AI for research of rare hematological diseases
The small number of patients with diseases considered rare results in a shortage or lack of data to study them in depth. Advancement in research can, however, be facilitated by artificial intelligence-based technologies that can create synthetic data.
SYNTHEMA aims to create a hub that can channel a large amount of data thus covering many variants and special cases in different conditions of rare hematological diseases, with the goal of improving and optimizing clinical/translational research and making a contribution to research.
Synthetic generation of hematological data over federated computing frameworks
Development of new data analysis systems in the field of hematological diseases
€ 7 mld
STRATEGY AND SOLUTION
To create a translational hub of health data, both clinical, therapy support data such as X-rays and imaging, and genomic data, for rare hematologic diseases (particularly Sickle Cell Disease and Acute Myeloid Leukemia). This framework aims to develop and validate innovative artificial intelligence-based techniques for clinical data anonymization and synthetic data generation to address data scarcity and fragmentation and expand know-how for research in compliance with GDPR.
The platform will be based on a privacy-preserving federated learning network, and equipped with SMPC (Secure Multi-Party Computation) and differential privacy (Differential Privacy) protocols, connecting healthcare data centers, academic research centers and industries.
Datawizard for SYNTHEMA
Data collection, harmonization, and interoperability for clinical use case design and definition of a common data model. Development of pipelines for anonymization and generation of synthetic data with consideration of data privacy and utility.
Development of the federated learning infrastructure to gather requirements, design and deploy the federated learning platform infrastructure and doterlo of SMPC and DP.
Coordination and supervision of project management from scientific and operational perspectives to ensure high quality standards and risk reduction through an ethical management framework for co-creation, monitoring and ethical evaluation.
Outreach, exploitation, and collaboration activities with communities of interest at the clinical, academic, industrial, and social levels by designing and implementing outreach and communication strategies, leveraging the results, and ensuring the sustainability of the platform.