Performance of Cell-Free DNA Screening for Fetal Common Aneuploidies and Sex Chromosomal Abnormalities: A Prospective Study from a Less Developed Autonomous Region in Mainland China

To evaluate the performance of noninvasive prenatal screening (NIPS) in the detection of common aneuploidies in a population-based study, a total of 86,262 single pregnancies referred for NIPS were prospectively recruited. Among 86,193 pregnancies with reportable results, follow-up was successfully conducted in 1160 fetuses reported with a high-risk result by NIPS and 82,511 cases (95.7%) with a low-risk result. The screen-positive rate (SPR) of common aneuploidies and sex chromosome abnormalities (SCAs) provided by NIPS were 0.7% (586/83,671) and 0.6% (505/83,671), respectively. The positive predictive values (PPVs) for Trisomy 21, Trisomy 18, Trisomy 13 and SCAs were calculated as 89.7%, 84.0%, 52.6% and 38.0%, respectively. In addition, less rare chromosomal abnormalities, including copy number variants (CNVs), were detected, compared with those reported by NIPS with higher read-depth. Among these rare abnormalities, only 23.2% (13/56) were confirmed by prenatal diagnosis. In total, four common trisomy cases were found to be false negative, resulting in a rate of 0.48/10,000 (4/83,671). In summary, this study conducted in an underdeveloped region with limited support for the new technology development and lack of cost-effective prenatal testing demonstrates the importance of implementing routine aneuploidy screening in the public sector for providing early detection and precise prognostic information.


Introduction
Noninvasive prenatal screening (NIPS) using circulating cell-free DNA (cf-DNA) in maternal plasma has been widely implemented as a routine screening method for fetal chromosomal aneuploidy in obstetric practice. Previous studies have indicated that NIPS showed a high accuracy in detecting common trisomies (Trisomy 21, Trisomy T18 and Trisomy T13) and the feasibility in reporting sex chromosome abnormalities (SCAs) by using massively parallel sequencing (MPS), chromosome-specific sequencing (CSS) and single nucleotide polymorphism (SNP)-based methods via various next-generation sequencing (NGS) platforms [1][2][3][4]. Currently, it is recommended for aneuploidy screening in all pregnancies, regardless of maternal age or other risk factors [5], by the American College of Obstetricians and Gynecologists (ACOG).
In China, NIPS has been recommended as a routine prenatal screening test to assess the risks associated with pregnancies [5]. Besides commercial companies, hospitals accredited for prenatal diagnosis have started offering standard NIPS, while expanded NIPS also becomes widely used in some developed areas of China [21]. However, most of the published data from China were collected from different centers. Due to the small sample size without follow-up data, the number of different types of abnormalities from each single center was insufficient for comparison [8,10,12], or represent the general population. Moreover, the utility of expanded NIPS in the general obstetric population is still on debate. Therefore, evaluating the performance and utility of standard NIPS in a prospective large sample size from a local center is necessary and urgent. We used less developed Guangxi Zhuang Autonomous Region in China as an example. For this study, we included Zhuang, the second largest ethnic group in China, and other minorities' groups. Guangxi region has a large genetic heterogeneity and a high incidence of birth defects. In perinatal period, the incidences of birth defects are about 156/10,000 births in Guangxi (Guangxi Birth Defect Prevention Report, 2020). Additionally, the expanded NIPS currently offered in Guangxi region is very expensive due to its higher sequencing read-depth (~0.15-0.3×; reagent cost as USD120 per case), compared with standard NIPS (~0.04-0.15×, reagent cost as USD60 per case). As shown in previous studies published in Chinese cohorts, a total of 1.2% fetal chromosome abnormalities were detected by NIPS [8] and 16.7-19.0% fetal chromosome abnormalities were diagnosed by chromosomal microarray analysis (CMA) and low-pass genome sequencing [8,22]. This demands a cost-effective approach, such as standard NIPS, for prenatal screening, mainly to detect common fetal aneuploidies.
Herein, it is a prospective study about the performance of screening for common aneuploidies and sex chromosomal abnormalities, using noninvasive assessment of the fetal genome from maternal serum-noninvasive prenatal screening test in Guangxi Zhuang Autonomous Region, in China. A large sample size of 86,262 consecutive pregnant women who received standard NIPS in a local single center were recruited for evaluating the performance of detecting different chromosomal abnormalities in different risk cohorts. In addition, outcome follow-up was successfully conducted in the majority of the cases (n = 83,671, 97%). Overall, this study provided valuable references for genetic counseling and clinical application. Eventually, its implementation could significantly reduce the rate of birth defects in this region. All patients were recruited with the following criteria, based on the guidelines of the American College of Medical Genetics and Genomics (ACMG) [23] and the National Health Commission of the People's Republic of China (2016): maternal age ≥16, singleton pregnancy ≥12 weeks of gestation, and no history of transfusion or transplantation during past years.

