Validation of the macroscopic scoring scheme according to Thompson for pathological changes in invertebral disc degeneration in canine cadaveric spines and correlation with imaging findings using low field magnetic resonance imaging
MetadataShow full item record
Intervertebral disc degeneration (IVDD) is a common cause of chronic back pain in humans and dogs. In human pathology the five-category grading scheme for gross pathological changes, according to Thompson (1), is most often used as the golden standard in IVDD research. The aim of this study was to validate the Thompson scheme in canines and to correlate the findings on gross pathology with imaging findings using low field (0.2 Tesla) magnetic resonance imaging (MRI). A total of 183 intervertebral segments, obtained from 19 randomly selected dogs older than one year of age, euthanized for various unrelated reasons were used for this study. Sagittal T2-weighted MRI of the thoracic and lumbar spine was performed within 24 hours post mortem. Immediately after MRI the spines were cut in the mid-sagittal plane and high resolution photographs were taken of each intervertebral segment (endplate-disc-endplate). The MR images and the photographs were entered into a computer program enabling randomized and blinded scoring by four individual scorers. The 183 segments were macroscopically scored according to Thompson (1). The MR images were scored according to the grading system by Pfirrmann (2). Cohen’s weighted kappa analysis was used for the inter- and intra-observer agreement of the Thompson score (resp. κ 0.90, and κ 0.833) and for the correlation between the results of MRI and macroscopic scoring (κ 0.70). It is concluded that the Thompson score can be used for grading canine IVDD with a high inter- and intraobserver agreement. Correlation between macroscopic grading of spine segments according to Thompson and grading of low field MR images according to Pfirrmann was substantial and therefore low field MRI can be used for clinical diagnosis of canine IVDD although there are some limitations.