A retrospective study of noninvasive prenatal testing for chromosome aneuploidies and subchromosomal copy number variations in 24359 single pregnancies

Background: With the development of whole-genome sequencing, small subchromosomal deletions and duplications could be found by non-invasive prenatal testing(NIPT). Our study is aimed to review the ecacy of NIPT as a screening test for aneuploidies and subchromosomal copy number variations (CNVs) in 24359 single pregnancies. Methods: A total of 24359 single pregnancies with different clinical features were retrospectively analyzed. Pathogenicity of abnormal NIPT results were assessed according to American College of Medical Genetics and Genomics(ACMG). Chromosome aneuploidies and subchromosomal CNVs were conrmed by karyotyping and chromosomal microarray analysis(CMA). Results: A total of 442 pregnancies (442/24359,1.9%) were with abnormal NIPT results. The positive predictive value (PPV) for trisomy 21(T21), trisomy 18 (T18), trisomy 13 (T13), and sex chromosome aneuploidies (SCAs) was 84.8%, 54.2%, 11.1% an 40.5% respectively. The PPV for subchromosomal CNVs was 59.0% (46/78). The clinical information, prenatal diagnosis results and follow-up results of 46 true positive cases, 6 cases with subchromosomal CNVs inconsistent with NIPT and 1 case of false negative were also demonstrated in detail. Conclusion: Our data have potential signicance in demonstrating the signicance of NIPT not only for common whole chromosome aneuploidies but also for subchromosomal CNV. Besides, the clinical information, prenatal diagnosis results and follow-up results of 52 cases with subchromosomal CNV and 1 case of false negative would provide important guidance for genetic counseling.


An overview of clinical data
Among the 24367 cases undergoing NIPT, 8 cases were not eligible for the next analysis due to the low concentration of fetal DNA, so the remaining 24359 cases were under-analyzed in the present study.

NIPT Results For Other Chromosome Aneuploidie
In 442 pregnancy with abnormal NIPT, there were 74 cases of other chromosome aneuploidie including 38 cases of trisomy 7 (T7), 9 cases of trisomy 20 (T20), 12 case of trisomy 16(T16)and so on. One pregancy with the result of T16 further chosen karyotyping and microarray for prenatal diagnosis. The result of karyotyping was normal but microarray showed that fetus was not only suspected to be a low proportion (< 20%) of trisomy 16 but also there was a loss of heterozygosity (LOH) about 15.3 Mb fragment in the 16q23.1q24.3 segment. LOH means that chromosomes of a homologous pair are inherited from the same parent, which increases the risk of imprinting disorder. Study has reported that LOH were often caused by trisomic rescue and one third of these trisomic rescues lead to a LOH [12] . Currently, there is no evidence for imprinted genes encoded by 16q23.1q24.3.
The imprinted genes has already been proven for chromosomes 6 We also analyzed indications of 78 prenatal diagnosis cases with subchromosomal CNVs, we found that 20 of 37 cases with high risks of serum screening were true positive. Interestingly, 20 of 35 cases without any indications were true positive (Table 5). Among 52 abnormal cases, 31 cases were correlated to microdeletion or microduplication syndromes suggesting that NIPT may be an important method to nd potential birth defect. Moreover, we compared fetal free DNA concentration between 46 true positive cases and 32 false positive cases. The result showed that true positive cases had higher fetal fraction that false positive cases (p = 0.013) (Fig. 2). In addition to fetal concentration, abnormal maternal chromosome also increases the false positive rate of the fetal chromosome. The clinical information, prenatal diagnosis results and follow-up results of 52 abnormal cases are shown in the table 6 and table  7 (sorted by CNV size) The chromosome region 22q11.2 has long been recognized as a hotspot for genomic rearrangement. This region contains several large segmental duplication/low copy repeats (LCRs) that function as mediators of non-allelic homologous recombination (NAHR) and predispose the genomic region to chromosomal rearrangements [13] . 22q11.2 deletion syndrome and 22q11.2 duplication syndrome are caused by genomic rearrangement of 22q11.2. Some researchers believe that the incidence of 22q11.2 deletion syndrome is higher than 22q11.2 duplication syndrome, with an estimated frequency of 1 in 4,000-6,000 live births [14] . However, retrospective analysis of our study suggested that 22q11.2 duplication syndrome occured more frequently. The clinical manifestation of ve fetuses with 22q11.2 duplication syndrome and one with 22q11.2 deletion syndrome were analyzed in detail follwed (table 8).

