Three-dimensional shape analysis of the scoliotic spine using MR tomography and rasterstereography.

Rasterstereography delivers only indirect information about the three-dimensional shape of the spine. Therefore, validation with radiologic methods is necessary, in particular in severe scolioses above 50 AE Cobb. Comparison with conventional a.p. radiographs yielded satisfactory results for scolioses up to 50 AE Cobb, but only in frontal view, i.e. in two dimensions. A true three-dimensional validation is possible only with MR tomography, however this is difficult due to severe technical limitations. In the present study 26 patients with Cobb angles between 26 AE and 116 AE were examined both with MR tomography and rasterstereography. Comparing the shape of the spinal midline as measured by the two methods we found that a) the gross 3-d shape of the spinal midline is well reproduced by the rasterstereographic model, b) the lateral rms error is about twice as large as with conventional radiographs and c) that the rms errors increase with increasing Cobb angle. This means that the rasterstereographic model reconstruction must be modified for severe scolioses.