Identification of aneuploidy or other major chromosomal structural anomalies by G-banded karyotype has been the standard approach to the prenatal genetic evaluation of fetal structural anomalies. Recently, however, more advanced genomic techniques have been developed that are capable of identifying clinically important chromosomal alterations beneath the resolution of metaphase-banded chromosomes. A recent National Institute of Child Health and Human Development prospective, blinded study in which chromosomal microarray analysis was compared with standard karyotype demonstrated that chromosomal microarray analysis identified clinically relevant copy number variants in 6.0% of anomalous fetuses with a normal karyotype.1 Similarly, a recent systematic review by Hillman et al2 showed that, in the presence of an abnormal fetal ultrasound scan result, relevant microarray findings other than aneuploidy occurred in 10% (95% confidence interval 8–13%) of cases.
The relative effect of chromosomal microarray analysis for anomalies of specific fetal systems remains uncertain.3 Such information is important because it would allow improved counseling and subsequent decision-making. In this study, we aimed to determine the association of copy number variants with single and multiple ultrasonographically detected anomalies of specific fetal organ systems.
MATERIALS AND METHODS
This was a planned secondary analysis of the multicenter National Institute of Child Health and Human Development microarray trial, which enrolled women at 29 centers.1 Institutional review board approval had been obtained from all sites, the data coordinating center, and the participating laboratories. In the primary study, 4,406 women had either chorionic villous sampling or amniocentesis and 4,340 women had both karyotype and chromosomal microarray analysis results available. Further information regarding the microarray laboratory procedures, confirmation, classification, and reporting of array results has previously been described.1 The indications for the procedures included advanced maternal age, positive aneuploidy screening results, structural anomalies detected on ultrasonography, and a previous child with or other family history of either a genetic or congenital disorder. In the present analysis, the rate of copy number variants for fetuses identified as having an ultrasound-detected abnormality and a normal karyotype was determined and compared with karyotypically normal fetuses without ultrasonographically detected anomalies whose only indication for prenatal diagnosis was advanced maternal age (Fig. 1).
For this analysis, all ultrasound reports in which structural anomalies of the fetus were the indication for invasive testing were reviewed centrally by study personnel and data regarding the anomalies were abstracted. In 20 cases, the original ultrasound reports were not available and the anomalies were ascertained using information obtained at the time of the invasive procedure and entered into the primary study data sheet by local investigators. Three of the 1,085 originally coded ultrasound-detected anomaly cases did not meet criteria for classification as an anomaly and were excluded. All details were entered into a nonhierarchical web-based database using Cartagenia BENCH software, which allowed the coding of 19 different anatomical and nonstructural categories based on the Human Phenotype Ontology. Fetal growth restriction was classified into three subcategories based on an estimated fetal weight being less than the 10th, the fifth, or the third centile for gestational age. Amniotic fluid was classified as oligohydramnios if the maximum vertical pocket was less than 2 cm and polyhydramnios if the maximum vertical pocket was greater than 2 standard deviations above the mean for gestational age. If specific amniotic fluid measurements were not recorded on the ultrasound report, the qualitative assessment of volume was accepted. Full details of the categories and the subcategories are available as Appendix 2, available online at http://links.lww.com/AOG/A520.
Fetuses with multiple anomalies were classified according to each system for which an abnormality was present. Minor soft markers used for aneuploidy screening such as isolated choroid plexus cysts, mild hydronephrosis (anteroposterior diameter between 5 and 10 mm), nuchal translucency less than 3.5 mm or echogenic cardiac foci were not included in this analysis. A nuchal translucency of 3.5 mm or greater or a nuchal fold of greater than 6 m or a cystic hygroma were included as ultrasound-detected anomalies in our series.
For the original study, copy number variants were classified as common benign, variants of unknown clinical significance, or known pathogenic. Frequently observed benign copy number variants present in our own databases of copy number variants detected in the course of postnatal analysis, in peer-reviewed publications, and in curated databases of apparently unaffected persons were classified as “common benign.” For purposes of this analysis, all copy number variants other than those classified as common benign were included. These include “pathogenic copy number variants” of any size encompassing a region implicated in a well-described abnormal phenotype and any copy number variant greater than 1 megabase regardless of location (n=61).1 In all cases, microarray analysis of DNA from maternal and paternal blood samples was used to determine whether copy number variants detected in the fetal samples were inherited or de novo. We confirmed all de novo array findings by a second method. Further information outlining this is available in Appendix 3, available online at http://links.lww.com/AOG/A521.
