Positive predictive value of non-invasive prenatal screening for fetal chromosome disorders using cell-free DNA in maternal serum: independent clinical experience of a tertiary referral center.
Published in BMC Medicine June 2015
Non-invasive prenatal screening (NIPS) for fetal chromosome abnormalities using cell-free deoxyribonucleic acid (cfDNA) in maternal serum has significantly influenced prenatal diagnosis of fetal aneuploidies since becoming clinically available in the fall of 2011. High sensitivity and specificity have been reported in multiple publications, nearly all of which have been sponsored by the commercial performing laboratories. Once results are returned, positive and negative predictive values (PPVs, NPVs) are the performance metrics most relevant to clinical management. The purpose of this report is to present independent data on the PPVs of NIPS in actual clinical practice.
Charts were retrospectively reviewed for patients who had NIPS and were seen March 2012 to December 2013 in a tertiary academic referral center. NIPS results were compared to diagnostic genetic test results, fetal ultrasound results, and clinical phenotype/outcomes. The PPV was calculated using standard epidemiological methods. Correlation between screen results and both maternal age at delivery and gestational age at time of screening was assessed using Wilcoxon’s rank sum test.
Of 632 patients undergoing NIPS, 92 % of tests were performed in one of the four major commercial laboratories offering testing. However, all four laboratories are represented in both the normal and abnormal results groups. There were 55 abnormal NIPS results. Forty-one of 55 abnormal NIPS results were concordant with abnormal fetal outcomes, 12 were discordant, and 2 were undetermined. The PPV for all conditions included in the screen was 77.4 % (95 % CI, 63.4 – 87.3). Of 578 patients with normal NIPS results, normal pregnancy outcome was confirmed for 156 (27 %) patients. This incomplete follow-up of normal NIPS results does not affect PPV calculations, but it did preclude calculations of sensitivity, specificity, and NPV. Maternal age at delivery was significantly lower for patients with abnormal discordant results, compared to patients with abnormal concordant results (P = 0.034). Gestational age at time of screening was not associated with concordance of screen results (P = 0.722).
The experience of using NIPS in clinical practice confirms that abnormal results cannot be considered diagnostic. Pre-test counseling should emphasize this. Diagnostic genetic testing should always be offered following abnormal NIPS results.