Clinical indications and Z-score-assisted NIPT testing: a new perspective in prenatal screening

Introduction

In 1997, Professor Dennis Lo of the Chinese University of Hong Kong identified the existence of cell-free fetal DNA (cffDNA) in maternal peripheral plasma which remains stable throughout gestation and swiftly diminishes postdelivery. This substance is regarded as the optimal choice for non-invasive prenatal diagnosis, hence propelling the advancement of non-invasive prenatal testing (NIPT).1 Since 2011, NIPT has been integrated into clinical practice2 and is now extensively used to screen for trisomy 21 (T21), trisomy 18 (T18) and trisomy 13 (T13), having demonstrated superiority as the most precise screening tool for common fetal aneuploidies. NIPT provides superior sensitivity and specificity in identifying T21, T18 and T13 relative to traditional biochemical and ultrasound testing.3 A multitude of studies has demonstrated that NIPT can diminish the occurrence of superfluous invasive procedures.4 Overall, NIPT, as an accurate screening instrument, enhances aneuploidy detection capabilities while diminishing dependence on invasive diagnostic methods. NIPT is recognised as the premier non-invasive technique for detecting common fetal aneuploidies; nonetheless, it functions solely as a screening tool since cffDNA predominantly derives from the placenta which may exhibit genetic discrepancies compared with the fetus.5 Therefore, false-positive and false-negative results may arise. In actual testing, false-positive results from NIPT are more prevalent than false-negative ones, frequently heightening anxiety levels throughout pregnancy.

For clinicians and patients, the positive predictive value (PPV) is more informative than conventional sensitivity and specificity measures.6 PPV indicates the proportion of patients with a positive NIPT result who genuinely possess the illness, functioning as an effective metric for genetic counselling. A low PPV of 10–20% indicates that more than 80% or 90% of abnormal results are likely false positives.7 Recent studies indicate significant variability in PPV values, with reported ranges of 65%–94% for T21, 47%–85% for T18 and 12%–62% for T13.8 In fact, the PPV performance of NIPT may be influenced by various factors, including fetal fraction, confined placental mosaicism, maternal copy number variations, testing procedures, classification algorithms and sequencing platforms.9 In addition to these factors, the clinical indications of pregnant women are also an important consideration in determining PPV. Xiang et al pointed out that among seven clinical screening indications, the PPV for T21 ranged from 50.62% to 73.09%, for T18 from 20.00% to 58.33% and for T13 from 4.17% to 47.37%.10 Other studies have demonstrated statistically significant variations in PPV for different clinical screening indications of T21, T18 and T13, with the highest positive rates and PPV for NIPT detection indications linked to ultrasonic anomalies, particularly structural abnormalities and elevated nuchal translucency (NT).11 12 Consequently, clinical indications for NIPT should be strictly controlled. Furthermore, PPV is also closely related to the Z-score obtained from NIPT testing. The Z-score in NIPT refers to the value obtained by comparing the test sample with a normal diploid control sample during the NIPT testing process. The Z-score represents the risk value in NIPT and serves as an important indicator for assessing whether the fetus is at high risk for chromosomal aneuploidy and is the predominant approach employed for screening fetal chromosomal aneuploidies.9 Numerous researches have reported a significant correlation between Z-scores and PPV. The most recent research report indicates that Z-scores are closely related to NIPT accuracy, with positive NIPT results at Z≥9 exhibiting more accuracy than those at 5≤Z<9 and 3≤Z<5.4.13 Therefore, laboratories providing NIPT must assess and evaluate the precision of Z-scores to guarantee the reliability of testing and the clinical utility of result interpretation for individual patients. Nonetheless, information about the evaluation of the precision of Z-scores for NIPT of T21, T18 and T13 is still scarce.

In this study, we collected NIPT data from 37 891 pregnant women and diagnostic results from 96 high-risk pregnant women between January 2021 and June 2024 to assess the performance of NIPT. Additionally, we grouped pregnant women with high-risk results to explore the correlation between NIPT and clinical screening indications as well as Z-scores, providing data and information for clinical genetic counselling and prenatal diagnosis.