Frequencies of ultrasound-detected fetal anomalies in the major anatomical systems were tabulated showing genetic abnormalities diagnosed by karyotype and additional copy number variants seen on array. The value of findings provided by microarray, referred to as “incremental yield,” was calculated as the percentage of patients with a normal karyotype who had microarray findings that were other-than-common benign. This was calculated by category and subcategory for each anomaly according to whether the anomaly was found in isolation or in the presence of anomalies in other organ systems. Chi square or Fisher’s exact test were used to compare the incremental yield of chromosomal microarray analysis in the group with ultrasound-detected anomalies with the control group whose indication was advanced maternal age. All tests were two-tailed and P<.05 was used to define nominal statistical significance. Because this is a secondary, exploratory analysis, unadjusted P values have been reported in the text and tables. However, when a Bonferroni correction for multiple comparisons is applied, the threshold for significance is P=.001 (0.05/49). The Bonferroni is based on independent tests; therefore, the count of tests includes comparisons for 11 anatomical systems and 38 anatomical subcategories for larger organ systems in tables listing ultrasonographically detected anomalies. SAS was used for all analyses.
There were 1,082 pregnancies with ultrasonographically detected structural abnormalities. In this cohort, gestational age at procedure ranged from 10 weeks to 38 weeks with a median of 18 weeks. Of these, 398 (36.8%) women had their invasive prenatal diagnosis before 14 weeks of gestation.
Seven hundred fifty-two fetuses (69.5%) had a normal karyotype. The frequency of other-than-common benign copy number variants in fetuses with ultrasonographically detected structural anomalies was significantly higher than in fetuses without anomalies (61 of 752 [8.1%] compared with 71 of 1,966 [3.6%], P<.001). The breakdown of copy number variants for the anomaly group and the advanced maternal age group is shown in Table 1. The full list of copy number variants included in this analysis and their associated ultrasound findings is available in the Appendix 3 (http://links.lww.com/AOG/A521).
Among the 752 fetuses with anomalies and a normal karyotype, 498 and 254 had an abnormality in single or multiple organ systems, respectively (Table 2). Twenty-eight fetuses (5.6%) with anomalies confined to a single organ system had an other-than-common benign copy number variant; this frequency was nominally statistically different from that in the advanced maternal age control group (3.6%, P=.04). However, the most frequent ultrasonographic abnormalities among our cohort of patients were abnormalities of the nuchal area used primarily for aneuploidy screening and included an increased nuchal translucency of 3.5 mm or greater, a nuchal fold of greater than 6 mm, or a cystic hygroma. When these findings were isolated (n=186), the frequency of copy number variants (N=7 [3.8%]) was comparable to that in the control group. When fetuses with these diagnoses were excluded from the overall analysis of isolated structural anomalies, the frequency of copy number variants (n=21 [6.7%]) was nominally significantly greater (P=.009) than in the control group (Table 2). Fetuses with anomalies in more than one system (N=254) had a 13.0% frequency of other-than-common benign copy number variants, which was significantly higher (P<.001) than that seen in the control group.
Table 3 shows the frequency of all ultrasonographically detected abnormalities detected in karyotypically normal pregnancies (n=752) classified by system and by whether the anomaly was isolated. The greatest incremental yield for chromosomal microarray analysis was seen for cardiac anomalies (15.6%, P<.001), facial abnormalities (15.2%, P<.001), and intrathoracic abnormalities (15.0%, P=.004), including all cases, whether single or multiple. Of abnormalities seen only in a single organ system, isolated renal and cardiac anomalies were associated with the greatest nominally significant incremental yield provided by chromosomal microarray analysis (15.0%, P=.036 and 10.6%, P=.012, respectively). Although isolated anomalies in other organ systems were associated with point estimates for the incremental benefit of microarray that were greater than that in the control group, the numbers in these groups were small and the differences between groups were not statistically significant.
Table 4 shows the frequency of ultrasound-observed abnormalities and genetic anomalies subclassified by abnormalities of each organ system. Of particular note, cardiac abnormalities were classified into subgroups depending on the ultrasound appearance of the anomalies: abnormalities of the four-chamber view, of the outflow tracts, and specific diagnoses. When outflow tract abnormalities were the only ultrasound finding, there was an incremental benefit of chromosomal microarray analysis of 30.0% (P=.005). Of note is that the chromosomal microarray analysis findings were not predominantly 22q11.2 deletions (which can be detected using karyotype analysis with targeted fluorescence in situ hybridization). Alternatively, 66.7% (16 of 24) of patients with cardiac defects had copy number variants other than a 22q11.2 deletion. When fetuses with a 22q.11.2 deletion were excluded from analysis, copy number variants were still significantly more common in fetuses with any cardiac defects (n=16 [11.0%], P<.001) or in those with isolated outflow tract abnormalities (n=3 [30.0%], P=.005) than in the structurally and karyotypically normal fetuses of women of advanced maternal age. Table 5 summarizes the frequency of the commonly seen copy number variants in our series. The majority of other-than-common benign copy number variants (50.8%) only occurred once.
