Agreement between magnetic resonance imaging, myelography, and surgery for detecting recurrent, thoracolumbar intervertebral disc extrusion in dogs.

OBJECTIVES Although magnetic resonance imaging (MRI) is reported to be superior to myelography to determine the location and site of first time disc herniation, comparison of these diagnostic methods in cases of recurrent intervertebral disc disease (IVD) herniation after a first surgery has not been evaluated. The objective was to compare the diagnostic accuracy of MRI and myelography in a series of dogs undergoing repeat surgical decompression for recurrent IVD extrusion when compared to the gold standard of surgery. METHODS Ten dogs with recurrent IVD herniation underwent MRI and myelography followed by surgical decompression. Three observers reviewed the images to determine the site and side of the first surgery and the recurrent lesion. Agreement was determined by calculating a kappa (κ) score. RESULTS Substantial interobserver agreement was noted for recurrent lesion site using MRI and myelography (κ = 0.77 vs. 0.73) and when comparing MRI and myelography to the reported surgical site (κ = 0.73 vs. 0.67). Interobserver agreement was greater with MRI for circumferential location compared to myelography (κ = 0.76 vs. 0.43), similar to what was found when comparing to surgical side (κ = 0.82 vs. 0.49). The previous surgical site in this study had no effect on ability to identify the new lesion. CLINICAL SIGNIFICANCE Despite the limitations of MRI, there was greater agreement between observers using MRI for both the recurrent and first lesion.