Evaluation of non‐invasive prenatal testing to detect chromosomal aberrations in a Chinese cohort

Abstract The aim of this study was to evaluate the clinical feasibility of non‐invasive prenatal testing (NIPT) to detect foetal copy number variations (CNVs). Next‐generation sequencing for detecting foetal copy number variations (CNVs) was performed on the collected samples from 161 pregnancies with ultrasound anomalies and negative NIPT results for aneuploidy. The performance of NIPT for detecting chromosome aberrations was calculated. The sensitivity and specificity of NIPT for detecting CNVs > 1 Mb were 83.33% and 99.34%; the PPV and negative predictive rate (NPV) were 90.91% and 98.68%. Non‐invasive prenatal testing can be performed to detect chromosomal aberrations in first trimester with high performance for CNVs, and occasional discordant cases are unavoidable.


| INTRODUC TI ON
Non-invasive prenatal testing (NIPT) with next-generation sequencing (NGS) on cell-free foetal DNA (cffDNA) in maternal plasma has been widely used for foetal aneuploidy screening because it can reduce unnecessary invasive procedures that may result in miscarriage or intrauterine infection. 1 Traditional prenatal diagnosis strategies have been altered since NIPT is considered as a primary prenatal screening method due to the high accuracy. [2][3][4] Copy number variation sequencing (CNV-seq) with higher resolution has been applied to diagnose foetal submicroscopic chromosome abnormalities invasively. 5,6 Accordingly, non-invasive detection of foetal CNVs was encouraged by the successful application of NGS on cffDNA in screening for aneuploidy. Li et al 7  for CNVs can be adopted as the first-tier prenatal approach. But the controversy on the utility of NIPT for detecting foetal CNVs is still remained, 12 and more evidences are needed to explore the clinical feasibility. In our study, we evaluated the screening effectiveness of NIPT to detect CNVs in pregnancies with abnormal ultrasound findings, which are common indications for invasive genetic testing.

| Study design
We gathered the pregnancies with ultrasound anomalies and negative NIPT results for aneuploidy. Maternal blood samples and foetal samples such as amniotic fluid or foetal tissues were collected. Noninvasive prenatal testing was performed on maternal plasma for detecting foetal CNVs, and diagnosing CNVs in foetal samples was by CNV sequencing (CNV-seq). The resolution of CNVs was no less than 1 Mb, and the pathogenicities of identified CNVs were evaluated following American College of Medical Genetics and Genomics (ACMG) guidelines. The karyotypes of foetuses and their parents were obtained by G-banding karyotyping. All participants were offered genetic counselling and gave informed consent. This study was approved by the institutional review board of Shengjing Hospital.

| Detecting foetal CNVs in maternal plasma with next-generation sequencing (NGS)
Five mL of maternal peripheral blood was collected in an EDTAcontaining Vacutainer tube (Becton Dickenson) and centrifuged at 1600 g for 10 minutes at 4°C The plasma and white blood cells were transferred to microcentrifuge tubes separately, and the plasma was centrifuged again at 16 000 g for 10 minutes, then

| CNV sequencing in foetal gDNA
Genomic DNA (gDNA) was extracted from amniotic fluid or foetal tissues using the Genomic DNA extraction kit (QIAGEN); then the gDNA was purified using the Purification DNA kit (Zymo Research).
The concentration of gDNA was quantitated using the Invitrogen Qubit 2.0 (ThermoFisher Scientific), with the standard of greater than 8 ng/μL. The library was constructed and purified, using the same methods which were performed in maternal plasma. Then DNA libraries were quantitated using the Kapa SYBR fast qPCR kit from Kapa Biosystems, with the standard of greater than 25 nmol/L.

The quantitated DNA libraries were pooled and loaded into Illumina
Nextseq CN500 flow cells, then were sequenced using the singleended 36 bp sequencing protocol. No less than 2.5 million unique reads were analysed by the software provided by Berry Genomics.
By comparing the CNVs results in Table 1, a false positive case (SJ023) and two false negative cases (SJ015 and SJ028) were confirmed ( Figure 1). Of the 10 true positive CNVs, the locations were coincident, but the sizes were slightly different. When the CNV-seq results were compared with corresponding karyotypes, 2 CNVs (SJ047 and SJ022) did not appear to be direct duplication or deletion at telomeres in the karyotypes. A 3.4 Mb deletion should not have been detected by karyotyping in case SJ103. Through the available parental karyotypes, we were able to trace five CNVs from derivative chromosomes in 3 cases, and 5 de novo CNVs. As described in Table 2, the performance of NIPT for detecting CNVs was calculated depending on the standards of CNV-seq results, achieving the sensitivity of 83.33%, the specificity of 99.34% and the PPV (positive predictive rate) of 90.91%. The sensitivity and specificity for CNVs between 1 Mb and 5 Mb were higher than those for CNVs ≥ 5 Mb.
As shown in Table 1, among the 12 CNVs detected by CNV-seq, 11 CNVs were pathogenic, and known CNVs syndromes were involved in 3 cases.

| D ISCUSS I ON
Cell-free foetal DNA (cffDNA) can definitely lead to occasional discordant NIPT results as the detection target of NIPT. Increasing evidences have shown that cffDNA in maternal plasma is derived predominantly from placental trophoblastic cells, 13 so NIPT reflects the genetic information of the placenta, not foetus. [14][15][16] Confined placental mosaicism (CPM) indicates that chromosome aberrations only exist in placenta and not in foetus, which is widely accepted as a cause of false positive NIPT results. 17 Confined placental mosaicism is also reportedly relevant to intrauterine growth restriction (IUGR) and an increased risk of perinatal morbidity and mortality. 18 Coincidentally, the later ultrasound findings in case SJ023 showed IUGR, which added evidence to our speculation. Maternal background is believed as another contributing factor to affect NIPT performance. In this case, maternal interference was excluded by maternal DNA sequencing. Placental mosaicism may have led to the false negative result in case SJ015, which meant that the chromosomal constitutions of the foetus and placenta were both abnormal.
The degree of mosaicism could vary greatly in different regions of placental tissue. If more cffDNA from the lower level mosaic region was released into the maternal plasma, the NIPT result would be negative. 19 As a biological factor, the mosaicism either in foetus or in placenta, is a limitation of NIPT accuracy that cannot be overcome. 11 CPM and placental mosaicism were usually used to explain the discordant result of detecting aneuploidy in previous studies, our theoretical speculation was hard to be confirmed due to the unavailable placental tissue. Further research will focus on finding the difference in the chromosomal constitutions of the foetus, placenta and the different placental tissue regions.
Low foetal fraction is a common reason for false negative NIPT results. 20  for aneuploidy, if the family history was unknown in advance. In addition, unlike Down's syndrome, the risk of pathogenic CNVs is not closely related to maternal age, and even younger women are liable to suffer from microdeletion in foetus. 24 If the detection range is expanded from aneuploidy to CNVs in first trimester, the detection rate of chromosome abnormalities will be higher, and the benefit population will be wider, especially for the pregnancies

ACK N OWLED G EM ENTS
This work was supported by the grants from National Key R&D Program of China. (2016YFC1000700, 2016YFC1000702)

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflicts of interest.