We have confirmed that chromosomal microarray analysis increases the detection of prenatally diagnosed genomic abnormalities in women with recognized fetal anomalies. The American College of Obstetricians and Gynecologists and the Society of Maternal-Fetal Medicine now recommend that chromosomal microarray analysis be performed in lieu of karyotyping in pregnancies with an anomalous fetus undergoing invasive testing.1,2,4,5 In this study, we have demonstrated that this increased detection depends on the number and type of anomaly. Copy number variants are more likely when the fetus has multiple anomalies and in those with isolated anomalies, the greatest yield occurs in cardiac and renal anomalies. This information should be of value in patient counseling and pregnancy management.
These findings extend the work of others.2 For example, in an analysis of chromosomal microarray analysis results from 2,828 women, clinically significant copy number variants were seen in 6.5% of fetuses with anatomic anomalies,4 a frequency similar to that seen in our study. In that series, chromosomal microarray analysis was particularly informative when craniofacial and cardiac malformations were seen. However, a direct comparison to this study is not possible because almost one fourth of cases in the report had large copy number variants but no available karyotype, making it impossible to quantify the incremental value of microarray.
One of the strengths of our analysis is the prospective data collection. All consecutive patients with anomalous fetuses were offered simultaneous microarray analysis at the time of invasive testing. Retrospective studies from referral genetic laboratories include selected patients, some of whom had chromosomal microarray analysis to improve the interpretation of a previously characterized karyotype.4 Another study strength was the expertise of the ultrasonographers. All 29 sites were American Institute of Ultrasound in Medicine–accredited and followed standardized guidelines.6 Additionally, a standardized array design was used for all cases.
Although our study is large, it was not powered to address the association of specific deletions with specific ultrasound-detected abnormalities. Indeed, more than half of the copy number variants detected occurred only once. To enhance the clinical usefulness of chromosomal microarray analysis, it is essential to continue to collect additional phenotype–genotype correlations. The large number of relatively rare copy number variants associated with anomalies demonstrates that use of targeted arrays containing only a limited number of well-described copy number variants is not sufficient. For example, when a cardiac defect is detected, limiting analysis to karyotyping and fluorescence in situ hybridization for the common 22q11.2 deletion would fail to identify more than two thirds of the genomic findings.
To quantify the incremental value of microarray testing in anomalous fetuses, we compared them with structurally normal fetuses being sampled for advanced maternal age. Because copy number variants are not age-related, this is an appropriate comparison group representing the population risk of a copy number variant. This advanced maternal age group had a rate of other-than-common benign copy number variants of 3.6%, much lower than those with structural anomalies. Because 61% of these participants were sampled in the first trimester, it is possible that some could have had anomalies undetected at the time of sampling. However, if this were the case, it would bias the study toward the null hypothesis; thus, the increase in information provided by chromosomal microarray analysis in the presence of a fetal anomaly is, if anything, underestimated.
In our analysis, all variants of uncertain significance were included rather than attempting to classify some as likely pathogenic and others as likely benign. We believe this is the best approach because the field and our knowledge of phenotype–genotype correlations is still evolving. Although this may slightly overestimate the frequency of clinically relevant copy number variants, it allows comparison of the two groups without concern for bias based on scan results.
Cytogenomic information revealed by chromosomal microarray analysis has clinical benefit. In many cases, the etiology of a structural anomaly is unknown and patients are left to make decisions on pregnancy management based on uncertainty as to the long-term prognosis. Identification of a copy number variant allows a more precise understanding of the medical and neurocognitive implications of the anomaly, which is important in decision-making about the pregnancy and in planning care for the child.
In summary, when a fetal anomaly is detected on ultrasound scan, chromosomal microarray analysis offers additional information compared with karyotype with the degree of benefit dependent on the type and number of organ systems involved. Further research using pooled databases is required to provide more precise point estimates of the frequency of copy number variants associated with specific types of anomalies. An ongoing long-term follow-up of the original National Institute of Child Health and Human Development study population will allow better correlation between prenatally detected copy number variants and the subsequent phenotype. There appears to be clear value of chromosomal microarray analysis in the evaluation of fetal structural anomalies, and this study confirms that it should be offered when anomalies are diagnosed.