The PhD defence will be in Domus Juridica (UiO), Auditorium 3. The host of the session will moderate the technicalities while the chair of the defence will moderate the disputation.
Ex auditorio questions: the chair of the defence will invite the audience to ask questions ex auditorio at the end of the defence.
Trial lecture
March 25th, time: 10:15 am, Domus Juridica (UiO), Auditorium 2.
The ethics of the use of AI in clinical decision support tools: The role of explainability, uncertainty quantification, and algorithmic fairness in building the right tools in the right way
Main research findings
Heart valve disease affects millions of people worldwide and represents a growing burden on public health. Among its many forms, certain conditions involving the mitral valve — one of four valves that regulate blood flow through the heart — can lead to dangerous irregular heartbeats and sudden cardiac death. Identifying which patients are at risk remains one of the key challenges in cardiology today, as current clinical tools lack the precision needed for reliable risk assessment.
This thesis addresses that challenge by developing and applying computational tools that combine machine learning and computer simulation to better understand mitral valve disease. First, a machine learning tool called DeepValve was developed to automatically detect the mitral valve structure from cardiac MRI scans. Building on this, statistical shape analysis of the heart revealed that specific geometric deformations of the left ventricle are significantly associated with arrhythmic events, suggesting a novel shape-based marker for risk. Computer simulations further demonstrated that mechanical stretching of muscle tissue can trigger sustained dangerous heart rhythms, particularly in the presence of scarring. Finally, patients with mitral valve disease were shown to exhibit a distinct and abnormal curling motion of the heart wall, now quantified for the first time across imaging modalities.
Together, these advances move the field closer to a more complete and quantitative picture of mitral valve disease.
Adjudication committee
- Professor Vicente Grau, Oxford University
- Software engineering manager Kristin McLeod, Microsoft Development Center Norway
- Professor Mikael Mortensen, University of Oslo
Supervisors
- Professor Mary Margot Maleckar, Simula Research laboratory
- Dr. Nickolas Forsch, Simula Research laboratory
- Professor Gabriel Balaaban, Kristiania University College
- Professor Kent-André Mardal, University of Oslo
Chair of defence
Head of Department Arne Bang Husby
Host of the session
Professor Mikael Mortensen, University of Oslo