The goal of the project is to provide clinicians a fast and reliable, velocity quantification method to explore abnormal blood flow in the most anatomically challenging area: cardio-thoracic vasculature. We propose to develop a highly undersampled radial-based 4D/5D Flow MRI sequence using a complete investigation of all aspects of development: simulation of MR physics, prototyping and implementation, and advanced physics-informed deep learning-based image reconstruction. The strength of the proposal lies in this comprehensive approach that will allow us to overcome current limitations. In collaboration with clinicians, experts in cardiovascular pathologies, we will explore the feasibility and added value of our proposed method in a clinical setting.