European Cancer Imaging Initiative. This action is a flagship under Europe's Beating Cancer Plan and aims to create a federated, cross-border digital infrastructure that connects cancer imaging data repositories across the European Union. The stated goals are to accelerate research, enable development and clinical validation of artificial intelligence driven tools, and ultimately improve diagnosis, personalised treatment and predictive medicine for cancer patients.

The initiative sits at the intersection of the European Data Strategy, the emerging European Health Data Space and EU efforts to scale up AI in healthcare. It is explicitly framed around two priorities. First, technological innovation that leverages large, harmonised datasets. Second, adherence to privacy, ethical and legal frameworks so that clinicians, researchers, innovators and patients can trust the system.

Core functions and promises

At a practical level the initiative seeks to do three things. It will connect scattered cancer imaging datasets into an interoperable, privacy preserving infrastructure. It will provide testing and benchmarking facilities so developers can validate AI solutions in realistic conditions. It will support research collaborations and multi-centre studies that are difficult to run when data are siloed.

Federated infrastructure explained:A federated system lets data stay with the original data holders while enabling algorithms to run across distributed datasets. This reduces the need to move sensitive patient level images and records across borders. Federated approaches typically rely on standardised metadata, harmonised data formats and secure computation methods.

Expected benefits for research and care:Researchers will gain access to larger and more diverse imaging sets which can improve statistical power and reduce bias. Innovators can benchmark and validate AI models at scale. Clinicians may see faster and more precise diagnostic support tools once clinical validation is complete. The initiative also aims to make it easier for citizens to contribute data voluntarily under data altruism arrangements.

Projects, funding and the technical backbone

The EU launched the initiative alongside two projects that provide the operational backbone. The EUCAIM project is the primary technical effort building the Cancer Image Europe platform. It is supported with €18 million of EU co-funding from the Digital Europe Programme. The AI Testing and Experimentation Facility for Health or TEF-Health is positioned to host experimentation environments where SMEs and other developers can trial AI tools against federated datasets.

EUCAIM and the AI4HI network:EUCAIM builds on previous Horizon 2020 work known as the AI for Health Imaging network, a cluster of five projects including Chaimeleon, EuCanImage, ProCancer-I, Incisive and Primage. These projects helped define standards for image annotations, data harmonisation and proof of concept federation. EUCAIM brings together about 95 consortium members from 17 countries to scale those efforts into a pan-European federated platform.

TEF-Health explained:TEF-Health is a Digital Europe Programme initiative to create testing and experimentation environments. It will allow small and medium sized enterprises to test AI solutions in representative clinical settings, which is a key step before clinical deployment. The European Digital Innovation Hubs will help innovators navigate legal and technical requirements and provide services like test-before-invest, training and access to finance.

Project / Mechanism	EU co-funding	Role and targets
EUCAIM (under Digital Europe)	€18 million	Develop the Cancer Image Europe federated platform. Target to include over 100,000 cancer cases and 60 million annotated images by 2025.
TEF-Health (Digital Europe)	Not specified in launch text	Testing and experimentation facility to validate AI tools in real care environments.
UNICA (EU4Health)	€3.9 million	Extend federated infrastructure with breast, lung and prostate screening imaging data across 12 centres including countries outside the EU.
BreastScan (EU4Health)	€3.6 million	Build a large scale breast screening dataset with 14 data holders across 9 EU countries to test AI tools for screening.

Milestones and delivery timeline

The Commission and project partners published a staged timeline for delivery. The plan aims to move from design and federated proof of concept to operational testing and full roll out within a three year window. That timetable creates pressure to align technical, legal and organisational elements across many countries and institutions.

Date	Milestone
December 2023	Design of the pan European digital infrastructure completed and collaboration mechanisms established
End of 2024	First version of the platform released
End of 2025	Final release of the platform expected
2026	Platform fully operational and running
September 2025	Reported platform status: 83 imaging datasets, nine cancer types, approximately 107,000 subjects and 50 AI tools available to users

Technical and regulatory design choices

The initiative emphasises interoperability, data harmonisation, privacy preservation and standardised annotation. It also aims to align with the European Health Data Space which sets rules for use of health data across the EU. Federated learning and distributed analysis will be central because they mitigate some legal risks associated with moving patient level data across borders.

