Ductus venosus pulsatility index measurement reduces the false-positive rate in first-trimester screening

Reference values for ductus venosus Doppler velocity indices between 11 and 13+6 weeks of gestation: A single-center prospective study in Iran

Background

This study aimed to investigate reference Doppler velocimetry indices (DVIs) of the fetal ductus venosus (DV) during 11-13 + 6 gestational weeks.

Materials and Methods

In a prospective observation over referrals to a single tertiary care center in a 2-year interval, normal singleton pregnancies with fetal crown-rump lengths (CRLs) of 43-80 mm were examined by a single experienced sonographer for their DV pulsatility index (DVPI), DV resistance index (DVRI), and S-wave maximum velocity/A-wave minimum velocity (S/A ratio). Multinomial and quantile regression functions were used to analyze the effect of gestational age (estimated by CRL) on reference values (5^th and 95^th percentiles of the distribution in each gestational day/week). P < 0.05 was considered significant.

Results

Over a sample of 415 participants with a mean/median gestational age of 12 + 1 weeks, no significant correlations were found between the CRL and DVIs using multinomial regression functions (linear model best fitted for all [DVPI: B coefficient = 0.001, P = 0.235] [DVRI: B coefficient = 0.001, P = 0.287] [DV S/A: B coefficient = 0.010, P = 283]). Quantile regression analyses of DVIs' reference values were nonsignificant across the CRL range except for the DVRI ([5^th regression line: coefficient = -0.004, P = 0.018] [95^th regression line: coefficient = -0.001, P = 0.030]).

Conclusion

Reference values for DVPI, DVRI, and DV S/A ratios were established as 0.80-1.39, 0.62-0.88, and 2.57-6.70, respectively. Future meta-analyses and multicenter studies are required to incorporate DV DVIs into an updated universal version of the practice.

Comparison of Multivariable Logistic Regression and Other Machine Learning Algorithms for Prognostic Prediction Studies in Pregnancy Care: Systematic Review and Meta-Analysis

Background

Predictions in pregnancy care are complex because of interactions among multiple factors. Hence, pregnancy outcomes are not easily predicted by a single predictor using only one algorithm or modeling method.

Objective

This study aims to review and compare the predictive performances between logistic regression (LR) and other machine learning algorithms for developing or validating a multivariable prognostic prediction model for pregnancy care to inform clinicians’ decision making.

Methods

Research articles from MEDLINE, Scopus, Web of Science, and Google Scholar were reviewed following several guidelines for a prognostic prediction study, including a risk of bias (ROB) assessment. We report the results based on the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. Studies were primarily framed as PICOTS (population, index, comparator, outcomes, timing, and setting): Population: men or women in procreative management, pregnant women, and fetuses or newborns; Index: multivariable prognostic prediction models using non-LR algorithms for risk classification to inform clinicians’ decision making; Comparator: the models applying an LR; Outcomes: pregnancy-related outcomes of procreation or pregnancy outcomes for pregnant women and fetuses or newborns; Timing: pre-, inter-, and peripregnancy periods (predictors), at the pregnancy, delivery, and either puerperal or neonatal period (outcome), and either short- or long-term prognoses (time interval); and Setting: primary care or hospital. The results were synthesized by reporting study characteristics and ROBs and by random effects modeling of the difference of the logit area under the receiver operating characteristic curve of each non-LR model compared with the LR model for the same pregnancy outcomes. We also reported between-study heterogeneity by using τ² and I².

Results

Of the 2093 records, we included 142 studies for the systematic review and 62 studies for a meta-analysis. Most prediction models used LR (92/142, 64.8%) and artificial neural networks (20/142, 14.1%) among non-LR algorithms. Only 16.9% (24/142) of studies had a low ROB. A total of 2 non-LR algorithms from low ROB studies significantly outperformed LR. The first algorithm was a random forest for preterm delivery (logit AUROC 2.51, 95% CI 1.49-3.53; I²=86%; τ²=0.77) and pre-eclampsia (logit AUROC 1.2, 95% CI 0.72-1.67; I²=75%; τ²=0.09). The second algorithm was gradient boosting for cesarean section (logit AUROC 2.26, 95% CI 1.39-3.13; I²=75%; τ²=0.43) and gestational diabetes (logit AUROC 1.03, 95% CI 0.69-1.37; I²=83%; τ²=0.07).

