Evaluation of the Z-score accuracy of noninvasive prenatal testing for fetal trisomies 13, 18 and 21 at a single center

Abstract

Objective: To assess the correlation between Z-scores of positive noninvasive prenatal testing (NIPT) results and the positive predictive value (PPV) of NIPT.

Methods: Pregnancies with positive NIPT results at Guangzhou Women and Children's Medical Centre between July 2017 and May 2020 were included in this study. Fetal karyotyping or microarray analysis was provided to patients with abnormal NIPT results for confirmatory testing. Logistic regression analyses was applied to study the relationship between the Z scores and the PPV performance. The optimal cutoff values for indicating fetal common trisomies were obtained based on receiver operating characteristic (ROC) curve analysis, and then the PPV were calculated in pregnancies with positive NIPT results at Z-score greater than or equal to cutoff value and in patients with a Z-score between 3 and cutoff value respectively.

Results: A total of 214 pregnancies with positive NIPT results for fetal common trisomies were validated by invasive prenatal diagnosis and follow up in this study. Of these, NIPT indicated trisomy 13 in 25 cases, trisomy 18 in 54 cases and trisomy 21 in 135 patients. Logistic regression analyses showed a significant association (p < 0.05) between the Z-scores and true positive results for T21 and T18. For T13, the significant association was not observed (p > 0.05). The ROC curve analysis showed that the optimal cutoff Z-score for indicating fetal trisomies 13, 18, and 21 were 6.889, 7.574 and 6.612 respectively, and the corresponding area under curve were 0.706, 0.916, and 0.954. In this cohort with abnormal NIPT results, the cutoff values revealed a sensitivity of 96.8% and a specificity of 90% for indicating trisomies 21, and a sensitivity of 88.9% and a specificity of 92.6% for trisomies 18. However, probably due to the sample size, the sensitivity and specificity for indicating trisomy 13 were lower (85.7% and 61.1%) than that for trisomies 21 and 18. The PPVs in pregnancies with positive NIPT results at Z-score greater than or equal to cutoff value were 99.18% (121/122) for trisomy 21, 92.31% (24/26) for trisomy 18 and 46.15% (6/13) for trisomy 13. In patients with a Z-score between 3 and cutoff Z-score, the PPV of NIPT for trisomies 21, 18, and 13 were 30.77% (4/13), 10.71% (3/28), and 8.33% (1/12) respectively. Moreover, by classifying Z scores as 3 ≤ Z < 5, 5 ≤ Z < 10, and Z ≥ 10, the majority of Z scores were above 10 with a PPV of 99% for T21 and just 5.2% were between 3 and 5 with a PPV of 14.3%. In contrast for T18, over a third of tests had Z scores between 3 and 5. The PPV in this group is just over 5%.

Conclusions: The present results show that the PPV performance of NIPT for fetal trisomies 13, 18, and 21 are closely associated with Z-score. The higher the Z-score, the greater the likelihood that the aneuploidy result is correct. Our experience in evaluating the Z-score accuracy of NIPT in this study could be of use in similar work.