Prenatal diagnostic accuracy in South Carolina demonstrated by autopsy

Fully 150 consecutive fetal/neonatal autopsies were reviewed to determine to what extent they confirmed or altered the impressions gained through prenatal ultrasonographic fetal examination. Distinctions were made between features that may or may not be assessable by prenatal ultrasound. Analyses of weights and measures were based on recently published regressions derived from a worldwide review of normative data. Our analysis indicated a high level of correspondence between prenatal ultrasound findings and later observations of independent persons at autopsy (85% positive predictive value and 44% sensitivity). We concluded that skills of maternal-fetal medicine specialists, located at several geographically divergent centers, are confirmed by a high level of correspondence between prenatal ultrasound and autopsy findings. The low sensitivity was due in large part to the relatively subtle nature of many autopsy findings that had not been predicted by prenatal examination.