Conclusions

Prediction models with the best performances across studies were not necessarily those that used LR but also used random forest and gradient boosting that also performed well. We recommend a reanalysis of existing LR models for several pregnancy outcomes by comparing them with those algorithms that apply standard guidelines.

Trial Registration

PROSPERO (International Prospective Register of Systematic Reviews) CRD42019136106; https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=136106

First trimester ultrasound tests alone or in combination with first trimester serum tests for Down's syndrome screening

BACKGROUND

Down's syndrome occurs when a person has three, rather than two copies of chromosome 21; or the specific area of chromosome 21 implicated in causing Down's syndrome. It is the commonest congenital cause of mental disability and also leads to numerous metabolic and structural problems. It can be life-threatening, or lead to considerable ill health, although some individuals have only mild problems and can lead relatively normal lives. Having a baby with Down's syndrome is likely to have a significant impact on family life.Non-invasive screening based on biochemical analysis of maternal serum or urine, or fetal ultrasound measurements, allows estimates of the risk of a pregnancy being affected and provides information to guide decisions about definitive testing.Before agreeing to screening tests, parents need to be fully informed about the risks, benefits and possible consequences of such a test. This includes subsequent choices for further tests they may face, and the implications of both false positive and false negative screening tests (i.e. invasive diagnostic testing, and the possibility that a miscarried fetus may be chromosomally normal). The decisions that may be faced by expectant parents inevitably engender a high level of anxiety at all stages of the screening process, and the outcomes of screening can be associated with considerable physical and psychological morbidity. No screening test can predict the severity of problems a person with Down's syndrome will have.

OBJECTIVES

To estimate and compare the accuracy of first trimester ultrasound markers alone, and in combination with first trimester serum tests for the detection of Down's syndrome.

SEARCH METHODS

We carried out extensive literature searches including MEDLINE (1980 to 25 August 2011), Embase (1980 to 25 August 2011), BIOSIS via EDINA (1985 to 25 August 2011), CINAHL via OVID (1982 to 25 August 2011), and The Database of Abstracts of Reviews of Effects (the Cochrane Library 2011, Issue 7). We checked reference lists and published review articles for additional potentially relevant studies.

SELECTION CRITERIA

Studies evaluating tests of first trimester ultrasound screening, alone or in combination with first trimester serum tests (up to 14 weeks' gestation) for Down's syndrome, compared with a reference standard, either chromosomal verification or macroscopic postnatal inspection.

DATA COLLECTION AND ANALYSIS

Data were extracted as test positive/test negative results for Down's and non-Down's pregnancies allowing estimation of detection rates (sensitivity) and false positive rates (1-specificity). We performed quality assessment according to QUADAS criteria. We used hierarchical summary ROC meta-analytical methods to analyse test performance and compare test accuracy. Analysis of studies allowing direct comparison between tests was undertaken. We investigated the impact of maternal age on test performance in subgroup analyses.

