Objective  To investigate the advantage and feasibility of three-dimensional computed tomography (CT) reconstruction technique for optic canal fracture evaluation.

  Methods  After the removal of soft tissues, 16 anatomical markers were selected from the optic canals of one Jackal dog for measurement. One ophthalmologist and one radiologist carried out the measurement independently. The 16-planar spiral CT (Sensation 16, Siemens) was adopted for CT reconstruction, using the interval of 0.1 mm. Reconstructive threshold series was built up based on anatomical results of one optic canal. The threshold series included width, level, brightness, opacity, and thickness et al, during multi-planar reconstruction (MPR), surface rendering technique (SRT), volume rendering technique (VRT), and InSpace technique. According to the threshold series from above, CT spatial measurement was made for all the measuring points agreed between two observers. Rotational projection method was used for measuring. According to common types, fracture models were made from 18 optic canals (9 Jackal dogs), including tear fracture, greenstick fracture, rupture fracture, sink fracture, and comminuted fracture. Surface shaded display (SSD), maximum intensity projection (MIP), VRT and InSpace techniques were applied in CT reconstruction. Comparison was made between the CT reconstruction result and the anatomy of the fracture models. CT virtual endoscopy (VE) technique was applied to reconstruct and fly through the optic canal dynamically, so as to evaluate the recurrence of three-dimensional anatomy.

  Results  InSpace technique combined with the rotating planar projection was a reliable method to measure the length of diameter line in space quantitatively. The measurement results using this method showed no significant difference with the anatomical measurement results (P=1.00). The whole spatial morphology of optic canal could be exhibited by SSD, VR, and InSpace technique. The sensitivity and accuracy of each technique were different depending on the fracture type. InSpace technique showed good result for most fracture types, except the sink fracture. SSD was excellent in displaying the sink fracture, although it presented worse for the sharp fracture compared with InSpace and VRT. MIP was not satisfactory in displaying most fracture types. VRT generally appeared to be the medium between InSpace and SSD in displaying optic canal fracture. Three-dimensional reconstruction combined with MPR remarkably improved the sensitivity for detection of fracture. VE technique presented the inner structure of optic canal clearly. Moreover, it could differentiate between fracture and variation. The effect of VE reconstruction was excellent in terms of spatial appearance and topography.

  Conclusion  Volume CT scanning and three-dimensional reconstruction technique is a reliable diagnostic method. The combination of different reconstructive methods can effectively evaluate the optic canal fracture.