Carroll Kristen L, Murray Kathleen A, MacLeod Lynne M, Hennessey Theresa A, Woiczik Marcella R, Roach James W
Salt Lake City Shrine Hospital for Children, Fairfax at Virginia, USA.
J Pediatr Orthop. 2011 Jun;31(4):e30-5. doi: 10.1097/BPO.0b013e31821adde9.
Numerous studies underscore the poor intraobserver and interobserver reliability of both the center edge angle (CEA) and the Severin classification using plain film measurements. In this study, experienced observers applied a computer-assisted measurement program to determine the CEA in digital pelvic radiographs of adults who had been previously treated for dysplasia of the hip (DDH). Using a teaching aid/algorithm of the Severin classification, the observers then assigned a Severin rating to these hips. Intraobserver and interobserver errors were then calculated on both the CEA measurements and the Severin classifications.
Four pediatric orthopaedic surgeons and 1 pediatric radiologist calculated the CEAs using the OrthoView TM planning system and then determined the Severin classification on 41 blinded digital pelvic radiographs. The radiographs were evaluated by each examiner twice, with evaluations separated by 2 months. All examiners reviewed a Severin classification algorithm before making their Severin assignments. The intraobserver and interobserver reliability for both the CEA and the Severin classification were calculated using the interclass correlation coefficients and Cohen and Fleiss κ scores, respectively.
The intraobserver and interobserver reliability for CEA measurement was moderate to almost perfect. When we separated the Severin classification into 3 clinically relevant groups of good (Severin I and II), dysplastic (Severin III), and poor (Severin IV and above), our interobserver reliability neared almost perfect.
The Severin classification is an extremely useful and oft-used radiographic measure for the success of DDH treatment. Our research found digital radiography, computer-aided measurement tools, the use of a Severin algorithm, and separating the Severin classification into 3 clinically relevant groups significantly increased the intraobserver and interobserver reliability of both the CEA and Severin classification. This finding will assist future studies using the CEA and Severin classification in the radiographic assessment of DDH treatment outcomes.
众多研究强调，使用X线平片测量中心边缘角（CEA）和Severin分类时，观察者内和观察者间的可靠性较差。在本研究中，经验丰富的观察者应用计算机辅助测量程序，在曾接受过发育性髋关节发育不良（DDH）治疗的成人数字骨盆X线片上测定CEA。然后，观察者使用Severin分类的教学辅助工具/算法，对这些髋关节进行Severin分级。接着计算CEA测量值和Severin分类的观察者内和观察者间误差。
4名小儿骨科医生和1名儿科放射科医生使用OrthoView TM规划系统计算CEA，然后对41张盲法数字骨盆X线片进行Severin分类。每位检查者对这些X线片进行两次评估，两次评估间隔2个月。所有检查者在进行Severin分级之前，均复习了Severin分类算法。分别使用组内相关系数以及Cohen和Fleiss κ评分，计算CEA和Severin分类的观察者内和观察者间可靠性。
CEA测量的观察者内和观察者间可靠性为中度至几乎完美。当我们将Severin分类分为3个具有临床相关性的组，即良好（Severin I和II）、发育不良（Severin III）和较差（Severin IV及以上）时，我们的观察者间可靠性接近几乎完美。
Severin分类是评估DDH治疗成功与否的一项极其有用且常用的影像学测量方法。我们的研究发现，数字X线摄影、计算机辅助测量工具、使用Severin算法以及将Severin分类分为3个具有临床相关性的组，显著提高了CEA和Severin分类的观察者内和观察者间可靠性。这一发现将有助于未来在DDH治疗结果的影像学评估中使用CEA和Severin分类的研究。
