J Manipulative Physiol Ther. 2001 Nov-Dec; 24(9): 560-8.

Posterior disk displacement: morphologic assessment and measurement reliability--cervical spine.

Cooley JR, Danielson CD, Schultz GD, Hall TA.

Department of Radiology, Southern California University of Health Sciences, USA.

BACKGROUND: Magnetic resonance imaging (MRI) is often used to assess for disk displacement after manipulation, but limited information about the true incidence of iatrogenic herniations exists. Preliminary data must be obtained concerning the size of different types of displacement to further assess this relationship. The reliability of chiropractic radiologists in assessing disks, and a comparison of measuring devices should be evaluated. OBJECTIVE: To identify average measurements for normal and displaced disks and to assess the reliability of measurements by chiropractic radiologists. STUDY DESIGN: Intraobserver and interobserver reliability study assessing cervical disk displacement on MRI scans. METHODS: Three evaluators assessed the disks on 106 MRI scans. Six categories were assessed and compared. Thirty-seven scans were reassessed for intraobserver comparisons. Interobserver and intraobserver variations and measurement-device correlations were determined. RESULTS: Interexaminer measurement reliability for the 2 devices was 0.80 to 0.84. Intraexaminer reliability ranged from 0.58 to 0.94. Interexaminer and intraexaminer agreement for the presence of disk displacement was 86% (kappa = 0.69) and 78% to 85% (kappa = 0.50-0.67), respectively; for the presence of osteophytes, 92% (kappa = 0.54) and 86% to 95% (kappa = 0.60-0.80); and for the classification of disk displacements, 76% (kappa = 0.53) and 73% to 80% (kappa 0.44-0.61). Distinguishing between normal versus bulged disks demonstrated the greatest classification disagreement. Clear size differences between the types of disk displacement were noted. The ruler and digitizer correlation coefficient was 0.96. CONCLUSIONS: Interexaminer and intraexaminer agreement were good to very good concerning measurements and fair to good concerning disk assessments. Different disk displacement types demonstrated obvious mean size differences. No significant mean difference in measurements between the ruler and the digitizer was noted.