Indications

The increased resolution of chromosomal microarray (CMA) over routine karyotyping results in an increased detection rate of (submicroscopic) chromosomal abnormalities. As in many other countries, we started with the introduction of CMA in cases with ultrasound abnormalities ( ) followed by the introduction of CMA in routine PND for all indications ( ).

It can be disputed that when patients opt for an invasive diagnostic procedure because of, for example, an increased risk of Down’s syndrome, microarray testing is preferable to karyotyping or rapid aneuploidy detection (for trisomies 13, 18, 21, and X,Y aneuploidy) (RAD stand-alone) to exclude/detect more unbalanced chromosome aberrations ( ). Especially since many submicroscopic abnormalities cause ID and/or anomalies not detectable by fetal ultrasound examination ( ). Are patients only interested in knowledge about fetal Down’s syndrome or do they also want to be informed about other severe early-onset diseases of their child? The discussion on choosing broad or targeted diagnostics is not a new issue in prenatal diagnostics ( ). showed that most pregnant women chose broader testing (karyotyping instead of RAD stand-alone) in a prospective study, and several years later, showed that if invasive testing is performed, most patients preferred higher resolution microarray testing (0.5 Mb) over microarray with 5 Mb resolution (comparable to postnatal karyotyping). In addition, showed that 62% of the women made an adequately informed choice, based on sufficient knowledge and attitude consistent with their choice of microarray resolution.

Pretest counseling

With the introduction of prenatal whole-genome array testing, it was anticipated that, compared to karyotyping, the frequency of genetic findings that are difficult to interpret and the frequency of diagnoses of late-onset (un)treatable disorders not related to the initial referral indication would be higher ( ). Therefore pretest genetic counseling was implemented as an obligatory element of microarray testing during pregnancy, which is reflected in several publications ( ) and clinical guidelines ( ). Routine application of prenatal microarray led to increasing numbers of patients choosing prenatal array without sufficient numbers of medical doctors specialized in clinical genetics available for pretest counseling. When we introduced microarray diagnostics in the prenatal setting, pretest counseling was performed by a clinical geneticist. Now that microarray diagnostics is our standard care—as also described by —pretest counseling is performed by gynecologists/ultrasonographers in close collaboration with clinical geneticists and the laboratory specialists. We have a brochure with information about prenatal diagnostics and the different genetic tests. To further facilitate the pretest counseling, predetermined types of findings patients wish to be informed about need to be discussed. Both patients and medical specialists need a clear classification with illustrative examples to understand the possible outcomes that are discussed during pretest counseling. described an example:

• •
  

  Pathogenic (for the proband, e.g., fetus):

  
  
  
  
  • •
    

    Causative findings: pathogenic findings explaining the phenotype or matching the indication, for example, 22q11 microdeletion in a proband with a tetralogy of Fallot.

    
    
  • •
    

    Unexpected diagnosis: pathogenic findings not explaining the phenotype or not matching the indication (early or late onset, treatable or not treatable), for example, a deletion in the DMD gene associated with Duchenne muscular dystrophy in a male fetus referred for PND because of an abnormal first-trimester screening.

    
    
  • •
    

    Susceptibility CNVs: variants associated with neurodevelopmental disorders, but of extreme phenotypic heterogeneity and/or variable expressivity, for example, 16p12.1 deletion in a patient with developmental delay.

  
  
  
  
• •
  

  VOUS.

  
  
• •
  

  Carrier status of recessive diseases, for example, deletion of CFTR gene.

  
  
• •
  

  IFs: abnormalities found by chance, unintentionally, in parents of probands, for example, mosaic Turner syndrome discovered on B-allelic frequency plot during quality control of the array profile of a pregnant woman referred for PND due to fetal ultrasound abnormalities.

When we started microarray analysis (in our case, single nucleotide polymorphism (SNP) array) in the case of fetal anomalies in clinical settings in 2009, we offered pregnant couples a choice regarding the extent to which they wished to be informed about predefined categories of genetic anomalies ( ). At first the choices were to be informed about (1) only an abnormal result that explains the ultrasound abnormality, (2) results that probably have an adverse health effect in infancy and childhood, and (3) results that probably have an adverse health effect in adulthood (for fetus and possibly the parent). In a study, 11.1% of parents chose only option 1 ( ). Later on, we abandoned these choices as we encountered cases where a severe early-onset disorder did not explain the ultrasound abnormality, and doctors would be posed for a huge dilemma if parents would have chosen option 1. An illustrative example is described by : a case of a fetus with a complex heart malformation in which a 5.0-Mb deletion in 15q11.2q13.1, associated with Angelman syndrome, was found. Where the parents had in mind to continue the pregnancy based on the ultrasound abnormality, they terminated the pregnancy after this additional information. In a study, we offered pregnant couples the additional choice of information on susceptibility CNVs. The majority of pregnant couples chose to be informed of any (susceptibility for) severe condition(s) of any age of onset ( ). Our lab in Rotterdam has completely replaced conventional karyotyping by a whole-genome SNP array for all indications for prenatal invasive testing since July 2012 ( ). Although specialists in PND have always been familiar with the phenomenon of UD, the number and nature of UD has expanded with the implementation of the SNP array when compared to conventional karyotyping ( ).

As described by , pretest counseling aids pregnant couples in handling the uncertainties that may accompany offering a broader scope of genetic PND using the array.

