Can we predict delivery date with obstetric ultrasound?

Estimated date of delivery with electronic medical records by a hybrid GBDT-GRU model

An accurate estimated date of delivery (EDD) helps pregnant women make adequate preparations before delivery and avoid the panic of parturition. EDD is normally derived from some formulates or estimated by doctors based on last menstruation period and ultrasound examinations. This study attempted to combine antenatal examinations and electronic medical records to develop a hybrid model based on Gradient Boosting Decision Tree and Gated Recurrent Unit (GBDT-GRU). Besides exploring the features that affect the EDD, GBDT-GRU model obtained the results by dynamic prediction of different stages. The mean square error (MSE) and coefficient of determination (R²) were used to compare the performance among the different prediction methods. In addition, we evaluated predictive performances of different prediction models by comparing the proportion of pregnant women under the error of different days. Experimental results showed that the performance indexes of hybrid GBDT-GRU model outperformed other prediction methods because it focuses on analyzing the time-series predictors of pregnancy. The results of this study are helpful for the development of guidelines for clinical delivery treatments, as it can assist clinicians in making correct decisions during obstetric examinations.

Predictive factors of labour onset using ultrasonography

We analysed the effectiveness of transvaginal ultrasonographic and foetal/maternal pulse Doppler findings as predictors of labour onset within 1 week. We included 22 single normal pregnancies and evaluated the one-point and short- and long-term differences in uterine artery pulsatility index (PI), umbilical artery PI, middle cerebral artery PI (MCA-PI), peak systolic velocity, and cervical length (CL). Presence of funnelling and membrane separation over the internal cervical os was evaluated. Significant changes were observed in the one-point measurement of and short-term and long-term differences in CL, the one-point measurement of and long-term difference in MCA-PI, and the presence of membrane separation (Grade 2). In multivariate analysis, the significant predictors were short-term differences in CL (odds ratio [OR]: 5.27), long-term differences in MCA-PI (OR: 13.3), and presence of membrane separation (Grade 2) (OR: 5.38). Transvaginal ultrasonographic and foetal pulse Doppler findings were effective predictors of labour onset within 1 week.Impact statementWhat is already known on this subject? Parameters reported to predict labour onset include the Bishop score, cervical length, decreased long-term cervical length, funnelling of the internal cervical os, and adrenal gland volume.What do the results of this study add? Short-term changes in cervical length, long-term changes in middle cerebral artery pulsatility index, and the presence of membrane separation Grade 2 were found to be useful predictive factors of labour onset in this study.What are the implications of these findings for clinical practice and/or further research? The prediction of labour onset enables clinicians to properly manage pregnancy and delivery considering maternal and foetal conditions.

Serial Change in Cervical Length for the Prediction of Emergency Cesarean Section in Placenta Previa

Purpose

To evaluate whether serial change in cervical length (CL) over time can be a predictor for emergency cesarean section (CS) in patients with placenta previa.

Methods

This was a retrospective cohort study of patients with placenta previa between January 2010 and November 2014. All women were offered serial measurement of CL by transvaginal ultrasound at 19 to 23 weeks (CL1), 24 to 28 weeks (CL2), 29 to 31 weeks (CL3), and 32 to 34 weeks (CL4). We compared clinical characteristics, serial change in CL, and outcomes between the emergency CS group (case group) and elective CS group (control group). The predictive value of change in CL for emergency CS was evaluated.

Results

A total of 93 women were evaluated; 31 had emergency CS due to massive vaginal bleeding. CL tended to decrease with advancing gestational age in each group. Until 29–31 weeks, CL showed no significant differences between the two groups, but after that, CL in the emergency CS group decreased abruptly, even though CL in the elective CS group continued to gradually decrease. On multivariate analysis to determine risk factors, only admissions for bleeding (odds ratio, 34.710; 95% CI, 5.239–229.973) and change in CL (odds ratio, 3.522; 95% CI, 1.210–10.253) were significantly associated with emergency CS. Analysis of the receiver operating characteristic curve showed that change in CL could be the predictor of emergency CS (area under the curve 0.734, p < 0.001), with optimal cutoff for predicting emergency cesarean delivery of 6.0 mm.

Conclusions

Previous admission for vaginal bleeding and change in CL are independent predictors of emergency CS in placenta previa. Women with change in CL more than 6 mm between the second and third trimester are at high risk of emergency CS in placenta previa. Single measurements of short CL at the second or third trimester do not seem to predict emergency CS.

Abnormal Middle Cerebral Artery Doppler Associates with Spontaneous Preterm Birth in Normally Grown Fetuses

OBJECTIVE

Our aim was to evaluate whether Doppler changes in the fetal middle cerebral (MCA) and umbilical arteries (UA) suggesting fetal hypoxemia precede the onset of spontaneous preterm birth (PTB).

METHODS

We studied 2,340 appropriate-for-gestational-age singleton pregnancies that had MCA and UA pulsatility indices (PI) recorded at 28-32 weeks. Values including the cerebroplacental ratio (CPR) were converted into multiples of the median and evaluated according to both gestational age at the onset of labor and the interval between ultrasound and labor. ROC analysis was used to calculate the ability in the prediction of spontaneous PTB before 32, 34 and 37 weeks' gestation.

RESULTS

While no correlations were observed for the UA PI and CPR, lower MCA PI values were associated with an earlier onset of labor (p < 0.001) and a shorter ultrasound-labor interval (p = 0.028). The ROC analysis at different gestational ages and intervals to labor indicated that MCA PI values were poorly predictive of spontaneous PTB (all areas under the curve <0.7).

CONCLUSIONS

Low MCA PI values at 28-32 weeks are associated with subsequent spontaneous PTB, indicating that fetal hypoxemia unrelated with placental disease might be implicated in the onset of labor. This association, however, is unlikely to be useful in the prediction of PTB.