Hediger Mary L, Fuchs Karin M, Grantz Katherine L, Grewal Jagteshwar, Kim Sungduk, Gore-Langton Robert E, Buck Louis Germaine M, D'Alton Mary E, Albert Paul S
Division of Intramural Population Health Research, Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health, Rockville, Maryland USA.
Department of Obstetrics and Gynecology, Columbia University Medical Center, New York, New York USA.
J Ultrasound Med. 2016 Aug;35(8):1725-33. doi: 10.7863/ultra.15.09087. Epub 2016 Jun 27.
To report on the ultrasound quality assurance program for the National Institute of Child Health and Human Development Fetal Growth Studies and describe both its advantages and generalizability.
After training on an ultrasound system and software, research sonographers were expected to capture blank (unmeasured) images in triplicate for crown-rump length, biparietal diameter, head circumference, abdominal circumference, and femur length. A primary expert sonographer was designated and validated. A 5% sample (n = 740 of 14,785 scans) was randomly selected in 3 distinct rounds from within strata of maternal body mass index (round 1 only), gestational age, and research site. Unmeasured images were extracted from selected scans and measured with the ultrasound software by an expert sonographer. Correlations and coefficients of variation (CVs) were calculated, and the within-measurement standard deviation (ie, technical error of the measurement), was calculated.
The reliability between the site sonographers and the expert was high, with correlations exceeding 0.99 for all dimensions in all rounds. The CV % values showed low variability, with the percentage differences being less than 2%, except for abdominal circumference in rounds 2 and 3, in which it averaged about 3%. Correlations remained high (>0.90) with increasing fetal size; there was a monotonic increase in technical errors of the measurement but without a corresponding increase in the CV %.
Using rigorous procedures for training sonographers, coupled with quality assurance oversight, we determined that the measurements acquired longitudinally for singletons are both accurate and reliable for establishment of an ultrasound standard for fetal growth.
报告美国国立儿童健康与人类发展研究所胎儿生长研究的超声质量保证计划，并描述其优势和可推广性。
在超声系统和软件上进行培训后，研究超声检查人员需针对头臀长、双顶径、头围、腹围和股骨长度采集一式三份的空白（未测量）图像。指定并验证了一名主要专家超声检查人员。从孕妇体重指数（仅第一轮）、孕周和研究地点的分层中，分三轮随机抽取5%的样本（14785次扫描中的740次）。从选定的扫描中提取未测量的图像，并由专家超声检查人员使用超声软件进行测量。计算相关性和变异系数（CV），并计算测量内标准差（即测量技术误差）。
各研究地点的超声检查人员与专家之间的可靠性很高，所有轮次中所有维度的相关性均超过0.99。CV%值显示出低变异性，除第二轮和第三轮腹围平均约为3%外，百分比差异均小于2%。随着胎儿尺寸增加，相关性仍然很高（>0.90）；测量技术误差呈单调增加，但CV%没有相应增加。
通过对超声检查人员进行严格培训，并辅以质量保证监督，我们确定，对于单胎纵向采集的测量数据，在建立胎儿生长超声标准方面既准确又可靠。