Study Population and Samples
According to prenatal screening protocols in the hospital: Ultrasound examinations were provided routinely for all pregnant women in the obstetric checkpoints (11-13 +6 weeks, 16-18 weeks, 22-24 weeks, 28-32 weeks and after 38 weeks) to measure and assess for fetal nuchal translucency (NT), structural abnormalities and growth and development. Apart from ultrasound examinations, (1) for first-trimester Down syndrome screening (DSS1) (11-13 +6 weeks), detection of pregnancy-associated plasma protein A and free β-chorionic gonadotrophin were offered; and (2) for second-trimester DSS (DSS2, 15-20 +6 weeks), detection of α-fetoprotein, unconjugated estriol and free β-chorionic gonadotrophin were provided. For women with intermediate/high-risk pregnancy indicated by either first-or second-trimester screening, NIPS was offered based on the guidelines of the National Health Commission of the People's Republic of China (2016). Those with structural abnormalities reported by ultrasound screening include cardiac malformations, cleft lip and palate, fetal hydrops, limb malformations, cystic hygroma, renal dysplasia, lung cystadenomas, etc. They are excluded in this study, resulting in a total of 86,262 pregnancies. All eligible pregnant women undertook pre-test counseling and provided written informed consents. A flowchart of this study is illustrated in Figure 1.
For each participant, 10 mL maternal peripheral blood was collected with Streck Cell-Free DNA BCT tubes (La Vista, New England, USA) and stored at room temperature (around 24 • C). Each sample was then proceeded (within 72 h after collection) for a twostep centrifugation, centrifuged at 1600× g for 15 min (4 • C) and 16,000× g for 10 min (4 • C), to isolate the supernatant plasma. Subsequently, the plasma sample was transferred to a 1.5 mL fresh EP tube, stored at −20 • C for the first day and at −80 • C afterwards, for further processing. Each plasma sample was frozen and thawed only once, to avoid DNA degradation.

DNA Extraction, Library Construction, Sequencing and Bioinformatics Analysis
NIPS was performed by following the manufacturer's instructions. Briefly, cf-DNA was extracted from 1.2 mL plasma, using magnetic beads-based plasma cf-DNA extraction kit (Berry Genomics, Beijing, China), subjected to library construction and purification, using the magnetic beads-based MPS DNA library construction and purify kit (Berry Genomics, Beijing, China). Libraries were quantified by Kapa SYBR fast qPCR kit (Kapa Biosystems, Wilmington, USA), on a StepOnePlus platform (ABI, Thermo Fisher Scientific, Waltham, MA, USA). Libraries of 96 samples with barcodes were then pooled together with equimolar and subjected for single-end sequencing (37 base-pairs with another 8 base-pairs as index) on a Nextseq-500 platform (Illumina, San Diego, CA, USA).
Sequencing reads were endured if sequencing quality value (Q30) was >85%, and GC content ranged from 38 to 42. Overall, after mapping to the human reference genome (GRCh37/hg19) by RUPA software developed by Berry Genomics [8], a minimum of 3 million uniquely aligned reads was obtained for each sample. The data were then processed by a Bambni Test Data Analysis System 3.7 (Berry), with parameters reported in previously published studies [8,24]. After GC correction, fetal fraction (FF) was estimated by using elastic net (ENET) algorithm [25]. Subsequently, chromosome aneuploidy was reported using the criteria of Z-score ≥ 3 (trisomy) or ≤ −3 (monosomy). When different fetal fractions were reported by two algorithms (ENET and chromosome Y-based), mosaic chromosome aneuploidy was considered. The analytical algorithm for CNVs was reported in previous studies [8,26], with a resolution of 5 Mb.
All samples failed in any of the following steps were requested for resubmission: (1) samples were requested for resampling (such as whole blood or plasma with poor quality, low FF, multiple failing factors of sequencing and failure in re-experiment) and processed by using the peripheral blood/plasma procedure, as described above; (2) samples were requested for re-examining/re-experiment due to failed quality control (QC), such as high or low concentration of DNA extraction, abnormal peak or low concentration of library construction, GC content bias, insufficient data amount, multiple chromosome abnormalities (Multis), borderline Z-score, multiple failing factors of sequencing, etc. A repeat of library construction, sequencing and data analysis was performed afterwards; the same procedure (DNA extraction, library construction, sequencing and bioinformatics analysis) was conducted as described above.

