Chromosomal microarray analysis for pregnancies with abnormal maternal serum screening who undergo invasive prenatal testing

Abstract Recently, chromosomal microarray analysis (CMA) has been implemented as a first‐tier test in pregnancies with ultrasound anomalies. However, its application for pregnancies with abnormal maternal serum screening (AMSS) only is not widespread. This study evaluated the value of CMA compared to traditional karyotyping in pregnancies with increased risk following first‐ or second‐trimester maternal serum screening. Data from 3973 pregnancies with referral for invasive prenatal testing following AMSS were obtained from April 2016 to May 2020. Routine karyotyping was performed and single nucleotide polymorphism array was recommended. The foetuses were categorized according to the indications as AMSS only (group A) and AMSS with ultrasound anomalies (group B). CMA was performed on 713 prenatal samples. The proportion of women opting for CMA testing in both groups increased over the years. The incremental yield of clinically significant findings for pregnancies with high risk of screening results was similar to that for the foetuses with ultrasound soft markers (P > 0.05), but significantly lower than that for the foetuses with structural anomalies (P < 0.05). The total frequencies of variants of unknown significance in groups A and B showed no significant difference (P > 0.05). CMA should be performed for pregnant women undergoing prenatal invasive testing due to AMSS, especially with high‐risk results, regardless of ultrasound findings.

funded, while traditional screening is publicly funded and included in population-based programme in China. Hence, in the vast majority of countries and territories, the serum screening test remains the predominant method, although the rate of serum screening has decreased after NIPT introduction. Regarding abnormal maternal serum screening (AMSS), further testing recommendations include invasive procedure such as amniocentesis, chorionic villus sampling or a secondary screening tool of high sensitivity like NIPT. Although women with AMSS results are optimal candidates for NIPT, 2 in practice, considering that confirmatory invasive testing is mandatory for cases with NIPT-positive results in the end, quite a few women opt for invasive diagnostic testing to obtain definite results.
It has been generally accepted that chromosomal microarray analysis (CMA) should be recommended in pregnancies with foetal structural abnormalities. 3,4 However, its use for AMSS pregnancies is not yet widespread. Maternal serum screening is mainly targeted to detect chromosomal aneuploidy, including trisomy 21, 18 and 13, and not to detect an increased risk of microdeletions/microduplication syndromes. Due to the absence of a consensus concerning the optimal testing method, pregnant women often hesitate to choose traditional karyotyping alone or to choose CMA testing concurrently undergoing invasive prenatal diagnosis. Several reports have emphasized the importance of CMA testing in populations with a low risk of microdeletions/microduplication syndromes. [5][6][7] To provide more data, we reviewed the CMA selection trends and testing results in pregnancies with AMSS combined with or without ultrasound findings in our centre for over 5 years.

| Patients and samples
This retrospective study reviewed 4070 consecutive singleton pregnancies referred for invasive prenatal diagnosis due to abnormal results on first-or second-trimester maternal serum screening, accompanied with or without other clinical indications. AMSS was defined as high risk for Down syndrome (>1/270), critical risk for Down syndrome (1/1000-1/270), high risk for Edwards syndrome (>1/350) and critical risk for Edwards syndrome (1/1000-1/350). Individuals who were chromosomal abnormality carriers and had undergone NIPT, as well as those with adverse reproductive history, were excluded. As a result, 3973 cases were enrolled, and the women were classified into two groups following additional indications: AMSS only (group A) and AMSS with ultrasound anomalies (group B).
Amniotic fluid was sampled from 3852 (97.0%) of these women during the 18th and 24th gestational weeks; umbilical cord blood was sampled for 94 (2.3%) women during the 25th and 32nd gestational weeks; and chorionic villi samples were collected from the remaining 27 (0.7%) women during the 10th and 14th gestational weeks. The demographic characteristics are presented in Table 1.
Conventional karyotyping was routinely performed in all cases.
CMA was also recommended to all, but it was performed according to the patients' willingness. Pregnancy outcome follow-up was conducted via clinical records and telephone communication. This study was approved by the local Ethics Committee of the Fujian Maternal and Child Health Hospital. Written informed consent to participate in the study was obtained from each patient.

| DNA extraction and CMA platforms
Genomic DNA was extracted from uncultured amniotic fluid, foetal cord blood and CVS using a QIAGEN kit (Qiagen) according to the manufacturer's instructions. Single nucleotide polymorphism array (SNP array) was performed using Affymetrix CytoScan 750K array (Affymetrix Inc.), which includes 200 000 probes for single nucleotide polymorphisms and 550 000 probes for copy number variations (CNVs) distributed across the entire human genome.

| CMA data interpretation
To analyse the results, Chromosome Analysis Suite software (Affymetrix) and human genome version GRCh37 (hg19) were used. A resolution was generally applied: gains or losses of ≥400 kb and loss of heterozygosity (ROH) ≥10 Mb. Uniparental disomy (UPD) was reported based on the identification of the region of homozygosity (ROH) covering the entire chromosome. UPDtool was used for genome-wide detection of UPD within the child-parent trios to confirm maternal or paternal UPD origin. All detected CNVs were compared with in-house and national public CNV databases as follows:

| Conventional karyotyping
This procedure was described in our previous study. 11
Statistical comparisons were performed using the chi-square test, and P < 0.05 was considered statistically significant.

| Yearly change for the rates of CMA testing in different groups over the 5 years
Of the 3973 cases, 713 (17.9%) were processed in parallel using conventional karyotyping and CMA testing. The rates of CMA  Figure 1, the yearly rate of CMA testing increased gradually for both group A and group B, from 3.4% in 2016 to 26.4% in 2020 (P < 0.05), and from 25.0% in 2016 to 79.3% in 2020, respectively (P < 0.05).

