Predicting the risk of Turner syndrome based on ultrasonographic markers in the first trimester of pregnancy

To the Editor: Turner syndrome is the most common sex-chromosome abnormality in females with an incidence of 1 per 2500 female live births, and it is associated with either complete or partial loss of one X chromosome.[1] Predicting the risk of Turner syndrome in the first trimester of pregnancy is of critical importance, in order to provide parturient with reproductive choices for affected fetuses at the earliest opportunity. The 2013 International Society of Ultrasound in Obstetrics and Gynecology (ISUOG) Practice Guidelines stated that fetal nuchal translucency (NT) thickness and absence of the nasal bone (NB) could be used to screen for chromosomal abnormalities in the first trimester of pregnancy. The detection rate for Turner syndrome is about 100% for fetuses with nuchal cystic hygroma but close to 0% for those without an increased NT.[2] The cutoff value of fetal NT thickening had been controversial in previous studies.[3] It has been observed that additional ultrasound markers such as NB, blood flow across the tricuspid valve (TV), and through the ductus venosus (DV) had the potential to improve the screening performance for Turner syndrome. However, only a few studies have focused on the evaluation of the above additional ultrasound markers in fetuses with Turner syndrome. The aim, therefore, of this study was to develop a model that incorporates maternal age and fetal ultrasonographic markers for predicting the risk of Turner syndrome in the first trimester of pregnancy. The study was approved by the Ethics Committee of Beijing Obstetrics and Gynecology Hospital, Capital Medical University (No. IEC-B-03-V01-FJ1). Written informed consent was signed by each participant before data collection. This was a retrospective study using fetal ultrasound and clinical information from two centers between January 2011 and May 2021. To build a robust prediction model as well as nomogram, participants from Beijing Obstetrics and Gynecology Hospital, Capital Medical University were used as the training set, and participants from Shanxi Provincial People's Hospital were used as the validation set. Sample size was compatible with the requirements of the study design. We selected singleton pregnancies at 11 to 13+6 weeks’ gestation. The criteria for selection were as follows: (a) fetal karyotype confirmed by amniocentesis or examination after birth and (b) for each fetus, the ultrasound and clinical information were complete. During the study period, 106 Turner syndrome cases were retrieved for inclusion in the training set, which were matched with 2120 normal fetuses that were selected using the same criteria as for Turner syndrome cases. At the same time, 441 participants (420 normal + 21 with Turner syndrome) from Shanxi Provincial People's Hospital were used as the validation set [Supplementary Figure 1, https://links.lww.com/CM9/B457]. Ultrasonographic markers. For each fetus, the following data were collected and evaluated: maternal age, crown-rump length, gestational age, fetal karyotype, fetal NT thickness, NB, and blood flow across the TV and through the DV. The ultrasonographic markers were evaluated using retrieved fetal images by a sonographer (G.X.) who was blinded to the karyotypic results. The fetal NB was defined as abnormal either if it is not visible or if its echogenicity is the same or lower than that of the skin over the NB. Pulsed Doppler ultrasound was used to evaluate the TV and DV flow. Abnormal TV flow (tricuspid regurgitation) was diagnosed when it lasts for at least half of the systole period with a velocity of >80 cm/s. DV flow was defined as abnormal when the a-wave was reversed. Prediction model and nomogram. In the training set, the prediction model for screening Turner syndrome was developed using logistic regression based on maternal age, fetal NT, NB, TV flow, and DV flow. For comparison, an additional model based on the combination of maternal age and fetal NT was developed. The predictive performance of the above two models were evaluated. In order to provide clinicians with an individualized tool to predict patient-specific probability for Turner syndrome, a nomogram was developed based on the prediction model (incorporates five markers) that constructed in the training set. Statistical analysis was performed using R software (version 4.1.0; R Foundation, Vienna, Austria). Data were presented as median (interquartile range) or frequencies (%). Difference for numerical and for categorical variables were calculated based on the t-test, Kruskal–Wallis test, or chi-squared test, as appropriate. Multiple logistic regression analysis was used to develop the prediction model. The detection rates of the Turner syndrome fetuses were determined for fixed false-positive rates (FPR) using logistic regression analysis. Additionally, the receiver operating characteristic (ROC) curves were also plotted and the area under the ROC curves (AUCs) were calculated. The DeLong test was used to compare ROC curves. P < 0.05 was considered statistically significant. There was