Case 24 And 29
Both of theses two cases were con rmed to had 22q11.21 microduplication and their parents did not accept CMA to verify the source this mutation. Fortunately, these two cases were clinically healthy after birth. It has been reported that the duplication of the 22q11.2 can cause different degrees of mental retardation, learning disabilities and growth retardation. However, there are also cases without any clinical phenotype, which indicates that 22q11.21 microduplication may have incomplete manifestation or performance differences [15] . Therefore, it is di cult to determine the pathogenicity of 22q11.21 microduplication. microduplication were inherited from their parents with normal phenotypes [16] .

Case 37
Case 37 was con rmed to had 22q11.21q11.23 microduplication by CMA and also con rmed to be clinically healthy after birth.

Case 34
Case 34 was con rmed to have the loss of heterozygosity in 22q12.1q12.2. clinical phenotypes was unavalibale. Fetal mother were also advisesed to have a array comparative genomic hybridization because of her mental retardation. CMA analysis showed her with 22q11.23 microduplication.
Case 35 The 22q11.21 deletion syndrome(DiGeorge syndrome/velocardiofacial syndrome) is characterized by congenital heart defects, immune de ciency, transient neonatal hypocalcemia, velopharyngeal insu ciency and a distinctive facial appearance but also by learning disabilities, behavioral anomalies and hypotonia. The majority of the 22q11.2 deletions are de novo [17] . Case 35 is con rmed to have a 22q11.21 deletion syndrome with hoarseness, atrial septal defect, patent ductus arteriosus and myocardial injury after birth. The origin of 22q11.21 deletion in case 35 was unknown.  [18] . In our study, case 46 chosen to perform prenatal diagnosis for further con rmation and the result showed that the fetus was female and this mutation is inherited from her normal mother. The NIPT result of case 53 with high risk of serum screening was normal. System structure screening (other hospital) in 24 week showed fetal growth retardation. The fetal was diagnosed as neonatal pneumonia, low weight, congenital heart disease and hyperbilirubinemia after birth. Karyotype detection of peripheral blood showed 46,XX,del(4)(p4). Thus, we improved the experimental method for better cffDNA enrichment and retrospectively tested case 53 [19] . The results showed increased cffDNA fraction from 6.5-15.1% and a 34Mbp deletion in 4p16.3-p15.1 region, which is related with Wolf-Hirschhorn syndrome (WHS).

