Potentially Missed Diagnoses in Prenatal Versus Postnatal Exome Sequencing in the Lack of Informative Phenotype:...

To investigate how many novel pathogenic (P) and likely pathogenic (LP) nonprotein-truncating or noncanonical splicing variants would be classified as variants of unknown significance (VUS) if they were detected in fetuses without abnormalities.

The study included 156 patients with neurodevelopmental disorders diagnosed through postnatal exome sequencing. Causative P/LP nonprotein-truncating and noncanonical splicing variants were retrospectively reclassified in cases without specific prenatal manifestations, disregarding postnatal symptoms.

Of the 156 patients, 72 had a nontruncating or noncanonical splicing variant. Six patients were excluded for having more than one possible causative variant. Twelve patients had prenatal malformations known to be associated with the diagnosed disorder; therefore, variant interpretation remained unchanged. In 33 of the 54 remaining cases, the variant had been previously reported as P/LP. Reclassification of the other 21 LP/P variants revealed that 16 would have been classified as VUS if detected prenatally.

In our cohort, ∼24% (16/66) of causative nonprotein-truncating/noncanonical splicing variants would have been classified as VUS if sequencing had been conducted during pregnancy. The potential for false-negative results, stemming from limitations in the phenotypic information available prenatally, should be discussed with prospective parents. The criteria for classifying and reporting variants in the prenatal setting may require adjustment.