AICARD
The AICARD Research Project ‘Transforming Cardiac Research: Visual Exploration and AI Prediction Modeling of Real-Life, Multi-Modal Data’ is an interdisciplinary research project between the Computational Imaging Research Lab (CIR) and the Department of Cardiology at the Medical University of Vienna and Visualization and Data Analysis Research Group (VDA) at University of Vienna. AICARD aims to transform cardiac research by exploring routine clinical data through advanced machine learning and visualization techniques. We will develop tools to enable more effective research, i.e. enabling medical professionals to explore and discover patterns in disease progression and to assess treatment response.
CIR Team: Philipp Seeböck, Pamela Zolda, Georg Langs
Cardiology Team: Ronny Schweitzer, Noemi Pavo, Manuel Mayr, Christian Hengstenberg
Visualization and Data Analysis (VDA) Team: Torsten Möller, Laura Koesten
Project Summary
Currently, research in cardiology and most other medical disciplines relies on a confined set of specific variables and relatively narrow, well-defined, patient cohorts. While this has yielded significant clinical and scientific advances, most patient encounters and data collected remain invisible to research. AICARD will change that. We will link routine clinical practice, machine learning, and visualization to make large, real-life patient cohorts accessible to research. As an initial step, we will expand an existing, prospective registry documenting routine patient encounters at the Department of Cardiology (kaRDA), by integrating high-dimensional imaging data and various other cardiac examinations. Natural language processing will transform unstructured clinical free-text into structured data. This will facilitate the development of multi-modal AI models and interactive visualization tools to analyze patient cohorts, trajectories, and AI predictions. These will enable medical professionals to discover data-driven patterns in disease progression and treatment response. In AICARD, we will explore clinically relevant questions related to heart failure with reduced ejection fraction (HFrEF). HFrEF is a condition with a wide range of treatment responses ranging from clinical remission to recurrent hospitalizations and cardiovascular death. Despite significant implications for patients and healthcare, the factors driving this variability remain poorly understood. By developing multi-modal AI models and visualization tools tailored to handle the heterogeneity of real-world data, we aim to explore our deeply phenotyped HFrEF cohort, and accurately predict outcomes such as worsening heart failure in a diverse patient population. The primary aim of AICARD is to foster a cutting-edge, data-driven research environment by leveraging routine clinical data through advanced machine learning and visual analysis techniques.
AICARD Project Webpage: https://www.ai-card.at/