MAIN RESULTS

We included 126 studies (152 publications) involving 1,604,040 fetuses (including 8454 Down's syndrome cases). Studies were generally good quality, although differential verification was common with invasive testing of only high-risk pregnancies. Sixty test combinations were evaluated formed from combinations of 11 different ultrasound markers (nuchal translucency (NT), nasal bone, ductus venosus Doppler, maxillary bone length, fetal heart rate, aberrant right subclavian artery, frontomaxillary facial angle, presence of mitral gap, tricuspid regurgitation, tricuspid blood flow and iliac angle 90 degrees); 12 serum tests (inhibin A, alpha-fetoprotein (AFP), free beta human chorionic gonadotrophin (ßhCG), total hCG, pregnancy-associated plasma protein A (PAPP-A), unconjugated oestriol (uE3), disintegrin and metalloprotease 12 (ADAM 12), placental growth factor (PlGF), placental growth hormone (PGH), invasive trophoblast antigen (ITA) (synonymous with hyperglycosylated hCG), growth hormone binding protein (GHBP) and placental protein 13 (PP13)); and maternal age. The most frequently evaluated serum markers in combination with ultrasound markers were PAPP-A and free ßhCG.Comparisons of the 10 most frequently evaluated test strategies showed that a combined NT, PAPP-A, free ßhCG and maternal age test strategy significantly outperformed ultrasound markers alone (with or without maternal age) except nasal bone, detecting about nine out of every 10 Down's syndrome pregnancies at a 5% false positive rate (FPR). In both direct and indirect comparisons, the combined NT, PAPP-A, free ßhCG and maternal age test strategy showed superior diagnostic accuracy to an NT and maternal age test strategy (P < 0.0001). Based on the indirect comparison of all available studies for the two tests, the sensitivity (95% confidence interval) estimated at a 5% FPR for the combined NT, PAPP-A, free ßhCG and maternal age test strategy (69 studies; 1,173,853 fetuses including 6010 with Down's syndrome) was 87% (86 to 89) and for the NT and maternal age test strategy (50 studies; 530,874 fetuses including 2701 Down's syndrome pregnancies) was 71% (66 to 75). Combinations of NT with other ultrasound markers, PAPP-A and free ßhCG were evaluated in one or two studies and showed sensitivities of more than 90% and specificities of more than 95%.High-risk populations (defined before screening was done, mainly due to advanced maternal age of 35 years or more, or previous pregnancies affected with Down's syndrome) showed lower detection rates compared to routine screening populations at a 5% FPR. Women who miscarried in the over 35 group were more likely to have been offered an invasive test to verify a negative screening results, whereas those under 35 were usually not offered invasive testing for a negative screening result. Pregnancy loss in women under 35 therefore leads to under-ascertainment of screening results, potentially missing a proportion of affected pregnancies and affecting test sensitivity. Conversely, for the NT, PAPP-A, free ßhCG and maternal age test strategy, detection rates and false positive rates increased with maternal age in the five studies that provided data separately for the subset of women aged 35 years or more.

AUTHORS' CONCLUSIONS

Test strategies that combine ultrasound markers with serum markers, especially PAPP-A and free ßhCG, and maternal age were significantly better than those involving only ultrasound markers (with or without maternal age) except nasal bone. They detect about nine out of 10 Down's affected pregnancies for a fixed 5% FPR. Although the absence of nasal bone appeared to have a high diagnostic accuracy, only five out of 10 affected Down's pregnancies were detected at a 1% FPR.

Cardiac defects, nuchal edema and abnormal lymphatic development are not associated with morphological changes in the ductus venosus

BACKGROUND

In human fetuses with cardiac defects and increased nuchal translucency, abnormal ductus venosus flow velocity waveforms are observed. It is unknown whether abnormal ductus venosus flow velocity waveforms in fetuses with increased nuchal translucency are a reflection of altered cardiac function or are caused by local morphological alterations in the ductus venosus.

AIM

The aim of this study was to investigate if the observed increased nuchal translucency, cardiac defects and abnormal lymphatic development in the examined mouse models are associated with local changes in ductus venosus morphology.

STUDY DESIGN

Mouse embryos with anomalous lymphatic development and nuchal edema (Ccbe1(-/-) embryos), mouse embryos with cardiac defects and nuchal edema (Fkbp12(-/-), Tbx1(-/-), Chd7(fl/fl);Mesp1Cre, Jarid2(-/-NE+) embryos) and mouse embryos with cardiac defects without nuchal edema (Tbx2(-/-), Fgf10(-/-), Jarid2(-/-NE-) embryos) were examined. Embryos were analyzed from embryonic day (E) 11.5 to 15.5 using markers for endothelium, smooth muscle actin, nerve tissue and elastic fibers.

RESULTS

All mutant and wild-type mouse embryos showed similar, positive endothelial and smooth muscle cell expression in the ductus venosus at E11.5-15.5. Nerve marker and elastic fiber expression were not identified in the ductus venosus in all investigated mutant and wild-type embryos. Local morphology and expression of the used markers were similar in the ductus venosus in all examined mutant and wild-type embryos.

CONCLUSIONS

Cardiac defects, nuchal edema and abnormal lymphatic development are not associated with morphological changes in the ductus venosus. Ductus venosus flow velocity waveforms most probably reflect intracardiac pressure.

Is there a role for fetal ductus venosus and hepatic artery Doppler in screening for fetal aneuploidy in the first trimester?

OBJECTIVE

The aim of the present study was to evaluate the utility of ductus venosus (DV) and hepatic artery (HA) doppler in pregnant women who have high risk for aneuploidy in first trimester combined screening.

METHODS

This prospective study was performed between February 2011 and February-2012, at a tertiary referral hospital. Singleton pregnancies with high risk for aneuploidy in combined screening test and normal nuchal translucency (NT) measurements were included in the study group. Measurements of DV Pulsatility Index of Veins (PIV) and HA Pulsatility Index (PI) were compared between the study group and controls.

