[Diagnostic value of three-dimensional reconstruction in CT of traumatic spinal fractures]

Rofo. 2003 Nov;175(11):1500-7. doi: 10.1055/s-2003-43405.
[Article in German]

Abstract

Purpose: Evaluation of the diagnostic value of three-dimensional CT-reconstruction in the pre-operative evaluation of traumatic fractures of the spine, compared with axial slices and two-dimensional reconstruction.

Materials and methods: The CT image data of 65 patients with 85 different acute traumatic spine fractures (C 2 through L 5) were collected in a period of 42 months. Retrospectively performed 2D- and 3D-reconstructions were analyzed independently by three CT-experienced readers using the Magerl classification and the readings compared with intraoperative findings or the final diagnoses.

Results: The fractures were classified according to the AO-classification as 56 compression, 16 distraction and 13 rotation fractures. Axial slices alone incorrectly classified 2 (1.2 %) of 168 (56 fractures x 3 readers) type A fractures. The CT classification was incorrect in 17 (35.4 %) of 48 (16 fractures x 3 readers) type B fractures and in 31 (79.5 %) of 39 (13 fractures x 3 readers) type C fractures. The 2D-reconstruction increased the percentage of incorrect diagnoses to 13 (7.7 %) type A, to 26 (54.2 %) type B and to 27 (69.2 %) type C fractures. The 3D-mode incorrectly classified 2 of 168 type A fractures (1.2 %), 6 of 48 type B fractures (12.5 %), and 1 of 39 type C fractures (2.6 %). The axial mode was superior to the 2D-reconstructions in type A and type B fractures, but inferior in type C fractures. The percentage of correctly classified type B fractures was significantly higher (p < 0.05) with 3D-reconstruction than with the 2D-mode. In type C fractures, this percentage was significantly higher (p < 0.001) with the 3D-mode than with axial slices or 2D-reconstruction. The interactive monitor analysis of the 3D-reconstructions with additional virtual cutting by the reader improved the analysis, especially the evaluation of the spinal canal.

Conclusion: The 3D-reconstruction in traumatic spine fractures significantly improves the diagnostic outcome, especially the visualization and classification of rotation fractures. Its disadvantages are a slight increase in artifacts and impaired visualization of small details. Interactive monitor analysis and post-processing of the 3D-mode proved to be advantageous.

MeSH terms

  • Adult
  • Diagnostic Errors
  • False Positive Reactions
  • Female
  • Humans
  • Image Processing, Computer-Assisted
  • Male
  • Reproducibility of Results
  • Retrospective Studies
  • Spinal Fractures / classification
  • Spinal Fractures / diagnostic imaging*
  • Time Factors
  • Tomography, X-Ray Computed / methods*