Clinical Outcome and Follow-Ups
Women with positive NIPS results were offered confirmatory diagnostic tests (karyotyping and/or SNP array) via invasive procedure and post-test counselling for pregnancy management. Clinical evaluation was carried out for those pregnancies kept to term birth. Follow-up information was obtained from all NIPS participants by phone or interview, including any ultrasound abnormalities and/or soft markers found during pregnancies, pregnancy outcomes, any dysmorphic features and abnormalities found in newborns.

Data Analysis
Open-source tool python (https://www.python.org/) was used for data plotting and statistical analysis. Mean, standard deviation, median and range were calculated for each of the biographic datum (such as gestational age, maternal age, weight and body mass index (BMI)), as well as FF. Student's t-test was used to compare the differences of FF between reportable results and no-call groups, and between second-trimester and third-trimester groups. Comparisons of FF and related factors (such as gestational age, maternal age, weight, BMI and Z-score) among multiple groups were performed by oneway ANOVA. Chi-Square test and Fisher's exact test (n < 5) was used to (1) compare the prevalence of pregnancy women opted for continuing pregnancies in groups, with or without confirmation by prenatal diagnosis after receiving a positive results by NIPS; (2) compare the prevalence of pregnancy in women who opted for termination of pregnancies (TOPs) in groups between common trisomies and SCAs, with confirmation by prenatal diagnosis after receiving a positive results by NIPS; and (3) analyze the difference of screenpositive rate (SPR) and PPV among different referral indications. A p-value < 0.05 was considered as statistically significant. Those no-call samples and low-risk cases without follow-up results were excluded from the study, when calculating the prevalence, sensitivity and specificity, PPV, negative predictive value (NPV), false positive rate (FPR), and false negative rate (FNR) of the test.

Characteristics of Study Population
A total of 86,262 singleton pregnancies were referred for NIPS testing during the study period, and NIPS yielded results in 86,193 (99.9%) cases. Maternal age ranged from 16 to 54 years, with a median of 33 years. Of the 86,193 cases, 37,387 pregnancies (37,387/86,193, 43.38%) were advanced maternal age (AMA). In addition, the median gestational week was 17 +3 weeks (ranged 12-38 weeks), while 1456 (1.69%) of pregnant women were with obesity (BMI ≥ 30 kg/m 2 ) observed [27]. Furthermore, 3.25% (2802/86,193) of women conceived with assisted reproductive technology (ART). The demographic characteristics of these participants are summarized in Table 1.

Testing Failure Rate and Results Distribution of NIPS
After sample submission, there were three QC steps; samples that failed in any step were requested for sample resubmission for NIPS test ( Table 2) (Table 2). Among the group with low FF (n = 125, mean FF as 2.95%), the average FF of the resubmitted samples was increased to 5.49% (range from 3.19 to 15.53%), and the average FF was significantly higher in the cases with yielded results (5.49%, n = 110), as compared to those without (or namely no-call, mean 2.82%, n = 15, p < 0.001 ( Table 2). The FF was significantly increased with gestational weeks (p < 0.001) but obviously decreased with maternal weight and BMI (both p < 0.001) ( Figure 2). We also found significantly higher FF in the second-trimester group (n = 78,231, 90.76%, cf-DNA =~12.57%) than that in the third-trimester (n = 584, 0.68%, cf-DNA =~23.0%,) (p < 0.001, not shown in the table). The results also showed a positive relationship between FF and Z-score of trisomy 21 (p < 0.001). The detailed results are shown in Table A1.  Among the 86,193 cases (99.9%) with a reportable NIPS result, there were 1160 fetuses reported to be high-risk (1.35%) by NIPS, of which 586 cases (50.52%) were common trisomies, 505 cases (43.53%) were SCAs and 69 cases (5.95%) were rare chromosomal abnormalities. Overall, 84.57% of them (981/1160) received an invasive/clinical diagnosis for further confirmation.
Among the 525 cases with common trisomies who received invasive confirmations, 444 were confirmed (330 cases of T21, 84 cases of T18 and 30 cases of T13). Among these, 17 cases had atypical cytogenetic findings, including mosaicism of trisomies (mosaic level: 7 to 58%), Robertsonian T21 and other chromosomal structural abnormalities. Therefore, the PPV for T21, T18 and T13 were 89.67%, 84.00% and 52.63%, respectively. By considering those common trisomies (three with T21 and one with T18) identified in NIPS low-risk pregnancies, the combined sensitivity, FPR, PPV and NPV for common trisomies were 99.11%, 0.10% and 84.57% and 99.995%, respectively. In comparison, among the 400 cases with SCAs receiving further confirmations, the combined FPR and FNR were 0.30% and 1.94%, respectively. It resulted in a combined PPV of 38.00%. Among three categories of SCAs, both FPR and FNR were highest for Monosomy X, resulting in a lowest PPV of 12.04%. In comparison, in the group which indicated decrease of X chromosome DNA in male fetuses, the FPR and FNR of 46,XY(delX) were both the lowest among three groups. However, the PPV of 46,XY(delX) was also the lowest, probably owing to the low SPR (0.04%). The detailed results are shown in Table 3.
Overall, NIPS yielded high sensitivity and specificity in detecting common trisomies and SCAs among different referral indications ( Table 3). The combined PPV for common aneuploidies and SCAs was 64.43%. For cases with Trisomy 21, the PPVs in cases with AMA, DSS high-risk and ultrasound cohorts were higher than that in the other cohorts (p = 0.07), while the PPVs for monosomy X varied in different referral indications (range from 0 to 50%, p = 0.06, Table 4); however, no significant differences were found for these PPVs.