Possible problematic array findings: Unexpected diagnoses/incidental findings

Next to pathogenic causative array findings (explaining the fetal phenotype and therefore fitting the referral reason), so-called UD or IFs can also be made. We have previously proposed to use the term “UD” for fetal pathogenic findings not explaining the phenotype or not matching the indication while we proposed to use the term “IFs” for abnormalities found by chance, unintentionally, in parents of probands ( ). However, internationally the term “IFs” is often used for unexpected pathogenic findings in the fetus as well as the parents ( ). Pathogenic findings that do not match the indication for testing or the fetal phenotype seen on ultrasound have always accompanied whole-genome cytogenetic testing ( ); however, because array testing has a higher resolution, they are now more frequently found. In a study on 1033 fetuses, an unexpected diagnosis was found in five cases (0.48%); one chromosomal abnormality (0.1%) concerned a microscopic finding, whereas four chromosomal abnormalities (0.39%) were submicroscopic findings ( ). Interpretation of microarray findings is also more complex in a prenatal setting than in a postnatal setting due to the fact that the prenatal phenotype is only described by ultrasound imaging and cannot be fully investigated. It is not always possible to assess whether the pathogenic CNV explains the fetal phenotype. Therefore it is not always easy to distinguish a causative array finding from an unexpected diagnosis. Subsequent (targeted) ultrasound examination, autopsy or postnatal examination, and follow-up may reveal additional abnormalities that may match the fetal genotype. Due to this limitation, it is difficult to determine the actual ratio between clearly pathogenic causative findings and UD in prenatal settings. Unexpected early-onset diagnoses of severe diseases may be considered an additional value of prenatal array testing if invasive sampling is performed.

Pathogenic unexpected array findings may be further subdivided into four subcategories: early-onset untreatable diseases, early-onset treatable diseases, late-onset untreatable diseases, and late-onset treatable diseases. There is no international consensus (even in postnatal settings) on what kind of UD in fetuses and IFs in parental samples should be reported, and caution should be taken in case of every late-onset (un)treatable disease. What should be reported seems to depend mainly on condition-specific factors such as disease severity, age of onset, and treatment availability and evidence indicating pathogenicity ( ). suggest to report pathogenic findings even if they seem not to match fetal phenotype and to provide extensive posttest genetic counseling. Postnatal guidelines suggesting release of all UD to patients and recommending a minimum list of genes that should be investigated have been published ( ). However, in prenatal and pediatric settings, severe ethical and psychosocial implications may arise if we would do so ( ), especially because some of them may lead to future discrimination of individuals known to carry late-onset diseases (both treatable and untreatable) and therefore may violate the child’s right to an open future ( ). The decision “to report or not to report” seems to be highly individual. The majority of genetic counselors and geneticists (~80%) believe that, after extensive pretest (or even preconception) counseling ( ), the patients should be given a choice as to what kinds of UD are returned to them ( ). Since many genetic diseases are very rare and the risk of such a particular finding is very low, an appropriate balance between informing about all possibilities and preventing unnecessary anxiety in the future parents has to be found ( ).

evaluated whether UD made by prenatal array testing contributed to pregnancy management. In the described cohort a UD was found in 19/4043 (0.5%) pregnancies. This report includes a case with a 428-kb partial deletion in the DMD gene, associated with Duchenne muscular dystrophy, in a male fetus in a pregnancy at increased risk for Down’s syndrome based on first-trimester screening. This finding was important information for the decision on pregnancy continuation. The parents chose to terminate the pregnancy. These are the types of cases, in which we and our patients perceive the SNP array to be of additional value over conventional karyotyping. Overall, the clinical usefulness of UDs in this report was assessed based on the couple’s responses during posttest counseling and their decisions. In 16/19 cases the UD was helpful either for the couples in making a decision about the course of their pregnancy, for perinatal management or family genetic counseling. This adds another motive for offering whole-genome array during pregnancy in patients who wish broad testing of their fetus.

Possible problematic array findings: Susceptibility copy number variants for neurodevelopmental disorders

CNVs associated with incomplete penetrance, variable expressivity, and heterogeneity of clinical features (e.g., 16p11.2 microdeletion or 22q11 microduplication; so-called susceptibility CNVs for neurodevelopmental disorders) may complicate posttest counseling, if reported, and may complicate classification of array findings as well. Although there are guidelines ( ) and such findings may be classified as pathogenic ( ), some still classify such CNVs as VOUS ( ). The phenotypes of carriers of these CNVs seem to vary from normal to severely affected, and in the past the phenotypes of control individuals carrying these CNVs were shown to be intermediate between affected carriers and noncarrier control individuals ( ), so the presence of an abnormal phenotype seems to be dependent on a “second hit” such as another genomic alteration ( ).

Most guidelines recognize these susceptibility CNVs as potentially problematic; however, recommendations vary between papers ( ). and suggest that these findings should be discussed during pretest counseling and the decisions on reporting should be given to patients. and suggest that only some of them should be reported, depending on the penetrance and fetal phenotype.

These CNVs, associated with neurodevelopmental disorders, but of extreme phenotypic heterogeneity and/or variable expressivity, are quite often found in prenatal settings. The literature data show a statistically significant higher frequency of these CNVs in fetuses with ultrasound abnormalities when compared to fetuses without ultrasound anomalies, 3.6% versus 0.8% ( ), 1.4% versus 0.55% ( ), and 2.6% versus 1.4% ( ), respectively.

