Abstract：
BACKGROUND: The estimation of fetal weight by fetal magnetic resonance imaging is a simple and rapid method with a high sensitivity in predicting birthweight in comparison with ultrasound. Several national and international growth charts are currently in use, but there is substantial heterogeneity among these charts due to variations in the selected populations from which they were derived, in methodologies, and in statistical analysis of data.
OBJECTIVE: This study aimed to compare the performance of magnetic resonance imaging and ultrasound for the prediction of birthweight using 3 commonly used fetal growth charts: the INTERGROWTH-21st Project, World Health Organization, and Fetal Medicine Foundation charts.
METHODS: Data derived from a prospective, single-center, blinded cohort study that compared the performance of magnetic resonance imaging and ultrasound between 36+0/7 and 36+6/7 weeks of gestation for the prediction of birthweight ≥95th percentile were reanalyzed. Estimated fetal weight was categorized as above or below the 5th, 10th, 90th, and 95th percentile according to the 3 growth charts. Birthweight was similarly categorized according to the birthweight standards of each chart. The performances of ultrasound and magnetic resonance imaging for the prediction of birthweight <5th, <10th, >90th, and >95th percentile using the different growth charts were compared. Data were analyzed with R software, version 4.1.2. The comparison of sensitivity and specificity was done using McNemar and exact binomial tests. P values <.05 were considered statistically significant.
RESULTS: A total of 2378 women were eligible for final analysis. Ultrasound and magnetic resonance imaging were performed at a median gestational age of 36+3/7 weeks, delivery occurred at a median gestational age of 39+3/7 weeks, and median birthweight was 3380 g. The incidences of birthweight <5th and <10th percentiles were highest with the Fetal Medicine Foundation chart and lowest with the INTERGROWTH-21st chart, whereas the incidences of birthweight >90th and >95th percentiles were lowest with the Fetal Medicine Foundation chart and highest with the INTERGROWTH-21st chart. The sensitivity of magnetic resonance imaging with an estimated fetal weight >95th percentile in the prediction of birthweight >95th percentile was significantly higher than that of ultrasound across the 3 growth charts; however, its specificity was slightly lower than that of ultrasound. In contrast, the sensitivity of magnetic resonance imaging with an estimated fetal weight <10th percentile for predicting birthweight <10th percentile was significantly lower than that of ultrasound in the INTERGROWTH-21st and Fetal Medicine Foundation charts, whereas the specificity and positive predictive value of magnetic resonance imaging were significantly higher than those of ultrasound for all 3 charts. Findings for the prediction of birthweight >90th percentile were close to those of birthweight >95th percentile, and findings for the prediction of birthweight <5th percentile were close to those of birthweight <10th percentile.
CONCLUSIONS: The sensitivity of magnetic resonance imaging is superior to that of ultrasound for the prediction of large for gestational age fetuses and inferior to that of ultrasound for the prediction of small for gestational age fetuses across the 3 different growth charts. The reverse is true for the specificity of magnetic resonance imaging in comparison with that of ultrasound.
OBJECTIVE: This study aimed to compare the performance of magnetic resonance imaging and ultrasound for the prediction of birthweight using 3 commonly used fetal growth charts: the INTERGROWTH-21st Project, World Health Organization, and Fetal Medicine Foundation charts.
METHODS: Data derived from a prospective, single-center, blinded cohort study that compared the performance of magnetic resonance imaging and ultrasound between 36+0/7 and 36+6/7 weeks of gestation for the prediction of birthweight ≥95th percentile were reanalyzed. Estimated fetal weight was categorized as above or below the 5th, 10th, 90th, and 95th percentile according to the 3 growth charts. Birthweight was similarly categorized according to the birthweight standards of each chart. The performances of ultrasound and magnetic resonance imaging for the prediction of birthweight <5th, <10th, >90th, and >95th percentile using the different growth charts were compared. Data were analyzed with R software, version 4.1.2. The comparison of sensitivity and specificity was done using McNemar and exact binomial tests. P values <.05 were considered statistically significant.
RESULTS: A total of 2378 women were eligible for final analysis. Ultrasound and magnetic resonance imaging were performed at a median gestational age of 36+3/7 weeks, delivery occurred at a median gestational age of 39+3/7 weeks, and median birthweight was 3380 g. The incidences of birthweight <5th and <10th percentiles were highest with the Fetal Medicine Foundation chart and lowest with the INTERGROWTH-21st chart, whereas the incidences of birthweight >90th and >95th percentiles were lowest with the Fetal Medicine Foundation chart and highest with the INTERGROWTH-21st chart. The sensitivity of magnetic resonance imaging with an estimated fetal weight >95th percentile in the prediction of birthweight >95th percentile was significantly higher than that of ultrasound across the 3 growth charts; however, its specificity was slightly lower than that of ultrasound. In contrast, the sensitivity of magnetic resonance imaging with an estimated fetal weight <10th percentile for predicting birthweight <10th percentile was significantly lower than that of ultrasound in the INTERGROWTH-21st and Fetal Medicine Foundation charts, whereas the specificity and positive predictive value of magnetic resonance imaging were significantly higher than those of ultrasound for all 3 charts. Findings for the prediction of birthweight >90th percentile were close to those of birthweight >95th percentile, and findings for the prediction of birthweight <5th percentile were close to those of birthweight <10th percentile.
