Strategies to Detect Chromosomal Anomalies Not Identified by NIPT

ABSTRACT

Introduction

Genome‐wide non‐invasive prenatal testing (gwNIPT) has screening limitations for detectable chromosomal conditions and cannot detect microdeletions/microduplications (MD) or triploidy. Thickened nuchal translucency (NT) only detects around 10% of these cases.

Methods

A 4‐year retrospective study of singleton pregnancies undergoing first‐line gwNIPT screening with subsequent CVS or amniocentesis. All MD cases, with or without gwNIPT screening, were also analyzed.

Results

Among 919 pregnancies with gwNIPT and invasive testing, 338 had a single chromosomal abnormality, with 9 false negative gwNIPT results (2.9%) and 26 undetectable abnormalities (18 MD, 8 triploidy) (7.7%). Twelve cases had a dual chromosomal abnormality and 4 returned a low‐risk gwNIPT result. Only three (9%) of the “missed cases” had a large NT and two of these also had a structural abnormality. Approximately 90% of chromosomal anomalies missed by gwNIPT were detected by invasive prenatal testing indicated by one or more of the following: failed NIPT (9%), low PAPP‐A (12%), early growth restriction (37%) and structural anomalies at pre‐NIPT, 13‐ or 20‐week ultrasounds (51%).

Conclusion

Most chromosomal abnormalities missed or unable to be found by gwNIPT are detected due to growth restriction or structural anomalies, not an enlarged NT. Failed NIPT and low PAPP‐A concentrations contributed to detection.

Summary.

• What is already known about this topic?

  • ◦Genome‐wide NIPT (gwNIPT) will have some false negative results for abnormalities it is designed to detect and it cannot detect microdeletions or triploidy.
  • ◦Enlarged nuchal translucency (NT) measurement appears to have a low detection rate for chromosomal anomalies not identified by gwNIPT.

• What does this study add?

  • ◦Approximately 90% of chromosomal anomalies not identified by gwNIPT can be detected by invasive prenatal testing indicated by one or more of the following: failed NIPT, low PAPP‐A, early growth restriction and structural anomalies at pre‐NIPT, 13‐week, or 20‐week ultrasounds.
  • ◦The strongest association is with multiple fetal structural abnormalities identified at 13‐ or 20‐weeks (or intracranial abnormalities if a single organ system is involved) and early fetal growth restriction.

1. Introduction

Genome‐wide non‐invasive prenatal testing (gwNIPT) analyses cell‐free DNA (cfDNA) to detect copy number variants (CNVs) larger than 7 megabases (Mb), which is a similar limit of resolution to karyotyping [1]. Detection rates (DR) are above 90% for the common autosomal trisomies (CAT), sex chromosome abnormalities (SCA) and rare autosomal trisomies (RAT) [2, 3]. The DR for segmental imbalances (SI) (CNVs > 7 Mb) was 77.4% [4]. The scope of gwNIPT is currently incomplete as it cannot detect triploidy or CNVs < 7 Mb (microdeletions or microduplications [MD]) [5, 6].

Nuchal translucency (NT) enlargement identified during combined first trimester screening (cFTS) has been suggested as a strategy to enhance detection of chromosomal abnormalities missed by NIPT [1, 7, 8]. Different NT threshold values have been employed, based either on static measurements across the 11–14 weeks’ gestation range, on gestational age centiles or on multiples of the NT gestational median (NT MoM) [9, 10]. A recent study found NT measurements above the 99th centile in 9% of cases missed by NIPT, and above the gestationally adjusted 95th centile in 15% [11]. Thus, NT screening alone at 11–14 weeks’ gestation will fail to detect over 80% of cases missed by NIPT.

Anomalous placental biochemistry, notably low levels of pregnancy associated plasma protein‐A (PAPP‐A), free‐beta human chorionic gonadotrophin (fbhCG) and placental growth factor (PlGF), identified at cFTS and first trimester preeclampsia screening, is associated with chromosomal abnormalities and could be an indication for invasive prenatal testing [12, 13, 14].

The 13‐week scan detects around 60% of major fetal structural anomalies which are associated with chromosomal abnormalities and are an indication for invasive testing with microarray analysis, due to the increased resolution compared with karyotyping [15, 16, 17]. Early fetal growth restriction is advocated as an indication for invasive testing with microarray analysis, despite a low‐risk gwNIPT and absence of a structural abnormality [18, 19].