Federated learning and benchmarking:Federated learning trains models by sending algorithms to local datastores rather than centralising data. This method can preserve data sovereignty and privacy but depends on consistent data schemas, high quality annotations and robust security. Benchmarking frameworks are needed to compare model performance across heterogeneous datasets.

Progress reported so far and network growth

By September 2025 EUCAIM reported that the infrastructure connected 83 imaging datasets across nine cancer types representing roughly 107,000 subjects. The experimentation platform reportedly hosted 50 AI tools and had 203 registered users from 16 countries. In January 2025 the EUCAIM consortium expanded to include partners from Estonia, Norway and Latvia. Additional EU4Health projects such as UNICA and BreastScan aim to supply screening imaging data to the federation.

Opportunities for industry and SMEs

The Commission frames TEF-Health and the European Digital Innovation Hubs as routes for companies, particularly SMEs and start ups, to access real world data and to test AI solutions under supervised conditions. The hope is that easier access to high quality datasets and testing capabilities will reduce barriers to market entry and accelerate clinical adoption of validated tools.

Risks, practical challenges and open questions

The initiative promises significant potential but it faces measurable challenges. Technical difficulties include harmonising imaging modalities and annotation standards across different hospitals. Data quality issues and annotation costs are substantial and can affect model generalisability. The federated approach reduces but does not eliminate legal complexity in cross border data processing. Alignment with the European Health Data Space remains work in progress and regulatory uncertainty could delay uptake.

Ethics, consent and data altruism:The initiative promotes voluntary data altruism as a means for citizens to allow their data to be used for research. In practice this raises questions about informed consent models, opt out mechanisms, the representativeness of altruistic pools and potential secondary uses of data. These are not purely technical issues and require transparent governance that maintains public trust.

Other unresolved questions include long term governance, who will sustain the platform outside EU project funding, how commercial and public interests will be balanced, and how clinical workflows will adopt AI tools in practice. Clinical validation and regulatory approval for AI based devices are time consuming and require carefully designed multi centre studies which the platform aims to support.

How this fits into the EU innovation ecosystem

The European Cancer Imaging Initiative is an example of the EU attempting to combine policy, regulation and targeted funding to create a data infrastructure that can underpin innovation. It draws funding and technical lineage from Horizon 2020, Digital Europe and EU4Health. If implemented well it could become a reusable model for other disease areas. However success depends on delivering sustained funding, resolving legal and governance issues and demonstrating clear clinical benefit that leads to adoption by hospitals and health systems across Member States.

What to watch next

Key indicators of progress will be whether the platform meets its dataset and image count targets, how many distributed data holders connect, the number and quality of multi centre clinical validation studies, and whether TEF-Health experiments lead to regulatory submissions or real world deployments. Observers should also monitor how the initiative aligns with the final European Health Data Space rules and how governance arrangements for data use and sustainability are resolved.

Quick reference: timeline and targets

Target	Date or milestone
Design of pan European infrastructure completed	December 2023
First platform release	End of 2024
Final platform release	End of 2025
Full operation	2026
Dataset ambition	More than 100,000 cancer cases and at least 60 million annotated images by 2025
Initial clinical sites	21 clinical sites from 12 countries at project start, aiming at least 30 data providers from 15 countries

The European Cancer Imaging Initiative seeks to create a valuable shared research and testing asset. The technical and organisational building blocks are in motion. Delivering on the stated clinical and innovation promises will require continued attention to data quality, legal clarity, public trust and sustainable governance. Observers and stakeholders should treat the initiative as a long term infrastructure effort rather than a quick fix for clinical AI adoption.

EIC Corporate Day with Vattenfall: 12 deep tech innovators pitched grid, heating and wind solutions to a major European energy group

2024-12-12Read Now