CONCLUSIONS: The sensitivity of magnetic resonance imaging is superior to that of ultrasound for the prediction of large for gestational age fetuses and inferior to that of ultrasound for the prediction of small for gestational age fetuses across the 3 different growth charts. The reverse is true for the specificity of magnetic resonance imaging in comparison with that of ultrasound.
摘要：
背景：通过胎儿磁共振成像（MRI）估计胎儿体重（EFW）是一种简单而快速的方法，与超声（US）相比，具有很高的预测出生体重（BW）的敏感性。目前正在使用几种国家和国际增长图，但由于以下原因，这些图之间存在很大的异质性：不同的方法和数据统计分析。
目的：本研究的目的是使用三种常用的胎儿生长图比较MRI和US预测BW的性能：INTERGROWTH-21st（IG-21）项目，世界卫生组织(世卫组织)和胎儿医学基金会（FMF）。
方法：数据来自前瞻性，单中心,我们重新分析了在妊娠36+0/7~36+6/7周(WG)之间比较MRI和US预测BW≥95百分位数的盲性队列研究.EFW被归类为>或<5th，>或<第十，>或<90，根据三个增长图表，>或<95百分位数。根据每个图表的BW标准类似地对BW进行分类。US和MRI预测BW<5的表现，<10th,>90，和>95百分位数使用不同的增长图进行了比较。用R软件4.1.2版分析数据。通过McNemar和精确二项测试进行敏感性和特异性的比较。P值<0.05被认为是统计学上显著的。
结果：2378名女性符合最终分析的条件。US和MRI在中位胎龄36+3/7WG进行，分娩发生在39+3/7WG的中位胎龄，和中位数BW为3380克。BW<第5百分位数和<第10百分位数的发生率在FMF图表中最高，在IG-21图表中最低。而BW>90百分位数和>95百分位数的发生率在FMF图表中最低，在IG-21图表中最高。在三个生长图中，EFW>95百分位数的MRI在预测BW>95百分位数的敏感性明显高于US。然而,其特异性略低于美国。相比之下,在IG-21和FMF图表中，EFW<10百分位数的MRI预测BW<10百分位数的敏感性明显低于US，而MRI的特异性和阳性预测值(PPV)明显高于US。BW>第90百分位数的预测结果接近于BW>第95百分位数的预测结果，BW<第5百分位数的预测结果接近BW<第10百分位数的预测结果。
结论：在三个不同的生长图中，MRI在预测胎龄大（LGA）胎儿方面的敏感性优于US，在预测胎龄小（SGA）胎儿方面的敏感性低于US。与US的特异性相比，MRI的特异性正好相反。
目的：本研究的目的是使用三种常用的胎儿生长图比较MRI和US预测BW的性能：INTERGROWTH-21st（IG-21）项目，世界卫生组织(世卫组织)和胎儿医学基金会（FMF）。
方法：数据来自前瞻性，单中心,我们重新分析了在妊娠36+0/7~36+6/7周(WG)之间比较MRI和US预测BW≥95百分位数的盲性队列研究.EFW被归类为>或<5th，>或<第十，>或<90，根据三个增长图表，>或<95百分位数。根据每个图表的BW标准类似地对BW进行分类。US和MRI预测BW<5的表现，<10th,>90，和>95百分位数使用不同的增长图进行了比较。用R软件4.1.2版分析数据。通过McNemar和精确二项测试进行敏感性和特异性的比较。P值<0.05被认为是统计学上显著的。
结果：2378名女性符合最终分析的条件。US和MRI在中位胎龄36+3/7WG进行，分娩发生在39+3/7WG的中位胎龄，和中位数BW为3380克。BW<第5百分位数和<第10百分位数的发生率在FMF图表中最高，在IG-21图表中最低。而BW>90百分位数和>95百分位数的发生率在FMF图表中最低，在IG-21图表中最高。在三个生长图中，EFW>95百分位数的MRI在预测BW>95百分位数的敏感性明显高于US。然而,其特异性略低于美国。相比之下,在IG-21和FMF图表中，EFW<10百分位数的MRI预测BW<10百分位数的敏感性明显低于US，而MRI的特异性和阳性预测值(PPV)明显高于US。BW>第90百分位数的预测结果接近于BW>第95百分位数的预测结果，BW<第5百分位数的预测结果接近BW<第10百分位数的预测结果。
结论：在三个不同的生长图中，MRI在预测胎龄大（LGA）胎儿方面的敏感性优于US，在预测胎龄小（SGA）胎儿方面的敏感性低于US。与US的特异性相比，MRI的特异性正好相反。
