Quantitative MRI in a non-surgical model of cervical spinal cord injury

Quantitative T-2 (qT2), diffusion tensor imaging (DTI), and histology were used to investigate a cervical model of spinal cord injury (SCI) in the rat. While quantitative MRI can significantly increase the specificity in the presence of pathology, it must be validated for each type of injury or disease. In the case of traumatic SCI most models are difficult to image, either due to the location of the injury, or as a result of damage to surrounding tissues resulting from invasive surgical procedures. In this study a non-surgical cervical model of SCI, produced using a combination of focused ultrasound and microbubbles, was used to produce pathology similar to that seen in models of contusive and compressive injuries. qT2 and DTI were performed at 24h and 1 and 2weeks following injury, and compared with H&E and luxol fast blue histology. In the injured spinal cord, in addition to intra/extracellular (I/E) water and myelin water in white matter, qT2 revealed a large component with very short T-2 of about 3ms, which was highly correlated with the presence of hemorrhage in both gray and white matter at 24h, and with the presence of hemosiderin in gray matter at 2weeks following injury. The T-2 of the I/E water peak was also elevated at 24h in both gray and white matter, which was correlated with the presence of vacuolation/edema on histology. Cystic cavities were only seen at the 1 or 2week timepoints, and were correlated with the presence of a water peak with T-2>250ms. No significant changes in diffusivity parameters were observed. Pathologies were often co-occurring, with opposite effects on the average T-2 in a given voxel, reducing the visibility of injured tissue on standard T-2-weighted MR images. Copyright (c) 2015 John Wiley & Sons, Ltd.