The aims of this study were to establish how chromosomal anomalies not identified by gwNIPT were ultimately detected, and to examine which strategy, or combination of strategies, may provide the most appropriate indication for invasive prenatal testing.

2. Methods and Materials

2.1. Participants and Setting

Data for all women with a live singleton pregnancy undertaking prenatal invasive diagnostic testing, either by chorionic villus sampling (CVS) or amniocentesis, were retrospectively collected over a 4‐year period between 1/01/2020 and 31/12/2023 from two multi‐site specialist obstetric ultrasound and prenatal screening practices in Australia's two largest cities. Both used gwNIPT as a first line screening test in the first trimester and acted as tertiary referral centers for women with high‐risk NIPT results obtained externally. Multiple pregnancies and those with failed NIPT results were excluded from the study.

2.2. Procedures

Both practices followed identical screening protocols sending blood samples to the same laboratory for gwNIPT using massively parallel shotgun sequencing with the Illumina platform Veriseq version 2 (Illumina Inc. California). It is capable of detecting common autosomal trisomies, sex chromosome abnormalities, rare autosomal trisomies and segmental imbalances greater than 7 Mb [20]. All singleton pregnancies undergoing gwNIPT screening were included in the study cohort, other than those with multiple aneuploidies. For the purpose of assessing microdeletion/microduplication cases, which cannot be detected by gwNIPT, all cases were included irrespective of whether or not NIPT screening was performed.

2.3. Prenatal Diagnostic Testing

All diagnostic testing was performed by the initial quantitative fluorescent polymerase chain reaction and single nucleotide polymorphism (SNP)‐array. The microarray results of either CVS or amniocentesis were included in the study. In instances where amniocentesis followed an aberrant CVS result in the same pregnancy, only the amniocentesis result was included. To illustrate, a high‐risk cfDNA result for trisomy 13 (T13) with a mosaic T13 result on CVS, but a later disomy 13 result on amniocentesis, would be classified as a false positive NIPT result for T13. An amniocentesis result identifying a MD, fetal mosaicism, or uniparental disomy was considered abnormal.

Variants of unknown significance and parentally inherited copy number variants not resulting in a phenotypic anomaly were considered false positive results for the purpose of this study.

2.4. Ultrasound

Pre‐NIPT ultrasound to assess fetal number, viability and gestational age was routinely performed [21]. In instances where structural abnormalities or enlarged NT (> 2.2 mm at 10–11 weeks) were identified, genetic counseling ensued to help determine if NIPT was undertaken or bypassed in favor of CVS [22, 23]. Pre‐eclampsia screening was routinely performed, including placental serology (fbhCG, PAPP‐A, PlGF). The late first trimester ultrasound to assess fetal growth and structure and NT measurement was typically performed around 13 weeks of gestation (i.e. 12.5–13.5 weeks) and the data entered into the Viewpoint software (GE, Solingen, Germany).

All 13‐week scans were performed by Fetal Medicine Foundation accredited sonographers using GE E10 platforms (GE Healthcare, Chicago, USA). The examinations were initially performed transabdominally with additional transvaginal imaging as necessary for optimal visualization. Where structural abnormalities and/or enlarged NT measurements (≥ 3.0 mm) were identified, invasive testing was offered, even with a low‐risk NIPT result, as an increased risk of a chromosomal abnormality not detected by the NIPT has been reported in this context [8]. Individualized management was applied in the context of anomalous placental serology results. The ultrasound examination was classified as normal when the fetus displayed appropriate growth and appeared structurally normal with no sonographic markers of aneuploidy evident. Structural anomalies did not include ultrasound findings that can resolve, such as nasal bone hypoplasia, dilated jugulo‐lymphatic sacs, cystic hygroma or generalized edema.

Dating of the pregnancy was from the IVF transfer date or from the last menstrual period confirmed by early dating or a pre‐NIPT scan, redated to be consistent with scan dates if there was a significant difference. First trimester growth restriction as an indication for invasive testing was classified as the later first trimester scan being less than established dates by more than 7 days. Second and third trimester growth restriction as an indication for invasive testing was classified as growth below the 10th centile. At the 20‐week morphology scan, this was usually greater than a 10‐day discrepancy from the established pregnancy dating by early scans.

