AI4LUNGS 2025 Wrap-up — Ethics, Validation & Collaboration Across Europe!
Greetings from the AI4LUNGS project, and welcome to the 4th edition of the Newsletter!
As we close 2025, many milestones have been achieved behind the scenes and on international stages, let’s dive into the latest updates!
4th Stakeholder Forum Meeting
<img src="https://media.licdn.com/dms/image/v2/D4D12AQG7wKM3Bir3Og/article-inline_image-shrink_1500_2232/B4DZsRRD_HLQAU-/0/1765521246201?e=1767830400&v=beta&t=h4IFp6urT5Fg0LslIT14L3-USOUdnSwRgN0vOubA-QQ" alt="Contenuto dell’articolo"/>
Exploring the ethical, legal, and social impact of AI4LUNGS
The 4th Stakeholders Forum meeting, led by partners Deloitte and Timelex, focused on the ethical, legal and social implications (ELSI) of integrating AI into clinical practice. The session presented the first Digital Ethics Risk Assessment and clarified how the GDPR and the EU AI Act shape the future development of AI4LUNGS.
The session highlighted important concerns about over-reliance on AI in clinical settings, emphasising the need to preserve human judgment and critical thinking as technology becomes more integrated into care. Participants discussed the value of interdisciplinary collaboration and the importance of ensuring clinicians are fully supported when adopting AI tools. Above all, the conversation reinforced the necessity of maintaining a human-centered approach as AI becomes an increasingly present part of clinical workflows.
Join the Stakeholder Forum now and contribute to discussions shaping responsible AI in healthcare!
Project Progress: Advancing AI4LUNGS Behind The Scenes
AI4LUNGS has made significant progress during this second half of the year! While we advocate and promote the mission and vision of integrating AI into the health sector, a group of legal, clinical, and technical experts are behind the scenes working hard to make the AI4LUNGS platform practical for clinicians; legally abiding regulations; guaranteeing upholding digital ethics.
Digital Ethics & Risk Management
Led by Deloitte, AI4LUNGS has successfully completed the first Digital Ethics Risk Assessment (DERA), marking a significant milestone in the integration of responsible and trustworthy AI within AI4LUNGS. Opportunities for improvement were identified, with targeted mitigation measures proposed. AI4LUNGS has already begun to prepare the second DERA report, which will reassess risks and evaluate mitigation impacts. This process strengthens the project’s dedication to ethical, safe, and future-proof AI in healthcare.
Algorithms Validation, Pilot Testing & Assessment
Led by our clinical partner MOR, AI4LUNGS is currently shaping a comprehensive validation plan ensuring that technical, clinical, UI/UX, and economic dimensions of the system are thoroughly evaluated. This work has involved close collaboration across all partners to build a reliable and meaningful evaluation framework. In parallel, preparations are underway for the pilot study protocols and the development of clinician training strategies.
GAIA, another AI4LUNGS’ clinical partner, has begun strengthening the synergy between technical and clinical teams through joint algorithm reviews, structured feedback loops, and continuous system refinement to ensure that AI4LUNGS remains aligned with real clinical practice.
AI4LUNGS Platform infrastructure and integration
Led by KPMG, the AI4LUNGS is preparing the platform’s initial setup, focusing on infrastructure and proof-of-concept development. This includes: 🫁deploying isolated cloud components (VNet, ADF, ADLS, Synapse), 🫁validating connectivity, and 🫁carrying out integration testing with dummy containers.
Next, the team will integrate models and workflows, set up CI/CD pipelines via Azure DevOps, ingest participant models through the Azure Container Registry, and expose API endpoints for frontend integration.
This structured and secure approach ensures the AI4LUNGS platform evolves into a scalable and clinically aligned digital ecosystem.
Scientific Contribution on International Stages
This year, 5 scientific papers of AI4LUNGS have been accepted and presented at various conferences and symposiums.
IEEE International Symposium on Systems Engineering (ISSE) 2025
The paper “Co-Designing Trustworthy Connected Medical Devices: A Vee-Model Framework for Usability and Regulatory Alignment,” authored by Anna Palaiologk and Thanos Arvanitidis from Future Needs (Dissemination Leader and Social Sciences and Humanities (SSH) partner), has been accepted and presented at the 2025 IEEE International Symposium on Systems Engineering (ISSE).
25th International Conference on Bioinformatics and Bioengineering
The paper “Incrementally Learning to Segment the Lungs: Similarities and Differences Across Institutions” authored by Joana Sousa, Tania Pereira, and Hélder P. Oliveira from INESC TEC has been accepted to the 25th International Conference on BioInformatics and BioEngineering. Presented in Athens by Joana and Helder, the study explores how continual learning, specifically Experience-Replay with a memory buffer, can address the challenges of lung segmentation caused by variability in CT data across institutions. By examining how memory size, domain similarity, and learning order impact catastrophic forgetting and model performance, this work offers important insights for building more adaptable and clinically reliable AI systems for lung imaging.
Attending OPASCA Users & Friends Meeting 2025 in Mannheim
AI4LUNGS participated in the OPASCA Users & Friends Meeting 2025 where clinics, researchers, and innovators came together to explore the future of digital healthcare. Representing the project, Dr. Alexander Scherrer from Fraunhofer ITWM delivered an invited talk on trustworthy and explainable AI in clinical workflows, highlighting how AI4LUNGS’ clinical decision support system can enhance safety, efficiency, and personalised care.
AI4LUNGS at the IAPP Conference Brussels: Discussing EHDS & AI Model Training
AI4LUNGS was represented at the IAPP Conference in Brussels on the 20th November, where our legal partner, Magdalena Kogut - Czarkowska from Timelex, delivered a presentation on the European Health Data Space (EHDS) and discussed how health data re-use supports AI model training within AI4LUNGS in a panel session alongside Maria Bardaji Cruz, Privacy Compliance EMEA from Johnson & Johnson and Julia Bonder-Le Berre, Head of Global Privacy from Iron Mountain for the session “EU Health Data Transformation: The Promise of the European Health Data Space.”
Read all our published papers here.
Building synergies across Europe
AI4LUNGS held its first synergy meeting with ONCOVALUE, Microb-AI-ome Project, dAIbetes_HorizonEurope, CoMPaSS-NMD, SPACETIME_EU Project, LUCIA EU Project, COMFORT-EU, and TOLIFE Project, bringing together a community of EU projects working at the intersection of AI and healthcare. During this collaborative session, each project introduced its mission, explored common goals, and identified opportunities to align efforts toward more impactful, patient-centered innovation. This gathering marks the beginning of a stronger, more connected ecosystem where shared knowledge and coordinated action can accelerate meaningful progress for clinicians, patients, and the European healthcare landscape. Check out our synergy projects.
AI4LUNGS hosting a panel session at dAIbetes' Workshop!
AI4LUNGS got invited to host a panel session during synergy project dAIbetes' workshop in Lisbon on the 14th of November. The project coordinator INESC TEC presented the AI4LUNGS project and joined the discussions in defining virtual twins in healthcare, data collection and its hurdles, and data security.
Stay Connected!
We would like to extend our warmest wishes to our audience. A big thank you to everyone who has supported us in our second year of project. We are excited to see what the third year will bring us.