Materials and methods

Results

Clinical performance of NIPT in testing T21, T18, T13

Of the 96 pregnant women identified with high-risk results for T21, T18 and T13, there were 59 cases of T21, 25 cases of T18 and 12 cases of T13 (figure 1). The screening positive rate was 0.25% (96/37 891). Following up, we identified two false-negative results. The overall sensitivity and specificity were 96.55% (56/58) and 99.89% (37 793/37 833), respectively.

Figure 1

Request permissions

Flowchart of NIPT results of pregnant women undergoing screening for fetal chromosome disorders between 2021 and June 2024. FN, false negative; FP, false positive; NIPT, non-invasive prenatal testing; T13, trisomy 13; T18, trisomy 18; T21, trisomy 21; TN, true negative; TP, true positive.

Table 2 shows the clinical performance of NIPT in testing T21, T18 and T13. Of the 59 women diagnosed with T21, 40 were confirmed as true positives, resulting in a PPV of 67.80%. The sensitivity and specificity for T21 were 95.24% (40/42) and 99.95% (37 830/37 849), respectively. In a cohort of 25 women with T18, 14 were identified as true positives, yielding a PPV of 48.00%. The sensitivity and specificity for T18 were 100% (25/25) and 99.97% (37 866/37 877), respectively. Among the 12 women diagnosed with T13, 2 were confirmed as true positives, leading to a PPV of 15.38%. The sensitivity and specificity for T13 were both 100% (12/12) and 99.97% (37 879/37 889), respectively. The PPVs of T21, T18 and T13 had significant differences (χ²=26.272, p<0.001).

Table 2

View inline•Open as popup

Clinical performance of NIPT in testing T21, T18, T13

Based on the years of testing, we conducted an in-depth analysis of the PPV values for NIPT. Table 3 shows the results. The peak PPV occurred in 2021 at 71.43%, followed by 2022 at 64.29%, 2024 at 52.63% and the lowest PPV was recorded in 2023 at 46.43%. The trend in PPV values for pregnant women with T21 corresponded to the overall results, declining from 92.31% in 2021 to 68.75% in 2022, 70.00% in 2024 and reaching a low of 50.00% in 2023. The PPV values for T18 were highest in 2022 at 75.00%, followed by 2021 and 2023, both at 50.00%, with the lowest value recorded in 2024 at 40.00%. In T13, true positives were identified solely in 2022 and 2024, yielding a PPV of 25.00%.

Table 3

View inline•Open as popup

PPV performance of T21, T18 and T13 in pregnant women in different years

Discussion

NIPT is a non-invasive and highly reliable screening test. NIPT, in contrast to invasive techniques, does not pose a danger of miscarriage or injury to the fetus.15 While NIPT is extensively used for identifying fetal anomalies, its outcomes fluctuate among various populations and there is insufficient data regarding the PPV of NIPT screening in different cohorts. This study presents an in-depth analysis of high-risk samples that underwent NIPT at our institution from 2021 to June 2024. The aim was to gain a deeper understanding of the detection capabilities of NIPT and to specifically focus on the PPV values of NIPT across various population subgroups, to provide valuable data and information for clinical genetic counselling and prenatal diagnosis.

This study, involving 37 891 pregnant women, determined that the overall sensitivity and specificity of NIPT were 96.55% and 99.89%, respectively. The PPVs of NIPT for detecting T21, T18 and T13 were 67.80%, 56.00% and 16.67%, respectively. Recent research data indicate that the sensitivity and specificity of NIPT for detecting T21, T18 and T13 exceed 98%, with associated PPVs of 86.81%, 56.81% and 18.18%, respectively. In our study, the corresponding PPVs showed a decreasing trend compared with these values, which may be attributed to a decrease in the incidence of these three conditions.16 Due to the fact that our data showed that the PPV was at its lowest in 2023, which was only 50.00%, it is possible that the low PPV in that year was a contributing factor to the overall fall in PPV. On reviewing our data, we found that nearly 80.00% of the high-risk samples in 2023 had Z-scores within the grey zone of 3–5 which presumably contributed significantly to the low PPV problem. Despite the relatively low PPV for T21,17 NIPT still has a superior detection rate and specificity for T21 compared with serum screening,18 making it a more accurate screening test in clinical practice.

