Frank te Nijenhuis

Within my PhD project, we are developing computer vision methods to automatically analyze cerebral Digital Subtraction Angiography (DSA) images. This includes automatic detection of vessel occlusions, quantifying vessel differences between the pre- and post-treatment DSA or delineating brain regions on DSA. The ultimate goal of these methods is to support clinical decision-making during mechanical thrombectomy procedures.

With colleagues from the ICAI Stroke Lab we are investigating potential use-cases for synthetic tabular data for prediction modelling. Specifically, we aim to mitigate class imbalance by generating synthetic samples for minority classes, thereby improving predictive performance.

Recent work on perfDSA demonstrates the potential of fully automatic and quantitative frameworks for perfusion analysis on DSA, analagous to more established modalities such as CTP and MRP. We are currently validating perfusion DSA to assess its clinical applicability, with the goal of enabling real-time decision-making during thrombectomy.

Carotid webs are shelf-like fibrous projections in the carotid artery that can disrupt blood flow and increase the risk of ischemic stroke. We are developing automated methods to detect these webs on CTA scans, aiming to improve early diagnosis and support clinical decision-making.