Advancing AI for fast MR-guided prostate biopsies

Abstract
Deep learning can accelerate MR-guided prostate biopsies by reconstructing undersampled images while maintaining accurate instrument tracking. Trained on over 8,000 clinical scans and validated prospectively, a spatiotemporal model achieved reliable needle guide localization with errors below 5 mm at up to 12x undersampling. These findings demonstrate the feasibility of AI-enabled real-time MR image reconstruction for interventional guidance.

Meeting details:
The Teams link will be available when you RSVP to this event.
14:30 - 15:30 (CET):
Stan Noordman
Advancing AI for fast MR-guided prostate biopsies

About the speaker
Constant (Stan) Richard Noordman studied at Utrecht University, where he obtained a Msc in Artificial Intelligence and in Medical Imaging. In 2021, Stan started as a PhD candidate working on deep-learning based prostate MRI reconstruction for diagnosis and intervention (as part of the FastMRI project at the Diagnostic Image Analysis Group, under supervision of Henkjan Huisman and Jurgen Fütterer. After he completes his PhD, he will start work as a Postdoc on the same topic, trying to push further towards real-time navigation of instruments during MR-guided minimally invasive interventions.