Blue Nathan R, Savabi Mariam, Beddow Meghan E, Katukuri Vivek R, Fritts Cody M, Izquierdo Luis A, Chao Conrad R
Department of Obstetrics and Gynecology, Division of Maternal-Fetal Medicine, University of New Mexico, Albuquerque, New Mexico, USA.
J Ultrasound Med. 2019 Mar;38(3):587-596. doi: 10.1002/jum.14725. Epub 2018 Sep 23.
To compare a traditional ultrasound (US) method for estimated fetal weight (EFW) calculation and fetal growth restriction diagnosis with 2 newer methods for the prediction of small for gestational age (SGA) at birth.
We reviewed deliveries at our institution from January 1, 2013, to March 31, 2017. Singleton, nonanomalous, well-dated fetuses with a US examination within 2 weeks of delivery were included. Estimated fetal weights and percentiles were calculated by a traditional method (Hadlock et al; Radiology 1991; 181:129-133) and 2 newer methods: Intergrowth-21st (INTG; Ultrasound Obstet Gynecol 2017; 49:478-486) and Salomon et al (Ultrasound Obstet Gynecol 2007; 29:550-555). We calculated each method's test characteristics to predict SGA (birth weight < 10th percentile) using both traditional (EFW < 10th percentile) and receiver operating characteristic (ROC)-derived fetal growth restriction cutoffs. Mean percentile discrepancies between EFW and birth weight measurements were calculated to compare method accuracy. We hypothesized that the INTG and Salomon methods would have superior SGA prediction compared with the Hadlock method.
Of 831 pregnancies with a US examination within 2 weeks of delivery, 138 (16.7%) were SGA at birth. Hadlock had the smallest US-birth weight percentile discrepancy (P < .001 versus both INTG and Salomon). When comparing ROC curves, the Hadlock and INTG methods performed comparably, with areas under the curve of 0.91 and 0.90 (P = .08) and optimal EFW cutoffs of the 15th and 22nd percentiles, respectively. The Salomon method performed less well, with an area under the curve of 0.82 (P < .001 versus both Hadlock and INTG methods).
In our study cohort, the Hadlock method predicted the birth weight percentile more accurately than the INTG or Salomon methods and performed comparably with INTG to predict SGA when ROC-derived cutoffs were used.
比较一种传统超声（US）方法用于估计胎儿体重（EFW）计算及胎儿生长受限诊断，与两种用于预测出生时小于胎龄儿（SGA）的新方法。
我们回顾了2013年1月1日至2017年3月31日在我院的分娩情况。纳入单胎、无畸形、孕周准确且在分娩前2周内接受过超声检查的胎儿。采用传统方法（Hadlock等人；《放射学》1991年；181:129 - 133）以及两种新方法：Intergrowth - 21st（INTG；《超声妇产科》2017年；49:478 - 486）和Salomon等人（《超声妇产科》2007年；29:550 - 555）计算估计胎儿体重和百分位数。我们使用传统方法（EFW <第10百分位数）以及基于受试者工作特征（ROC）得出的胎儿生长受限临界值，计算每种方法预测SGA（出生体重<第10百分位数）的检验特征。计算EFW与出生体重测量值之间的平均百分位数差异以比较方法的准确性。我们假设INTG和Salomon方法在预测SGA方面优于Hadlock方法。
在831例分娩前2周内接受超声检查的妊娠中，138例（16.7%）出生时为SGA。Hadlock方法的超声 - 出生体重百分位数差异最小（与INTG和Salomon方法相比，P <.001）。比较ROC曲线时，Hadlock方法和INTG方法表现相当，曲线下面积分别为0.91和0.90（P = 0.08），最佳EFW临界值分别为第15和第22百分位数。Salomon方法表现较差，曲线下面积为0.82（与Hadlock方法和INTG方法相比，P <.001）。
在我们的研究队列中，Hadlock方法预测出生体重百分位数比INTG或Salomon方法更准确，并且在使用基于ROC得出的临界值时，与INTG方法在预测SGA方面表现相当。