2.5. Statistical Analysis

Categorical variables are expressed as numbers and proportions and continuous variables in means and standard deviation (SD) or medians and interquartile ranges (IQR) depending on the frequency distribution. Proportions were compared between groups, depending on size, using either a chi‐squared test or a two‐tailed Fisher’s exact test.

The gwNIPT detection rate for multiple chromosomal abnormalities was compared with that for the larger single chromosomal abnormality identified during this study and with relevant data from published meta‐analyses [2, 3, 4] by a modified binomial exact test. Statistical analysis was performed in Stata version 15.1 (StataCorp. 2017, College Station, Texas), with α = 0.05 used to define statistical significance.

2.6. Ethical Considerations

This study was approved by the South Eastern Sydney Local Health District Human Research Ethics Committee (2022/ETH00890).

3. Results

3.1. Study Population

Over the 4‐year study timeframe, 44,087 patients with a single live fetus were referred for NIPT. Late first trimester ultrasound was performed in 44,957 patients and used to determine the centile of the suggested NT measurement threshold values. At the LFTU, the median crown‐rump length (CRL) was 71.4 mm (IQR 67.4–75 mm) and 78% of cases were examined between 12.5 and 13.5 weeks of gestation (CRL 63–78 mm). The median maternal age was 33 years (range 19–56 years, IQR 30–36 years), median body mass index (BMI) was 25 kg/m² (range 17–35.5 kg/m², IQR 21–27 kg/m²) and ethnicity was recorded as Caucasian (68.5%), East Asian (19.1%), South Asian (9.5%), mixed (1.2%), Middle Eastern (1.2%), or Black (0.5%).

3.2. Chromosomal Abnormalities

During the study period, 1467 patients underwent CVS and/or amniocentesis, of whom 919 (62.6%) underwent preliminary NIPT screening which constituted the study cohort. In that cohort there were 338 fetuses (37%) with single chromosome abnormalities. GwNIPT did not achieve the diagnosis in 35 single chromosomal abnormality cases (10.4%), with false negative results in nine anomalies (CAT [n = 3], monosomy X [MX] [n = 1], other SCA [n = 1], SI [n = 4]) and a further 26 cases with abnormalities undetectable by gwNIPT (MD [n = 18], triploidy [n = 8]), as detailed in Table 1. The detection rate was 98%–100% for the common autosomal trisomies (CAT), RAT and SCA, 90% for MX and 84% for SI.

TABLE 1.

NIPT detection rate for single chromosomal abnormalities.

Chromosomal abnormality	Number of cases	Detected by NIPT	Not detected by NIPT	DR (%)
T21	168	165	3	98
T18	37	37	0	100
T13	10	10	0	100
SCA	53	52	1	98
MX	10	9 a	1	90
RAT	9	9	0	100
SI	25	21	4	84
MD	18	0	18 b	0
Triploidy	8	0	8	0
Totals	338	303	35	90

Abbreviations: DR, detection rate; MD, microdeletion/duplication (< 7 Megabases); MX, monosomy X; NIPT, non‐invasive prenatal testing; RAT, rare autosomal trisomy; SCA, sex chromosome abnormality; SI, segmental imbalance (> 7 Megabases); T13, trisomy 13; T18, trisomy 18; T21, trisomy 21.

^a 3 of these cases were mosaic MX.

^b 3 cases were false positive for RAT, SI, or SCA, respectively and 1 case was identified due to gender discordance (NIPT XY, ultrasound features of female external genitalia, with dupYp11.2 identified).

Further 12 fetuses (0.8%) had dual chromosomal abnormalities, most with an additional SI or MD (Table 2). In 8 of these cases, one of the chromosomal abnormalities was not detected on gwNIPT screening and in 4 cases both the chromosomal abnormalities were not detected. In 4 cases the NIPT result was high risk of an unrelated single chromosomal abnormality to that found with invasive testing. In 24 cases, the NIPT failed to produce a result and this group included 3 cases with chromosome abnormalities (T13 [n = 2], triploidy [n = 1]).

TABLE 2.

Genome‐wide non‐invasive prenatal test (gwNIPT) results in cases with combined chromosomal abnormalities (major chromosomal abnormality/minor chromosomal abnormality).

