Evaluation of traumatic spinal canal stenosis in thoracolumbar burst fractures. A comparison of three methods for measuring the percent canal occlusion.

OBJECTIVES: Various percent canal occlusion measurement methods have been used to assess the spinal canal stenosis following a thoracolumbar burst fracture. However, there are no studies performed comparing these techniques. The objective of this study was to determine the reliability of three techniques used to measure the spinal canal stenosis in thoracolumbar burst fractures on lateral radiograph and CT scans, and to identify the most reliable and practical method of assessment.

METHODS: Fifty preoperative lateral radiographs and axial computed tomography (CT) scans of thoracolumbar burst fractures were analyzed by five observers in two sessions. The three measurement techniques were tested: the ratio of sagittal diameter of spinal canal at the injury level to that at the adjacent level on lateral plain radiographs (Ratio 1) and on axial CT scans (Ratio 2); and the ratio of cross-sectional area of spinal canal at the injury level to that at the adjacent level on axial CT scans (Ratio 3). Interobserver and intraobserver reliability for each measure was assessed.

RESULTS: Greater coefficient of variation (CV) of measurement was founded for the Ratio 1 obtained from lateral plain radiographs than the other two ratios (Ratio 2 and 3) from CT scans (P<0.05), while no differences of CV were noted between the Ratio 2 and Ratio 3 with CT scans (P>0.05). There was no significant improvement (P>0.05) of the interobserver agreement after a 3-month interval for all three methods. A high significant positive correlation was observed between Ratio 1 and Ratio 2, Ratio 1 and Ratio 3, and Ratio 2 and Ratio 3 (P<0.05 for all).

CONCLUSIONS: CT scans are more reliable than plain radiographs in the evaluation of spinal canal occlusion in thoracolumbar burst fractures although the agreement might be acceptable for the latter. Measurement of sagittal diameter of spinal canal using CT scan might be a more reasonable choice than of cross-sectional area of spinal canal because of its simplicity and comparable measurement reliability.