Goldman Ashton H, Hoover Kevin B
Department of Orthopaedic Surgery, Virginia Commonwealth University, Richmond, VA, USA.
Department of Radiology, Virginia Commonwealth University, 1250 E Marshall St. 3rd Floor, PO Box 980615, Richmond, VA, 23298, USA.
Skeletal Radiol. 2017 Apr;46(4):477-481. doi: 10.1007/s00256-017-2571-3. Epub 2017 Jan 22.
Multiple radiographic acquisition techniques have been evaluated for their effect on measurements of acetabular morphology. This cadaveric study examined the effect of two acquisition parameters not previously evaluated: beam center position and source-to-detector distance. This study also evaluated the effect of reader differences on measurements.
Following calibration of measurements between two readers using five clinical radiographs (training), radiographs were obtained from two cadavers using four different source-to-detector distances and three different radiographic centers for a total of 12 radiographic techniques (experimental). Two physician readers acquired four types of measurements from each cadaver radiograph: lateral center edge angle, peak-to-edge distance, Sharp's angle, and the Tonnis angle. All measurements were evaluated for intra-class correlation coefficient (ICC), kappa statistics for hip dysplasia, and factors that resulted in measurement differences using a mixed statistical model.
After training of the two physician readers, there was strong agreement in their hip morphology measurements (ICC 0.84-0.93), agreement in the presence of hip dysplasia (κ = 0.58-1.0), and no measurement difference between physician readers (p = 0.12-1.0). Experimental cadaver measurements showed moderate-to-strong agreement of the readers (ICC 0.74-0.93) and complete agreement on dysplasia (κ = 1). After accounting for reader and radiographic technique, there was no difference in hip morphology measurements (p = 0.83-0.99).
In this cadaveric study, measurements of hip morphology were not affected by varying source-to-detector distance or beam center. We conclude that these acquisition parameters are not likely to affect the diagnosis of hip dysplasia in a clinical setting.
已对多种X线摄影采集技术对髋臼形态测量的影响进行了评估。本尸体研究考察了两个此前未评估过的采集参数的影响：束中心位置和源-探测器距离。本研究还评估了阅片者差异对测量结果的影响。
使用五张临床X线片在两名阅片者之间进行测量校准（培训）后，使用四种不同的源-探测器距离和三个不同的X线摄影中心从两具尸体获取X线片，共12种X线摄影技术（实验）。两名医生阅片者从每张尸体X线片中获取四种测量值：外侧中心边缘角、峰边距、夏普角和托尼角。使用混合统计模型对所有测量值进行组内相关系数（ICC）、髋关节发育不良的kappa统计以及导致测量差异的因素评估。
在对两名医生阅片者进行培训后，他们在髋关节形态测量方面有很强的一致性（ICC 0.84 - 0.93），在髋关节发育不良的存在情况上有一致性（κ = 0.58 - 1.0），且两名医生阅片者之间的测量无差异（p = 0.12 - 1.0）。实验尸体测量显示阅片者之间有中度到高度的一致性（ICC 0.74 - 0.93），在发育不良方面完全一致（κ = 1）。在考虑阅片者和X线摄影技术后，髋关节形态测量无差异（p = 0.83 - 0.99）。
在本尸体研究中，髋关节形态测量不受源-探测器距离或束中心变化的影响。我们得出结论，这些采集参数在临床环境中不太可能影响髋关节发育不良的诊断。