Prenatal diagnosis is the gold standard for chromosomal abnormalities. Some pregnant women assert that a low Z-score from NIPT for aneuploidy negates the need to undergo further examination.19 However, in this study, we identified two cases of false negatives, where NIPT results indicated low risk, yet prenatal diagnosis confirmed T21. Previous research indicates that low fetal DNA concentration, fetal cellular and chromosomal abnormalities, mosaicism, maternal copy number variations and statistical fluctuations in Z-scores during detection may contribute to false negatives in NIPT.20–23 Among these, low fetal DNA concentration and placental mosaicism are the two principal contributors.24 25 False negatives are more probable when fetal DNA concentration is<2%. Hu et al successfully increased the average concentration of cfDNA from 10% to 20% using cfDNA enrichment technology, significantly reducing the incidence of false negatives.26 Our analysis of the two pregnant women who experienced false negative results in this study, along with their baseline information and the original NIPT data, revealed that both of the missed samples met the quality control criteria for detection with a cffDNA concentration of>3%. Nevertheless, they were still missed, prompting the hypothesis that placental mosaicism could be responsible for these two false-negative cases. CffDNA originates from the placenta and discrepancies between the placental karyotype and the fetal karyotype can exist. Analysing the placenta could help verify this hypothesis. Furthermore, in examining the causes of false negatives, it is crucial to not only analyse the raw data and retest the original samples but also to ascertain the authenticity of the specimens; specifically, to confirm that the specimens originated from the child’s mother rather than another pregnant woman. Both false-negative cases subsequently showed abnormal ultrasound findings, highlighting the significance of prenatal ultrasound. Given the constraints of NIPT, the interpretation and communication of NIPT results should be more comprehensive to guarantee that both pregnant women and healthcare providers fully understand the limitations and potential risks of NIPT, facilitating more reasonable medical decisions. Consequently, performing comprehensive clinical consultations prior to testing is essential.

Our research findings demonstrate that the PPV of NIPT for detecting T21, T18 and T13 exhibits significant variability across pregnant women with diverse clinical indications. The results demonstrate that the highest positive rates for NIPT screening indications for T21, T18 and T13 correlate with advanced maternal age and Xiang et al reported a maximum PPV of 73.09% for T21 in “advanced age” pregnancies.10 In addition, Cai et al also showed that the PPV of common trisomy was significantly higher in “advanced age” pregnant women.27 The results indicate that advanced maternal age is associated with an increased likelihood of chromosomal abnormalities in fetuses, necessary for diagnostic testing. The PPV escalates with maternal age due to the heightened occurrence of fetal chromosomal aneuploidy, which arises during mitosis or meiosis, as maternal age increases. Nevertheless, certain studies have demonstrated that the PPV for T21, T18 and T13 does not exhibit a substantial increase throughout specified age categories (5-year intervals), indicating that positive results from NIPT should be regarded with similar importance by both providers and patients, irrespective of maternal age.5 In high-risk pregnant women, NT thickening has a PPV of 100%, consistent with previously reported data by Wang et al.12 As NIPT testing is conducted after NT screening, this serves as a reminder to NIPT testers to exercise heightened caution in pregnant women exhibiting ultrasonic structural abnormalities. Additionally, other studies have found that the highest positive rate and PPV for NIPT indicators are associated with ultrasonic abnormalities, particularly significant structural abnormalities and NT thickening.11 The findings suggest that pregnant women exhibiting aberrant ultrasound results have a higher likelihood of carrying fetuses with chromosomal abnormalities. Despite low-risk NIPT results, subsequent prenatal examinations, such as the comprehensive abnormality scan, should meticulously assess fetal condition. Among pregnant women at high risk for T21, T18 and T13, the clinical indication of “voluntary request” has the second-highest positive rate following advanced maternal age. For pregnant women at high risk for T21 and T18 with the clinical indication of “voluntary request”, the PPVs are 63.16% and 42.86%, respectively. Despite pregnant women with the screening indication of “voluntary request” being classified as low-risk, conducting NIPT screening in low-risk populations may diminish the sensitivity and PPV of NIPT. Nevertheless, our research findings indicate that this group warrants attention. Furthermore, for pregnant women with abnormal serum screening, Li et al’s study found that among all screening indications, the abnormal maternal serum screening group had the highest prevalence of chromosomal abnormalities.28 Nevertheless, owing to the limited number of individuals in this group within our study, we were unable to reach the same conclusion. Wei et al demonstrated that the probability of chromosomal abnormalities in twins is higher than in singletons.29 However, in our study, we only observed two cases of screening-positive twin pregnancies, indicating a low efficacy of screening for abnormalities in twin pregnancies. Compared with a previous study,11 there was no significant difference in PPV values between different clinical indications in our study which may be due to a small number of NIPT-positive cases. Therefore, in future research, it is necessary to increase the sample data size in order to obtain more in-depth results.

