Objective
We sought to report on laboratory and clinical experience following 6 months of clinical implementation of a single-nucleotide polymorphism–based noninvasive prenatal aneuploidy test in high- and low-risk women.
Study Design
All samples received from March through September 2013 and drawn ≥9 weeks’ gestation were included. Samples that passed quality control were analyzed for trisomy 21, trisomy 18, trisomy 13, and monosomy X. Results were reported as high or low risk for fetal aneuploidy for each interrogated chromosome. Relationships between fetal fraction and gestational age and maternal weight were analyzed. Follow-up on outcome was sought for a subset of high-risk cases. False-negative results were reported voluntarily by providers. Positive predictive value (PPV) was calculated from cases with an available prenatal or postnatal karyotype or clinical evaluation at birth.
Results
Samples were received from 31,030 patients, 30,705 met study criteria, and 28,739 passed quality-control metrics and received a report detailing aneuploidy risk. Fetal fraction correlated positively with gestational age, and negatively with maternal weight. In all, 507 patients received a high-risk result for any of the 4 tested conditions (324 trisomy 21, 82 trisomy 18, 41 trisomy 13, 61 monosomy X; including 1 double aneuploidy case). Within the 17,885 cases included in follow-up analysis, 356 were high risk, and outcome information revealed 184 (51.7%) true positives, 38 (10.7%) false positives, 19 (5.3%) with ultrasound findings suggestive of aneuploidy, 36 (10.1%) spontaneous abortions without karyotype confirmation, 22 (6.2%) terminations without karyotype confirmation, and 57 (16.0%) lost to follow-up. This yielded an 82.9% PPV for all aneuploidies, and a 90.9% PPV for trisomy 21. The overall PPV for women aged ≥35 years was similar to the PPV for women aged <35 years. Two patients were reported as false negatives.
Conclusion
The data from this large-scale report on clinical application of a commercially available noninvasive prenatal test suggest that the clinical performance of this single-nucleotide polymorphism–based noninvasive prenatal test in a mixed high- and low-risk population is consistent with performance in validation studies.
Since becoming clinically available in late 2011, cell-free DNA (cfDNA)-based noninvasive prenatal testing (NIPT) for fetal aneuploidy has seen an unprecedented rapid adoption into clinical care. This followed multiple publications on methodologies, validation, and test performance, all demonstrating improved sensitivities and lower false-positive (FP) rates than current screening methods. Opinion statements by national and international professional societies support the clinical use of NIPT in pregnant women, with most recommending use restricted to women at high risk for fetal aneuploidy.
Two approaches to NIPT have been developed and commercialized. In the first approach, fetal chromosome copy number is determined by comparing the number of sequence reads from the chromosome(s) of interest to those from reference chromosomes. The second approach entails targeted amplification and sequencing of single-nucleotide polymorphisms (SNPs). This approach requires a sophisticated informatics-based method to compute aneuploidy risk through SNP distribution. Validation of the SNP-based NIPT method at 11-13 weeks’ gestation was recently reported, demonstrating high sensitivity and specificity for detection of trisomy 21, trisomy 18, trisomy 13, Turner syndrome (monosomy X), and triploidy.
Despite hundreds of thousands of tests already having been performed worldwide, there are few large-scale reports describing performance of NIPT in actual clinical settings, with most studies reporting on <1000 total patients. Here, laboratory and clinical experience of >31,000 women who received prenatal screening with a SNP-based NIPT is reported.
Materials and Methods
This is a retrospective analysis of prospectively collected data on 31,030 cases received for commercial testing from March through September 2013. This study received a notification of exempt determination from an institutional review board (Albert Einstein College of Medicine Institutional Review Board: no. 2014-3307). Samples were classified as out of specification and excluded in cases of gestational age <9 weeks, multiple gestation, donor egg pregnancy, surrogate carrier, missing patient information, sample received >6 days after collection, insufficient blood volume (<13 mL), wrong collection tube used, or if the sample was damaged.
