Résumé : ObjectivesWe aim to introduce a new technique for minimally invasive autopsy: the Post‐Mortem Ultrasound (PM‐US). Our purpose is to demonstrate its feasibility and sensitivity in detecting major congenital abnormalities as compared to conventional autopsy.MethodsIn this prospective validation study, we undertook all PM‐US at Brugmann University Hospital between March 1; 2012 and Sept 30, 2013. We selected a sequential population of fetuses, for which parents accepted conventional and virtual autopsy. We excluded cases with no available written consent. Fetuses underwent PM‐US and than conventional autopsy. The study had institutional approval. We recruited 88 consecutive fetuses, from 11 to 40 weeks gestation (WG), coming from TOP, miscarriages and IUFD. We performed a PM‐US using different transducers and data were compared with conventional autopsy. Conventional autopsy was performed according to the Societé Francaise de Foetopathologie (SOFFET, France) guidelines, conducted by experienced perinatal pathologists, blinded to the ultrasound data. The ultrasound and autopsy data were recorded in the same database.Resultsa complete virtual autopsy by ultrasound was possible in 95% of the cases. Sensitivity in detecting brain abnormalities was 91%, with a specificity of 90%. In 20% of cases, a neuropathological examination was not possible due to extreme maceration. The sensitivity of PM‐US in detecting thoracic abnormalities was 88% and the specificity was 92%. Concerning abdominal anomalies, we also found a sensitivity of 87% and a specificity of 94%.ConclusionsThis pilot study confirms the feasibility of PM‐US for a virtual autopsy as early as 11 weeks gestation. This new technique shows a high sensitivity and specificity in detecting congenital structural abnormalities as compared to conventional autopsy.