Additional Findings
Apart from common trisomies and SCAs, there were 69 cases with other chromosomal abnormalities reported by NIPS, accounting for 5.95% (69/1160) of the overall positive cases. Thirteen cases were CNVs (≥5 Mb), and 56 cases with RATs. For CNVs, CNV sizes ranged from 6 to 32.5 Mb for deletions (n = 8) and 5.5 to 99 Mb for duplications (n = 9), respectively. Twelve of these 13 cases (92.3%) underwent invasive diagnosis, and the results confirmed four CNVs (4/12, 33.3%). In addition, among pregnancies that sought invasive prenatal diagnosis in the NIPS low-risk group, there were an additional 16 CNVs (≥5 Mb) identified (Table 3).
For RATs, there were 17 cases with increased dosage of chromosomes (likely trisomy), 15 cases with decreased dosage of chromosomes (likely monosomy) and 24 cases with Multis by NIPS. However, only nine of them (9/44, 20.5%) were confirmed by invasive testing. In addition, an additional RAT (mosaic T22) was reported in NIPS low-risk group (Table A2). For these RATs with Multis (n = 24), 75% (18/24), cases were reported to have more than three chromosome abnormalities, and most of them included common trisomies. There were only five cases confirmed by further invasive testing (5/22), and three of them, in fact, were common trisomies.
In addition, in one case with increased dosage of chromosome 18 by NIPS (Table A2, No. 46), CMA showed uniparental disomy 18, which might be resulted from trisomy rescue. We further followed up those cases with normal fetal karyotypes (n = 17), and maternal aplastic anemia and gallstone were identified in two cases, respectively. Interestingly, the diagnostic results presented a high combined PPV (87.5%, 7/8) of segmental deletion in chromosome 18 (chr18-) that was reported in 10 cases with CNVs and RATs (Table A2).
Overall, for CNVs and RATs, including Multis, the PPVs were 33.33% and 20.45%, respectively. The combined PPV, NPV and FPR for these additional abnormalities were 23.21%, 99.98% and 0.05%, respectively. In total, when NIPS detection scope were expanded from common trisomies and SCAs to genome-wide abnormalities (CNVs and RATs), the combined PPV was only slightly decreased from 64.43% to 62.08%, and the FPR was also slightly increased from 0.40 to 0.45% (Table 3), both of which were similar to a previous study [12].

Clinical Outcome and Follow-Up
Among the 82,511 NIPS reported low-risk cases with pregnancy outcomes available, 1242 cases (1.51%) had birth defects and 1163 cases (1.41%) had adverse pregnancy outcomes; most of them (99.0%, 2381/2405) were not affected by the chromosomal abnormalities that were within our study scope (Tables A3 and A4). Among the 24 false negative cases, four cases with common trisomies (three T21 and one T18) presented abnormal phenotype in their fetal or neonatal periods, and three cases with mosaicism of monosomy X. In addition, antenatal and/or postnatal anomalies were reported among the 12 of the other 17 cases with other chromosomal abnormalities detected (70.59%).
For 179 cases with NIPS high-risk results but that did not undergo invasive confirmations, follow-up was successfully conducted in all cases. Sixty-eight (38.0%) opted for termination, and 111 (62.0%) continued pregnancies, respectively. In addition, among these 179 cases without invasive confirmations, 82.9% (87/105) of cases reported to have SCAs by NIPS opted for continuing pregnancies, which was significantly higher than that with confirmation of SCAs (47.4%, 72/152, p < 0.001; Table A4).