The most problematic counseling issue is that the exact risk for developing a specific disease/phenotype, if such a finding is discovered prenatally, is still largely unquantifiable ( ) and, therefore, the prognosis for the fetus is unsure. The available data are based on postnatal observations ( ), and there are only a few population studies showing both early- and late-onset phenotypes in the carriers of susceptibility CNVs in comparison to population matched noncarriers (and not to laboratory control population). studied the 22q11.2 locus as an example of the clinical challenge in genomics. They provided not only the true population prevalence of 22q11 microduplication but also showed the population-unbiased developmental trajectories of psychiatric disorders of both rearrangement types from 1 to 32 years of age. The authors have seen marginal differences in psychiatric disorders between 22q11 microdeletion and 22q11 microduplication carriers, which suggest that, for many of the psychiatric disorders, 22q11.2 microduplication carriers need the same careful and persistent clinical attention as 22q11 microdeletion patients. The posttest counseling is also facilitated by comparing these cohort results with the risks in the general population, to provide information on the possible course of the disease when a 22q11 microduplication is found prenatally.

Although many of the susceptibility CNVs are associated with phenotypes that cannot be diagnosed in utero, some may be prenatally actionable as they involve a structural abnormality (e.g., heart anomaly in case of 1q21.1 duplication). In such cases an additional ultrasound investigation should be offered ( ). underlined that the current diagnostic ICD-10 system was not developed to capture the milder clinical manifestations of neurodevelopmental disorders, so they presumed that individuals presenting milder symptoms may not be referred for clinical evaluation. In the case of an increased risk for a neurodevelopmental disorder, a medical follow-up may be advised to assure early interventions, when the first mild symptoms are noticed.

So far there are not many prenatal cohort evaluations. performed a study of 57 cases with a PND of a susceptibility CNV for neurodevelopmental disorders. Rapid and immediate consultation by a clinical geneticist experienced in PND was applied in all cases. They found that the severity of the ultrasound anomalies and not the diagnosis of risk for neurodevelopmental disorders was the most important factor contributing to the decision on pregnancy continuation. In a large majority of cases with milder or no ultrasound anomalies, the pregnancy was continued and a normal outcome at birth was observed. In addition, most patients were interested in both prenatal and postnatal actionability of susceptibility CNV and could cope with uncertainty related to the CNV. Follow-up studies on cases diagnosed with a susceptibility CNV are necessary to further facilitate posttest counseling. showed a significant difference in communication and personal–social development between children with a reported susceptibility CNV during pregnancy and the control population. Unfortunately, their study group was very small and more studies are needed.

suggest that the information of susceptibility CNVs may represent a burden to couples and offering such knowledge should be accompanied by careful pre- and posttest counseling. Concerns have been raised about the possible emotional harm indeterminate findings may cause to pregnant couples ( ). Another study stated that uncertain test results do not enable choice, and, therefore, the burden of the anxiety and stress of uncertain test results is not justified ( ). Some have argued that the burden of generating susceptibility CNVs likely outweighs the possible benefits of detecting more pathogenic findings ( ). Others argue that it is paternalistic to withhold any outcomes from PND ( ). Instead of withholding outcomes, it has been suggested that an improved approach to reduce parental uncertainty is needed in counseling ( ).

investigated the psychological impact of PND and disclosure of susceptibility CNVs in several parents who had received such a result. The majority of participants indicated they would like to be informed about the susceptibility CNV again. Most had no enduring worries. Participants unanimously indicated that pregnant couples should have an individualized pretest choice about susceptibility CNVs (non)disclosure. No negative psychological impact with the PND and disclosure of these CNVs on participants was observed. A key factor in mitigating parental anxiety with susceptibility CNV disclosure indeed appears to be posttest genetic counseling.

There is no internationally recognized policy to report susceptibility CNVs or not. Both classification and reporting of susceptibility CNVs might be questionable, but these are relatively frequent findings and they should be further investigated. It would be useful to do long-term follow-up of prenatally ascertained cases to provide more insight into the yet unquantifiable risk for an abnormal phenotype when a susceptibility CNV is prenatally detected in an uneventful pregnancy in an apparently healthy family as postnatal studies may be biased. The risks may be different in families with no history of neurodevelopmental disorders and in those with multiple family members with mild symptoms even if they were not referred for medical advice before.

Possible problematic array findings: Variants of unknown clinical significance

The question “what to report” becomes even more complicated when it concerns VOUS. There is no consensus in reporting VOUS ( ). Although some guidelines and recommendations are available ( ), the interpretation of array data remains complex ( ), and there is a certain gray zone between potentially pathogenic VOUS and truly pathogenic findings, which is caused by limited (open access) knowledge and rarity of genetic diseases.

Several publications show that at least some VOUS are prenatally reported; however, there are different conditions for reporting. Some reported all VOUS ( ), some suggest reporting VOUS, but only after extensive pre- and posttest counseling ( ). Others report only VOUS they deem relevant ( ) and others report VOUS after multidisciplinary decisions ( ). Some authors strongly recommend only report pathogenic array findings in prenatal settings ( ). Reporting CNVs classified as benign or polymorphic, and reporting true VOUS, from which the causality is often overestimated (even if de novo) ( ), does not contribute to pregnancy management. In addition, if reported, these findings may even complicate the decision to continue or terminate the pregnancy (illustrated by the example described by . Bernhardt and colleagues presented some women to whom uncertain results were given and who continued to worry after delivery and regretted about having the test. Most patients did not even recall discussing uncertain variants during pretest counseling and many were blindsided by normal preliminary results (RAD and karyotyping) ( ). Moreover, showed that some of the patients with a VOUS actually believed that a pathogenic CNV was found in their child. Therefore careful consideration whether a VOUS is potentially pathogenic and worth reporting (actionable) or not should be ideally made in a multidisciplinary setting ( ). If there is a local policy to report VOUS in the prenatal setting, patients should be informed about the possibility of finding and reporting VOUS ( ). When reporting VOUS, pre- and posttest counseling is obligatory in guiding patients through the uncertainty ( ).

