Evaluation of the clinical utility of extended non-invasive prenatal testing in the detection of chromosomal aneuploidy and microdeletion/microduplication

Birth defects are the main cause of neonatal deaths and disabilities due mainly to congenital structural abnormalities or genetic diseases, leading to severe clinical symptoms or complications [1]. Among these, 80% of birth defects are caused by genetic factors alone or in synergy [2]. For example, trisomy 21, 18 and 13 syndromes, especially trisomy 21 (Down syndrome), are the most common chromosomal abnormalities found in neonatal birth defects. Some children with trisomy 21 have no obvious deformity phenotype in prenatal ultrasonography and cannot be screened effectively by the technique. Although screening for Down syndrome is now free and widely available to the public, there remains a high rate of missed detection [3]. Sex chromosome aneuploidy (SCA) is a relatively common SCA disorder with twice the incidence of trisomy 21, affecting approximately 1 in 4000 newborns. In addition, chromosomal copy number variation (CNV) is an important factor affecting human disease and phenotypic variation [4]. Some CNVs can lead to severe microdeletion and microduplication syndrome (MMS) [5], while some MMSs are even more likely to occur than Down syndrome. Recent studies have shown that the incidence of MMS in the foetuses of chromosomally normal pregnant women is as high as 1–1.7% [6]. For example, Williams syndrome, DiGeorge syndrome and Prader–Willi syndrome are common high-incidence syndromes. In addition, MMS causes about 12% of unexplained intellectual disabilities, various deformities and developmental delays [7].

Effective prenatal screening and prenatal diagnosis techniques are particularly important for preventing birth defects. Cytogenetic karyotype analysis has always been used as the gold standard for prenatal diagnosis. It allows the visual observation of abnormalities in terms of chromosome number and large fragment structure, but it cannot fully detect microdeletions, microduplications or uniparental diploidy. As the two main CNV detection techniques, chromosomal microarray analysis (CMA) and low-depth CNV sequencing provide high-resolution analysis of microdeletions and microduplications [8, 9]. In particular, CMA includes array-based comparative genomic hybridisation and a single-nucleotide polymorphism (SNP) array, which is a high-resolution genome-wide screening technology that can be used to analyse both chromosomal aneuploidy and CNVs [10]. Chromosomal array analysis has become the main molecular detection technology in addition to karyotyping. Chromosomal microarray analysis (CMA) as a diagnostic technique.

With the development of next-generation sequencing technology, non-invasive prenatal testing (NIPT) technology based on cell-free foetal DNA (cffDNA) in maternal peripheral blood has been shown to outperform traditional screening methods and is rapidly becoming the primary screening test in clinical practice [11]. This method has been validated in multiple clinical cohorts, confirming that NIPT is highly sensitive and specific for T13, T18 and T21 aneuploidy risk screening [12]. Other chromosomal abnormalities can be revealed by the low-coverage whole-genome sequencing of maternal and placental DNA fragments [13]. The antenatal extended NIPT (NIPT-PLUS) technique is gradually being applied clinically. Compared with NIPT detection technology, NIPT-PLUS increases the output of sequencing data. However, due to the low incidence of individuals with chromosomal abnormalities and the use of sequencing platforms, data analysis parameters and types of chromosomal abnormalities in clinical units, the accuracy of NIPT-PLUS for the detection of RATs and CNVs is still unclear, limiting its wide clinical application. Extended NIPT (NIPT-PLUS) technique as a screening technique.

In recent years, most research has focused on retrospective studies of large-sample NIPT assays, which focus on positive predictive values (PPVs) and accuracy analyses of aneuploidy. However, few studies have analysed the detection rate and missed detection rate of NIPT-PLUS for detecting chromosomal aneuploidy and CNV. In this study, we compared 452 pregnant women with a prenatal diagnosis who underwent CMA and NIPT-PLUS to compare and analyse the results of the two testing methods. The detection rate and optimal detection range of NIPT-PLUS for common chromosomal aneuploidy, SCA, RAT and MMS were evaluated, respectively. Simultaneously, the key factors causing false-negative and false-positive NIPT results were analysed, and the clinical utility of the method was evaluated further.

