On Tuesday November 14 Maarten Terpstra defended his PhD thesis entitled: “Real-time MRI-guided radiotherapy with deep learning”. The thesis focused on the development of deep learning models to enable real-time adaptive MRI-guided radiotherapy using an MRI-Linac. This includes the development of MODEST, a deep learning model for fast respiratory-resolved 4D-MRI acquisition and reconstruction, and models enable real-time tracking for 2D and 3D MRI in real-time. Moreover, a new loss function was developed for image registration and complex-valued image reconstruction, called ⊥-loss.
A digital copy of his thesis can be obtained from here. More information and pictures are available on his social media page.