The data from our study indicate that in the T21 and T18 groups, an increase in the Z-score corresponds with an increase in the PPV, signifying a substantial correlation between PPV and Z-score. Similar to previous research data, the performance of PPV is closely related to the Z-score.30 The PPV for T21, T18 and T13 in the high Z-score group is significantly higher than that in the low Z-score group.14 25 In the study by Junhui et al, receiver operating characteristic curve analysis showed that the optimal cut-off Z-scores for fetal T13, T18 and T21 were 6.889, 7.574 and 6.612, respectively. At these cut-off values, the sensitivity for T21 and T18 was 96.8% and 88.9%, respectively, and the specificity was 90% and 92.6%, respectively,9 This finding also supports our study’s result that T21 and T18 have the highest PPV when the Z-score exceeds 10. However, due to the limited number of samples for T13 in our study, there were no high-risk samples with Z>10, necessitating further investigation in the future. These data suggest that a higher Z-score is associated with a greater likelihood of a true positive result for aneuploidy. Additionally, a retrospective analysis of the Z-scores of false-positive cases for T13, T18 and T21 revealed that false-positive cases primarily had Z-scores<5 which falls within the grey zone according to the kit manual. In our study, when the Z-score ranged from 3 to 5, the PPV for high-risk T21 populations was less than 20%. Theoretically, the likelihood of a false positive is greater than 80%. Studies have shown that the occurrence of placental mosaicism may slightly increase the NIPT Z-score to above 3 but below 5, thereby affecting the accuracy of NIPT.13 Therefore, for pregnant women with Z-scores between 3 and 5, they can be informed of the possibility of a false positive to alleviate anxiety. Given the substantial link between Z-score and PPV, Yang et al25 conducted a rank correlation analysis between Z-score and maternal age, gestational age, fetal fraction and body mass index, revealing that the concentration of cffDNA significantly increases with the Z-scores of T21, T18 and T13. Thus, when a case has a fetal DNA concentration of about 10% and a Z-score of about 10, it suggests a potential true positive case; if the sample shows a fetal DNA concentration higher than 10% but a lower Z-score, around 5, it may be suspected as a false-positive case. Consequently, we advocate for the incorporation of the fetal DNA fraction value in the NIPT result report to furnish clinicians with additional information.

Our research further supports that NIPT is an effective screening technique to detect aneuploidy disorders and signifies a groundbreaking progression in prenatal diagnosis. Nevertheless, our study also possesses certain limitations. First, with the development of assisted reproductive technologies and the growing percentage of older pregnant women, the incidence of twin pregnancies has significantly risen.31 32 NIPT is an advanced prenatal screening method for twin pregnancies, demonstrates excellent sensitivity and specificity in detecting fetal aneuploidy.33 However, there is limited research data on the performance of NIPT in twin pregnancies. Our investigation identified only two cases of twin pregnancies with positive screening results, necessitating future research to assess the efficacy of NIPT in twin pregnancy detection. Second, all cases in our study were from a tertiary hospital in Beijing and data from more medical institutions are required to better clarify the prognostic efficacy of NIPT. Third, due to the limited incidence of T13, more cases need to be studied to assess the efficacy of NIPT in detecting T13.