Introduction to GenoMed4All
The GenoMed4All project is revolutionizing the approach to diagnosis, prognosis, and treatment of haematological diseases through the use of Artificial Intelligence (AI). Haematological diseases, comprising up to 450 disorders related to blood cells, lymphoid organs, and coagulation factors, present a significant public health challenge. Most of these diseases have a genetic background, and they can cause chronic health problems or be life-threatening.
The Challenge of Data Scarcity and Fragmentation
Data scarcity remains a pressing issue in the field of haematological diseases. The number of available samples is small and characterized by a high level of fragmentation, largely due to the sensitive nature of the data. The European Haematology Association estimated the financial burden of blood disorders on European society to be approximately €22.5 billion per year in 2016.
GenoMed4All’s Federated Learning Platform
To address these challenges, GenoMed4All is developing a Federated Learning (FL) platform, a shared, distributed space where clinicians and researchers can collaborate on the definition, development, testing, and validation of AI models. This platform allows experts to access and analyze distributed clinical data without physical transfer, ensuring patient data remains secure on-premises.
Benefits for Clinicians and Researchers
For clinicians, the platform acts as a local decision support system, providing insights from an ever-learning model. For researchers, it offers an AI sandbox to explore datasets, develop, train, and benchmark new AI models on real-world data.
Ethics-First Approach to Data Sharing
GenoMed4All emphasizes an ethics-first approach to data collection, harmonization, and cross-border sharing. The project has developed robust ethical agreements, data protection management protocols, and risk assessment plans. It has also produced recommendations to guide the ethical development and deployment of AI in clinical settings, aligning with the EU AI Act.
Validating AI in Blood Disorders: Use Cases
The project targets three use cases: Myelodysplastic syndromes (MDS), Multiple Myeloma (MM), and Sickle Cell Disease (SCD). For MDS, GenoMed4All identified a genomic signature predicting the risk of progression to Acute Myeloid Leukaemia (AML). For MM, the team combined genomic data with imaging techniques to increase prediction accuracy of overall and leukaemia-free survival. For SCD, a novel algorithm was developed to detect silent infarcts using patient MRI data.
Legacy and Future Directions
GenoMed4All’s integrative approaches can be extended to other blood disorders and potentially to other areas of medicine. The project’s federated platform unlocks interoperability, data reusability, and knowledge transfer across Europe, paving the way for precision medicine and a common European Health Data Space (EHDS).
