Discussion
Previous studies revealed that common aneuploidies and recurrent pathogenic copy number variations are the main etiologies of birth defects (Grati et al., 2015; Feldkamp et al., 2017). Therefore, to prevent birth defects, it is critical to confirm these abnormalities in fetuses during pregnancy. Currently, a definitive prenatal diagnosis can only be made through invasive testing to obtain fetal tissue or cells for genetic analyses. In many countries, karyotyping is a criterion test for all patients undergoing prenatal diagnosis and is regarded as the gold standard for numerical chromosomal abnormalities, mosaicisms, and structural rearrangements. However, karyotyping has several limitations: it is expensive, labor-intensive, low-throughput, and has limited ability to detect minor structural rearrangements (Farra et al., 2020). Therefore, the question remains open: What additional assays should be offered to patients that can detect pathogenic variations in a high-throughput, accurate, and timely manner?
In 2016, the American College of Obstetricians and Gynecologists (ACOG) and the Society for Maternal-Fetal Medicine (SMFM) recommended CMA for prenatal diagnosis of cases with one or more fetal structural abnormalities (Vora et al., 2016). In 2018, the Society of Obstetricians and Gynecologists of Canada (SOGC) and the Canadian College of Medical Geneticists (CCMG) recommended that rapid aneuploidy detection (RAD) be performed prior to CMA when multiple fetal malformations are detected or nuchal translucency (NT) is ≥ 3.5 mm (Armour et al., 2018). In 2020, the ACOG-SMFM proposed that for women who wish to evaluate their pregnancy for submicroscopic chromosomal change, prenatal diagnostic testing with CMA is recommended, however, no specific clinical indications for the application of CMA was explicitly proposed in this practice bulletin (Rose et al., 2020). Thus, for patients without fetal ultrasound abnormalities, there is no clear consensus on whether it is necessary to routinely recommend genetic tests other than karyotyping. On the other hand, a recent study (Hay et al., 2018) reported that for patients undergoing prenatal diagnosis with structurally normal fetuses, 2.5% of fetuses with clinically significant chromosomal abnormalities would be missed if only karyotyping had been performed. Meanwhile, the authors also acknowledged the two major barriers to the use of CMA for women with structurally normal fetuses, the VUS results appeared frequently (5.4% in this study) and the high costs. This result suggests that for cases without fetal ultrasound abnormality, novel, rapid genetic detection methods that are superior to conventional RAD approaches are needed to improve the detection efficiency of pCNV while circumventing the negative impact of performing whole-genome CNV tests.
The PNBoBs™ assay could be one such method. Compared to other rapid aneuploidy detection techniques such as Quantitative fluorescence PCR (QF-PCR) and Fluorescent in situ hybridizations (FISH), despite its slightly higher overall costs, the PNBoBs™ retain the advantages of rapidity and accuracy while expanding the target disease from common aneuploidies to frequent microdeletion syndromes. Over the last decade, several studies have confirmed that combining PNBoBs™ with karyotyping can effectively improve prenatal pCNV detection. However, none of these studies specified whether it should be applied to patients with different referral indications, or how to use the results to make clinical decisions (Grati et al., 2015; Fang et al., 2018; Huang et al., 2018; Li et al., 2018; Miao et al., 2019; Tao et al., 2019; Farra et al., 2020). With the clinical application of PNBoBs, different perspectives have emerged; some have suggested that it can replace traditional karyotyping (Piotrowski et al., 2012), arguing that PNBoBs™ can detect most clinical cases of pCNV and that it possesses advantages over traditional karyotyping. However, a more recent, alternative view is that PNBoBs™ should be replaced by genome-wide CNV analytical techniques, such as CNV-seq (Xu et al., 2020), regardless of the indication for prenatal diagnosis. These different views reflect the complexity of and lack of consensus on application criteria for PNBoBs™ in prenatal diagnosis.
Current guidelines recommend CMA as a first-tier test, but only when a fetal structural anomaly has been detected by ultrasound. However, with other indications for prenatal diagnosis, there are currently no sufficient evidence-based guidelines to support the diagnostic value of CMA. Several studies have reported the detection rate of the PNBoBs™ assay in patients with different indications for prenatal diagnosis, but their classification methods differed. Some classified all fetal ultrasound findings as ultrasound anomalies (Fang et al., 2018; Huang et al., 2018; Tao et al., 2019; Wang Y. et al., 2020), while others further divided findings into ultrasound structural malformations (FUSA) and soft-marker abnormalities (FUSMA) according to existing guidelines (Rosenfeld et al., 2014; Grati et al., 2015). We adopted the latter classification method for this study. Three cases of pCNV in the FUSA group that were missed by karyotyping were detected by PNBoBs™. Thus, the pCNV detection rate in our FUSA group [83% (10/12)] was much higher than that reported by Rosenfeld et al. [16.4% (108/660)] (Rosenfeld et al., 2014) and Grati et al. [18.7% (360/1922)] (Grati et al., 2015). This difference may be attributed to the small size of the FUSA group in our cohort. In the HR-NIPT group, 53% (10/19) of the cases were positive using either method. This finding was consistent with the rates reported by Huang et al. [50% (3/6)] (Huang et al., 2018)and Tao et al. [54% (20/43)] (Tao et al., 2019), indicating the efficacy of PNBoBs™ for rapid exclusion or confirmation of common fetal aneuploidies in pregnant women positive for NIPT. Of the five structurally normal cases of pCNV with normal karyotypes detected by PNBoBs™, three were in the HR-MSS group, with one case each in the FUSMA and AMA groups. These data suggest that PNBoBs™ is helpful for improving the detection rate of pathogenic variations in prenatal diagnosis for cases without abnormal structural features detectable by ultrasound.
