Drerup B, Hierholzer E
Institut für Experimentelle Biomechanik, Universität Münster, Germany.
Clin Biomech (Bristol). 1994 Jan;9(1):28-36. doi: 10.1016/0268-0033(94)90055-8.
Video rasterstereography is a method for back surface measurement comprising automatic back surface reconstruction and shape analysis. It is particularly appropriate for the examination of scoliosis. In this application shape analysis includes model-based calculations of vertebral rotation (determined from surface rotation) and of the spinal midline in three dimensions. The results are delivered in quasi-real time (computing time < 5 min). The aim of the present study was to validate the method by comparison of rasterstereographic and radiographic data. Anteroposterior radiographs and rasterstereographs (478) of 113 scoliosis patients were analysed, each pair taken on the same day. Matching the radiographic midline of the spine to its rasterstereographic equivalent, the deviations between the two curves are properly expressed by their root mean square (r.m.s.) deviation. A r.m.s. deviation in the order of 4 mm was found. Similarly, the r.m.s. deviation of vertebral rotation from surface rotation was about 3°. No systematic difference of vertebral and surface rotation, as reported by other authors, could be found. This may be attributed to our method of data evaluation, consisting of a sophisticated analysis of surface curvature and shape asymmetry. These mathematical procedures are made possible by the high sampling density and resolution of video rasterstereography. Conventional scoliosis parameters (e.g. Cobb angle, apical rotation, apex height, etc.) can be estimated with limited accuracy from the reconstructed midline. The relevant standard deviations are given.
视频光栅立体摄影术是一种用于后表面测量的方法，包括自动后表面重建和形状分析。它特别适用于脊柱侧弯的检查。在这种应用中，形状分析包括基于模型的椎体旋转（由表面旋转确定）和三维脊柱中线的计算。结果以准实时方式提供（计算时间<5分钟）。本研究的目的是通过比较光栅立体摄影数据和放射摄影数据来验证该方法。分析了113例脊柱侧弯患者的前后位X线片和光栅立体照片（478张），每组均在同一天拍摄。将脊柱的放射摄影中线与其光栅立体摄影对应线进行匹配，两条曲线之间的偏差通过其均方根（r.m.s.）偏差得到恰当表示。发现均方根偏差约为4毫米。同样，椎体旋转与表面旋转的均方根偏差约为3°。未发现其他作者所报道的椎体和表面旋转的系统差异。这可能归因于我们的数据评估方法，该方法包括对表面曲率和形状不对称性的精细分析。这些数学程序因视频光栅立体摄影术的高采样密度和分辨率而成为可能。可以从重建的中线以有限的精度估计传统的脊柱侧弯参数（例如Cobb角、顶椎旋转、顶点高度等）。给出了相关的标准差。