Analysis was performed for all samples on chromosomes 13, 18, 21, X, and Y, and included detection of trisomy 21, trisomy 18, trisomy 13, and monosomy X. All samples were processed and analyzed at Natera Inc’s Clinical Laboratory Improvement Act (CLIA)-certified and College of American Pathologists (CAP)-accredited laboratory (San Carlos, CA). Laboratory testing was performed as previously described using validated methodologies for cfDNA isolation, polymerase chain reaction amplification targeting 19,488 SNPs, high-throughput sequencing, and analysis with the next-generation aneuploidy test using SNPs (NATUS) algorithm. Samples were subject to a stringent set of quality-control metrics. A second blood draw (redraw) was requested if total input cfDNA, fetal cfDNA fraction, or signal-to-noise ratio did not meet quality metrics, or for poor fit of the data to the model. In cases of large regions (>25%) of loss of heterozygosity or suspected maternal or fetal mosaicism, redraw was not requested. Reports included a risk score for the 4 aneuploidies; when requested, reports included fetal sex. Risk scores were calculated by combining the maximum likelihood estimate generated by the NATUS algorithm with maternal and gestational age prior risks. All samples with a risk score ≥1/100 were reported as high risk for fetal aneuploidy and samples with risk scores <1/100 were considered low risk. For the purposes of this study, the high-risk results were further divided into a maximum-risk score of 99/100 or an intermediate-risk score of ≥1/100 and <99/100. The presence of >2 fetal haplotypes (indicative of either triploidy or multiple gestation) was reported only when the confidence was >99.9%. Additional sex chromosome aneuploidies (XXX, XXY, and XYY) were reported from June 2013. The following patient characteristics were requested for each sample: maternal date of birth, maternal weight, gestational age, and whether a paternal sample was included.
Patients with available International Classification of Diseases, Ninth Revision ( ICD-9 ) codes ( Appendix ; Supplementary Table 1 ) were categorized into 3 subcohorts: (1) “low risk” if aged <35 years and no aneuploidy-related high-risk codes; (2) “at risk” for fetal aneuploidy based solely on maternal age ≥35 years; or (3) “high risk” for fetal aneuploidy by ICD-9 code, regardless of maternal age. High-risk indications included positive screening tests, ultrasound anomalies, and relevant family history. Patients without reported ICD-9 codes were categorized by maternal age as low risk (<35 years) or high risk (≥35 years).
Follow-up information on high-risk results was obtained by telephone and recorded in an internal database. Clinical follow-up was completed on June 14, 2014, at which time all pregnancies were completed. Two partner laboratories accounting for 38.1% of the total 31,030 cases were responsible for their own follow-up efforts and were excluded from outcome calculations. Providers were encouraged to share information about false-negative (FN) results. Samples were categorized as follows: (1) “true positive” (TP) included high-risk samples that were confirmed by prenatal or postnatal diagnostic testing, or based on clinical evaluation at birth; (2) “FP” included high-risk samples that were shown to be euploid by follow-up testing or based on clinical evaluation at birth; (3) “suggestive” included samples where prenatal ultrasound detected at least 1 structural anomaly and 1 soft sonographic marker consistent with NIPT findings, but karyotype confirmation was not obtained; (4) “pregnancy loss” where the patient experienced spontaneous abortion and karyotype confirmation was not obtained; (5) “termination” where the patient elected to end the pregnancy without karyotype confirmation; (6) “no follow-up” included samples where information was unavailable; and (7) “FN” included NIPT low-risk samples that were reported as aneuploid by the provider. When placental and fetal karyotypes were both available and determined to be discordant, NIPT findings were considered TP if they matched the fetal karyotype, and FP if they did not match the fetal karyotype. Pregnancies were considered mosaic when chromosome analysis revealed either placental or fetal mosaicism or there was discordance between placental and fetal karyotypes.
Patient and sample characteristics were expressed as means, SD, medians, and ranges. Linear regression analysis was used to determine the relationship between fetal fraction and gestational age, between fetal fraction and maternal weight, and between fetal/maternal cfDNA and maternal weight; a reciprocal model was used when determining the relationship between fetal fraction and gestational age or maternal weight. For comparison of euploid and aneuploid calls, fetal fractions were expressed as multiples of the median (MoM) relative to low-risk calls weighted by week of gestation, and significance determined using a Mann-Whitney rank sum test. The 2 FN results were included in the appropriate aneuploid category, and FP calls were excluded from aneuploidy fetal fraction analyses. The benefit of a paternal sample on redraw rates and differences in aneuploidy incidence between the a priori risk groups were determined using a χ 2 test. The Kruskal-Wallis 1-way analysis of variance on ranks test was used to evaluate maternal age and gestational age differences for the different risk groups. Positive predictive value (PPV) ([TP]/[TP + FP]) was calculated for cases with known cytogenetic analyses. SigmaPlot 12.5 (Systat Software, San Jose, CA) was used for all statistical analyses. P < .05 was considered statistically significant.