Most retrospective research articles have mainly analysed the PPVs of NIPT-PLUS for aneuploidy and CNV, and few have analysed the sensitivity and miss rate of NIPT-PLUS for aneuploidy and CNV. The sensitivity and miss rates are important to enable pregnant women to correctly understand this means of testing and genetic counselling. Our data show that through the parallel detection of CMA and NIPT-PLUS, the optimal appropriate range for NIPT-PLUS detection and the sensitivity and miss rate for CNV detection can be obtained. A study by Maya et al. [19] found that the residual risk of clinically significant CNVs in pregnancies without structural sonographic anomalies is appreciable and depends on the extent of non-invasive prenatal screening and maternal age. However, it is not clear what the detection rate of NIPT-PLUS is in pregnant women with indications for a prenatal diagnosis, such as ultrasound abnormalities. In our study, the selected samples were from pregnant women with indications for a prenatal diagnosis. Indications for a prenatal diagnosis were advanced maternal age (15.93%), abnormal Down syndrome screening results (14.16%), structural anomalies on ultrasonography (63.27%) and other indications (13.05%).

Studies have shown that compared with traditional screening methods, NIPT-PLUS has higher sensitivity and specificity in detecting common foetal chromosomal aneuploidies [20]. The PPV of T21, T18 and T13 screening was 80–94%, 60–85% and 20–63%, respectively [21, 22]. This study did not conduct a retrospective analysis but conducted a direct prenatal diagnosis based on indications for a prenatal diagnosis. This is not the same advantage as in the other studies. Pregnant women who chose CMA testing were tested directly by NIPT-PLUS to evaluate the detection efficiency of the technique for chromosomal aneuploidy and CNVs. Nevertheless, the analysis of NIPT-PLUS data from this study showed that the PPVs of NIPT-PLUS for target chromosomal aneuploidy T21, T18 and T13 were 90.32%, 87.50% and 25.00%, respectively. This is consistent with the PPVs of common trisomy reported in other studies. The detection rates of T21, T18 and T13 were 96.55%, 100% and 100%, respectively, which was also consistent with the rates in the published literature [23].

The literature reports that the PPV of NIPT-PLUS for SCA and MMS is not as high as that for commonly seen T21, T18 and T13 [24, 25]. Studies have shown that the PPV of SCA detected by NIPT-PLUS is about 46.7% [26]. The data of this study showed that the PPV of SCA detected by NIPT-PLUS was 88.89%, which could be related to the fact that this study was not a large-scale retrospective study but rather, it observed a specific obstetric population. However, this study provides good reference data for the detection rate of SCA by NIPT-PLUS. From our data, it can be seen that the sensitivity of NIPT-PLUS to 47, XXY; 47, XXX and 47, XYY was 100%, but the sensitivity to 45, XO was only 90%, which is also consistent with the conclusions reported elsewhere, i.e. NIPT-PLUS can detect XXX, XXY and XYY better than 45, XO [22, 27]. Moreover, the comparison with the results of the prenatal diagnoses shows that NIPT-PLUS could not accurately indicate partial mosaicism 45, XO; partial X chromosome deletion or the X isoarm chromosome, although it could be used as a first-line screening technology, which is enough to prompt pregnant women and clinicians to consider invasive prenatal diagnosis. By analysing the case of 45, XO, it can also be seen that CMA technology has disadvantages in detecting low-ratio mosaicism and chromosomal translocation. The FISH technique and cytogenetic karyotyping are better for detecting chimerism, and cytogenetic karyotyping has an irreplaceable role in detecting chromosomal translocation. Therefore, for abnormal results of aneuploidy detected by NIPT-PLUS, in addition to a molecular-level diagnosis, cytogenetic karyotype analysis may be also an indispensable test.

Furthermore, RAT is also relatively common in the NIPT-PLUS test, but most of the time it is considered to be a false-positive result. According to the genome content on the chromosome, the types of RAT mainly reported by NIPT-PLUS prenatal screening in our laboratory were T9, T15, T16, T20 and T22. According to the present study’s data, the sensitivity of NIPT-PLUS for RAT detection reached as high as 80%, but there were many false positives, and the PPV was only 50%. This may be related to the non-invasive false positives caused by confined placental mosaicism [28]. The cffDNA in maternal peripheral blood detected by NIPT-PLUS is believed to be mainly from the apoptosis of placental trophoblast cells [29], which cannot fully represent the foetus. The chromosomal abnormality that occurs only in the placenta and not in the foetus is called restricted placental mosaicism [30], and studies have shown that its incidence is about 1–2% [31]. Because it is often assumed that foetuses with rare chromosomal trisomy are generally miscarried or stop developing at an early stage, there are insufficient data to support the detection rate and PPV of RAT using the NIPT-PLUS technique.

