Translational computing in cardiology
In medicine, too much and often conflicting information is available to make it easy to integrate it all meaningfully into clinical practice, but sensible computing can help by providing an intuitive, disease-oriented interpretation
Cardiovascular diseases remain a major cause of death and loss of quality of life. While current prevention and treatment strategies have delayed symptoms and improved life expectancy, this has resulted in an increasing number of patients now requiring careful follow-up and subsequent therapy.
To optimally treat these chronic patients, a detailed assessment and a thorough understanding of how their heart is working is essential. Imaging is often used for this but the images have to be analysed and summarised into relevant information to enable personal decision-making.
Given the wealth of available imaging and other diagnostic data, computational tools such as image processing, computational modelling and machine learning are required to enable processing of all the information for the treating clinician.
The group aims at developing and investigating computing for understanding each individual heart, in order to improve diagnosis, prognosis and therapy. We work in an interdisciplinary way combining biomedical engineering, physics, and mathematics with basic research, as well as clinical expertise, in order to link technological and theoretical science with physiological and clinical knowledge.
We combine medical imaging with computer models representing all relevant details of the working heart as well as with knowledge about therapies. Additionally, the wealth of information available from any single individual and groups of patients from routine practice and clinical trials can be integrated in an understandable way by using the latest machine learning and artificial intelligence methods.
Modern cardiology relies on the detailed (imaging) assessment of the function and structure of the heart for selecting targeted prevention and drug, device or surgical therapies. A wealth of data is available in clinical practice but integrating it into a consistent representation of an individuals’ heart is highly challenging. However, this is crucial in order to choose the best personalised care and is also needed to improve our understanding of cardiovascular diseases and their impact in different groups of individuals.
We aim at providing computing approaches to help this data integration as well as interpretation for both individuals and populations alike. These methods include image analysis and interpretation and machine learning for data integration and decision-making, as well as tools for better and more targeted data acquisition.