RESULTS

Within the study period, 104 women with singleton pregnancies were evaluated for DV and HA measurements and among these, 64 women met the inclusion criteria. A control group that comprised 40 women with similar gestational age, normal NT measurements and low-risk in first trimester combined tests was generated. DV-PIV measurements were significantly higher (p = 0.03), whereas HA-PI measurements were similar (p > 0.05) in women who had high-risk for aneuploidy in first trimester combined test.

CONCLUSION

We concluded that the addition of DV-PIV and HA-PI measurements to the first trimester combined screening might increase the accuracy for Down syndrome detection.

Reference range for the pulsatility index ductus venosus Doppler measurement between 11 and 13 + 6 weeks of gestation in a Brazilian population

OBJECTIVE

To establish reference range for the pulsatility index (PI) ductus venosus (DV) Doppler measurement between 11 and 13 + 6 weeks of gestation in a Brazilian population.

METHODS

A retrospective cross-sectional study was performed with 430 singleton pregnancies. The PI DV Doppler measurement was performed as routine during the first trimester screening. DV was identified by color Doppler and the pulsed Doppler gate was placed in the distal portion of the umbilical sinus. When at least three typical DV waveforms were obtained, PI DV was measured manually in one waveform. Polynomial regression was used to obtain the best fit using PI DV Doppler measurement and crown-rump length (CRL) with adjustments by the determination coefficient (R(2)). 5th, 50th and 95th percentiles for the PI DV Doppler measurements at each gestational were determined.

RESULTS

The mean of PI DV Doppler measurement was 1.1 ± 0.2 (range 0.8-3.7). The linear regression was the best fit: PI DV measurement = 1.288-0.0034*CRL (R(2 )= 0.03).

CONCLUSIONS

Reference range for the PI DV Doppler measurement between 11 and 13 + 6 weeks of gestation in a Brazilian population was established.

Cumulative sum plots and retrospective parameters in first-trimester ductus venosus quality assurance

OBJECTIVE

This study aimed to evaluate the application of two quality assurance methods to the ductus venosus pulsatility index (DVPI), as a first-trimester aneuploidy marker, including retrospective assessment of distribution parameters and cumulative sum (CUSUM) plots.

METHODS

The DVPI was measured in 14 444 singleton fetuses at 11+0 to 13+6 weeks in two Fetal Medicine centers during a 4-year period. Sonologist-specific quality assurance distribution parameters, previously described for nuchal translucency, were assessed: the median multiples of the median (MoM), the logarithmic standard deviation of DVPI MoMs and the weekly DVPI percent decrease. Quality assurance results were compared between median MoMs and MoM-based CUSUM plots.

RESULTS

When sonologist-specific DVPI distribution parameters were retrospectively applied for quality assurance, a 1.0 median MoM, a 0.1 median logarithmic standard deviation and a 3.4 median weekly DVPI drop percentage were observed. CUSUM plots showed good agreement with 0.9-1.1 MoMs range for median MoM, in the assessment of sonologist-specific performances.

CONCLUSION

Retrospective and prospective DVPI quality assurance methods appear to be applicable to DVPI at 11+0 to 13+6 weeks. Its use should be encouraged if DVPI is to be added to first-trimester Down syndrome or cardiac defects screening.

Screening performance for trisomy 21 comparing first trimester combined screening and a first trimester contingent screening protocol including ductus venosus and tricuspid flow

OBJECTIVE

To compare the standard first trimester combined risk assessment for trisomy 21 with a contingent screening protocol including tricuspid flow and ductus venosus flow.

MATERIAL AND METHOD

Women with singleton pregnancies and a first trimester combined risk assessment>1:1000 were included. They all had additional assessment of the ductus venosus and the tricuspid flow. We compared screening performance in two screening strategies: (a) First trimester combined screening strategy based on the individual risk results from the routine screening test and (b) Contingent screening strategy based on a combination of the routine test results and additional ultrasound markers.

RESULTS

We included 917 women in the study, 894 in the euploid group and 23 in the trisomy 21 group. Using a contingent screening strategy resulted in a significant decrease in screen positive rate from 48.3% to 17.7% (p<0.001) in the studied population. There was no statistical difference in detection rate between the two screening strategies.