| Overall CMA results
A total of 85 aberrations were recorded in 82 out of 713 cases processed in parallel using conventional karyotyping and CMA testing. Clinically significant aberrations were observed in 59

| CMA findings in pregnancies with AMSS only
This group consisted of 3540 foetuses with high-risk maternal serum screening results and 60 foetuses with critical risk.
Amongst the 476 cases analysed using CMA, 43 genetic mutations were observed in 42 cases, and 28 (65.1%) of them detected in 28 cases were clinically significant aberrations. Amongst them, 20 cases were karyotype-detectable, and 8 were CMAdetectable only (  anomalies was 8.6% (7/111), which was significantly higher than that for pregnancies with solely AMSS (8.6% vs 1.7%, P < 0.05). Details are presented in Table 3.

| CMA findings in pregnancies with AMSS and ultrasound abnormalities
The total frequency of VOUS in this group was 2.5% (6/237). The values for pregnancies with ultrasound soft markers and with other ultrasound anomalies were 1.9% and 3.7%, respectively; there were no significant differences between AMSS only and AMSS with ultrasound anomaly groups.

| D ISCUSS I ON
In the past years, many studies have emphasized the diagnostic utility of CMA in low-risk pregnancies. 6,7,12 The American

Congress of Obstetricians and Gynecologists recommends that
CMA be made available to all patients choosing to undergo invasive diagnostic testing. 13 In this study, we evaluated the rate of CMA testing and detection results in pregnancies with AMSS for 5 years. The yearly rates of CMA testing increased over the years.
This may be partly associated with patients' increased acceptance of new technology and physicians' improved ability to cope with varying CMA results.
Predictably, trisomy 21 and 18 were the most frequently detected aberrations in pregnancies with AMSS. In addition, 1.7% and 3.8% additional submicroscopic aberrations with clinical significance were found by CMA in group A and group B, respectively. The value of 1.7% was similar to that previously reported for patients with normal ultrasound examinations and normal karyotypes. 5,11,14,15 It is noteworthy that all the karyotype-undetectable aberrations in group A were detected from pregnancies with high risk rather than critical risk on maternal serum screening. An abnormal serum screening result frequently indicates a need for NIPT. 16,17 As demonstrated in this study, if NIPT was offered to pregnant women with abnormal traditional maternal serum screening but without ultrasound anomalies, 13/28 abnormal results with clinical significance would have been missed. However, for concerning pregnancies with critical risk for Down syndrome (1/1000-1/270), NIPT is more suitable than invasive testing.
Soft markers are generally considered either risk factors for underlying foetal aneuploidy, or only statistical markers for Down syndrome. 18,19 Therefore, they are actually low-risk factors for CNVs. In the current study, the proportion of CMA testing for pregnancies with soft markers was much higher than that for women with AMSS only, which indicated that pregnant women with soft markers are more likely to be recommended and willing to undergo CMA testing; however, their incremental yield of clinically significant findings was comparable, although both were much lower than that for women having other foetal abnormalities that had increased chances of being associated with CNVs.
Therefore, we believe that the indication of AMSS should receive the same attention as soft markers.
The major challenge of CMA in the prenatal setting is the incremental detection of uncertain results, including VOUS and susceptibility loci (SL), especially in the setting of normal ultrasound.
In our laboratory, we generally categorized SL with approximately 10% penetrance as VOUS, 10 and penetrance of more than 20% as a likely pathogenic variant. Studies exploring choices of women who are undergoing prenatal CMA in a clinical setting regarding these types of findings indicate that most women opt for maximal information on their foetuses and many wish to avoid uncertain results. [20][21][22] Our policy is to report all findings to the patients to circumvent lawsuits to some extent. 23 It is worth mentioning that the detection rates of VOUS in AMSS pregnancies with and without ultrasound anomalies showed no significant difference.

However, the proportion of clinically significant variants in AMSS
with ultrasound anomalies except for soft markers was twice as high as that of VOUS. For the groups with normal ultrasound or with soft markers, the proportions were similar, which was con- The purpose of prenatal diagnosis is not only to make choices about pregnancy, but also to make early predictions of the symptoms that may appear in the foetus after birth and play an essential role in driving early intervention. In our study, we noted aberrations that were considered pathogenic or likely pathogenic but with little clinical significance detected in pregnancies with or without ultrasound anomalies, such as XLI, a recessively inherited disorder of cutaneous keratinization. Male patients always have polygonal, semitransparent and fine scales on the skin at birth or soon after birth, and the scales gradually become deep dark and rough due to a recurrent microdeletion in the stearyl sulfatase gene (STS). 29 Postnatal follow-up confirmed the diagnosis of ichthyosis in two male foetuses and revealed normal dermatological manifestations in a female foetus with ichthyosis. 30 The limitations of this study include its retrospective design with a small sample size; moreover, not all patients with AMSS accepted CMA which may cause some data bias.
In summary, pregnant women with AMSS who underwent invasive prenatal testing only were subjected to the selection process twice via either NIPT or invasive testing, as well as karyotyping only or with CMA currently. The decision to opt for invasive prenatal testing reflected their desires for more definite results for their foetuses.
Considering (a) the efficiency to detect additional 1.7% aberrations with clinical significance, (b) the evidence that not all chromosomal anomalies are associated with ultrasound abnormalities, and (c) the continuous accumulation of professional consulting experience for variants of uncertain significance, we conclude that CMA should be offered to all pregnant women with AMSS, especially those with high-risk results, undergoing prenatal invasive testing, regardless of ultrasound findings.

This work was supported by Research Fund Project of Fujian
Provincial Maternity and Children's Hospital grant (YCXQ 2018-32).

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

DATA AVA I L A B I L I T Y S TAT E M E N T
The data described in this study is available upon request from the corresponding authors.