Discussion
Our study provides the largest sample size of NIPS pregnancy outcome data (n = 83,671) from a single center. This prospective population-based study, demonstrated that standard NIPS is a cost-effective method for detecting fetal common aneuploidies in a less developed region. This is further supported by high quality of pregnancy outcome follow-up data. Among the 86,193 pregnant women received a NIPS result with the follow-up tests, the PPVs for T21, T18, T13 and SCAs provided by standard NIPS were at 89.7%, 84.0%, 52.6% and 38.0%, respectively, which were comparable to those reported by higher sequence read depth expanded NIPS (95%, 82%, 46% and 47%) [8]. The DR, PPV, NPV and FPR of detecting common trisomies were also similar to those reported in previous reports [1,2,8]. In addition, we further evaluated the NIPS performance for common trisomies among different risk cohorts. PPVs of Trisomy 21 demonstrated a higher accuracy among cohorts with ultrasound soft markers (95%), DSS high-risk result (94.81%) and AMA (92.82%) cohorts, as comparable to other risk cohorts (ranged 75-85%). Our data also supported the ACOG to advocate the use of NIPS for common aneuploidies in all pregnancies regardless of any risk, especially for high-risk population (higher PPV), and demonstrated the robustness of the NIPS platforms even in less developed autonomous region. Our study not only enhanced the accuracy of evaluating the FNR for each type of chromosomal abnormalities by NIPS, but it also provided a baseline of birth defects/adverse pregnancies (2.92%, 2405/82,511) in less developed autonomous region. The data demonstrated that most of NIPS negative cases with birth defects (99.83%, 2401/2405) were not affected by the common chromosomal abnormalities.
We also compared the combined PPV for SCAs (38.00%), which was also comparable to that reported by expanded NIPS (46.7%) [8]. Overall, the PPV (12.04%) for monosomy X was lower than previously reported from two sequencing platforms (19.39% to 28.57%) [3]. One of the reasons might be that the PPV in the AMA cohort was extremely low (4.29%) compared to the other groups (Table 4). However, we cannot exclude the low PPV contributed by confined placental mosaicism [28] or maternal monosomy X mosaicism [29]. PPVs for SCAs reported by NIPS (lower PPV compared with the one for common trisomies) are extremely important for the pregnant women's decision-making. For cases with suspected Monosomy X, apart from reporting a simple Monosomy X, complex chromosomal abnormalities involving X chromosome such as translocations, isochromosomes, rings were identified in eight cases by further invasive testing. Observation of such complex rearrangements was similar to the previous study [30]. For instance, among the 179 cases with high-risk result by NIPS but did not pursue invasive confirmation; 82.9% (87/105) of cases reported to have SCAs that opted for continuing pregnancies, significantly higher than that with confirmation of SCAs (47.4%, 72/152, p < 0.001, Table A4). Another reason might be that most of the fetuses with SCAs except monosomy X would have less lethality, as compared to the ones with common trisomies.
In our study, we also explored the applicability of standard NIPS in identifying other chromosomal abnormalities. First, standard NIPS was able to provide a 33.3% PPV for reporting CNVs (>5 Mb), which was also comparable to the one reported by expanded NIPS (40.8%). However, the number of reported CNV cases (13/86,193) was much less than the one reported by expanded NIPS in a cohort with similar sample size (163/94,085) [8].
Despite low FPR for these extra abnormalities, the relatively low PPVs obtained might lead to increase the risks of parental anxiety and the unnecessary invasive procedure. In addition, previous prenatal CMA findings indicated that 84.0% of known pathogenic CNVs are less than 5 Mb [31], which were unlikely to be detectable even by expanded NIPS. For the segmental chromosomal abnormalities, fetuses would commonly present with ultrasound anomalies, which warrants an invasive test for comprehensive genetic evaluation via CMA or genome sequencing (low-pass or high read-depth) [22,32,33].
In addition, our study also showed low-to-moderate PPVs for RATs (20.45%), which were also likely comparable to those reported by expanded NIPS (6 to 28.6%) [8,10]. For Multis, confound factors such as maternal tumor [34] and systemic lupus erythematosus [35] were known, while two cases with maternal blood disease and gallstone were found in our study (2/18 with further follow-up). In the context that current guidelines on reporting such events (CNVs and RATs) are still controversial [36]. As in China, Guangxi is an economically underdeveloped area, with limited resources (including limited support in new technology development; lack of cost-effective prenatal testing, such as NIPS, available; and absence of technical staff supports). Although expanded NIPS have been developed internationally, the application is restricted by the limited resources. In addition, the charge fee of standard NIPS (USD221 per case), including USD60 as reagent cost, is comparable to the routine G-banded chromosome analysis (USD230 per case). Therefore, standard NIPS is still regarded as an affordable prenatal screening test for common aneuploidies and SCAs in the Guangxi region.