CONCLUSION

There is increasing evidence in favour of using additional ultrasound markers as part of contingent screening protocols in the first trimester. We do suggest performing further studies in routine clinical settings to provide validation of the available risk algorithms.

Long-term effects of in utero Doppler ultrasound scanning--a developmental programming perspective

Ultrasound scanning has been used as a diagnostic and screening tool in obstetric practice for over 50 years. There is no evidence of immediate or long-term harm to the developing fetus from exposure to B mode ultrasound. However, exposure to high levels of Doppler ultrasound during early development is increasingly common, and the full safety implications of this exposure are not clear. Doppler ultrasound exposure in utero gives rise to increased apoptosis in animal models, and there is evidence of the effects of exposure to Doppler ultrasound persisting throughout life, with increased non-right-handedness observed in human epidemiological studies. We consider the idea that there may be long-term developmental implications for fetuses exposed to Doppler ultrasound early in gestation. These effects may be mediated via thermal or mechanical disruption to the developing conceptus, giving rise to free radical damage. Excess free radical exposure early in gestation is a strong candidate for the final common pathway underlying developmental programming effects, and gives rise to concern that fetuses exposed to high levels of ultrasound are at risk of a developmental programming effect. It is suggested that there is a need for animal studies of developmental programming using exposure to Doppler ultrasound scanning as the exposure of interest, and for more observational data to be collected in the clinical setting. While these data are collected, it seems prudent to continue to adhere to the principle of 'as low as reasonably achievable' (ALARA) when exposing first-trimester fetuses to Doppler ultrasound.

Ductus Venosus: A Love Story of 14 Years

Ductus venosus is a tiny vessel with a central role in fetal circulation. Combining B-mode with color and pulsed Doppler is feasible to identify this vessel and evaluate the blood flow waveform at 11 to 13 weeks. The higher prevalence of abnormal A-wave in fetuses with abnormal karyotype and/or cardiac defects turned DV evaluation into a useful marker for chromosomal abnormalities and cardiopathies. Even when combined with nuchal translucency (NT) or biochemical markers, DV blood flow evaluation contributes to an increase in sensitivity and reduces false-positive rate. Abnormal ductal flow is also related to a worse fetal and perinatal outcome. In monochorionic twin pregnancies, in addition to NT measurement at 11 to 14 weeks, the Doppler assessment of DV blood flow increases relevantly the performance of screening for those at higher risk of developing twin-to-twin transfusion syndrome. This story of 14 years surely contributed to change the way first trimester screening is being implemented.

First-trimester measurement of the ductus venosus pulsatility index and the prediction of congenital heart defects

OBJECTIVE

This study was carried out to evaluate the additional predictive value of ductus venosus pulsatility index for veins (DV-PIV) in the identification of congenital heart defects (CHDs) in fetuses with an enlarged nuchal translucency (NT) and a normal karyotype.

METHODS

All chromosomally normal fetuses referred to our Fetal Medicine Unit between September 1996 and December 2008 with known NT, DV-PIV and ductus venosus (DV) a-wave measurements were included. Intrafetus variation in DV-PIV was overcome by averaging three recordings. Follow-up included special focus on CHD. The odds of CHD at any NT and DV-PIV value were evaluated using logistic regression analysis.

RESULTS

Of 792 fetuses included, the NT was enlarged (equal to or above the 95(th) percentile (P95)) in 318 (40.2%). The DV-PIV was abnormal (≥ P95) in 41.8% of the fetuses with an enlarged NT and the a-wave was abnormal (negative or reversed) in 29.9%. CHD was diagnosed in 35 fetuses, 33 of which had an enlarged NT. Amongst the fetuses with an enlarged NT, the sensitivities for CHD of abnormal DV-PIV and DV a-wave were 73% and 55%, with specificities of 62% and 73%, respectively. Logistic regression analysis showed that in this risk group the DV-PIV multiple of the median (MoM) (as a continuous variable) was significantly associated with the risk of CHD (odds ratio = 2.4), independent of the degree of NT enlargement, whereas the DV a-wave did not significantly add to the prediction of CHD.

CONCLUSION

Two-thirds of fetuses with an enlarged NT, a normal karyotype and CHD have an increased DV-PIV. DV-PIV can be used as continuous variable in combination with NT to increase specificity in the identification of CHD and to refine the individual risk assessment.