The ProCAncer-I together with the other AI4HI funded projects organised a workshop with the title “Developing open, standard-based, interoperable Cancer Imaging Repositories in Europe: Issues, Experiences and Challenges” within the IEEE BHI-BHS Conference in Ioannina (27-30 September 2022)
Some of the topics covered:
Strategic and Methodological issues related to AI in Health Imaging
- FUTURE-AI guidelines for trustworthy and deployable AI
- The concept of a Model’s Passport
- AI explainability paradigms
- Federated learning & Inference Services
- Approaches for AI model development in health imaging and initial results
- AI approaches for image harmonization
Ethical/Legal
- GDPR compliance and governance models
- Ethical and legal aspects from technology point of view
- Ethical and legal aspects from technology point of view
Project data sharing, policies, approaches, challenges Technical
- Data annotation & image segmentation
- Data interoperability through common data models
- Common data models for non-imaging data
- Extending OMOP-CDM for imaging data
- Infrastructures for storage and management of health imaging data.
The AI4HI Network includes five large EU-funded projects on big data and AI in cancer imaging (CHAIMELEON, EUCANIMAGE, INCISIVE, ProCΑncer-I, PRIMAGE) and has been organised into 8 working groups (Ethical and legal issues, Metadata interoperability, Data storage and management, Data annotation, AI development, AI validation, Clinical Working Group and Outreach Working Group), each consisting of 15 experts representing the five projects and a wide range of stakeholders, perspectives, approaches and disciplines. This workshop focused on presenting the results delivered by these working groups analyzing the existing landscape, solutions and challenges based on a concrete set of clinical use cases related to a number of cancer types (lung, breast, liver, colorectal, prostate, brain, etc).
1
2