Chromosomal abnormality found with invasive testing	gwNIPT result	What gwNIPT failed to detect	gwNIPT capable of detecting abnormalities (both/one)
MX/SI	MX	SI	Both
MX/MD	MX	MD	One
MX/SCA	MX	Abnormal Y chromosome	Both a
MX/SCA	MX	Abnormal Y chromosome	Both a
SI/MD	NAD	SI/MD	One
SI/MD	NAD	SI/MD	One
SI/MD	SI	MD	One
SI/MD	SI	MD	One
T18/MD	NAD	T18/MD	One
T18/MD	T18	MD	One
T21/RAT	NAD	T21/RAT	Both
T21/RAT	RAT	T21	Both

Abbreviations: MD, microdeletion/microduplication (< 7 Megabases); MX, Monosomy X; SCA, sex chromosomal abnormality; SI, segmental imbalance; T18, trisomy 18; T21, trisomy 21.

^a NIPT often detects some of the Y chromosome.

On definitive prenatal testing, a MD was found in combination with another chromosomal abnormality in 7 cases (T18 [n = 2], MX [n = 1] and SI [n = 4]), with four of these cases undertaking invasive testing after a high‐risk NIPT result for the additional major chromosomal abnormality, all of which were true positive results. The other dual chromosome abnormality cases had invasive testing due to other indications. In one case, a female fetus with a MD (Yp11.2 duplication) was detected on invasive testing with the preliminary NIPT result indicating XXY.

3.3. Pre‐NIPT Ultrasound and Chromosomal Abnormality

Abnormalities were identified in 217 pre‐NIPT ultrasound examinations, with 111 women still undertaking NIPT and the remaining 106 choosing to bypass NIPT in favor of CVS, finding 51 (48%) chromosomally abnormal pregnancies. Of those cases undertaking initial NIPT, 58 (52%) returned a high‐risk result and all but 6 cases were confirmed by CVS (with euploid findings in those at high risk for MX [n = 2], RAT [n = 2], T13 [n = 1] and SI [n = 1]). In the remaining 53 cases invasive testing occurred despite a low‐risk NIPT result revealing two (4%) chromosomal abnormalities, including a case each of triploidy and SI. Overall, 50% (n = 109) of the cases with anomalies seen at the pre‐NIPT ultrasound (NT enlargement, generalized edema, or structural anomalies) had a chromosomal abnormality and these findings helped identify 29% of all chromosome abnormalities in those undertaking gwNIPT.

3.4. Indication for Invasive Testing

There were 35 single chromosomal abnormalities and 12 combined chromosomal abnormalities not correctly diagnosed by NIPT (47/338 [13.9%]). Four single abnormality cases had a false positive gwNIPT for a different chromosomal abnormality and 8 of the combined cases had a high‐risk gwNIPT for one of the two chromosomal abnormalities. In 31 single chromosomal abnormality cases and 4 cases with a combined chromosomal abnormality, there was a low‐risk gwNIPT result and the indication for invasive testing was fetal structural anomaly (first trimester [n = 11] or second trimester [n = 15]), fetal growth restriction (first trimester [n = 1], second trimester [n = 3]), first trimester growth and structural anomaly, with an enlarged NT (n = 1), enlarged NT measurement combined with a first trimester structural anomaly (n = 1), fetal sex discordance (n = 2) or PAPP‐A < 0.3 MoM (n = 1).

In cases with a low risk NIPT result, abnormal ultrasound findings were seen at the late first trimester ultrasound in 178 cases (revealing 13 chromosomal abnormalities) and at the second trimester morphology ultrasound in 208 cases (revealing 18 chromosomal abnormalities), as detailed in Table 3. If there was an abnormality in more than one organ system, the case was classified as “multiple”. First trimester anomalies of the brain, heart, face and multiple anomalies led to the identification of chromosomal anomalies with a positive predictive value (PPV) ranging from 4% to 28%. No chromosomal abnormalities were found in cases with aberrant right subclavian artery (ARSA), nasal bone hypoplasia, heterotaxy, skeletal, or renal anomalies. In the second trimester, anomalies involving the brain, heart, skeleton, multiple anomalies and growth restriction had underlying chromosomal abnormalities identified with a PPV ranging from 13% to 36%. No chromosomal abnormalities were found in cases with anomalies of the diaphragm, face, genitalia, kidneys, markers of aneuploidy (including ARSA, nasal bone hypoplasia and nuchal fold enlargement), hydrops, or fetal death in utero.