Some CMA guidelines discourage reporting VOUS ( ). To minimize the need for analyzing and reporting VOUS, advise  not to report VOUS except for deletions of >500 kb and duplications of >1 Mb with emerging evidence for pathogenicity. Trio-analysis is advised to aid interpretation and decision-making concerning reporting ( ). Finally, suggest that patients should be given the choice, which results are reported. Thus there is no international consensus on reporting VOUS, which may be the reflection of cultural differences how the local society deals with both uncertainty in medicine and paternalistic medical guidelines ( ).

Posttest counseling in practice

When prenatal array analysis reveals a chromosomal abnormality, this result should be discussed in a (daily) multidisciplinary meeting with laboratory specialists and clinical geneticists involved in prenatal care. The clinical geneticist then notifies the gynecologist who did the invasive procedure and informs the patient. In our settings, in this first phone call, the clinical geneticist shortly describes the chromosomal abnormality and invites the patient to the clinic for a meeting the same day or the day after. Unexpected bad news is not communicated to patients on Friday afternoon—unless further action is needed in the weekend—to avoid a stressful weekend prior to appropriate counseling. The official laboratory report is not sent to the general hospital system until the gynecologist and patient have been informed about the result to make sure that no one can accidently mention the result to the patient before the clinical geneticist had discussed the result with the patient. During the meeting in the clinic the chromosomal abnormality and the expected phenotype in the fetus are discussed. Comparison to prenatal cohorts is necessary and counseling solely based on postnatal biased reports is not recommended. The phenotypic diversity of chromosomal abnormalities is larger than we thought before. We also discuss the possibilities to continue or terminate the pregnancy. Depending on the individual emotional state of the patient, the procedure of termination and the emotional period after the termination are mentioned. In addition, the recurrence risk and options for next pregnancy can be explained if the patient is ready to think about perspectives and already has these questions. The patients are given written information and/or contact information for any further questions. An appointment with the gynecologist is made for further management of the pregnancy. Furthermore, we offer counseling by a psychologist to further deliberate the options, prepare for the termination and saying goodbye, or come to terms with the prospect of a child with a disability (see section 2).

Diagnostic yield

The technological innovation in prenatal diagnostics—from karyotyping to CMA and nowadays from limited targeted DNA analysis to WES—has substantially improved the diagnosis of previously undetectable genetic anomalies ( ). Currently, some countries are introducing WES in prenatal genetics in cases of fetal malformations ( ), generating a vast amount of genetic information of the unborn child next to karyotyping or even CMA. wrote a review on studies on prenatal exome sequencing and found a diagnostic yield varying between 6.2% and 80%. These were 16 cohort studies with five or more tested fetuses with an isolated structural abnormality, multiple abnormalities, or an increased nuchal translucency in the first trimester. These differences in diagnostic yield were largely due to inclusion criteria and partially due to trio versus singleton approaches to sequencing. The data suggested that the diagnostic yields will be greater in fetuses with multiple anomalies or in cases preselected by clinical geneticists. When looking at data of ongoing pregnancies, the diagnostic yield in fetuses with isolated or multiple abnormalities varied from 19% to 42% ( ).

When testing fetuses with a single structural anomaly, the diagnostic yield may be different in a particular organ system. For instance, when looking at so far published cases with an isolated structural abnormality and normal array results, a diagnostic yield of 4.5%–11.5% was found in cases with heart abnormalities ( ), a diagnostic yield of 7.3%–7.9% was found in cases with congenital anomalies of the kidney and urinary tract ( ), while a diagnostic yield of 15% was found in 46 cases with a (partial) corpus callosum agenesis ( ). In cases with an increased nuchal translucency, the diagnostic yield seems to increase with an increasing thickness of the nuchal translucency ( ). One case with an increased nuchal translucency and a relevant, but noncausative pathogenic variant did not have ultrasound abnormalities later in pregnancy ( ). A very high diagnostic yield has been found in cases with a skeletal dysplasia with a molecular diagnosis in 75%–88.9% of cases ( ).

Indications: Is it feasible to select fetuses for prenatal whole-exome sequencing based on the prenatal phenotype?

After discussions on the indications for microarray diagnostics ( ), nowadays, whole-genome microarray analysis is recommended in all cases with structural anomalies on expert ultrasound scans ( ), allowing an additional diagnostic yield as compared to karyotyping of up to 10% ( ). The currently expected diagnostic yield of WES would be 8.5% in an unselected cohort and 15% in multisystem anomalies ( ); therefore we suggest that prenatal WES may also be recommended in all cases of fetal structural anomalies.