no statistically significant difference in NT thickness (median of 1.40 mm in training set and 1.30 mm in validation set) and Turner syndrome distribution (95.24% in both training and validation sets) between the training and validation sets. Fetal NB, TV flow, and DV flow were abnormal in 57 (2.56%), 45 (2.02%), and 34 (1.53%) fetuses of training set, respectively [Supplementary Table 1, https://links.lww.com/CM9/B457]. In the validation set, the above markers were abnormal in 31 (7.03%), 34 (7.71%), and 17 (3.85%) fetuses, respectively. The predictive performance in screening for Turner syndrome of the model based on maternal age and fetal NT and the model based on five markers are presented in Table 1. For an FPR of 3%, the detection rate of the model based on maternal age and fetal NT was 80.02% and 80.95% in the training and validation sets, respectively. Adding the three ultrasound markers fetal NB, TV, and DV flow increased it to 89.62% and 90.48% in the training and validation sets, respectively [Supplementary Table 2, https://links.lww.com/CM9/B457]. Table 1 - Comparison of predictive performance of two models in screening for Turner syndrome. Model Sensitivity (%) Specificity (%) Accuracy (%) AUC (95% CI) Model 1 Training set 80.19 98.82 97.93 0.869 (0.810–0.928) Validation set 90.48 94.29 94.10 0.899 (0.771–1.000) Model 2 Training set 90.57 96.98 96.68 0.945 (0.908–0.982) Validation set 90.48 98.57 98.19 0.943 (0.849–1.000) Model l included two markers: maternal age and fetal NT thickness. Model 2 included five markers: maternal age, fetal NT thickness, abnormal NB, tricuspid regurgitation, and reversed DV flow.AUC: Area under the receiver-operating-characteristics curve; CI: Confidence intervals; DV: Ductus venosus; NB: Nasal bone; NT: Nuchal translucency. In the training set, the model based on maternal age, fetal NT, NB, TV flow, and DV flow showed significantly higher AUC when compared with the model based on maternal age and fetal NT (0.945, 95% confidence interval [CI]: 0.908–0.982 vs. 0.869, 95% CI: 0.810–0.928; P = 0.002), and there was no statistical difference between the above two models in the validation set (0.943, 95% CI: 0.849–1.000 vs. 0.899, 95% CI: 0.771–1.000; P = 0.354) [Supplementary Figure 2, https://links.lww.com/CM9/B457]. Each independent marker included in the five-marker model is reflected in the nomogram, which is a visual scoring system to calculate the estimated probability of Turner syndrome [Supplementary Figure 3, https://links.lww.com/CM9/B457]. In this study, we developed and validated a prediction model-derived nomogram for the individualized prediction of Turner syndrome in the first trimester of pregnancy, and it was superior to the model based on maternal age and fetal NT. Our study confirmed that advanced maternal age is associated with Turner syndrome, which may be because of oocyte aneuploidy associated with maternal aging caused by recombination failure, cohesion deterioration, spindle assembly checkpoint dysregulation, abnormalities in post-translational modification of histones and tubulin, and mitochondrial dysfunction. Wagner et al[2] found that median fetal NT was 1.8 mm in euploid fetuses and 10.0 mm in fetuses with Turner syndrome. We found a similar results; this may be explained by that Turner syndrome fetuses showed higher expression of chondroitin sulfates (CS-6 and CS-4) and lower expression of biglycan in the first trimester of pregnancy.[4] It is a rare study that only focused on the screening for Turner syndrome in the first trimester based on ultrasound markers. Wagner et al[2] built the model that consisted of maternal age, fetal NT thickness, NB, blood flow across TV and through DV, they reported a detection rate of 90.1% for fetuses with trisomy 18, trisomy 13, triploidy, and Turner syndrome based on the model for the fixed FPR of 3.0%. And a recent study of 3856 euploid and 17 Turner syndrome fetuses found similar results.[5] These were consistent with the current study. If this model is used as the primary screening method, more Turner syndrome fetuses would be found earlier. The invasive confirmation test is recommended for fetuses who are suspected of Turner syndrome by the prediction model. The current study has several limitations. First, fetuses with incomplete information were excluded, which may have contributed to bias in our findings. Second, during the study period, diagnostic instruments and the accuracy have increased with considerable scientific and technological progress, which may also contribute to bias in our findings. In conclusion, this study has presented a promising nomogram that incorporates maternal age and ultrasonographic markers. The nomogram can potentially be utilized as a convenient and effective tool in screening for Turner syndrome in the first trimester of pregnancy. Funding This work was supported by grants from the National Natural Science Foundation of China (Nos. 81971619, 82202182), the China Postdoctoral Science Foundation (No. 2020TQ0207), and Postdoctoral Foundation provide by Beijing Obstetrics and Gynecology Hospital, Capital Medical University. Conflicts of interest None.