Taking advantage of the disruptive impact of cell‐free fetal DNA screening for aneuploidy

Ultrasound is an important diagnostic tool that is universally applied in obstetric care. One of the advantages of the technology is that it can be applied to a number of potential obstetric complications. The precise role for ultrasound in pregnancy continues to evolve as obstetric practice changes. At present, most Australasian women are offered routine obstetric ultrasound scans at 11–13 and 18–20 weeks’ gestation. The timing of these investigations has been dictated largely by the ability to identify congenital abnormalities. Maternal serum AFP was first used to identify neural tube defects in universal screening programmes and performs best at 16–18 weeks’ gestation.1 The efficacy of the test relies on accurate pregnancy dating – and the added value of ultrasound as a diagnostic tool for neural tube defects, rather than merely being a dating tool, was soon recognised.2 Ultrasound performed better at a slightly later gestational time point – and the morphology scan has been embedded as a 18‐ to 20‐week test ever since.

The development of the 12‐week ‘nuchal translucency’ scan has also been closely associated with maternal serum screening. Soon after the introduction of routine neural tube defect screening using AFP, researchers recognised that serum markers were also altered in pregnancies affected by chromosomal abnormality, leading to the development of multiple marker algorithms that screen for trisomy 21 and other common forms of aneuploidy.3, 4 Ultrasound markers for trisomy 21 were also identified and could be assessed using similar statistical techniques to provide women with a personalised risk of carrying an affected fetus.5, 6, 7

For the last 20 years, we have been in ‘steady state’, reporting a 11‐ to 13‐week ‘nuchal translucency’ scan as a component of combined first trimester screening (cFTS) for chromosomal abnormality and an 18‐ to 20‐week scan to detect structural abnormality.8, 9 At 11–13 weeks, we have been able to detect 90% of pregnancies affected by trisomy 21 and other common chromosomal abnormalities (albeit with a 5% false‐positive rate) and at 18–20 weeks, we have had variable success – detecting up to 95% of fetuses affected by neural tube defects but only 50% of fetuses that have major cardiac defects.10, 11, 12 While we might be comfortable, even a little complacent with our performance, all this will now change.

The development of genomic technologies that detect and accurately quantify small amounts of cell‐free fetal DNA (cffDNA) in maternal plasma has proven to be a game changer. Norton et al. recently demonstrated that this technology is highly effective in screening for trisomy 21 – quoting sensitivity, specificity and positive predictive values of 100% (95% CI: 90.7–100), 99.94% (95% CI: 98.99–99.97) and 80.9% (95% CI: 66.7–90.9) respectively.13 The test can be applied at an earlier gestation (from 9 or 10 weeks’ gestation) and samples can be flown across the world to specialist labs – so there is no need for extensive local operator training. In the world of aneuploidy screening, the combined first trimester test has few advantages: cffDNA screening fails in 3% of cases, is not as effective for trisomies 18 and 13, does not detect any atypical aneuploidies and is currently poorly validated in multiple pregnancies.13, 14, 15 cffDNA is also expensive and, with universal screening, it would currently cost more than $1.0 million to pick up a trisomy 21 pregnancy missed by cFTS.16

Those of us involved in obstetric ultrasound are at crossroads: We can either retrain and do something else – perhaps molecular genomics – or use this opportunity to focus on other important aspects of obstetric care. There are certainly plenty of other opportunities. A review of the perinatal mortality data for Australia and New Zealand show that while deaths related to congenital abnormality account for almost one‐third of pregnancy losses, there are other potentially identifiable and preventable adverse obstetric outcomes that are implicated in a significant proportion of these cases (Table 1).17, 18 The New Zealand data set is easier to dissect as data for late termination of pregnancy, stillbirth and neonatal death are presented separately.18 Of late terminations, 78% is due to congenital abnormality – many of which could be avoided through earlier ultrasound identification. About 31.2% of stillbirths is due to hypertensive disease, fetal growth restriction and antepartum haemorrhage – all associated with placental insufficiency that may be predicted through cFTS including ultrasound assessment of uterine artery blood flow. Exactly, 32.2% of neonatal deaths are due to spontaneous preterm delivery; a complication that can be predicted by transvaginal cervical screening at the 18‐ to 20‐week scan.

Table 1.

Perinatal mortality in Australia and New Zealand

PSANZ‐PDC classification	Australia (2012)17	New Zealand (2013)18
	Percentage of all perinatal deaths	Percentage of all perinatal deaths	Percentage of all tops >20 weeks	Percentage of all stillbirths	Percentage of neonatal deaths
Congenital abnormality	29.1	26.4	77.7	6.8	19.1
Perinatal infection	2.7	3.3	1.4	3.3	5.3
Hypertension	3.2	2.2	1.4	2.6	2.0
Antepartum haemorrhage	5.9	12.4	3.6	14.3	16.4
Maternal conditions	7.9	5.7	2.9	7.2	5.3
Specific perinatal conditions	7.3	10.5	7.2	12.7	9.2
Hypoxic peripartum death	2.2	1.8	–	1.0	5.3
Fetal growth restriction	5.5	8.0	1.4	14.3	1.3
Spontaneous preterm delivery	16.5	13.4	4.3	8.1	32.2
Unexplained antepartum death	15.2	15.2	–	29.6	–
No obstetric antecedent	1.4	1.0	–	–	3.9