TABLE 3.

Fetal structural anomalies detected by ultrasound and their association with chromosomal abnormalities.

First trimester structural finding (organ system)	Total	CR abnormality	No CR abnormality	PPV of finding a CR abnormality (%) and the CR category
2 vessel cord	3	0	3	0
ARSA	3	0	3	0
Bowel echogenic	1	0	1	0
Cardiac	25	1	24	4 (MD)
Cystic hygroma/NT	66	1	65	2 (T18/MD)
Diaphragmatic hernia	1	0	1	0
Face	6	1	5	17 (MD)
General edema	1	0	1	0
Growth	2	2	0	100 (MD, triploidy)
Heterotaxy	2	0	2	0
Intracranial	7	2	5	28 (SI, triploidy)
Nasal bone	19	0	19	0
Omphalocele/gastroschisis	5	0	5	0
Pentalogy of cantrel	2	0	2	0
Renal/bladder	4	0	4	0
Skeletal	4	0	4	0
Multiple	27	6	21	22 (MD, 5 triploidy)
Total	178	13	165	7

Second trimester structural finding (organ system)	Total	CR abnormality	No CR abnormality	PPV of finding a CR abnormality (%) and the CR category
ARSA	17	0	17	0
Brain	25	3	22	12 (MD, T21, SI/MD)
Cardiac	50	1	49	2 (SI)
Diaphragmatic hernia	2	0	2	0
Face	6	0	6	0
FDIU	2	0	4	0
Growth	9	3	6	33 (3MD)
Hypospadias	2	0	2	0
Hydrops	1	0	1	0
Lung mass	3	0	3	0
Markers of aneuploidy	16	0	16	0
Nasal bone	21	0	21	0
Nuchal fold	10	0	10	0
Renal	3	0	3	0
Skeletal	24	2	22	8 (MD, SI/MD)
Multiple	17	4	13	24 (2SI, MD, T21/RAT)
Total	208	13	195	6

Abbreviations: ARSA, aberrant right subclavian artery; CR, chromosomal; DR, detection rate; FDIU, fetal death in utero; MD, microdeletion/duplication (< 7 Mb); PPV, positive predictive value; RAT, rare autosomal aneuploidy; SI, segmental imbalance (> 7 Mb); T18, trisomy 18; T21, trisomy 21; SI/MD, combined segmental imbalance and microdeletion.

3.5. NT Measurement and Chromosomal Abnormalities Not Detected by gwNIPT

NT measurements and NT MoM values were recorded in 33 of the cases with a chromosomal abnormality not detected by NIPT, which are detailed in Supporting Information S1: Table S1. The NT measurement only exceeded 3.0 mm in three (9%) of these cases, two of which also had fetal structural anomalies as indicators for invasive prenatal testing.

3.6. Triploidy

NIPT using massively parallel shotgun sequencing (MPSS) cannot detect triploidy [6]. All 8 cases were detected in the first trimester with definitive testing performed due to first trimester structural anomalies (n = 6), both a structural anomaly and early growth restriction (n = 1), or an extremely low PAPP‐A level (n = 1).

3.7. Microdeletions/Microduplications

Of the cases having a gwNIPT, 25 cases had a MD and just over a quarter of these (n = 7) were in combination with a larger CNV/major chromosomal abnormality. These major chromosomal abnormalities typically have a very high NIPT DR in isolation (Table 1) [2, 3, 4]. However, in 3 of 7 cases (43%) where the combined abnormality involved a MD, the major chromosomal abnormality was not detected. The resultant detection rate for major chromosome abnormalities when in combination with an MD was lower than when in isolation, compared to both the current study (p = 0.03) and the published meta‐analyses (p = 0.05) [2, 3, 4].

The current gwNIPT platform cannot detect deletions or duplications smaller than 7 Mb, which was confirmed in the current study data. To maximize MD case numbers for analysis, all cases with a MD found by CVS or amniocentesis were reviewed, irrespective of whether they had preliminary NIPT screening. There were 28 single MD cases (Supporting Information S1: Table S2A) and 7 cases where the MD was combined with another larger CNV (Supporting Information S1: Table S2B). The most common indication for invasive testing in all MD cases was a structural anomaly in either the late first trimester (n = 7) or the second trimester (n = 11). In four of the single MD cases, the NT was above 3.0 mm, which was an isolated finding, representing 16% of the cases where NT was measured (n = 25). GwNIPT was not performed in any of these four cases.

