62 Multicenter prospective study of SNP-based cfDNA screening for aneuploidy with genetic confirmation in 18,496 pregnancies

Abstract

To evaluate performance of single nucleotide polymorphism (SNP)-based cell-free DNA (cfDNA) screening for detection of trisomies 13, 18, and 21 in a large unselected cohort with genetic confirmation. We hypothesized that performance in all risk groups is equivalent to published studies without genetic confirmation. Unselected pregnant women who had SNP-based cfDNA for T13, T18 and T21 were recruited at 21 centers in 6 countries. Genetic confirmation was obtained for all pregnancies from prenatal or newborn DNA samples. Sensitivity, specificity, PPV and test failure (no call) rates were calculated and compared between women at low and high prior risk for aneuploidy. High risk was defined as having positive serum screening, nuchal translucency >3mm, major fetal anomaly on ultrasound or >35 years old with no prior screening. Performance of an updated cfDNA algorithm, blinded to pregnancy outcome, was also assessed. Of 20,887 women enrolled, 18,496 (88.5%) had genetic confirmation and comprised the study cohort. Mean gestational age at enrollment was 13.2 weeks. 133 (0.72%) cases of trisomy were identified: n=100 T21, n=18 T18 and n=15 T13. 18,205 women received a cfDNA result while 291 (1.57%) were no call. Screen positive rate was lower in low vs high risk cases (0.26% vs 2.1%, p<.0001). Sensitivity and specificity were similar between risk groups. (Table) PPV for all trisomies combined was lower in low risk women (74.3% vs 94.1%, p=.003), while it was not different for any specific trisomy. Performance of the updated algorithm was not different than in the prospective cohort but had lower no call rate after both the first (1.46% vs. 3.3%, p<.0001) and second draw (0.60% vs 1.57%, p<.0001). Based on genetic confirmation in an unselected cohort of 18,496 pregnancies, SNP-based cfDNA high sensitivity and specificity for the common trisomies was corroborated in all risk groups. The PPVs in women that are low risk for aneuploidy are >50% for all trisomies. An updated algorithm decreased the no-call rate while maintaining performance.