The diagnostic yield is highest in fetal cases with MCA, and therefore in MCA there is a clear indication for WES as recognized by the professionals ( ). However, in many such cases, the parents do not need additional genetic information urgently for making decisions on the pregnancy course; for example, in a fetus with short extremities, a small thorax and polydactyly the molecular diagnosis of Ellis–van Creveld syndrome does neither change the prognosis of the fetus nor perinatal management of the pregnancy. This is supported by the study of on the diagnostic yield of WES in a large unselected prenatal cohort, since 20/52 (38.5%) diagnoses were made in pregnancies that were terminated based solely on the ultrasound abnormalities. In our multidisciplinary team, we believe that the detection of genetic variants as a syndromic cause of fetal anomalies may be especially valuable in prenatal cases with a less severe phenotype or cases with an apparently isolated anomaly in which a molecular diagnosis has a larger impact on decision-making during pregnancy. In such cases, parents are facing difficult decisions based on an incomplete phenotype. It is well known that many serious syndromes have postnatal features that cannot be detected prenatally on routine or expert ultrasound scans, for example, ID or hypotonia. An example is the case of with an atrioventricular canal defect, where a pathogenic variant in the ANKRD11 gene causing KGB syndrome (OMIM 148050 ) was detected. An ID in addition to a structural cardiac defect can significantly influence the prospective parents’ decision on continuing or terminating the pregnancy. found a variant in the L1CAM gene in a male fetus with agenesis of the corpus callosum. In our practice, we have diagnosed several severe syndromic disorders in fetuses presenting with milder or isolated anomalies, for example, Cornelia de Lange syndrome in a fetus with mild intrauterine growth restriction and hypospadias ( ) and Warsaw Breakage syndrome in a fetus with apparently isolated intrauterine growth restriction in midpregnancy ( ). In cases of a treatable physical anomaly, the association with or absence of ID is most important for the parents in their decision to continue or terminate the pregnancy ( ). Several papers show a large clinical impact of prenatal WES in ongoing pregnancies on either decision-making potentially resulting in a termination of pregnancy or on pre- or perinatal management ( ).

Based on our clinical experience, we propose to offer rapid prenatal WES testing to prospective parents who would only consider termination of pregnancy when a severe underlying genetic condition is identified and to prospective parents who continue the pregnancy regardless of the genetic anomaly. In the last case a molecular diagnosis can substantially influence perinatal management. A “routine” WES test could be offered to all parents of fetuses with multiple anomalies detected by ultrasound who decide to terminate the pregnancy, but for whom a genetic diagnosis is mainly valuable to assess the recurrence risk and the diagnostic protocol in future pregnancies. In such cases, WES results do not change the pregnancy management and, therefore, can be discussed postpartum.

Pretest counseling

Pretest counseling for prenatal WES needs to be delivered in an intelligible fashion to patients from all educational religious and cultural backgrounds. Due to these challenges, this is best performed by trained genetic professionals who have affinity with prenatal testing ( ).

Realistic expectations about the likelihood of a diagnosis and the possibility of not obtaining a result before birth should be part of the pretest counseling and informed consent ( ). It should be communicated what outcomes are to be expected, what type of results are possible, how likely they are, and which of these results will be reported ( ).

In fact, pretest counseling for prenatal WES is not that different from pretest counseling for prenatal array. The classification with illustrative examples of array results can be adapted to possible WES outcomes:

• •
  

  Pathogenic (for the proband, e.g., fetus):

  
  
  
  
  • •
    

    Causative findings: pathogenic findings explaining the phenotype or matching the indication, for example, Cornelia de Lange syndrome in a fetus with a diaphragmatic hernia.

    
    
  • •
    

    Unexpected diagnosis: pathogenic findings not explaining the phenotype or not matching the indication (early or late onset, treatable or not treatable), for example, X-linked hemophilia in a male fetus with a cleft lip.

  
  
  
  
• •
  

  VOUS.

  
  
• •
  

  Carrier status of recessive diseases, for example, pathogenic variant in CFTR gene.

  
  
• •
  

  IFs: abnormalities found by chance, unintentionally, in parents of probands, for example, maternal pathogenic variant in BRCA1 gene.

In addition, limitations of the test should be addressed ( ). WES and CMA may not completely cover all genes and both can uncover variants that may not be interpretable yet. It should be explained to families that not finding a causative variant for the primary indication does not mean there is no genetic cause for this primary indication ( ). Concerning unexpected diagnosis, pretest counseling should inform on possible test outcomes. Only in this way, possible unexpected results can be made expected. It is important to be aware that the expectations that are set in the pretest counseling are not only crucial for the pregnant couples but also for the clinicians who thus create the boundaries in which they feel comfortable to convey unexpected findings.

Also, it should be noted that while a normal result of WES in the case of a mild prenatal phenotype is reassuring for prospective parents, one can never exclude the presence of an underlying syndrome ( ). After all, not all pathogenic variants can be detected by WES and some of the variants cannot be classified as (likely) pathogenic in the absence of the postnatal phenotype and functional studies. Obvious pathogenic variants that cannot be detected by WES are methylation defects. In a retrospective study on the potential diagnostic yield of WES in pregnancies complicated by fetal ultrasound anomalies, we found three cases of hypomethylation of H19 causing Silver–Russell syndrome ( ).

Perhaps in time—in line with pretest counseling for our standard care microarray diagnostics—pretest counseling can be performed by the gynecologists/ultrasonographers from the tertiary center in close collaboration with the clinical geneticists and laboratory specialists.

Pretest counseling for genomic test in practice

In many guidelines the importance of pre- and postcounseling is underlined. However, how this pre- and postcounseling is or should be done is rarely mentioned ( ). Here we describe our current practice.

