Study On Predicting Birth Weight Of Full-term Neonate In Southern China By Ultrasonic Measurement Parameters In 21-23 Gestational Weeks

part I Correlation analysis between fetal ultrasound measurement parameters in 21-23 gestational weeks and neonatal birth weight.ObjectivesTo investigate the correlation between neonatal birth weight and fetal growth parameters measured by ultrasound in 21-23 gestational weeks.And further exploring the predictive value of a single ultrasonic parameter for macrosomia and low birth weight infants.Methods1092 pregnant women with full-term singleton pregnancy in southern China were included in this study.To analyse the two-dimensional and three-dimensional ultrasound measurement parameters of fetal growth.Two-dimensional ultrasound parameters includes:biparietal diameter(BPD),head circumference(HC),abdominal circumference(AC),femur length(FL),humerus length(HL),middle upper arm circumference(Amid),mid-thigh circumference(Tmid);three-dimensional ultrasound parameters includes:fetal limb volume(fractional upper arm volume(Avol),fractional thigh volume(Tvol)).The 4D view software calculates the fetal limb volume offline.We randomly select 40 cases to calculate the inter and intra-class correlation coefficient of the same operator and between operators and to test its reliability,and follow up the neonatal birth weight,gender and gestational age.We investigate the correlation between two-dimensional and three-dimensional ultrasound parameters of 1013 fetals in 21-23 gestational weeks and neonatal birth weight by pearson correlation analysis.The ROC curve is further used to explore the sensitivity and specificity of various ultrasound parameters in predicting macrosomia and low birth weight infants.And the cut-off value is then calculated.ResultsThe inter and intra-class correlation coefficient of each parameter measured by the same operator and between two operators are both higher than 0.8.The percentage of differences in repeated measurements are less than 5%.The correlation coefficients of each ultrasonic parameter under the same gestational age and the neonatal birth weight at term are as follow,rTvol=0.381,rAC=0.341,rAvol=0.292,rTmid=0.277,rHL=0.267,rHC=0.247;rFL=0.233,rBpD=0.212,rAmid=0.208,p<0.05.ROC curve analysis shows that inr predicting macrosomia,the AUCs of Tvol,AC,Tmid,and Avol in the second trimester are 0.807,0.793,0.754,0.742 respectively.As for predicting low birth weight infants,the AUCs of AC,Tvol,Avol,and Tmid in the second trimester are 0.682,0.667,0.646,0.637 respectively.ConclusionThere is a strong correlation between neonatal birth weight and fractional thigh volume(Tvol),abdominal circumference(AC),and fractional upper arm volume(Avol)in the second trimester.They can be used as an effective indicator for neonatal birth weight prediction at term.Tvol,AC,Tmid,and Avol of the second trimester are more relevant in the prediction of macrosomia resulting in high sensitivity and specificity.While for the prediction of low birth weight infants at term,AC,Tvol,Avol,Tmid of the second trimester have worse predictive performance,and the sensitivity and specificity are relatively lower.part II A new birth weight prediction model for newborns at term based on the combination of maternal body mass index and two-dimensional,three-dimensional ultrasound parameters in 21-23 gestational weeksObjectivesTo investigate the correlation between maternal BMI and neonatal birth weight,and to combine the conclusions of the first part to establish a neonatal birth weight prediction model based on maternal BMI and ultrasound parameters in 21-23 gestational weeksMethodsWe record maternal height and weight(prenatal,mid-pregnancy and at delivery),and measure two-dimensional and three-dimensional ultrasound data in 21-23 gestational weeks.We follow the birth weight,gender,and gestational age of the infants.Then,we divide all the participants into the EFW model developing group(n=700)and the validation group(n=313)randomly.Single linear regression analysis was used to find out factors that are significantly related to neonatal birth weight(p<0.05).Stepwise multiple linear regression analysis was used to develop the equation for term neonatal birth weight prediction.Finally,the accuracy of the prediction model was tested by the validation group.ResultsMaternal BMI before pregnancy and at delivery,fractional thigh volume,abdominal circumference of the fetus and the gestational age of the mid-term ultrasound scan are related to the neonatal birth weight(p<0.05).EFW(estimated fetal weight)can be calculated by the equation obtained by step to step multiple linear regression analysis.This prediction model predicts 58.15%of the neonatal birth weight within the absolute error of±250g and the ones which have a relative error below 10%between the actual account for 70.29%,which is higher than that of using two-dimensional ultrasound parameters to predict alone.ConclusionWhen we use ultrasound measurement parameters in 21-23 gestational weeks to predict the birth weight of newborns,combining the fractional thigh volume which represents the development of the fetal soft tissue obtained by the three-dimensional ultrasound,can result in a more accurate estimation.