Fujimori Takahito, Iwasaki Motoki, Nagamoto Yukitaka, Kashii Masafumi, Takao Masaki, Sugiura Tsuyoshi, Yoshikawa Hideki
Department of Orthopedic Surgery, Osaka University Graduate School of Medicine, Osaka, Japan.
Clin Spine Surg. 2017 Feb;30(1):E64-E75. doi: 10.1097/BSD.0b013e3182a357ad.
Reliability and agreement study.
To assess the reliability of intraoperative 3-dimensional imaging with a mobile C-arm (3D C-arm) equipped with a flat-panel detector.
Pedicle screws are widely used in spinal surgery. Postoperative computed tomography (CT) is the most reliable method to detect screw misplacement. Recent advances in imaging devices have enabled surgeons to acquire 3D images of the spine during surgery. However, the reliability of these imaging devices is not known.
A total of 203 screws were used in 22 consecutive patients who underwent surgery for scoliosis. Screw position was read twice with a 3D C-arm and twice with CT in a blinded manner by 2 independent observers. Screw positions were classified into 4 categories at every 2 mm and then into 2 simpler categories of acceptable or unacceptable. The degree of agreement with respect to screw positions between the double readings was evaluated by κ value. With unanimous agreement between 2 observers regarding postoperative CT readings considered the gold standard, the sensitivity of the 3D C-arm for determining screw misplacement was calculated.
A total 804 readings were performed. For the 4-category classification, the mean κ value for the 2 interobserver readings was 0.52 for the 3D C-arm and 0.46 for CT. For the 2-category classification, the mean κ value for the 2 interobserver readings was 0.80 for the 3D C-arm and 0.66 for CT. The sensitivity, specificity, positive predictive value, and negative predictive value of intraoperative imaging with the 3D C-arm were 70%, 95%, 44%, and 98%, respectively. With respect to screws with perforation ≥4 mm, the sensitivity was 83%. No revision surgery was performed.
Intraoperative imaging with a 3D C-arm was reliable for detecting screw misplacement and helpful in decreasing the rate of revision surgery for screw misplacement.
可靠性与一致性研究。
评估配备平板探测器的移动C型臂（3D C型臂）术中三维成像的可靠性。
椎弓根螺钉在脊柱手术中广泛应用。术后计算机断层扫描（CT）是检测螺钉误置最可靠的方法。成像设备的最新进展使外科医生能够在手术过程中获取脊柱的三维图像。然而，这些成像设备的可靠性尚不清楚。
连续22例接受脊柱侧弯手术的患者共使用了203枚螺钉。由2名独立观察者以盲法用3D C型臂和CT对螺钉位置各读取两次。螺钉位置每2毫米分为4类，然后再分为可接受或不可接受这2个更简单的类别。通过κ值评估两次读数之间螺钉位置的一致程度。以2名观察者对术后CT读数的一致意见作为金标准，计算3D C型臂确定螺钉误置的敏感性。
共进行了804次读数。对于4类分类，3D C型臂两次观察者间读数的平均κ值为0.52，CT为0.46。对于2类分类，3D C型臂两次观察者间读数的平均κ值为0.80，CT为0.66。3D C型臂术中成像的敏感性、特异性、阳性预测值和阴性预测值分别为70%、95%、44%和98%。对于穿孔≥4毫米的螺钉，敏感性为83%。未进行翻修手术。
3D C型臂术中成像在检测螺钉误置方面可靠，有助于降低螺钉误置的翻修手术率。
