OP09.04: Agenesis of ductus venosus

Abstract

Short oral presentation abstractsgestation. The 3D volumes were segmented using a previously developed automatic atlas-based segmentation algorithm guided by manual landmarks and US-specific image features (image 1). After visual inspection, individual meshes were automatically generated, rescaled and aligned. A statistical shape model was estimated to describe morphological variability within a common framework (image 2). Results: The model was able to provide a morphological reference that included the normal shape variability (image 3) with green and yellow highlighting areas with greatest variance and to simulate the average longitudinal changes, corrected for size, observed in the healthy fetal population (image 4). Conclusions:We demonstrate that a 3D shape model can be extracted from a 3D US volume allowing objective quantification of craniofacial morphology and determine a reference for the healthy population. If facial morphological differences in genetic conditions can be isolated as distinct from normal faces, this may provide an objective means to diagnose in-utero abnormalities from 3D US scans.