Moreover, among 10,541 pregnancies with high-risk results by DSS (mainly enriched by DSS2 cohort), only 1.9% (201/10,541) of samples yielded high-risk results by NIPS, which can potentially prevent about 98% pregnant women from obtaining a further invasive prenatal diagnosis. Similarly, in previous study [1], NIPS yielded an extremely lower FPR (nearly 100 times) and higher PPV (80.9%) for common trisomies, as compared with standard screening (with measurement of nuchal translucency and biochemical analytes, 3.4%). Therefore, the performance of standard NIPS from a less developed autonomous region in Mainland China is superior to DSS methods, with reporting lower FPR and higher PPV. In additional, standard NIPS yielded high sensitivity (>90.48%) and specificity (>99%) in detecting common trisomies and SCAs (XXY, XXY, XYY) among different referral indications; however, relatively lower sensitivity of trisomy 18 in ultrasound soft markers group (n = 4, sensitivity = 75%) and monosomy X in low-risk cohort (n = 33, no clinical indications group, sensitivity = 75%) were observed. While the PPVs for trisomy 21 (75-95%), SCAs (XXY, XXY, XYY, 47-78%) and monosomy X (0-50%) were variable, there were no significant differences in these referral indications (p > 0.05) ( Table 4). Therefore, the results demonstrated the use of standard NIPS detection of common trisomies and SCAs is feasibility in any risk cohorts, especially when there were no other screening options available in this less developed region. Overall, the information will be useful for clinical counseling, to help and support pregnant women in the Guangxi region to go for NIPS.
The strengths of our study include (1) a large sample size and outcome follow-up data (n = 83,671, 97.07%) from (2) a single center and local cohort in a less developed region and (3) consecutive pregnancies from a prospective cohort. However, there are limitations we need to face: (1) There was a lack of samples with expanded NIPS for comparison in study, as it is expensive and not allowed for a large-sample size application.
(2) Despite low FPR for these abnormalities, relatively low PPVs (such as Monosomy X, CNVs and RATs) were still existed and might lead to increase the risks of parental anxiety and the unnecessary invasive procedure. (3) Almost all false negative cases were confirmed by karyotyping/SNP microarray; however, there were still a large number of cases with abnormal phenotypes (such as abnormal ultrasound, birth defects, stillbirth and miscarriage) lacking further genetic diagnosis requesting for follow-up genetic diagnosis with advanced methods. (4) Because chromosomal abnormalities are known to cause variable phenotype presentations, especially for a phenotype with atypical abnormality and intellectual disability, which is difficult to identify during the pregnancy, they could lead to the loss of a follow-up. This raises important issues regarding genetic counseling, which is underdeveloped in Guangxi. (5) Lastly, the majority of our cases was from the second-trimester pregnancy (n = 78,231, 90.76%), while there was only 0.68% of cases from the third-trimester pregnancy, which is relatively small (n = 584, 0.68%, cf-DNA =~23%), preventing a proper comparison; therefore, a larger sample size of cases from the thirdtrimester pregnancy is warranted for comparison in the near future.
In order to improve clinical management for pregnant women choosing NIPS, there are efforts we must take: entirely complete monitoring information (ultrasonography, postnatal CT and MRI) and establish workflow and management for prenatal diagnosis and follow-up, especially improving abilities for specialized counseling in local region. Since completed diagnosis and follow-up data can be recorded and collected from a specific system in Guangxi, this, in turn, will promote clinical practice and further study of NIPS in local regions.

Conclusions
In summary, our study with sufficient follow-up data (97.07%, 83,671/86,193), demonstrated that standard NIPS can provide a sensitive screening method in the detection of fetal common aneuploidies and sex chromosomal abnormalities. Standard NIPS may be a cost-effective method for routine prenatal screening in a less developed region in Mainland China.