Haematological diseases include up to 450 disorders caused by problems with blood cells, lymphoid organs, or blood clotting factors. They are usually divided into cancerous (oncological) and non-cancerous types. Most of these diseases have genetic causes. They pose a serious public health challenge: blood cancers make up about 5% of all cancers, many patients face long-term health issues, and some conditions can be life-threatening. In 2016, the European Haematology Association (EHA) estimated that blood disorders cost European society around €22.5 billion every year. Yet, data on these diseases remains scarce and fragmented, mainly because of the sensitive nature of health information.
To tackle these challenges, the European Union has launched GenoMed4All, a project aimed at transforming diagnosis, prognosis, and treatment of blood diseases using Artificial Intelligence (AI). By combining existing clinical data with advanced genomic profiling, GenoMed4All seeks to develop new ways to diagnose and treat haematological disorders. The project gathers multiple types of data, including clinical records, multi-omics data (such as genomics and metabolomics), and real-world patient information (like nutrition and environment) into a secure platform. This will improve the ability to diagnose, choose treatments, and predict outcomes tailored to each patient.
The project connects Europe’s main blood disorder data repositories, enabling standardized, cross-border data sharing while fully respecting privacy laws and ethical standards. Over the past four and a half years, GenoMed4All has united 23 partners from seven European countries, covering clinical research, regulation, healthcare, technology, and digital services. It also works closely with ten clinical partners from ERN-EuroBloodNet, the European Reference Network for rare haematological diseases. This network is supported by several European health organizations and patient advocacy groups.
A key challenge in using AI in healthcare is accessing sensitive clinical data securely. GenoMed4All addresses this by developing a Federated Learning (FL) platform. This platform lets clinicians and researchers collaborate on AI models without patient data ever leaving their local hospitals. This privacy-by-design system not only protects patient information but also helps overcome data fragmentation by linking multiple sources securely. Clinicians can use this platform to support their daily decisions with AI-driven insights, while researchers can explore large datasets to develop and test new AI models, benefiting from data across institutions without compromising privacy.
The project emphasizes ethics in data sharing and AI use. Healthcare data holds huge potential but is often underused due to inconsistent formats and incomplete records across hospitals. GenoMed4All promotes data quality, transparency, and reliability. The team has developed strong data protection protocols and risk assessments to ensure ethical compliance, especially important for rare diseases, where data bias and availability are ongoing challenges. The project follows EU regulations and has created guidelines to ensure AI tools are explainable, accountable, and trustworthy.
To support data integration, GenoMed4All has built common data models and bioinformatics pipelines to harmonize and anonymize data from different clinical sites. They apply the FAIR principles—Findable, Accessible, Interoperable, and Reusable—to make sure data can be shared safely and used efficiently. The project also set genomic standardization guidelines to enable future data sharing and expansion by research and clinical communities, enhancing care for blood disorders across Europe.
GenoMed4All focuses on validating AI models through real clinical use cases. The team works closely with clinicians to ensure AI tools address real needs and fit clinical workflows. Three key haematological diseases guide the project: Myelodysplastic syndromes (MDS), Multiple Myeloma (MM), and Sickle Cell Disease (SCD).
MDS affects mainly older adults and causes bone marrow failure, leading to low blood cell counts. It can progress to acute myeloid leukemia (AML), a fatal blood cancer if untreated. GenoMed4All identified genetic markers predicting progression to AML and used synthetic data to simulate early relapse, improving risk prediction.
Multiple Myeloma is a cancer of plasma cells that harms bones and kidneys. By combining genomic data with imaging like PET and CT scans, the project enhanced predictions for patient survival and disease progression.
Sickle Cell Disease is a hereditary blood disorder causing misshapen red blood cells that block blood flow, leading to serious complications. GenoMed4All developed an AI algorithm to detect silent brain lesions in SCD patients using MRI scans, aiding early diagnosis.
With roughly 30 million people living with rare diseases in Europe and 300 million worldwide, personalized medicine is vital. GenoMed4All’s approaches in these use cases can extend to other blood disorders and even other diseases like lung cancer, diabetes, and neurodegenerative conditions. The project is working to include more clinical sites and broaden its impact through networks like ERN-EuroBloodNet.
The technical platform GenoMed4All built offers a foundation for data interoperability and knowledge sharing across Europe. It connects hospitals and research centers, bridging gaps between clinicians and researchers. This paves the way toward precision medicine and the European Health Data Space, focusing on security, usability, and scalability.
Ultimately, GenoMed4All aims to transform clinical care with AI tools that are secure, reliable, explainable, and privacy-preserving. By supporting clinical decisions with advanced data analysis, it seeks to improve outcomes and bring patient-centered, data-driven healthcare to people with blood disorders across Europe.
Related Topics