As advised by the guidelines of the ISPD 2018 ( ), we now offer WES in selected prenatal cases in the setting of a multidisciplinary team. After expert ultrasound imaging (and before or after chorion villus biopsy/amniocentesis) in a tertiary center, pregnant women/couples with fetal ultrasound abnormalities are referred to a clinical geneticist specialized in prenatal genetics. During WES pretest counseling, after discussing the differential diagnosis, we explain the basics of genetics [chromosomes, DNA, and when relevant inheritance patterns (e.g., dominant, recessive, X-linked)] and discuss the possibilities of further DNA testing. Rapid aneuploidy detection, karyotyping, and array analysis can be put on a scale of detection level. Most often simplified pretest counseling for array analysis has already been done by the gynecologist, but we refer to it and to chromosome analysis done by array. At this moment, we offer prenatal exome sequencing in a form of a broad gene panel [including 3357 genes associated with congenital abnormalities (MCA) and ID] analysis. Concerning WES, we discuss the different possible outcomes: no abnormalities found (with the currently used techniques), a (likely) pathogenic variant in a gene that fits the phenotype, a variant of unknown significance [which in general is not enclosed prenatally but might turn out to be a (likely) pathogenic variant after birth when more is known about the phenotype] and an unexpected diagnosis/IF. Our current policy is not to mention variants of unknown significance prenatally unless in exceptional cases, the multidisciplinary team deems a specific variant of unknown significance relevant for the pregnancy management or when a functional test is available (see the example of CPT2). Considering UD, we always reveal variants associated with early-onset or a treatable disorder, while we do not reveal variants associated with an untreatable late-onset disorder. Carriership is revealed in certain cases, for example, when a risk of at least 25% of passing the disease is found.

As examples of UD, we discuss a predisposition to (breast)cancer or a cardiomyopathy (which we always reveal to parents in line with our current policy), adult-onset dementia (which is not revealed to parents in line with our current policy), and carriership. After oral and written information has been provided, an informed consent form is signed by both parents. It is important to take time, show empathy, and be available for additional questions during this diagnostic process as well as after diagnosis ( ). Such an extensive pretest counseling is very time-consuming but enables informed choice by patients and assures understanding of the complexity or unexpected results. It is, therefore, appreciated by patients ( ).

Possible problematic results from whole-exome sequencing: Variants of unknown clinical significance

The disadvantage of WES is that many findings can be associated with uncertainty (e.g., VOUS) ( ). When comparing to microarray, there are much more hits, which cannot be evaluated manually. Automatic filtering results are necessary to discard polymorphic variants and to limit this uncertainty. A whole-exome analysis can be performed instead of analyses of selected gene panels, which increases the chance of a diagnosis, however, also further enhances the chance of finding problematic VOUS ( ).

Reporting variants of unknown significance (VOUS) in the prenatal setting is generally not desirable. The guidelines specifically describing prenatal settings advise reporting (likely) pathogenic variants and only some VOUS that may contribute to the abnormal fetal phenotype ( ). How to handle VOUS is often not evident and depends on local practice ( ). What practice or protocol is in place should be clear for the laboratory scientists, clinicians, and finally also to patients. If there is a local policy to report VOUS in the prenatal setting, patients should be clearly informed about the possibility of finding and reporting VOUS ( ). The consent document should include the types of variants reported ( ). A multidisciplinary committee must be available ad hoc to discuss difficult cases ( ). state that VOUS should not be used in clinical decision-making. However, in some rare cases, reporting a VOUS can be justified. describe a prenatal case in which they found an IF: a homozygous VOUS in UQCRC2 associated with a mitochondrial complex III deficiency (OMIM 615160 ) with severe neonatal metabolic acidosis. As they could not exclude the presence of a metabolic phenotype in the fetus and because screening for metabolic acidosis would be straightforward after birth, they discussed the variants with the parents, indicating a low likelihood of pathogenicity and recommended screening directly after birth (which was negative) ( ).

To minimize the amount of problematic VOUS, data sharing of single nucleotide variants  (and CNVs) and fetal phenotypes is of great importance to extend the knowledge on variants and to improve variant classification ( ). Furthermore, functional studies to clarify certain VOUS can be performed after birth when the fetal/postnatal phenotype is not yet explained and certain VOUS could potentially be relevant.