In the present study, all 29 common aneuploidies and one case of translocation Down syndrome [46,XY,+21,der (21; 21) (q10; q10)] were quickly detected using PNBoBs™ within 48 h of sampling. Consistent with other studies (Tao et al., 2019; Xu et al., 2020), we found that PNBoBs™, at least for common aneuploidies, was highly concordant with karyotyping. Some have recommended that a positive RAD result, without karyotyping, can be directly used as a basis for medical decision-making in clinical management (Ogilvie et al., 2005; Comas et al., 2010). Our results also support this recommendation; however, it remains to be demonstrated that karyotyping helps rule out the possibility of translocation trisomies to provide accurate genetic counseling regarding recurrence risk. Moreover, 38 cases of balanced structural rearrangement, three cases of autosomal mosaicism, two cases of supernumerary marker chromosome, and two cases of unbalanced structural variation in our study were detected only by karyotyping. Among these, six cases showed pathogenic variations. These results indicate that the PNBoBs™ assay cannot completely replace karyotyping in prenatal diagnosis.
In a recent study comparing the detection performance of genome-wide copy-number analysis to PNBoBs™ in prenatal diagnosis (Xu et al., 2020), 10 cases of ambiguous results and two cases of sex-chromosome mosaicism were classified as false negatives by PNBoBs™. However, comparable types of karyotypes were accurately detected using PNBoBs™ in our study, which may be due to our improved data interpretation method for the PNBoBs™ assay (Jiang et al., 2021); If this hypothesis holds, then the rate of diagnostic agreement between CNV-seq and PNBoBs™ in Xu’s study would be 99.8% (1,873/1,876), indicating that the performance of PNBoBs™ may not be inferior to that of genome-wide copy number analysis, without distinguishing clinical indications. Moreover, unexpected findings and/or variants of unknown significance (VUS) identified by genome-wide CNV testing are still very challenging in terms of diagnostic assessment and genetic counseling (Durham et al., 2019). In Xu’s study, VUS accounted for 1.65% of the 1,876 cases assessed using CNV-seq (Xu et al., 2020). In contrast, PNBoBs™ provided explicit results, with clear genotype-phenotypic correlations (Vialard et al., 2011); thus, it is not likely to create ambiguity or confusion for genetic counseling. Furthermore, unlike PNBoBs™, none of the whole-genome CNV tests have regulatory approval for diagnostic use. Due to access restrictions, for prenatal diagnostic cases with indications for whole-genome CNV testing, most prenatal diagnosis centers in China must send samples to a third-party medical testing laboratory. The long-range transport of such samples inevitably increases the reporting turnaround time (approximately 3 weeks), increasing the risk of detection failure. Therefore, at present, we do not recommend whole-genome copy-number analysis for all patients undergoing invasive prenatal diagnosis. Nevertheless, because PNBoBs™ cannot accurately define the range of CNV, we recommend a validation test for CNV classified as incidental findings.
The types of samples tested with PNBoBs™ in previous studies included amniotic fluid (Gross et al., 2011; Vialard et al., 2012; García-Herrero et al., 2014; Fang et al., 2018; Huang et al., 2018; Li et al., 2018; Miao et al., 2019; Tao et al., 2019; Wang Y. et al., 2020; Farra et al., 2020; Xu et al., 2020), umbilical cord blood (Vialard et al., 2012; García-Herrero et al., 2014; Huang et al., 2018; Wang Y. et al., 2020; Farra et al., 2020), and chorionic villi (Huang et al., 2018; Wang Y. et al., 2020; Xu et al., 2020). We tested two types of samples: 2,364 amniotic fluid and four umbilical cord blood samples with specific features for fetal ultrasound abnormalities. The reason why we did not include chorionic villi samples (CVS) is that they carry a higher incidence of mosaicism owing to confined placental mosaicism (Wang et al., 1993). For CVS, if a positive result is obtained by PNBoBs™, further experimental verification is needed from an amniocentesis (Zhong et al., 2020), which will lead to increased medical expenses and a longer detection period. For umbilical cord blood, visual inspection is not suitable for assessing MCC; this will increase the risk of erroneous conclusions (Nagan et al., 2011). In contrast, although it is not always reliable, visual inspection is still considered a convenient way to assess MCC in amniotic fluid (Winsor et al., 1996). Therefore, to ensure the reliability and timeliness of the PNBoBs™ assay and leave enough time for follow-up pregnancy management, amniotic fluid collected in the second trimester of pregnancy is the ideal biological specimen for PNBoBs™ testing.
In conclusion, this study addressed the applicable populations, appropriate sample type, and clinical pathway for the PNBoBs™ assay in invasive prenatal diagnosis. We found that combining karyotyping with PNBoBs™ can improve both the diagnostic yield and efficiency of prenatal diagnosis and could be recommended in the second trimester in all patients without fetal structural anomalies who undergo invasive prenatal diagnosis, as a supplement to the parts not covered by the current guideline consensus, However, one limitation that needs to be noted is that the origins of all CNVs detected by PNBoBs™ could not be confirmed because of the refusal of some patients’ family members. Moreover, further investigations are required to better understand the phenotype-genotype correlations of microduplications in target regions with detailed clinical follow-up.