AI-based Multimodal and Multicentric Cardiac Signal Processing

Details

Contents of the Talk

The talk will delve into latest advancements in medical signal processing and artificial intelligence, focusing on several transformative areas. It will explore medical signal and image synthesis using Generative AI, showcasing how synthetic data can enhance diagnostics and research. The session will also address Multi-Modal AI, emphasizing the concurrent analysis of multi-modal datasets (e.g. Xray, ECG, medication, lab values), including irregularly sampled temporal signals, to deliver more robust insights. Another key topic will be Federated Learning between Hospitals, highlighting secure and collaborative AI development while preserving patient privacy. This infrastructure was recently installed within DZHK, allowing multicentric AI models to be trained over different hospital clients.

Short CV

Sandy Engelhardt, PhD is a Computer Scientist and leads the Group Artificial Intelligence in Cardiovascular Medicine in her role as Assistant Professor at Heidelberg University Hospital. The main research goal of her group is to leverage AI for cardiovascular precision medicine.

Her dissertation won the BVM-Award 2017 for the best PhD thesis awarded from the German Medical Image Processing Community. Together with her team, she planned, integrated and runs a multicentric Federated Learning Initiative in DZHK between eight medical centers in Germany. She is speaker of the Multi-dimensionAI consortium in the area of HFpEF funded by the Carl Zeiss Foundation by 5 Mio €. Furthermore, she is Program Chair of the ESC Digital and AI Summit 2025 organized by the European Society of Cardiology.

Her group develops cutting-edge AI methods and strengthens this key technology across disciplines. The research group focuses on the analysis of cardiovascular imaging and diagnostic procedures (MRI, CT, ultrasound, endoscopy, ECG, etc.) that cover different points in time along the treatment path. This includes research questions on automated risk prediction after cardiovascular interventions, geometric modeling and the characterization of movement patterns of the heart – areas where multi-modal deep learning methods offer promising and highly efficient solutions. Within the area of generative AI she currently focusses on 2D/3D/4D image synthesis.

Since 2024, she has been active in several European medical societies. She holds the Chair for AI in the EACTS Innovation Committee. As part of the ESC Digital Cardiology and AI Committee, she is Subgroup Lead for Education and Outreach. As Program Chair for the ESC Digital and AI Summit 2025 conference, she is primarily responsible for the event's program and for linking the areas of computer science/AI and cardiovascular research. She is also active in the ESC Program Committee and the EHRA Digital Medicine and mHealth Committee. In the University Medicine Network (NUM), she was elected as a spokesperson for FOSA Cardiology. She supports the University of Heidelberg as an advisory member of the AI Board. As DZHK Principal Investigator, she supports the strategic definition of the AI platform.