Our Educational Program is back!
The second wave of our training webinar series is coming up soon – key experts in the application of Artificial Intelligence in the field of Hematological Diseases will be present, don’t miss out!
Our Training Program – Looking back on Wave #1
The GenoMed4All & ERN-EuroBloodNet Educational program on Artificial Intelligence in hematology introduced the public to GenoMed4All’s mission and explore the various applications of Artificial Intelligence in biomedicine. This program covered a range of topics, starting from defining precision medicine and problem-solving methodologies to discussing the utilisation of multi omics (including Genomics, Metabolomics, Proteomics, and Radiomics) in clinical research. Specific use cases such as Myelodysplastic syndromes (MDS), Sickle Cell disease (SCD), and Multiple myeloma (MM) were highlighted.
Additionally, the program delved into strategies for standardising data and establishing cross-border connections between repositories. It also explored the application of Artificial Intelligence in diagnosis, early risk assessment, and predicting disease progression using a multi-modal approach.
Here is a summary of all 4 previous sessions. Follow us on our social media channels for the latest updates!
Session 1: ‘GenoMed4All & ERN-EuroBloodNet for Precision Medicine in Hematology’
How can Artificial Intelligence help advance personalised medicine in diagnosing and treating haematological diseases?
Our training series started with an introduction to the concept of precision medicine, highlighting its promise in tailoring medical treatment to individual characteristics. We discussed how advancements in genomics, proteomics, and other ‘omics’ technologies help pave the way for personalised approaches to diagnosis, treatment and prevention. It also tackled the obstacles that can be encountered in this process, such as data transfer and interoperability.
Session 2: ‘Uses cases challenges: MDS, SCD and MM’
In terms of the practical use of AI, how can it help them manage and extract insights from clinical data? How can the data be validated from both a legal and ethical standpoint? How will GenoMed4All’s 3 disease use cases minimise the negative impact of unharmonised, scattered and incomplete data?
Due to their rarity, many healthcare professionals may lack awareness and understanding of certain hematological diseases like Myelodysplastic Syndromes (MDS), Sickle Cell Disease (SCD) – this session provided answers on how our use cases approach these critical issues.
Session 3: ‘Data Standardization & Linkage‘
What is the current situation? How could rare disease clinical data disparity be a problem but also an opportunity? Find out the new paradigm: think outside the box. Have you heard of a Federated Learning model?
In this session, we learned that Federated Learning is a Machine Learning approach in which, instead of gathering all the data in one centralised location, the model is trained across multiple decentralised devices or servers holding local data samples. These devices or servers collaboratively learn a shared model while keeping the data localised. This approach offers privacy benefits since the data remains on the device or server where it originates, and only model updates, not raw data, are exchanged between the devices and the central server. This way, sensitive data can remain on the local device, such as in the hospitals, ensuring privacy and security. This is particularly useful in scenarios where data privacy is a concern, such as in healthcare. It enables collaborative model training across a network of devices while keeping sensitive data decentralised and secure. FL represents an innovative way to harness the collective intelligence of distributed data sources while respecting privacy and data locality constraints.
Session 4: ‘Data Integration & Analysis (Artificial Intelligence)’
How to achieve faster advances in medical research?
Data integration and analysis play a critical role in the machine learning pipeline. Before training a model, data must be integrated and analysed to identify relevant features, pre-process input data, and engineer new features that may enhance the model’s performance. Moreover, analysing the model’s predictions and performance on test data is essential for evaluating its effectiveness and identifying areas for improvement. In this session, we explored how researchers and practitioners can implement data integration strategies and the different approaches applied in hematology when using AI tools, specifically in MDS and MM use cases. In GenoMed4All, model updates from different devices are aggregated to create a global model without sharing raw data, and furthermore, different models and scenarios are analysed aiming to get explainable results. This way, we can keep data private, foster global collaboration, provide treatment personalisation, and reduce costs.
Through this first webinar series, our aim was to provide a comprehensive overview of how AI is transforming personalised medicine and healthcare, opening up with new opportunities to use real-world data to advance research and improve patient care in specific diseases such as MDS, SCD and MM.
All in all, we hope that attendees gained a deeper understanding of the challenges and opportunities associated with AI-driven healthcare innovation, and we look forward to continuing the conversation and exploring further advancements at the intersection of AI and rare disease healthcare in future webinars.