Compared with ordinary NIPT, NIPT-PLUS increases the volume of effective data and can detect more microdeletions and microduplications [32]. The literature shows that the PPV of NIPT-PLUS for CNV is 60–100% [33], and the latest literature shows that the PPV is about 50% [26]. This study is different from the previous retrospective analysis of the NIPT-PLUS test. Instead, the deletions detected by CMA were larger than 500 kb, fragments with repeats larger than 1000 kb were selected, and NIPT-PLUS detection was performed simultaneously to evaluate the detection effect of NIPT-PLUS on CNV. The data showed that the sensitivity of NIPT-PLUS to microdeletion and microduplication increased with the increase of CNV fragment size. For CNVs < 3 Mb, the detection rate of NIPT-PLUS was only 49.69%, of which 78.75% were suggested to be from the mother, which means that if the mother does not carry these small CNV fragments, it is very likely that they will be missed. Among the fragments > 10 M, NIPT-PLUS did not suggest that it was from the mother, but the missed detection rate reached 23.08%. On the one hand, the missed detection of this large deletion duplication may be related to the previously described false-negative and false-positive restricted placental mosaicism that causes SCA [34, 35]. On the other hand, this finding may be related to the low-depth sequencing of NIPT and limited data volume. Our data show that the detection rate of NIPT-PLUS for 5–10-Mb CNVs reached 100%, which is consistent with previously published data detection results [36]. In their study, Pei et al. [36] also showed that the PPV of CNVs < 10 Mb detected by NIPT was significantly higher than that of CNVs > 10 Mb. However, the detection rate of CNV detected by NIPT-PLUS in this study is more worthy of consideration for clinical application.

Similarly, we conducted statistical analyses on the detection rate of MMS commonly found in the clinical detection range of NIPT-PLUS. The data revealed that the detection rate of NIPT-PLUS for syndromes < 5 Mb, 1p36 microdeletion, cri-du-chat syndrome and chromosome 9p deletion (duplication) syndrome was 100%, while the detection rate for the most common DiGeorge syndrome reached 92.31%, and the detection rate for Angelman/Prader–Willi syndrome was 66.60%. Of course, these results need to be supported by large-sample data. However, the premise is that the size of the syndrome detected by CMA in this study was mainly 3–5 Mb, the detection rate of 15 cases of RACD syndrome (< 2 Mb) was 73.33%, and 8 of the 11 cases detected were suggestive of maternal origin. The detection rate of Williams–Beuren syndrome was the lowest, at about 33.3%. The current retrospective study showed that the PPV of NIPT-PLUS for MMS detection was 32% (CNVs ≥ 10 Mb) and 19% (CNVs < 10 Mb), and the PPV for common MMS was 93% (DiGeorge), 68% (22q11 0.22 microreplication), 75% (Angelman/Prader–Willi) and 50% (cri du chat) [23]. However, there are few reports on the detection rate of NIPT-PLUS for MMS detection. The detection rate and missed detection rate suggest the necessity and importance of NIPT-PLUS as a first-line clinical screening technique, with follow-up genetic counselling. It is suggested that although the PPV of CNV detection by NIPT-PLUS reached 50%, which was already higher than that of common prenatal screening techniques, it is also recognised that this technique has a missed detection rate for CNV and an optimal range for the detection of CNV fragment size. As mentioned above, we believe that NIPT-PLUS has better detection efficiency for CNVs > 5 Mb and is worthy of further clinical promotion, while it is not recommended for screening CNVs < 3 Mb due to its high missed detection rate. For CNVs between 3 and 5 Mb, the need to further inform clinics and pregnant women can be determined based on the pathogenic syndrome screened.

