EP09.13: Congenital heart diseases unrecognised in the prenatal period: is the quality of obstetrical ultrasound related to the type of heart disease?

To conduct an audit of the images present in the prenatal ultrasound reports of newborns and children in the South France region who were diagnosed with congenital heart disease (CHD) postnatally. This is a retrospective analysis of images from prenatal ultrasound reports of children born during 1 year in the France South region with a postnatal diagnosis of CHD requiring intervention in the first year of life through catheterisation and/or cardiac surgery. A quantitative and qualitative analysis of these images was performed by a junior screening physician according to predefined image quality scores and visualisation of anatomical landmarks. CHD were classified according to: ESC prognosis recommendations / the postnatal management period / type of treatment: catheterisation - surgery - hybrid / feasibility of prenatal screening: possible / not possible / not always possible. 104 children were born with a postnatal diagnosis of CHD. All these children were able to undergo a biventricular repair program. After examination, 60/104 cases were retrieved and analysed. Among them, 48/60 (80%) had both T2 and T3 ultrasounds. The analysis of images associated with prenatal ultrasound reports was classified as follows: incomplete (43/60, 72%); complete (17/60, 28%) but of poor quality (7/60, 12%); complete and of good quality (10/60, 16%). There was no significant difference between the 3 groups regarding the intrinsic image quality (zoom and resolution). However, there was a significant difference in anatomical quality scores between the incomplete and complete BQ groups (p = 0.00004). 15% (9/60) of CHDs were visible after rereading, with 45% and 55% in the incomplete and complete groups, and 20% and 80% in the poor and good quality subgroups respectively. The intrinsic quality of the obtained images (zoom and resolution) is satisfactory in all subgroups. However, the quality of obtaining anatomical landmarks is suboptimal overall. 15% of CHDs could also have been identified.