Dataset Information

Early pregnancy ultrasound measurements and prediction of first trimester pregnancy loss: A logistic model.

ABSTRACT: Our objective was to prospectively validate the use of gestational sac (GS), yolk sac (YS) diameter, crown-rump length (CRL), and embryonal heart rate (HR) dimensions to identify early pregnancy loss. This was a prospective cohort study of first trimester pregnancies. GS and YS diameter, CRL, and HR measurements were serially obtained in singleton and twin pregnancies from 6 through 10 weeks' gestation. Non-parametric tests and logistic regression models were used for comparisons of distributions and testing of associations. A total of 252 patients were included, of which 199 were singleton pregnancies, 51 were twins, and 2 were triplets (304 total fetuses). Fifty-two patients had 61 losses. We built nomograms with the changes of the parameters evaluated in ongoing, as well as in pregnancy loss. In the pregnancies which failed, all the parameters showed significant changes, with different temporal onsets: GS and YS were the first to become abnormal, deviating from normality as early as 6 weeks' gestation (OR 0.01, 95% CI 0.0-0.09, and OR 3.36, 95% CI 1.53-7.34, respectively), followed by changes in HR, and CRL, which became evident at 7 and 8 weeks (OR 0.96, 95% CI 0.92-1.0, and OR 0.59, 95% CI 0.48-0.73, respectively). Our observations showed that, after 5 complete weeks' gestation, a small GS and a large YS reliably predicted pregnancy loss. The YS reliably identified the occurrence of a miscarriage at least 7 days prior its occurrence. CRL and HR became abnormal at a later time in pregnancy and closer to the event. These findings have important implications for patient counseling and care planning, as well as a potential bearing on cost effectiveness within early pregnancy care.

SUBMITTER: Detti L

PROVIDER: S-EPMC6994659 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:BackgroundPregnancy loss prediction based on routinely measured ultrasound characteristics is generally aimed toward distinguishing nonviability. Physicians also use ultrasound indicators for patient counseling, and in some cases to decide upon the frequency of follow-up sonograms. To improve clinical utility, allocation of cut-points should be based on clinical data for multiple sonographic characteristics, be specific to gestational week, and be determined by methods that optimize prediction.ObjectivesTo identify routinely measured features of the early first trimester ultrasound and the gestational age-specific cut-points that are most predictive of pregnancy loss.Materials and methodsThis was a secondary analysis of 617 pregnant women enrolled in the Effects of Aspirin in Gestation and Reproduction (EAGeR) trial; all women had 1-2 previous pregnancy losses and no documented infertility. Each participant had a single ultrasound with a detectable fetal heartbeat between 6 weeks 0 days and 8 weeks 6 days. Cut-points for low fetal heart rate and small crown-rump length were separately defined for gestational weeks 6, 7, and 8 to optimize prediction. Identity and log-binomial regression models were used to estimate absolute and relative risks, respectively, and 95% confidence intervals between jointly categorized low fetal heart rate, small crown-rump length, and clinical pregnancy loss. Adjusted models accounted for gestational age at ultrasound in weeks. Missing data were addressed using multiple imputation.ResultsA total of 64 women experienced a clinical pregnancy loss following the first ultrasound (10.4%), 7 were lost to follow-up (1.1%), and 546 women (88.5%) had a live birth. Low fetal heart rate and small crown-rump length (≤122, 123, and 158 bpm; ≤6.0, 8.5, and 10.9 mm for gestational weeks 6, 7, and 8, respectively) were independent predictors of clinical pregnancy loss, with greatest risks observed for pregnancies having both characteristics (relative risk, 2.08; 95% confidence interval, 1.24-2.91). The combination of low fetal heart rate and small crown-rump length was linked to a 16% (95% confidence interval, 9.1-23%) adjusted absolute increase in risk of subsequent loss, from 5.0% (95% confidence interval, 1.5-8.5%) to 21% (95% confidence interval, 15-27%). Abnormal yolk sac diameter or the presence of a subchorionic hemmhorage did not improve prediction of clinical pregnancy loss.ConclusionIdentified cut-points can be used by physicians for patient counseling, and in some cases to decide upon the frequency of follow-up sonograms. The specified criteria should not be used to diagnose nonviability.

