1347375 - Relationship between 3D shape and blood flow of the great arteries using fetal CMR in fetuses with suspected coarctation of the aorta

Antenatal diagnosis of duct-dependent coarctation of the aorta (CoA) remains challenging, posing a burden on both parents and healthcare systems. Fetal cardiac magnetic resonance imaging (CMR) has shown potential to accurately predict neonatal CoA based on morphometric and blood flow measurements [1,2]. We have recently shown that statistical shape analysis techniques can be applied to 3D fetal CMR, allowing the range of complex spatial relationships between the aorta and arterial duct to be encapsulated in a single score. The aim of this study was to investigate the relationship between this ‘shape z-score’ and blood flow in cases with suspected CoA.


Methods:

Pregnant women carrying a fetus with suspected CoA were offered a fetal MRI with multiple two-dimensional single-shot fast spin echo sequences of the fetal thorax and phase-contrast magnetic resonance imaging (PC-MRI) sequences of the following fetal vessels: ascending aorta (AAO), superior vena cava (SVC), descending aorta (DAO), main pulmonary artery (MPA), arterial duct (AD), and the umbilical vein (UV). PC-MRI sequences were retrospectively gated [3]; sequences with severe motion were excluded [4]. If multiple sequences were available, mean measurements were used for analysis. Aortic isthmus (AoI) flow was derived from AAO – SVC. Blood flow was indexed to fetal MRI weight. The shape of the great vessels was characterised using a single numeric ‘shape z-score’ ranging from -3SD (false positive (FP) phenotype) to +3SD (CoA phenotype). These scores were generated from principal component analysis and linear discriminant analysis applied to 3D meshes [5] generated from motion-corrected black-blood 3D volumes [1]. Outcome was defined as duct-dependent CoA requiring surgical repair < 28 days age. Monochorionic twin pregnancies, cases with abnormal ventriculoarterial connections, external compression (i.e. diaphragmatic hernia) or missing outcome were excluded.

Results:

Based on the above criteria, 80/134 cases scanned between 2016 and 2021 were included for analysis (median gestational age 32 weeks, IQR 31-33). Neonatal CoA was confirmed in 31/80 (39%). Extreme shape phenotypes of CoA and FP are shown in Fig1. There was a significant difference in mean shape score (FP -0.56, CoA 0.92, p< 0.001), mean AAO flow (FP 140ml/kg/min, CoA 111ml/kg/min, p=0.008) and mean UV flow (FP 143ml/kg/min, CoA 121ml/kg/min, p=0.05). A significant negative correlation was found between shape score and both AAO and AoI flows (Spearman correlation coefficient -0.257, p=0.042 and -0.287, p=0.026 respectively).

Conclusion:

Reduced ascending aorta and isthmal flows are correlated with the shape z-score that predicts neonatal CoA. In agreement with previous fetal CMR findings [2], the phenotype of subjects that will develop CoA includes isthmal displacement, arch hypoplasia, and superior insertion of the isthmus on the arterial duct. Future work is required to define the value of these findings for antenatal prediction of CoA in a prospective cohort.