Diagnostic Value of Digital Estimates of Trophoblastic Mosaicism in Genome-Wide Cell-Free Fetal DNA Screening

Cell-free fetal DNA (cffDNA) screening exhibits higher sensitivity and lower false-positive rates (FPRs) compared to combined first-trimester screening for common trisomies. However, additional anomalies on other autosomes detected by genome-wide assays cause anxiety and trigger increased rates of amniocentesis. Here, we evaluated the diagnostic value of software-generated digital estimates of trophoblastic mosaicism to identify false-positive screenings and inform posttest counseling.

A total of 32 175 pregnancies were screened using a commercial genome-wide cffDNA screening assay alongside an open-source bioinformatics pipeline for fetal fraction and aneuploidy analysis. Amniocentesis outcomes were utilized to calculate positive predictive values (PPV) for common trisomies, sex chromosomal aneuploidy (SCA), rare autosomal trisomies (RAT), and subchromosomal alterations (SA). Receiver operating characteristics and regression analysis were performed to identify read count metrics that minimize FPRs.

The assay showed 100% sensitivity for detecting trisomies 21, 18, and 13, with PPV ranging from 17% (+13) to 75% (+21). For SCA, PPV varied from 50% (45, X) to 89% (47, XXY). PPV for SA was 37%. RAT were detected in 0.33% of cases with a PPV of 6.6%. A proprietary digital estimate of trophoblastic mosaicism (mosaic ratio) exhibited good diagnostic performance (area under the curve 0.83-0.95) for false positives: mosaic ratio below 0.50, observed in 64% of RATs, 24% of SAs, 17% of common trisomies, and 7% of 45,X cases, was never associated with fetal aneuploidy.

Digital estimates of trophoblastic mosaicism hold clinical significance for informing posttest counseling. Implementing these recommendations could reduce invasive testing for RAT from 0.33% to 0.11%.