Convention MRI is the imaging modality of choice to demonstrate the anatomical location and extent in spinal cord injury (SCI) following trauma. However, quantitative and qualitative lesion parameters such as lesion length, maximum spinal cord compression, spinal cord swelling, and presence of hemorrhage within the cord are of limited use in predicting patient neurological outcomes^1,2. Diffusion tensor imaging (DTI) in the c-spine has shown great promise in detecting severity of spinal cord injury³, which may provide valuable biomarker for patient recovery. In this prospective longitudinal study, we aimed to investigate if acute admission DTI parameters can predict patient neurological outcome at 1 year, using the American Spinal Injury Association (ASIA) motor scores.

The study included 30 blunt cervical SCI patients (24 male, 6 female; age: 51.7±17.5) admitted to our trauma center. All patients had a cord contusion seen on conventional MRI, at levels between C2 to C7. MRI imaging was performed acutely within 24 hours of injury. 16 patients were followed up to 1 year and ASIA motor scores were measured. 15 volunteers (12 male, 3 female; age: 45.67 ± 12.94) served as controls.

All MRI were performed on a 1.5T Siemens Avanto scanner. MR imaging included sagittal T2, FLAIR, and axial T2, T2* and 3D susceptibility weighted imaging (SWI) images. DTI were acquired using single-shot EPI sequence (TE/TR=87/2800ms, resolution=128 × 128, FOV=20cm, sl.thick=4mm) with 20 directions at b-value of 700 s/mm² and 3 averages. A total of 20 axial slices were acquired centered at the location of the contusion for patients. For volunteers, three acquisitions were performed with slices placed at upper (lower brainstem to C2), mid (C3-C5) and lower (C6 to T1) sections.

DTI data were reconstructed offline using Diffusion Toolkit (www.trackvis.org) for mean diffusivity (MD), fractional anisotropy (FA), axial diffusivity (AD) and radial diffusivity (RD) maps. An experienced radiologist drew region of interest (ROI) to include the contusion based on T2 and STIR sequences and avoided areas of hemorrhage based on T2* and SWI sequences using Trackvis tool. Hemorrhage within SCI was identified based on T2* and SWI sequences (HC: hemorrhagic contusion, non-HC: non-hemorrhagic contusion). In volunteers, three ROIs were drawn on upper, mid and lower sections of the C-spine. Due to the potential diffusion tensor differences based on c-spine location, the DTI values within the lesion were further normalized by control values based on lesion location. T-tests were performed to compare regional values of patients and controls. Stepwise regression analysis was performed to find relevant parameters (normalized DTI values, age, gender, HC or non-HC) that correlated with patient ASIA score at 1yr.

Figure 1 showed the patient DTI values as compared to different sections of normal volunteers (Lower, Mid, Upper). In volunteers, MD and RD reduces from lower to upper sections, while FA increases from lower to upper sections. AD only differs between the lower and mid sections. Significantly lower FA, AD and higher RD was observed in SCI. Stepwise regression found age and gender had no effect for outcome, while presence of hemorrhage within the SCI (HC) and specific DTI parameter as important factors in patient outcome (p<0.0001). Among the four DTI parameters, including AD in the regression model provided the best results (r²=0.93, p = 0.0019). Other significant DTI parameters included MD (r²=0.89, p = 0.0027) and RD (r²=0.86, p = 0.014). FA was not a significant factor. Figure 2 showed the correlation between DTI parameters with 1 yr ASIA motor score for HC and non-HC patients. Both non-HC and higher MD, AD or RD are indicative of higher ASIA score at 1 yr. Furthermore in HC patients, significant correlation was seen where higher MD, AD and RD were correlated with higher ASIA scores in HC patients.The variation in the DTI values within controls for different sections, as well as altered DTI values in the SCI, are consistent with earlier findings in acute traumatic c-spine injury^3,4. Reduced FA was mostly driven by reduced AD and increased RD, which are indicative of axonal injury and vasogenic edema, respectively. Both hemorrhagic vs. non-hemorrhagic contusion and DTI are strong factors in patient outcome in our group of patients. Among all DTI measures, AD showed strongest effect in predicting outcome, while not RD. As AD is a specific parameter for axonal injury, this may indicate that axonal injury in the cord is the main factor affecting patient recovery. Our study demonstrates acute DTI measurements are a valuable biomarker in predicting patient outcome following blunt cervical SCI.