3.8. First Trimester Placental Serology

Using our data, the 1st and 5th centiles for PAPP‐A were 0.27 MoM and 0.42 MoM, respectively, and for fbhCG were 0.21 MoM and 0.31 MoM, respectively.

Serology was performed in 21 of the chromosomally abnormal cases with a low risk NIPT, nine of which had a MD and six had triploidy. PAPP‐A levels were under the 5th centile in nine cases (43%) and under the 1st centile in seven cases (33%). Five of the six triploidy cases had a PAPP‐A level below 0.27 MoM and they also had hCG levels below 0.4 MoM.

First trimester serology was available in 16 of the MD cases. Only one had a PAPP‐A level below the 5th centile but above the 1st centile. None of the MD cases had a hCG below the 5th centile.

4. Discussion

4.1. Summary of Study Findings

In this retrospective study of women undertaking invasive prenatal testing for a variety of indications, 919 had preliminary gwNIPT screening, which, as expected, had high sensitivity for the common aneuploidies (98%–100%) and slightly lower sensitivities for monosomy X (90%) and cases with segmental imbalance (84%). Importantly, in 35 cases (10.4%) with a chromosomal abnormality, the gwNIPT test returned a low‐risk result, including nine false negative results for abnormalities the test was designed to detect and 26 MD and triploidy cases that are currently undetectable by gwNIPT. The indication for definitive prenatal testing in these cases was overwhelmingly an ultrasound finding of fetal structural or growth anomaly (85%), of which half were identified in the first trimester.

4.2. Comparison With Relevant Literature

Prospective pooled detection rates for trisomies 21, 18, and 13 are 99.2%, 95.6% ,and 97.2% respectively. For other aneuploidy, the detection rates are 88.9% for monosomy X [24], 92%–99% for other sex chromosome abnormalities [25], 87.2%–100% for RATs [3] and (77.4%) for segmental imbalances [4]. The detection and prevalence rates identified during the current study (Table 1) are in line with the literature. While positive prediction rates and false positive rates are well‐described in the literature, accurate false negative rates are not yet available [26]. The number of chromosomal abnormalities not detected by gwNIPT will be a combination of the prevalence and the detection rates of each chromosomal abnormality.

4.3. Clinical Implications of the Findings

4.3.1. Isolated NT Enlargement at the 13‐Week Scan has a Low Yield as an Indicator for Invasive Testing When the NIPT Aneuploidy Risk Result Is Low

Enlarged nuchal translucency at the 13‐week scan had a low yield (9% PPV) for detecting chromosome abnormalities missed by gwNIPT and when isolated, without any fetal structural abnormalities, the yield dropped to 3%. Lowering the NT threshold measurement would only slightly increase the DR, but at the cost of a much higher FPR, and a large reduction of the PPV [11]. Analyzing all MD cases, with and without gwNIPT, 4 (16%) had a NT measurement above 3.0 mm as the only finding.

4.3.2. The Pre‐NIPT Scan Effect

The poor PPV of NT enlargement at 13 weeks in the current study is likely a product of the pre‐NIPT ultrasound performed in the current study. Abnormalities were identified in 217 of the pre‐NIPT scans, and approximately half of these patients bypassed NIPT screening in favor of definitive testing, of whom 48% had chromosomal abnormalities (including 5% not detectable by gwNIPT) which accounted for 15% of all chromosome abnormalities in this cohort.

4.3.3. Fetal Structural Abnormalities and Growth Restriction Are the Keys to Identifying Chromosomal Abnormalities Not Detected by gwNIPT

In our study, all but one chromosomally abnormal case with a low‐risk gwNIPT had either first‐ or second trimester anomalous ultrasound findings. In the first trimester, growth restriction and intracranial and multiple structural anomalies had the strongest association with cases missed by gwNIPT (Table 3). In the second trimester, the strongest association was seen with growth restriction and then multiple anomalies, often including intracranial anomalies (Table 3). Sonographic aneuploidy “markers” such as nuchal fold enlargement, absent/hypoplastic nasal bone and an aberrant right subclavian artery were common indications for invasive testing but did not result in any extra chromosomal abnormalities being detected. Isolated anomalies of the brain, cardiac and skeletal systems in the second trimester were associated with an increased incidence of MDs (10%–15%), which increased to 25% when there were multiple anomalies, which accords with the findings of previous studies [27, 28].

