Boekhoff TM, Flieshardt C, Ensinger EM, et al.
SUMMARY OF BACKGROUND DATA: The mechanisms of injury in spinal cord injury in dogs are similar to those in human patients and the dog is considered to be a valuable translational model for new treatment modalities. Studies regarding the quantitative characteristics of magnetic resonance imaging (MRI) findings in spinal cord injury in a uniform cohort of patients are lacking. OBJECTIVE AND STUDY DESIGN: The aim of this retrospective study was to evaluate the quantitative MRI signal characteristics of the spinal cord in T2-weighted (T2W) sequences, degree of spinal cord compression, and functional outcome in paraplegic dogs with thoracolumbar disk herniation proving the usefulness of imaging before treatment studies. METHODS: MR images of 63 paraplegic dogs with intact or absent deep pain perception due to thoracolumbar disk herniation examined between January 2005 and June 2009 were reviewed blinded to clinical information. The presence and degree of spinal cord compression, and the presence and length of an intramedullary hyperintensity on T2W images were correlated to clinical signs and outcome. RESULTS: A statistically significant correlation was seen between the neurological status before surgery and the presence and extent of the intramedullary hyperintensity adjacent to the disk herniation in T2W sequences. In dogs with a longer duration of clinical signs, the degree of spinal cord compression was statistically significantly higher. The extent of hyperintensity and the degree of spinal cord compression presented a positive correlation, whereas improvement in the neurological score for each grade was faster with no or a smaller size of T2W intramedullary hyperintensity. CONCLUSIONS: In conclusion, a direct correlation between neurological status and MRI signal intensity and extent was proven. Moreover, the presence and extent of T2W hyperintensity can help determine the prognosis before surgery and to decide, whether new therapeutical strategies in dogs as a translational model should be evaluated.