Results
Patients and samples
Patient and sample characteristics for the 31,030 cases received during the study period are detailed in Table 1 . Mean maternal age was 33.3 years, with 51.4% (15,952) aged ≥35 years at the estimated date of delivery. Mean gestational age was 14.0 weeks, with 64.5% (20,001) of samples drawn in first trimester and 33.8% (10,479) in the second trimester.
| Demographic | Whole cohort, n = 31,030 | Follow-up cohort, n = 17,885 |
|---|---|---|
| Maternal age, y a | ||
| Mean | 33.3 ± 6.0 | 33.7 ± 6.1 |
| Median | 35.0 | 35.0 |
| Range | 14.0–60.0 | 14.0–52.0 |
| Gestational age, wk | ||
| Mean | 14.0 ± 4.4 | 14.5 ± 4.7 |
| Median | 12.6 | 13.0 |
| Range | 3.1–40.9 | 9.0–40.9 b |
| Maternal weight, lb c | ||
| Mean | 158.4 ± 39.2 | 157.2 ± 37.9 |
| Median | 149.0 | 148.0 |
| Range | 83.0–425.0 | 83.0–385.0 |
| Fetal fraction, % | ||
| Mean | 10.2 ± 4.5 | 10.8 ± 4.4 |
| Median | 9.6 | 10.1 |
| Range | 0.6–50.0 | 3.7–50.0 b |
a At estimated date of delivery
b As the follow-up cohort does not include any out-of-specification cases, or any cases that failed to receive a noninvasive prenatal testing result, minimum gestational age and fetal fraction are higher than in the whole cohort–however, mean values and SD are equivalent between the 2 cohorts
c Analysis of maternal weight was limited to centers and laboratories that provided this information, and samples originating from United States to avoid inconsistent weight units.
Figure 1 depicts the study flow chart. Samples from 325 (1.0%) patients were excluded as being outside of the specifications for testing ( Supplementary Table 2 ) and 1966 samples failed quality-control metrics ( Supplementary Table 3 ), mostly due to low fetal fraction, leaving 28,739 cases with NIPT results.
In 21,678 cases from clinics linking patient samples to a single case identification, 386 first draws did not meet requirements, thereby allowing analysis of redraw rates in 21,292 cases. A redraw was requested from 95.4% (1572/1648) of cases without a first draw result, 56.5% (888/1572) submitted a redraw, and 64.3% (571/888) of redraws were reported; 12 (2.1%) resolved redraws received a high-risk call. Redraw rates declined steadily over the reporting period ( Figure 2 ); the most recent first sample redraw rates were 9.4% at 9 weeks’, and 5.4% at ≥10 weeks’ gestation. Around 30% of patients given the opportunity to submit a paternal sample chose to do so, and inclusion of a paternal sample was associated with a lower redraw rate, with a similar decline over the study period ( Figure 2 ). This effect was more pronounced in women weighing >200 lb, where inclusion of a paternal sample reduced the redraw rate from 27.5% to 16.1% ( P < .001). The average turn-around time was 9.2 calendar days (95% confidence interval [CI], 9.16–9.23 calendar days), but significant improvements over the study period led to an average turn-around time in the last month of 6.7 calendar days (95% CI, 6.68–6.76 calendar days).
Fetal fractions
The average fetal fraction was 10.2% ( Table 1 ). Regression analysis, using the reciprocal of the independent variable (gestational age or maternal weight), revealed a positive correlation between fetal fraction and gestational age (r 2 = 0.05, P < .001) ( Figure 3 , A), and a negative association between fetal fraction and maternal weight (r 2 = 0.16, P < .001) ( Figure 3 , B). Furthermore, with increasing maternal weight, there was an increase in maternal cfDNA ( P < .001) and a decrease in fetal cfDNA ( P < .001) ( Figure 4 ). Fetal fractions when stratified by aneuploidy were decreased for trisomy 13 (0.759 MoM, P < .001), trisomy 18 (0.919 MoM, P = .012), and monosomy X (0.835 MoM, P < .001), and increased for trisomy 21 (1.048 MoM, P = .018) samples.
