HDZ project

The diagnosis of myocarditis in children is complex and has so far been strongly orientated towards diagnostic criteria for adults. In this interdisciplinary cooperation project, we are investigating whether myocarditis in children can be reliably diagnosed on the basis of MRI images using artificial intelligence (AI). A particular focus is on the integration of cardiac function parameters and the use of explainable AI (XAI) in order to make the decision-making processes of the AI comprehensible and to gain new insights into diagnostically relevant features.

Research questions

Q1: Can myocarditis in children be diagnosed on the basis of MRI using AI? -

Q2: What influence do functional parameters of the heart have on the AI-supported diagnosis?

Q3: Does the selected XAI method increase the reproducibility of the diagnosis made by AI?

Q4 : How can the influence of functional parameters on the diagnosis by AI be explained?

Relevance: The project contributes to the improvement of non-invasive diagnostics in children and supports physicians through transparent AI-based recommendations. It strengthens confidence in AI-supported medicine and promotes the understanding of complex relationships between functional parameters and myocarditis.

Team: The project is an interdisciplinary endeavour with expertise from medicine, AI and human-technology interaction between the HDZ and UniBi:

- Dr Tamino Huxohl (medical image processing, AI)

- Prof Wrede (explainable AI, interaction technologies)

- Dr Birte Richter (participatory development, human-technology interaction)

- PD Dr Laser (paediatric cardiological MRI diagnostics)

- Prof Schubert (head of the MYKKE register)