Discussion
In recent years, NIPT has been widely used in clinical practice for prenatal screening of T21, T18 and T13.
While screening for the whole genome, NIPT also found some chromosomal deletions and duplications at the submicroscopic level. The researches about the feasibility of using NIPT for screening of chromosome CNVs have been performed. In 2011, Peters et al reported for the rst time that wholegenome sequencing of fetal free DNA from plasma of pregnant women detected a deletion of approximately 4.2 Mb in the p11.22p12.1 region with a large amount of sequencing data (243M) [20] .
Since then, many scholars have detected chromosomal microdeletions and microduplications by NIPT. These researches all have high sequencing depth and the detection resolution can be as low as 300 kb [21][22] .
In this study, we reviewed the use of NIPT in the context of screening for common chromosome aneuploidies as well as subchromosomal CNV within a cohort of 23373 single pregnancies. The PPV for T21, T18, T13 and SCAs was 84.8%, 54.2%, 11.1% an 40.5% respectively. At the same time, we also found 125 positive cases with subchromosomal CNVs. 78 of 125 chosen for further detection, with 46 (59.0%, 46/78) true-positive cases and 32 false positive cases(41.0%, 32/78) validated by chromosome karyotyping and CMA. Previous researches have shown that the extent of chromosomal abnormalities presented in plasma of women with aneuploidy pregnancies is linearly correlated with the cfDNA fraction, thus the test accuracy of NIPT largely relies on the fetal fraction [23] . In our study, the content of fetal free DNA fragments in the plasma of pregnant women varied greatly and can reached 35% and the cell-free fetal DNA between the true positive cases and false positive cases showed signi cant difference(p < 0.05). Possible reasons causing the false positives also include abnormal maternal chromosome and con ned placental mosaicism(CPM) were also reported to cause the false positive cases.
chromosomal microdeletions and microduplications have become increasingly recognized as signi cant contributors to human diseases [24] , which are present in approximately 1.7% of structurally normal pregnancies [25] . Other studies have suggested that PPV of microdeletions detected by NIPT have not demonstrated a su ciently low false-positive rate to be deemed practical [26] . However, for our studies, NIPT had higher PPV (59%) for microdeletions and microduplications. Among the 46 true positive cases and 7 abnormal false positive cases, 31 cases were correlated to microdeletion or microduplication syndromes with 6 cases inherited from parents, 3 cases de novo and other 22 cases unavailable. Early detection of pathogenic and potentially pathogenic CNVs by NIPT has good bene t in prenatal screening.
22q11.2 microduplication is the most frequent in our research. The phenotype of the ve patients with 22q11.2 microduplications are diverse, with symptoms ranging from being normal to mental retardation and congenital heart malformation. 22q11.2 microdeletions is the second most common chromosomal abnormality secondary to Down syndrome [27] . However, the occurrence of 22q11.2 microduplications is more frequent than 22q11.2 microdeletion in our study, which is contrary to previous research conclusion. The rare occurrence of 22q11.2 microduplication cases may be explained by the absence of a de ned phenotype and incomplete penetrance [28] .
Studies have demonstrated that there is a small chance of a false negative result for NIPT [29] . In our study, there is a false negative case in 79 validated NIPT. The most common factor associated with these false negative results is the low fetal fraction, which are often affected by maternal weight, gestational age and extraction method [30][31] . In our research, extraction method for cffDNA enrichment was the main reason for the false negative cases. Therefore, improved extraction method for elevating fetal fraction were immediately used in December 2018, which may be the potential reason for improved the overall performance of NIPT and high PPV in this research. Faas BH et al in 2012 clari ed that cell free fetal DNA in the maternal plasma originates from cytotrophoblastic cells derived from trophoblast of the blastocyst [32] . The karyotypes of cytotrophoblast and fetus may be different due to fetus are derived from the inner cell mass (ICM) of the blastocys [29] . Other reasons for false negative results may be CPM and maternal mosaicism [33][34][35] .

Conclusions
This study demonstrates that the PPV for T21, T18, T13 and SCAs was 85%, 54%, 11% and 41% respectively. The PPV for CNVs was 59.0%. The PPV for CNVs < 10 Mb is 70.6% and for ≥ 10 Mb 37.1% respectively. Our data have potential signi cance in demonstrating the usefulness of NIPT not only for common whole chromosome aneuploidies but also for CNVs.

Patients
From 2015 to July 2019, 24359 pregnant women opted for NIPT to screen fetal chromosome aneuploidies. Informed written consent was obtained from all pregnant women who agreed to receive NIPT. Pregnancies with high risks were divided into advanced maternal age, ultrasound abnormalities, poor fertility history, positive serum screening, and other groups in this study.

NIPT sequencing
Maternal peripheral blood (5 ml) was collected in an ethylenediaminetetraacetic acid (EDTA) tube. The blood sample was stored at 4˚C immediately after collection. Afterwards, cfDNA extraction, library construction, quality control, and pooling were performed according to the JingXin Fetal Chromosome Aneuploidy (T21, T18, T13) Testing Kits (CFDA registration permit No. 0153400300). Sequencing reads were ltered and aligned to the human reference genome (hg19). Combined GC correction and Z-score testing methods were used to identify fetal autosomal aneuploidies. A cut off value of Z-score > 3 was used to determine whether the ratio of the chromosomes was increased. Here, each chromosome with an absolute value of the Z-score greater than 3 was marked with chromosome aneuploidies or microdeletions/ microduplications.

Chromosome karyotype analysis
Banding cytogenetics was performed on G-banded metaphase chromosomes of cultured peripheral blood lymphocytes using routine techniques.  Figure 1 The positive rate of NIPT for aneuploidy and subchromosomal CNV increases with maternal age and the number of positive cases in the ve age groups.