NIPT results
The combined rate of high-risk calls for all 4 indications was 1.77% (508/28,739); including 324 trisomy 21, 82 trisomy 18, 41 trisomy 13, and 61 monosomy X ( Table 2 ). One sample was not assigned a risk score for chromosome 21 due to a maternal chromosome 21 partial duplication but was accurately identified as fetal trisomy 21 by the laboratory. Of 20,384 samples evaluated for additional sex chromosome aneuploidies, other than monosomy X, there were 14 (0.07%) identified: 6 XXX, 6 XXY, and 2 XYY. Fetal sex was reported in 24,522 cases. There were no reports of gender discordance from women receiving low-risk reports. For women receiving high-risk reports, confirmation of fetal sex was available for 109 cases, of which 108 (99.1%) were correct; the single discordant case was reported as high-risk for monosomy X ( Supplementary Figure ) but cytogenetic testing revealed a 46, XY fetus. Although cases with known multiple gestations were excluded, the NATUS algorithm identified 127 (0.4%) samples as having >2 fetal haplotypes, indicative of either unreported twins, vanishing twin, or triploidy.
| All cases, N = 28,739 a | Trisomy 21 | Trisomy 18 | Trisomy 13 | Monosomy X |
|---|---|---|---|---|
| Risk ≥99/100 | 298 b | 78 b | 26 | 53 |
| 1/100 ≤ Risk <99/100 | 25 | 4 | 15 | 8 |
| Total | 324 b,c | 82 b | 41 | 61 |
| Prevalence, 1 in: | 88 | 349 | 697 | 467 |
a Total number of cases with reported result at ≥9 wk of gestation
b Trisomy 21 and trisomy 18 totals include a single case of double-aneuploidy
c Includes 1 case with a detected partial maternal chromosome 21 duplication, the fetus was determined to be high risk for trisomy 21 but the algorithm did not calculate a risk score.
ICD-9 codes were associated with 19.0% (5468/28,739) of women: 16.6% were low-risk, 44.1% were high-risk based only on advanced maternal age (≥35 years), and 39.3% had high-risk codes. As expected, the incidence of aneuploidy calls was smallest in the low-risk group (0.7%), followed by advanced maternal age women (1.6%), and largest in the high-risk group (3.4%) ( Table 3 ). Results for the 23,271 samples without ICD-9 codes showed a similar difference in aneuploidy calls between women aged <35 years (1.0%, 117/11,629) and those aged ≥35 years (2.4%, 274/11,642).
| Variable | Cases with ICD-9 codes, n = 5468 | Cases without codes, n = 23,271 | |||
|---|---|---|---|---|---|
| Low risk, age <35 y (n = 909) | AMA only, age ≥35 y (n = 2411) | High risk, all ages (n = 2148) | Low risk, age <35 y (n = 11,629) | High risk, age ≥35 y (n = 11,642) | |
| Maternal age, y a Median (range) | 28.2 ± 4.4 | 37.8 ± 2.4 | 31.3 ± 5.8 | 28.4 ± 4.5 | 37.9 ± 2.5 |
| 29.0 (15.0–34.0) | 37.0 (35.0–48.0) | 32.0 (15.0–47.0) | 29.0 (14.0–34.0) | 37.0 (35.0–52.0) | |
| Gestational age, wk a Median (range) | 14.1 ± 4.4 | 13.3 ± 3.5 | 15.8 ± 5.0 | 14.7 ± 4.9 | 13.4 ± 3.9 |
| 12.4 (9.0–33.3) | 12.4 (9.0–38.1) | 14.4 (9.0–37.0) | 13.0 (9.0–38.0) | 12.1 (9.0–40.9) | |
| Euploid | 903 | 2368 | 2073 | 11,457 | 11,293 |
| Trisomy 21 | 2 | 27 b | 50 | 57 | 188 |
| Trisomy 18 | 1 | 5 b | 13 | 21 | 42 |
| Trisomy 13 | 1 | 5 | 3 | 11 | 21 |
| Monosomy X | 2 | 2 | 6 | 28 | 23 |
| Total aneuploids | 6 | 38 | 72 | 117 | 274 |
| Monosomy X prevalence, % | 0.22 | 0.08 | 0.28 | 0.24 | 0.20 |
| Trisomy prevalence, % | 0.44 | 1.49 | 3.07 | 0.77 | 2.16 |
| Overall prevalence, % | 0.66 c | 1.58 c | 3.35 c | 1.01 d | 2.35 d |
a Mean ± SD, there was a significant difference between risk groups ( P < .001) for both maternal age and gestational age, as determined by the Kruskal-Wallis 1-way analysis of variance on ranks test
b Trisomy 21 and trisomy 18 totals include single case of double-aneuploidy
c Significant difference in aneuploidy call rate among 3 groups with ICD-9 codes ( P < .001), as determined by χ 2 test
d Significant difference in aneuploidy call rate between 2 groups without ICD-9 codes ( P < .001), as determined by χ 2 test.
