Reanalysis of unsolved prenatal exome sequencing for structural defects: diagnostic yield and contribution of postnatal/postmortem features

In 30-40% of fetuses with structural defects, the causal variant remains undiagnosed after karyotype, chromosomal microarray and exome sequencing. This study presents the results of a reanalysis of unsolved prenatal ES (pES) and investigates how postnatal/postmortem phenotyping contributes to identify relevant variants. Prospective reanalysis of pES data was performed in undiagnosed fetuses enrolled in the AnDDI-Prénatome cohort study. Postnatal/postmortem data up to 3 years were merged with prenatal features using HPO terms. The reanalysis involved updating the bioinformatic processing of raw data and periodic querying of the raw pES data using a GREP query. The reanalysis was performed in 58/94 (62%) unsolved pES, including 8 variants of unknown significance. Clinical examination at birth was available for all live newborns, and postmortem examination was available in 12 terminated fetuses. Additional features were identified in 27/58 fetuses (44%). The contribution of updated postnatal/postmortem phenotyping to the identification of relevant variants was limited. Only one new diagnosis was made, but thanks to the new implication of the SNAPC4 gene in the human disease rather than the additional clinical data provided by the postnatal update of the phenotype. Only one of the three additional VUS identified by reanalysis was also considered thanks to additional features provided by postmortem examination. In conclusion, in cases of unsolved pES, the benefit of prospective reanalysis of unsolved pES appears to be limited, even over time. It may be more appropriate to offer postnatal genome sequencing rather than pES reanalysis with postnatal/postmortem phenotyping update to establish a causal diagnosis.