Cho Hee Young, Jung Inkyung, Kim Young Han, Kwon Ja-Young
Division of Maternal-Fetal Medicine, Department of Obstetrics and Gynecology, Institute of Women's Life Medical Science, Yonsei University Health System, Seoul, Korea.
Department of Obstetrics and Gynecology, CHA Bundang Medical Center, CHA University, Seongnam, Korea.
Obstet Gynecol Sci. 2019 Mar;62(2):87-92. doi: 10.5468/ogs.2019.62.2.87. Epub 2019 Jan 28.
To evaluate the reliability of the Society for Fetal Urology (SFU) and Onen grading systems for fetal hydronephrosis in prenatal ultrasound according to the level of experience of the examiner.
We reviewed the prenatal ultrasound images of 146 fetuses (292 kidneys) that were diagnosed as having hydronephrosis between January 2005 and December 2014. One expert and two trainees assessed the prenatal renal ultrasound images using the SFU and Onen grading systems. The three examiners independently assessed each ultrasound image with both grading systems and reassessed the same images after 7 to 14 days. Cohen's kappa statistic was used to estimate intra- and inter-observer reliability in prenatal ultrasound images according to training level.
The intra-observer reliability of the SFU grading system (κ 0.873-0.945) showed almost perfect agreement and that of the Onen grading system (κ 0.749-0.913) showed substantial to almost perfect agreement. The overall inter-observer reliability of the SFU grading system (κ 0.620-0.825) showed substantial to almost perfect agreement and that of the Onen grading system (κ 0.618-0.724) showed substantial agreement. The weighted kappa value of inter-observer agreement was 0.223 to 0.400 for SFU grade 1 and 0.064 to 0.346 for SFU grade 3. For Onen grading, the inter-observer agreement was 0.012 to 0.214 for grade 2 and 0.193 to 0.334 for grade 3.
Both the SFU and Onen grading systems showed good intra-observer agreement in prenatal ultrasonography. The inter-observer agreement was decreased in SFU grades 1 and 3 and Onen grades 2 and 3. Therefore, more focus should be given to SFU grades 1 and 3 and Onen grades 2 and 3 for trainees.
根据检查者的经验水平，评估胎儿泌尿外科学会（SFU）和奥嫩（Onen）分级系统在产前超声检查中对胎儿肾积水诊断的可靠性。
我们回顾了2005年1月至2014年12月期间诊断为肾积水的146例胎儿（292个肾脏）的产前超声图像。一名专家和两名实习生使用SFU和Onen分级系统评估产前肾脏超声图像。三名检查者使用这两种分级系统独立评估每张超声图像，并在7至14天后重新评估相同的图像。采用Cohen's kappa统计量根据培训水平评估产前超声图像中观察者内和观察者间的可靠性。
SFU分级系统的观察者内可靠性（κ 0.873 - 0.945）显示几乎完全一致，Onen分级系统的观察者内可靠性（κ 0.749 - 0.913）显示实质性至几乎完全一致。SFU分级系统的总体观察者间可靠性（κ 0.620 - 0.825）显示实质性至几乎完全一致，Onen分级系统的观察者间可靠性（κ 0.618 - 0.724）显示实质性一致。SFU 1级观察者间一致性的加权kappa值为0.223至0.400，SFU 3级为0.064至0.346。对于Onen分级，2级观察者间一致性为0.012至0.214，3级为0.193至0.334。
SFU和Onen分级系统在产前超声检查中均显示出良好的观察者内一致性。SFU 1级和3级以及Onen 2级和3级的观察者间一致性有所降低。因此，对于实习生应更多关注SFU 1级和3级以及Onen 2级和3级。