There have been significant improvements in detection of structural abnormalities at the time of the first trimester scan. Examples include the detection of major cardiac defects and the detection of spina bifida. Cardiac defects can be detected using the same ultrasound‐based screening tools as those used in screening for trisomy 21; namely nuchal translucency thickness and abnormal haemodynamics of the ductus venosus and/or across the tricuspid valve.19, 20, 21 These are indirect markers of risk for structural cardiac defects. Together they identify 58% of fetuses affected by major cardiac defects.20 Pregnancies deemed to be at high risk should be referred for a formal echo at 13–15 weeks’ gestation – which, using high‐resolution 2D grey scale and colour Doppler imaging is diagnostic.22 Indeed, many severe defects can be defined by direct cardiac assessment at the time of the 11‐ to 13‐week scan. Spina bifida has been difficult to detect at 11–13 weeks’ gestation until the recent recognition of changes in the posterior fossa caused by tethering and downward rotation of the cerebellum.23 These features are assessed using the same mid‐sagittal section of the head and thorax that is used for nuchal translucency assessment – measuring the intracranial translucency and the cisterna magna. A recent prospective screening study that involved 16,164 fetuses including 11 cases with open spina bifida reported 100% sensitivity for this anomaly.24 The first trimester scan has the potential to be a powerful tool in the detection of major structural anomalies, but this will only occur if we recognise the need for a systematic and structured approach to the review of anatomy in the same way as is now performed at 18–20 weeks.25

Another major advance in first trimester screening that has been implemented by a minority of groups involves screening for pre‐eclampsia and intrauterine growth restriction (IUGR).26, 27 The Fetal Medicine Foundation published an algorithm screening for pre‐eclampsia that includes markers currently used in screening for aneuploidy (PaPP‐A and PlGF) as well as novel biomarkers [mean uterine artery pulsatility index (PI) and mean arterial pressure] that are easily added to the current 11‐ to 13‐week screening test.28 Initial reports of 90% sensitivity for early‐onset pre‐eclampsia have been reproduced and the algorithm has been validated in an Australian population.29 Recognition of being at ‘high‐risk’ for pre‐eclampisa allows preventive intervention (150 mg Aspirin, nocte) resulting in a 90% reduction in prevalence of pre‐eclampsia in the treatment group.30 The combination of screening and intervention allows an 80% reduction in the prevalence of disease, the prospect of a significant reduction in preterm delivery and neonatal admission and the potential for improved long‐term maternal cardiovascular health.31, 32, 33 Aspirin is much more effective at 11–13 weeks than at 18–20 weeks and ultrasound, through first trimester assessment of uterine artery PI, lies at the heart of this preventive programme.28, 29, 34

First trimester prediction of preterm birth is less effective and there are currently no data to demonstrate the advantage of early intervention.35 This contrasts with the findings of predictive testing at 18–20 weeks, where transvaginal assessment of cervical length is recognised to be an important tool in screening for early (<34 weeks) preterm delivery and where meta‐analysis of randomised controlled trials demonstrate a significant (40%) reduction in spontaneous preterm delivery and a consequent reduction in neonatal morbidity.36, 37 The value of cervical screening was first demonstrated 20 years ago – and we have been slow to recognise that this needs to be performed transvaginally for maximal effect.38 This has, in part, being due to the additional resources that are needed for screening; as it takes extra time to perform a transvaginal scan. We should, however, recognise that spontaneous preterm birth is the most prevalent cause of neonatal death and consequently completion of transvaginal cervical screening should be a priority of the 18–20 week scan. While we are beginning to see this being recognised in State‐based initiatives aiming to reduce the prevalence of preterm birth, current advice from professional organisations about prediction and prevention of preterm birth is less emphatic.39, 40, 41

It is time to recognise that the dynamics of screening and therefore of the 11‐ to 13‐week and 18‐ to 20‐week scans are changing. If ultrasound is to remain relevant to obstetric care, we need to adapt to offer appropriate and high‐quality screening for a wider range of obstetric complications. In the world of cffDNA screening, there will be less obsession with accurate NT measurement. It will still be important to recognise the large (>3.5 mm) nuchal – as this is associated with atypical chromosomal abnormality and invasive testing and molecular karyotyping is indicated rather than cffDNA screening.14 In contrast, the need for precision for smaller NT measurements that are used for cFTS risk assessment will be gone. Quality assurance in measurement will still be important – but in accurate assessment of the intracranial translucency, for assessment of spina bifida and for accurate measure of uterine artery PI, for screening for pre‐eclampsia and IUGR. At the 18‐ to 20‐week scan, we need to spend less time worrying about markers that have no place in screening for aneuploidy when either cFTS or cffDNA has reduced the prevalence of disease so effectively. This time would be better spent performing a proper transvaginal cervical assessment (a measurement that should also be formally quality assured due to its role in risk assessment) and further management for prevention of spontaneous preterm labour.

While there are good international data to support the introduction of these novel ultrasound‐based screening tools, we are behind the curve in validation and implementation in Australasia. Preterm birth is expensive and a good economic argument can be made for its prediction and prevention. It is time we recognise the need for and embrace change in our practice – it will benefit our patients.