Performance of an AI algorithm for quality control of routine fetal ultrasound

17,169 screening fetal US images, performed in all trimesters (T1, T2, T3), were collected from 7 North-American and European centres. The performance (sensitivity and specificity) of the software was assessed by comparing the algorithm's blinded assessment to experts' assessment. The concordances between experts' and the algorithm's assessments for the presence of a standard plane or structure was defined as true positive for the corresponding plane/structure. Concordance in the absence of the plane or visible structure was defined as a true negative. Reported in table 1. This study shows that AI can efficiently and automatically address main aspects of quality control for fetal ultrasound, while reducing unnecessary repetitive tasks such as labelling. Such a software could improve the workflow of practitioners and imaging quality by ensuring completeness of examinations and that expected anatomical structures are visible on standard planes.