In general, our policy is not to report variants of unknown significance prenatally. However, we have encountered a few cases, in which we were faced with a difficult dilemma. In a few cases the multidisciplinary team decided to report the variant of unknown significance. We have, for example, encountered a similar case as . Around a gestational age of 20 weeks, enlarged, dense kidneys were seen in a small fetus (head circumference and femur length p0,2, abdominal circumference p0,7). In addition, the stomach was relatively large with a striking proximal duodenum, and there was a single umbilical artery. An amniocentesis was performed. The next day pretest WES counseling was done, including information about nondisclosure of variants of unknown significance. Array analysis showed no pathogenic variant. WES using a broad gene panel with ca. 3300 genes associated with MCA/ID was applied, and a homozygous VOUS in the CPT2 gene was found. Pathogenic variants in CPT2 are associated with carnitine palmitoyltransferase II (CPT II) deficiency (OMIM 600650 ), which varies in severity from a stress-induced myopathic form to a lethal neonatal form. As prenatally detectable enlarged polycystic kidneys can be seen in the lethal neonatal form, the laboratory specialist decided to discuss this VOUS in the multidisciplinary team. In this multidisciplinary setting, we decided to inform the parents about this variant, and we advised metabolic screening and consultation by a pediatrician specialized in metabolic disorders directly after birth. We explained that the VOUS could have no meaning at all or could cause CPT II deficiency. We said that the renal abnormalities could be caused by CPT II deficiency, but that it did not explain the other ultrasound abnormalities. Furthermore, if the VOUS was disease causing, we could not predict which form of CPT II deficiency the fetus would have. Concluding, it was a result with a lot of uncertainty. On enquiry the patient said that it was good that this VOUS was reported as it could be very relevant directly after birth, although it felt a little vague now as it was not useful during pregnancy. However, she agreed one could not hold back the result until after birth. The patient decided to continue the pregnancy as in her eyes, the ultrasound abnormalities were not extremely severe and she would only consider a termination of pregnancy in case of extreme severe abnormalities. At a gestational age of 28 weeks, unfortunately additional ultrasound abnormalities were seen: severe intrauterine growth restriction, enlarged polycystic kidneys, oligohydramnios, and a hypoplastic thorax. We discussed the possibility of open exome analysis, but the patient refused this. A few weeks later, the ultrasound also showed cardiomegaly with pericardial effusion, intracerebral ventriculomegaly, and anhydramnios. Around a gestational age of 36 weeks the parents were devastated to find out that the pregnancy ended in a fetal demise. A pathology exam showed a severe growth restriction, lung hypoplasia, a disproportionate large and heavy heart with an atrial septal defect, and disproportionate large and heavy kidneys. Unfortunately, the fibroblasts culture (from a skin biopsy) failed, so metabolic analysis on fibroblasts was not possible. Metabolic analysis in amniocytes showed a clear decrease in CPTII activity in the patient’s amniocytes compared to control amniocytes, while VLCAD activity was normal compared to control amniocytes showing that the patient’s amniocytes have a CPTII deficiency. Kidney abnormalities, cardiomyopathy, structural heart anomalies, and ventriculomegaly have been described in CPTII deficiency ( ). The combination of cardiomyopathy with pericardial effusion, kidney abnormalities with anhydramnios, and, in particular, the severe growth restriction can explain the fetal death. Based on the ACMG criteria, the variant was reclassified to a likely pathogenic variant. Both parents are carriers, and now invasive prenatal diagnoses can be offered for the recurrence risk of 25%. As not the whole phenotype can be explained by the CPTII deficiency, further genetic analysis was offered. Full exome analysis showed no additional abnormalities.

Possible problematic results from whole-exome sequencing: Unexpected diagnoses/incidental findings

Any test result can reveal a pathogenic variant unrelated to the primary indication, but known to be associated with an abnormal phenotype ( ). We have previously proposed to use the term “UD” for fetal pathogenic findings not explaining the phenotype or not matching the indication, while we proposed to use the term “IFs” for abnormalities found by chance, unintentionally, in parents of probands ( ). However, internationally the term “IFs” is often used for unexpected pathogenic findings in the fetus as well as the parents ( ). recommend reporting these IFs that are indicative of serious health problems and (clinically) actionable. Hegde and colleagues recommend that only known and (likely) pathogenic IFs should be included in the test report, which is in line with the recommendation of not reporting VOUS. IFs without health implications or VOUS should not be reported ( ).

There are several guidelines and recommendation papers that advise on the management of such findings. It should be clear to the patient what kind of IFs are excluded from reporting (e.g., adult-onset conditions), which are included ( ), and what the psychological (and potential insurance) risks are of receiving such findings ( ). Information on the chance of identification of IFs should be given ( ) by a medical geneticist or genetic counselor during pretest counseling ( ). recommend providing multiple sources for this information in the form of written leaflets and online information. Many authors recommend that there should be a protocol or policy in place related to disclosure of IFs ( ). It is supported for patients to be able to opt-in or opt-out of receiving IFs and the informed consent should clearly indicate the availability of this choice ( ). Exceptions need to be discussed multidisciplinary on a case-by-case basis ( ). Not all recommendations are easy to employ in clinical practice, especially when having a patient with low educational status, when a counselor is doubting whether the choice is made based on sufficient knowledge and understanding. However, recommend to respect a patients’ choice to opt-out, even if the IF may be relevant to their health.

With regards to minors a policy should be in place that takes into consideration both the child’s autonomy and interests and the parents’ rights and needs, which they also believe applies prenatally ( ). When it concerns children (postnatally), recommend to (1) report IFs that have health implications for the proband and are actionable; (2) discuss on a case-by-case basis whether adult-onset conditions with significant health implications (for the proband and the family) should be returned, taking into account the health risk, actionability, and the informed consent of the parents; and (3) report carrier status in line with the informed consent, due to reproductive options.

It should be noted that finding an actionable/treatable IF is not a rarity. found that 2.7% of healthy individuals in an outbred population of European descent had a (likely) pathogenic variant in 1 of 59 medically actionable dominant disease genes for which the American College of Medical Genetics and Genomics recommends disclosure ( ).

In addition to treatable/not treatable, early-onset/late-onset disorders, trio-WES can also reveal unforeseen issues such as nonpaternity or parental consanguinity, but this is not unique to WES as this can also be found when performing a SNP array.