Project description:BackgroundClinical models to predict first trimester viability are traditionally based on multivariable logistic regression (LR) which is not directly interpretable for non-statistical experts like physicians. Furthermore, LR requires complete datasets and pre-established variables specifications. In this study, we leveraged the internal non-linearity, feature selection and missing values handling mechanisms of machine learning algorithms, along with a post-hoc interpretability strategy, as potential advantages over LR for clinical modeling.MethodsThe dataset included 1154 patients with 2377 individual scans and was obtained from a prospective observational cohort study conducted at a hospital in London, UK, from March 2014 to May 2019. The data were split into a training (70%) and a test set (30%). Parsimonious and complete multivariable models were developed from two algorithms to predict first trimester viability at 11-14 weeks gestational age (GA): LR and light gradient boosted machine (LGBM). Missing values were handled by multiple imputation where appropriate. The SHapley Additive exPlanations (SHAP) framework was applied to derive individual explanations of the models.ResultsThe parsimonious LGBM model had similar discriminative and calibration performance as the parsimonious LR (AUC 0.885 vs 0.860; calibration slope: 1.19 vs 1.18). The complete models did not outperform the parsimonious models. LGBM was robust to the presence of missing values and did not require multiple imputation unlike LR. Decision path plots and feature importance analysis revealed different algorithm behaviors despite similar predictive performance. The main driving variable from the LR model was the pre-specified interaction between fetal heart presence and mean sac diameter. The crown-rump length variable and a proxy variable reflecting the difference in GA between expected and observed GA were the two most important variables of LGBM. Finally, while variable interactions must be specified upfront with LR, several interactions were ranked by the SHAP framework among the most important features learned automatically by the LGBM algorithm.ConclusionsGradient boosted algorithms performed similarly to carefully crafted LR models in terms of discrimination and calibration for first trimester viability prediction. By handling multi-collinearity, missing values, feature selection and variable interactions internally, the gradient boosted trees algorithm, combined with SHAP, offers a serious alternative to traditional LR models.

Project description:ObjectivesThere are no international standards for relating fetal crown-rump length (CRL) to gestational age (GA), and most existing charts have considerable methodological limitations. The INTERGROWTH-21(st) Project aimed to produce the first international standards for early fetal size and ultrasound dating of pregnancy based on CRL measurement.MethodsUrban areas in eight geographically diverse countries that met strict eligibility criteria were selected for the prospective, population-based recruitment, between 9?+?0 and 13?+?6?weeks' gestation, of healthy well-nourished women with singleton pregnancies at low risk of fetal growth impairment. GA was calculated on the basis of a certain last menstrual period, regular menstrual cycle and lack of hormonal medication or breastfeeding in the preceding 2 months. CRL was measured using strict protocols and quality-control measures. All women were followed up throughout pregnancy until delivery and hospital discharge. Cases of neonatal and fetal death, severe pregnancy complications and congenital abnormalities were excluded from the study.ResultsA total of 4607 women were enrolled in the Fetal Growth Longitudinal Study, one of the three main components of the INTERGROWTH-21(st) Project, of whom 4321 had a live singleton birth in the absence of severe maternal conditions or congenital abnormalities detected by ultrasound or at birth. The CRL was measured in 56 women at <?9?+?0?weeks' gestation; these were excluded, resulting in 4265 women who contributed data to the final analysis. The mean CRL and SD increased with GA almost linearly, and their relationship to GA is given by the following two equations (in which GA is in days and CRL in mm): mean CRL?=?-50.6562 + (0.815118 × GA) + (0.00535302?× GA(2) ); and SD of CRL?=?-2.21626?+?(0.0984894?×?GA). GA estimation is carried out according to the two equations: GA?=?40.9041?+?(3.21585?×?CRL(0.5) )?+?(0.348956 × CRL); and SD of GA?=?2.39102?+?(0.0193474?×?CRL).ConclusionsWe have produced international prescriptive standards for early fetal linear size and ultrasound dating of pregnancy in the first trimester that can be used throughout the world.

Dataset Information

Early pregnancy ultrasound measurements and prediction of first trimester pregnancy loss: A logistic model.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets