Automated Deep Learning Pipeline for Pulse Wave Velocity Measurement in UK Biobank MRI Data

Motivation: MRI-based pulse wave velocity (PWV) is a golden marker of arterial stiffness, but the acquisition and manual segmentation are time-intensive. This study introduces an automated pipeline designed to accelerate PWV measurement process. Goal(s): We aim to present an automated pipeline for measuring aortic arch PWV from MRI localizers and phase contrast MR (PCMR). Approach: We trained two deep learning models: one to generate 3D aorta segmentations from 2D localizers for accurate 3D aortic length measurement, and another to segment PCMR for flow curves and transit time calculation. Results: Our model proved to be able to produce automated PWV measurement on UK Biobank dataset. Impact: Our model enables automatic aortic arch PWV measurement for UK Biobank subjects, which can be used in the investigation of arterial stiffness and prediction of cardiovascular disease for a large population.