We have encountered a few cases with an unexpected diagnosis, for example, a susceptibility to cancer or cardiomyopathy. Our policy has been to report these as discussed during pretest counseling and stated in the informed consent form. Patients are counseled about the unexpected diagnosis and referred for appropriate health monitoring for the particular disease.

Our local policy is not to report late-onset, untreatable disorders. However, sometimes an unexpected diagnosis is not so unexpected when looking at the family history. Using array, we encountered a maternal deletion of 17p associated with hereditary neuropathy with liability to pressure palsies (HNPP, OMIM 162500 ) in a fetus of a pregnant woman who was also undergoing neurological screening because of complaints. In such a case, disclosing a late-onset, untreatable disorder can enable rapid diagnosis for the patients’ current neurological problems. It is thus important to investigate the family tree as a part of pretest genetic counseling and to discuss these cases in a multidisciplinary team and to be able to make tailor-made decisions about specific cases. The clinical information on both the parents of the fetus and family can be crucial for decisions on disclosing both UD and potentially pathogenic VOUS that match the current medical problems in the fetus or family.

Posttest counseling

Posttest counseling by a clinical geneticist is of utmost importance in case of an abnormal result. When conveying abnormal results, this should be done by the multidisciplinary team, which should include laboratory specialists, clinical geneticists, and relevant clinical specialists, and needs to include the significance of the finding, details of the diagnosed condition, including the prognosis and potential therapies ( ). The logistics of reporting an abnormal result is rather similar to the logistics of informing patients and a referring fetal medicine specialist on an abnormal array result, which is described in the section on posttest counseling of abnormal array results. As we experience that the situation and information load is highly influencing patients, who first have to cope with the current situation and decide on the course of the current pregnancy, it is advisable to have a follow-up meeting to discuss the details of the recurrence risk and future reproductive options.

In case of a normal result, it should be explained to prospective parents that a genetic cause or syndrome is not entirely excluded. Also after the child is born, new phenotypic information may become available urging for reevaluation of the WES data that may—together with new literature data—reveal a pathogenic finding. When the prenatal WES is more common in clinical practice, perhaps one could argue that pretest counseling, including this point, may be sufficient in case of normal results.

A changing landscape of prenatal counseling

Because of the technological advances, prenatal genetic counseling has changed considerably. In the days of karyotyping, large chromosomal aberrations that were found in fetuses with ultrasound anomalies were per definition classified as pathogenic and were predicted to cause an abnormal fetal phenotype. However, on the other hand, some results also carried uncertainty. Some fetuses carried marker chromosomes of unknown origin with uncertain clinical significance. Clinical geneticists were used to counsel on such uncertainties, and in many cases, subsequent parental studies were needed for the risk assessment. Incidental diagnoses in the form of, for example, familiar balanced chromosomal aberrations were also encountered. However, with the introduction of microarrays, the number of patients with uncertain or unexpected results was magnified due to the detection of susceptibility CNVs and in a lower percentage UD/IFs.

With the introduction of prenatal WES the chance/risk of finding an unexpected diagnosis/IF has further increased. Therefore the higher incidence of challenging counseling sessions may be experienced as a burden by clinical geneticists. While this might be a challenge for the counselor, it also conveys health benefits for patients: knowledge of carriership can be relevant for reproductive options and health monitoring can be arranged in case of treatable/actionable disorders.

Another problematic issue is the fact that due to high-throughput techniques and therefore enormous growth of data, the need of updating the knowledge and current guidelines is urgent ( ). Continuous education is necessary for physicians and patients to understand the changes in molecular interpretation ( ). In the past a normal karyotype was very seldom assessed as abnormal after birth, and almost all cases could be explained by the differences in banding resolution or tissue-specific aberrations. Nowadays, the development of new techniques and rapid accrual of knowledge in genomic medicine made the chance of changing the patients’ diagnosis after reanalysis of existing genomic data much higher, when compared to karyotyping. A VOUS discovered by microarray or WES testing can be reclassified to a (likely) pathogenic finding based on additional information after birth or new knowledge in time. New diagnoses can not only be made due to the discovery of new genes but also by reclassifying already reported variants ( ). When it happens, it may cause distress to clinicians and patients. If this issue is not sufficiently addressed during pre- or posttest counseling, it may cause an uncomfortable feeling of misdiagnosis during pregnancy. Due to the complexity of the techniques that are nowadays available in clinics, awareness of the test limitations and good communication with patients became much more important than before.

Future perspectives

We anticipate that, similar to microarray, prenatal WES will be offered in all cases (with ultrasound anomalies) that undergo invasive sampling. In addition, as the costs of sequencing fall, and bioinformatic and analytic capabilities improve and rapid sequencing becomes available, we may switch from prenatal WES and microarray approaches toward whole-genome sequencing (WGS) ( ). Because of improved ability to detect a wider range of genomic abnormalities, including, for example, noncoding variants, this will improve diagnostic capabilities further ( ). It is likely that, in the future, more data will become available on the use of prenatal WGS ( ). WGS is also more suitable for CNV analysis, so we presume that, in the near future, the request for microarray diagnostics and WES test will be replaced by a request for whole-genome genotyping, and the results of both variant and CNV analysis will be done based on whole-genome fetal sequencing. This will reduce pretest counseling to one session, and we predict that it will be beneficial for patients who will not be blinded by the normal initial results from RAD and microarray. Posttest counseling can then concentrate on final results without the need for “in-between” counseling for multiple laboratory tests, and the whole stressful “diagnostic roller-coaster” will be simplified.