Follow-up of high-risk calls
From 17,885 cases in the follow-up cohort, outcome information was sought for the 356 high-risk calls; 152 high-risk calls from the whole cohort described above were not contained within the follow-up cohort.
Information regarding invasive testing uptake was available for 251/356 (70.5%) cases that received a high-risk result: 39.0% (139) elected invasive testing and 31.5% (112) declined invasive tests, and of the remaining 105 (29.5%), 39 had a spontaneous demise or elective termination. Within the 356 high-risk calls, there were in total 58 reported spontaneous abortions, including 16 cases categorized as TP, 2 FP, 4 with ultrasound findings suggestive of aneuploidy, and 36 with unconfirmed outcomes. There were 57 reported elective terminations, including 30 cases categorized as TP, 5 with ultrasound findings suggestive of aneuploidy, and 22 elective terminations with unconfirmed outcomes.
At the conclusion of clinical follow-up, 62.4% (222/356) of high-risk calls had karyotype information or at-birth confirmation: 184 confirmed affected pregnancies (TP) and 38 unaffected pregnancies (FP) ( Table 4 ). Eight cases showed placental or fetal mosaicism: 5 fetal mosaics (TP) were confirmed by amniocentesis (2 trisomy 21, 2 trisomy 18, 1 monosomy X), and 3 cases were considered FP because of confined placental mosaicism (CPM). Two CPM cases were high risk for trisomy 13 and were identified as mosaics by chorionic villus sampling (CVS), one was determined to be euploid by amniocentesis, and the other did not have a follow-up amniocentesis but ultrasound at 20 weeks was read as normal. In the third CPM case, at-birth testing revealed a 100% trisomy 18 placenta and a euploid child. Two FN results (both trisomy 21) were reported to the laboratory following amniocentesis due to other indications.
| N = 17,885 a | Trisomy 21 | Trisomy 18 | Trisomy 13 | Monosomy X | Total |
|---|---|---|---|---|---|
| High-risk calls | 233 b | 55 b | 30 | 38 | 356 |
| Confirmed outcomes | |||||
| True positive | 140 c | 27 | 8 | 9 | 184 |
| False positive | 14 d | 2 e | 13 f,g | 9 | 38 |
| Unconfirmed outcomes | |||||
| Suggestive h | 8 | 9 | 0 | 2 | 19 |
| Pregnancy loss i | 18 | 6 | 3 | 9 | 36 |
| Termination j | 14 | 3 | 0 | 5 | 22 |
| No follow-up k | 39 | 8 | 6 l | 4 | 57 |
| Low-risk calls | |||||
| Confirmed outcomes | |||||
| False negative | 2 | 0 | 0 | 0 | 2 |
a Total number of cases with reported result at ≥9 wk of gestation from participating centers
b Trisomy 21 and trisomy 18 totals include single double-aneuploidy case
c Includes 13 cases reported as trisomy 21 based on at-birth clinical evaluation
d Includes 3 cases reported as normal based on at-birth clinical evaluation
e Includes 1 confined placental mosaicism case
f Includes 2 confined placental mosaicism cases (1 confirmed and 1 unconfirmed)
g Includes 1 case reported as normal based on at-birth clinical evaluation
h Patients declined invasive testing but ultrasound findings were consistent with noninvasive prenatal testing findings (see “Materials and Methods” section)
i Patients experienced spontaneous abortion and did not obtain karyotype confirmation
j Patients chose to terminate pregnancy without diagnostic testing
k Follow-up information was not available
l One sample tested as high-risk (1/7.6) for fetal aneuploidy, analysis of second sample indicated that patient was at low-risk, follow-up information was not available.
