Archimedes and the Biomedical Research Foundation Share Latest Findings in AI and Medicine

Archimedes and the Biomedical Research Foundation of the Academy of Athens successfully hosted a special collaborative session at the Panhellenic Working Group Seminars of the Hellenic Society of Cardiology. This session focused on innovative applications of artificial intelligence in medicine, with a particular emphasis on advancements in cardiology. Held on Friday, February 7, 2025, in Room MC2 of the Megaron Athens International Conference Centre, the event brought together leading experts to explore how artificial intelligence is transforming cardiovascular medicine.

Distinguished speakers included Prof. Dr. Panos Vardas, M.D., from Athens, who delivered an insightful lecture on "The present and future of artificial intelligence," and Asst. Prof. Maria Vakalopoulou, who presented groundbreaking research on "Analysis of Medical Images and Medical Accuracy from Imaging Data." Prof. Sotirios Tsaftaris addressed the important topic of "Causal Artificial Intelligence for Medical Accuracy in Cardiology." The session concluded with a discussion on future extensions and collaborations, chaired by Prof. Dr. Dionysios Kokkinos, M.D., and Prof. Timos Sellis, with closing remarks by Dr. Dimitris Thanos, PhD, and Prof. Dr. Dimitrios Farmakis, M.D.

Three lead researchers from Archimedes delivered engaging presentations on current findings and future trends in artificial intelligence, medical image analysis, AI accuracy in cardiology, and potential collaborations.

Asst. Prof. Maria Vakalopoulou (Centrale Supélec, University Paris-Saclay, France and Lead Researcher at Archimedes, Athena RC) discussed the role of medical imaging and AI in healthcare and cardiology in particular. She focused on the use of recent deep learning-based foundation models and their applications in medical and clinical questions. Foundation models are huge models that have been trained in huge amounts of data, and they are currently achieving very good performance in a variety of applications. Their adaptation, however, to medical applications is not straightforward, and there are currently a lot of challenges and limitations. During the talk, the work of members of Archimedes was presented, towards the development of a holistic foundation model for chest X-Rays. More specifically, RayDINO was presented, a powerful and effective generalist model that outperforms the current state of the art models for a wide range of radiological tasks. IRayDINO’s performance for cardiology tasks and in particular, identification of rare diseases such as calcification of the aorta or tortuous aorta, as well as its capabilities for automatic report generation, has been discussed.

Prof. Sotirios Tsaftaris (The University of Edinburgh, UK and Lead Researcher at Archimedes, Athena RC) began by extolling the benefits of AI in terms of addressing needs of an ageing population which puts pressure in a tired healthsystem with rising costs and diminishing workforce. While readily finding correlations in big data is easy, AI systems often lack robustness, reliability, and fairness, and are susceptible to biases. Deep learning, though powerful, can lead to harmful outcomes if these issues aren't addressed. A key advancement is Causal AI, which moves beyond mere correlation to establish causal relationships. Causal inference, the process of determining a phenomenon's true effect within a system, is central to this. Causal AI uses this grounding for better inferences, especially crucial for personalized medicine in cardiology. Understanding disease across populations is paramount before individualizing treatment. This involves removing confounding factors, then reintroducing them while incorporating diverse patient data (clinical, genomic, imaging, etc.) and learning from past examples. Essential ingredients for this include Large Language Models (LLMs), patient registries, and multimodal data for which the speaker urged towards unified and coordinated action.

Prof. Maria Papadopouli (University of Crete, Greece, Research Associate at the Institute of Computer Science, FORTH, and Lead Researcher at Archimedes, Athena RC) discussed the potential use of AI and augmented-reality (AR) systems to enrich health-care support for emergency situations in remote places as well as in modern cardiology education by enhancing visualization, interactivity, and personalized learning. For example, AI-powered simulations could help medical students and professionals analyze complex cardiac conditions through predictive models and real-time diagnostics. AR applications enable immersive experiences, such as interactive 3D models of the heart, allowing users to explore anatomy, pathology, and surgical procedures in a hands-on manner. These technologies improve training efficiency and provide risk-free environments for practice. The development of energy-efficient reliable AI models that analyze multi-modal information in real-time, at the edge, to enable immersive experiences involves exciting multidisciplinary research.