Validation of Prenatal Aortic Arch Angle Measurements in the Diagnosis of Neonatal Coarctation of the Aorta.

ABSTRACT

Prenatal prediction of coarctation of the aorta (CoA) is challenging. Methods identifying prenatal CoA have high sensitivity with significant false positives. We previously derived prenatal aortic arch angles for identifying CoA with high sensitivity and specificity and aim to validate these angles and compare them with a model utilizing ascending aorta (AAo) and isthmus (Aoi) measures. Retrospective case/cohort study of fetuses with prenatal suspicion for CoA. 35 fetuses were included. Measurements included ascending-descending aortic angle (AAo.DAo), transverse-descending aortic angle (TAo.DAo); diameters and z-scores of Aoi from sagittal (Aoi-sag), three-vessel (Aoi-3VV) view and AAo. Discriminant functions for the 5 variables were compared using histograms and positive/negative predictive values (PPV/NPV). CoA was confirmed in 28/35 neonates. The PPV and NPV for angle measures were 100% and 77%. The AAo + Aoi-3VV model PPV and NPV were 92% and 80% and Aoi-sag + Aoi-3VV model were 82% and 71%. A linear discriminant model utilizing the 3 most predictive variables improved NPV to 90% and PPV to 100%. In conclusion, we validate that angle measures are superior to standard models of predicting CoA. An optimized 3 variable model maintains accuracy of identifying CoA while eliminating false positives.