3D MRI atlases of congenital aortic arch anomalies and normal fetal heart: application to automated multi-label segmentation

Abstract

The black blood contrast T2w ssFSE fetal MRI along with application of 3D slice-to-volume registration methods allows reconstruction of high-resolution 3D images of the heart that provide superior visualisation of fetal cardiovascular anomalies. However, there is a lack of formalisation of the MRI appearance of fetal cardiovascular anatomy and standardisation of vessel segmentation protocols. In this work, we present the first 3D fetal MRI atlases defining normal and abnormal fetal cardiovascular anatomy. We also perform evaluation of the feasibility of atlas-guided registration and deep learning methods for automated 3D multi-label vessel segmentation.