4.3.4. Detecting the (Currently) Undetectable

Smaller copy number variants (CNV) increase as a proportion of all chromosome aberrations with advancing gestation, reflecting reproductive biology and ascertainment differences [29]. Many MD syndromes do not have a prenatal phenotype and are unlikely to be detected prenatally. Up to 45% of genetic abnormalities identified after birth are pathogenic CNV [29].

In a recent Danish national study, pathogenic CNV accounted for 11.2% of all chromosome abnormalities, of which 79% were smaller than 5Mb [30]. The study concluded that approximately 20% of all chromosome aberrations (including pCNV < 5Mb, atypical SCA, triploidies and about 50% of true fetal mosaicism cases) would not be identified by current gwNIPT platforms, reinforcing the need for comprehensive prenatal screening strategies as outlined in the current study.

The most common indication for invasive testing in MD cases was first‐ and second‐trimester fetal structural and growth anomalies or sex discordance between the scan and the NIPT. An isolated increased NT measurement was seen in up to 16% of all MDs in our cohort, similar to a recent study [8].

As all our patients were offered preeclampsia screening, PAPP‐A and PlGF levels were usually available. Of all MD cases, a PAPP‐A level below the 5th centile (0.42 MoM) could detect 12%, which is similar to the DR using a larger NT. PAPP‐A was particularly low in most of the triploidy cases. Using the first centile would find 33% of all the triploidy cases missed by NIPT, which is similar to the proportion found by a first trimester structural abnormality. PAPP‐A levels tend to be lower in triploidy which is digynic, rather than diandric which is the less common type of triploidy and also more likely to have higher hCG levels and larger NTs [31]. When both a structural abnormality and a very low PAPP‐A occurred in the same pregnancy, the incidence of a chromosomal abnormality was higher.

4.3.5. Presence of a MD May Mask Larger CNVs and Adversely Affect NIPT Detection

A quarter of the MD cases were combined with a larger chromosomal abnormality which gwNIPT can detect, but the DR of the larger CNV was significantly lower when combined with a MD. The mechanism by which a MD may reduce the ability of the NIPT to detect an additional larger chromosomal abnormality is unclear and further investigation on larger numbers would be of benefit.

4.4. Strengths and Weaknesses of the Study

The strengths of this study include the large cohort with SNP‐microarray prenatal diagnostic testing across two multi‐site specialist screening practices, where most patients undertook a pre‐NIPT scan, gwNIPT and a 13‐week ultrasound to assess fetal growth and structures, irrespective of the NIPT result. All ongoing pregnancies had mid‐trimester morphology ultrasounds. Potential weaknesses of the study include an older, private practice patient cohort, where not every patient underwent gwNIPT or a detailed 13‐week ultrasound. It is likely that some cases with a chromosomal abnormality not identified by NIPT had a miscarriage, invasive testing elsewhere, or were not detected antenatally, making our figures an underestimate. The lack of neonatal follow‐up is acknowledged as a weakness but, as the prenatal screening clinics involved in this study are tertiary referral centers and not the primary obstetrics carers, we have no information on the newborn’s name, date of birth and place of birth, making linkage to the various State postnatal databases extremely difficult.

5. Conclusions and Areas for Future Research

Genome‐wide NIPT is a screening test that has some false negative results for abnormalities it is designed to detect. It cannot currently detect microdeletions/duplications or triploidy, which accounted for three‐quarters of the chromosomal abnormalities not detected by gwNIPT in the current study. Strategies to help identify these anomalies and minimize false negative results are based around three ultrasound time‐points: the pre‐NIPT scan, often at 10–11 weeks, which appears to be of increasing importance along with abnormal placental growth factors often available due to routine preeclampsia screening; the detailed 13‐week early anatomy scan, when particularly multiple findings have a higher predictive value; and the mid‐trimester fetal morphology and growth assessment. Enlarged NT measurement identified in isolation at the 13‐week ultrasound in the context of a low risk gwNIPT result has a low yield for chromosome abnormality.

Consent

All patients gave verbal consent for pregnancy screening and data collection for research.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting Information S1

Data Availability Statement

The data are available on request.