Deployment of cutting-edge multi-modal AI-based solutions in medical imaging

Expected Outcome

• New data services and user tools for the Cancer Image Europe platform covering data curation / inclusion by data providers, data quality assurance and compliance checks, and risk/security management in alignment with agreed standards, procedures and requirements, including the framework of the European Health Data Space.
• Multi-centre validation of AI-driven solutions in the field of medical imaging, including also evidence generation for clinical utility and cost efficiency.
• Deployment of multi-modal, cutting-edge AI-driven solutions in the field of medical imaging in healthcare and research settings, leveraging different types of data (at least one imaging modality and one other data type e.g. genomics, other omics, laboratory results, real-world data etc.) and building on the achievements of the Cancer Image Europe platform.
• Training and upskilling of medical imaging personnel and/or healthcare professionals for the deployed technology and further use.

Objective

• Accelerate the uptake of EU AI-driven solutions that are ready to be deployed in healthcare settings for patient care and for research purposes. This will facilitate the paradigm shift in the digital transformation of healthcare towards personalised medical solutions;
• Facilitate the deployment of EU cutting-edge AI-driven solutions in medical imaging, combined with other health data, for increased efficiency and better patient outcomes, leveraging the Cancer Image Europe platform;
• Further develop the data, testing and validation services and user tools of Cancer Image Europe platform in alignment with the legal and technical framework of the European Health Data Space, also towards supporting the development and uptake of EU cutting-edge multi-modal AI-based solutions in medical imaging (including Deep learning and Generative AI solutions) for healthcare;
• Ensure alignment and inter-operability of the Cancer Image Europe platform with the HealthData@EU infrastructure of the EHDS;
• Conduct multi-centre validation studies of promising medical imaging-based and multi-modal AI solutions for screening, early detection, diagnosis and care, generating evidence for clinical utility and cost effectiveness of tested solutions.

Scope

By building on, extending and leveraging the Cancer Image Europe platform, this action is expected to facilitate the uptake of EU AI-driven solutions in medical imaging (including Machine Learning and Generative AI), towards their deployment in clinical settings. It should also include upskilling of healthcare professionals and evidence generation to evaluate the performance of the deployed AI-driven solutions and engage patients.

The action should also contribute to upgrading and enhancing the uptake of the Cancer Image Europe platform through deployment of tools, application programming interfaces (APIs) and interfaces enabling high-quality data services for the users of the Cancer Image Europe platform. The solutions should cover, at least, data quality assessment, including security, data access and analysis, as well as services for AI developers supporting their regulatory compliance pathway, including the EHDS Regulation, if applicable. Moreover, data services will cover comprehensive data quality assessment - including security, data access, and analysis. Specifically, tools will be deployed to ensure that datasets federated in Cancer Image Europe are fully compliant with EHDS, utilizing Health DCAT-AP for metadata descriptions and adhering to the data quality and utility label standards set by the EHDS Regulation.

Moreover, the action should consider synergies and cross-fertilisation of tools with other relevant health data infrastructures: HealthData@EU, 1+Million Genomes (1+MG), Intensive Care Unit data infrastructure, the European Virtual Human Twins Initiative, and UNCAN.eu platform to further facilitate the integration and processing of different types of health data. Also, the outcomes of EU-funded projects dealing with other imaging modalities, such as digital pathology (e.g. BigPicture project), should be considered. Finally, synergies with the ongoing AI related activities of the Health Data Access Bodies (HDABs) in the EHDS, the SHAIPED project and the TEF-Health project should be leveraged.