For the sex chromosome aneuploidies XXX, XXY, and XYY, 7 of the 14 high-risk calls were within the follow-up cohort. Clinical follow-up revealed 4 cases with known outcomes: 2 TP (1 XXX, 1 XXY) and 2 FP (both XXX).
Based on the cases with cytogenetic confirmation, women with an intermediate-risk score were more likely to have a FP result (19/24, 79.2%) than women with a maximum-risk score (19/198, 9.6%, P < .001). For the 36 cases that experienced spontaneous abortion and did not obtain karyotype confirmation, 33 (91.7%) had a maximum-risk score. All 22 patients who elected to terminate the pregnancy without confirmation had a maximal-risk score.
Positive predictive value
Based only on cases with cytogenetic diagnosis ( Table 4 ), the PPV was 90.9% for trisomy 21 and 82.9% for all 4 cytogenetic abnormalities combined ( Table 5 ). A theoretical PPV was also calculated under the 2 boundary conditions that all unconfirmed high-risk cases were either FP or TP ( Table 5 ). This provided a range for the PPV of 60-94% for trisomy 21 and 52-89% for all abnormalities combined.
| Variable | Trisomy 21 | Trisomy 18 | Trisomy 13 | Monosomy X | Total |
|---|---|---|---|---|---|
| Cytogenetically confirmed cases | |||||
| TP/(TP + FP) (PPV) | 140/154 (90.9%) | 27/29 (93.1%) | 8/21 (38.1%) | 9/18 (50.0%) | 184/222 (82.9%) |
| All unconfirmed cases considered as FPs (lower bound) | |||||
| TP/(TP + FP) (PPV) | 140/233 (60.1%) | 27/55 (49.1%) | 8/30 (26.7%) | 9/38 (23.7%) | 184/356 (51.7%) |
| All unconfirmed cases considered as TPs (upper bound) | |||||
| TP/(TP + FP) (PPV) | 219/233 (94.0%) | 53/55 (96.4%) | 17/30 (56.7%) | 29/38 (76.3%) | 318/356 (89.3%) |
Among women without ICD-9 -coded indications, 63 women aged <35 years received high-risk calls, of which 39 (60.9%) had diagnostic testing and 34 were TP, a PPV of 87.2% (95% CI, 72.6–95.7%). Of 176 women ≥35 years with high-risk calls, 105 (59.7%) had confirmatory karyotyping and 87 were TP, a PPV of 82.9% (95% CI, 74.3–89.5%).
Results
Patients and samples
Patient and sample characteristics for the 31,030 cases received during the study period are detailed in Table 1 . Mean maternal age was 33.3 years, with 51.4% (15,952) aged ≥35 years at the estimated date of delivery. Mean gestational age was 14.0 weeks, with 64.5% (20,001) of samples drawn in first trimester and 33.8% (10,479) in the second trimester.
| Demographic | Whole cohort, n = 31,030 | Follow-up cohort, n = 17,885 |
|---|---|---|
| Maternal age, y a | ||
| Mean | 33.3 ± 6.0 | 33.7 ± 6.1 |
| Median | 35.0 | 35.0 |
| Range | 14.0–60.0 | 14.0–52.0 |
| Gestational age, wk | ||
| Mean | 14.0 ± 4.4 | 14.5 ± 4.7 |
| Median | 12.6 | 13.0 |
| Range | 3.1–40.9 | 9.0–40.9 b |
| Maternal weight, lb c | ||
| Mean | 158.4 ± 39.2 | 157.2 ± 37.9 |
| Median | 149.0 | 148.0 |
| Range | 83.0–425.0 | 83.0–385.0 |
| Fetal fraction, % | ||
| Mean | 10.2 ± 4.5 | 10.8 ± 4.4 |
| Median | 9.6 | 10.1 |
| Range | 0.6–50.0 | 3.7–50.0 b |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