This study’s results contain cases that suggest that the results of CMA and NIPT-PLUS are not completely consistent. Among them, 4 cases of CMA suggested ROH and uniparental disomy, 3 of which were non-invasive and showed no abnormality, and 1 case of 16 ROH non-invasive suggested T16. In two other cases, the CMA suggested 47, XYY, and the NIPT-PLUS suggested 45, XO. The remaining 15 cases were mainly false positives for aneuploidy or large-fragment CNVs detected by NIPT-PLUS and false negatives for CNVs < 3 Mb that were missed. The influencing factors of this inconsistent result are related mainly to confined placental mosaicism (discussed above) and to the fact that most of the cffDNA detected by NIPT-PLUS was maternal fat metabolism DNA.

In the enrolled cases, NIPT-PLUS aneuploidy false negatives were also detected, which have also been reported in many studies. A recent review article suggests that the concentration of cffDNA in the peripheral blood of pregnant women, maternal characteristics, foetal–placental characteristics, experimental factors and calculation methods can all affect NIPT-PLUS test results, resulting in false negatives or false positives [37]. It seems that false positives are more acceptable than false negatives because the birth of an aneuploid foetus greatly impacts the family. In this study’s data, a total of 3 cases of false-negative aneuploidy were detected, and 1 pregnant woman underwent CMA testing due to NT thickening; the result indicated that it was 45, XO, and the NIPT-PLUS results both indicated that the risk was low. The post-induction placenta of the pregnant woman was obtained for verification, and the MLPA results of the placental subsurface and maternal surface showed a very low proportion of 45, XO mosaicism, while the umbilical cord showed complete 45, XO. One case was chimeric 45, XO due to the abnormal ultrasound detection of CMA, and both NIPT-PLUS tests were low risk. Another false-negative pregnant woman with trisomy 21 underwent CMA testing due to nasal bone deletion, and the result suggested chimeric T21. The chimeric ratio of FISH verification was 13.57% (19/140).

Non-invasive prenatal testing-PLUS suggests low risk. Hints of foetal or placental mosaicism and confined placental mosaicism are the most critical factors leading to false negatives in NIPT-PLUS. This is also consistent with the description in previous literature [36]. Similarly, NIPT-plus false positives are also common. Copy number variation false-positive results are due mostly to maternal chromosome imbalance, while false positives of aneuploidy are associated with foetal or placental mosaicism and restricted placental mosaicism, twins and single pregnancies, maternal benign or malignant fibroids and Z value data analysis [28, 38].

Few studies provide a detailed description of the detection range, sensitivity or miss rate of NIPT-PLUS. Our data suggest that NIPT-PLUS has a certain miss rate for the detection of chromosome aneuploidy and CNV, especially for small-fragment CNV. Additionally, few studies have conducted specific data analyses on the detection range of NIPT-PLUS, while our data were synchronised with CMA experiments according to the CNV size, obtaining NIPT-PLUS as the best detection range for first-line screening. These data show that NIPT-PLUS, as the main screening technology in first-line clinical practice, has very high sensitivity for aneuploidy detection. However, cytogenetic karyotype analysis may also be needed in some situations. The NIPT-PLUS technique has low sensitivity for CNV detection, especially at < 3 Mb, and it has a relatively high missed detection rate. Therefore, it is not recommended for clinical use as a first-line screening standard. For CNVs > 5 Mb, the sensitivity increases and the missed detection rate decreases. Accordingly, NIPT-PLUS can be performed appropriately in clinical testing, but it needs to be closely combined with comprehensive perinatal care, such as ultrasound, to avoid the occurrence of birth defects. In addition, the test results suggest that confined placental mosaicism and foetal mosaicism are the key factors causing false negatives in NIPT-PLUS. Maternal chromosomal abnormalities and confined placental mosaicism are key contributors to false positives, so appropriate genetic counselling is especially important for pregnant women before and after NIPT-PLUS testing.

This study has some limitations. First, the sample size of this study was small, which did not allow the sensitivity of NIPT-PLUS to be calculated for common syndromes. To obtain better statistical reliability, more studies with greater numbers of specimens are required. Second, there was a lack of familial genetic information. During the prenatal diagnoses, we did not conduct a family survey to determine whether the CNV was inherited from the mother in the foetus. In future studies, we will explore the clinical effects of NIPT-PLUS combined with other diagnostic methods.