Diffusion Tensor Imaging (DTI)

Diffusion Tensor Imaging (DTI) is a novel imaging technique, which is an extension of diffusion weighted imaging. It has the potential to identify intact nerve fibre tracts and has been used to image the brain for a variety of conditions. It is currented used mainly as a research tool when imaging the spine and has not been widely implemented in mainstream clinical practice. However, early studies have shown that it holds considerable promise in predicting the severity of spinal cord injury.

Discussion

Studies have showed that the diagnostic value/psychometrics of different DTI measures such as apparent diffusion coefficient, fractional anisotropy, radial diffusivity, axial diffusivity, mean diffusivity, relative anisotropy, and volume ratio are highly statistically significant. In 2005, Facon et al. (2005) found that there was a statistically significant difference in the fractional anistrophy values between healthy controls and those with SCI groups. In particular, fractional anistrophy had a much higher sensitivity (SE=73.3%) and specificity (SP=100%) in spinal cord abnormalities detection compared with T2-weighted FSE imaging (SE=46.7%, SP=100%) and ADC (SE=13.4%, SP=80%). Koskinen et al., (2013) reveal that DTI revealed SCI pathology, which was undetectable using conventional MRI. Numerous studies including those by Choe et al. (2017), D’Souza et al. (2017), Shanmuganathan et al. (2008), Shanmuganathan et al. (2017), Kim et al., (2015), Koskinen et al., (2013) demonstrate a statistical difference in fractional anistrophy, axial diffusivity, radial diffusivity, apparent diffusion coefficient, relative anisotropy, and volume ratio values between healthy controls and those with SCI. Shanmuganathan et al. (2008) demonstrates that whole cord relative anisotropy, fractional anistrophy, and relative anisotropy were significantly lower in SCI with hemorrhage compared to controls, while volume ratio was significantly higher. Interestingly, they note that there was no significant differences in fractional anistrophy when comparing SCI to healthy controls. Overall, DTI parameters show considerable diagnostic value, in particular, in its high sensitivity and specificity and in detecting pathology not seen on MRI.

The evidence behind using DTI as a prognostication tool is mixed. Various studies have examined different DTI measures such as apparent diffusion coefficient, fractional anisotropy, radial diffusivity, axial diffusivity, mean diffusivity and their relationships with a number of functional and injury classification scores including International Standards for Neurological Classification of SCI, Functional Independence Measure, Spinal cord independence measure III, Frankel grading system score, American Spinal Injury Association motor score, modified Barthel index score and Abbreviated Injury Scale. Fractional anistrophy has shown promise as a prognostication tool.

In a 2013 case control study by Koskinen et al. the fractional anistrophy values in the SCI group were positively correlated with the motor and sensory scores of The International Standard of Neurological Classification for SCI. Moreover, the fractional anistrophy values in the same group were positively correlated with the motor subscale of Functional Independence Measure.

In 2017, D’Souza et al. found that there was a statistically significant positive correlation between fractional anistrophy values at the level of injury and Frankel grading system score. In contrast, there was no significant correlation between mean diffusivity at the level of injury and Frankel grading system score.

Shanmuganathan et al. (2017) demonstrate that axial diffusivity was significantly correlated with presence of hemorrhagic contusion, The International Standard of Neurological Classification for SCI, and spinal cord independence measure III. Mean diffusivity, axial diffusivity, and radial diffusivity were significant predictors of The International Standard of Neurological Classification for SCI motor score at the 1-year follow-up for participants with or without hemorrhage spinal cord injury. However, fractional anisotropy was not a significant predictor in the model. Moreover, mean diffusivity, axial diffusivity, fractional anistrophy, and radial diffusivity were not significant predictors of spinal cord independence measure III at 1-year follow-up for both hemorrhagic and non-hemorrhagic SCI.

Meanwhile, Choe et al. (2017) found that there were no significant relationships between DTI indices and total International Standard of Neurological Classification for SCI scores from different spinal cord columns.

Conclusions

There is level 3 evidence (from one case control study: Choe et al. 2017) that there may be no significant relationship between DTI incidences and total International Standard of Neurological Classification for SCI scores either at baseline or follow-up in individuals with SCI.

There is level 3 evidence (from one case control study: D’Souza et al. 2017) that there may be a significant positive relationship between fractional anisotropy and Frankel grading system scores, but no relationship between mean diffusivity and Frankel grading system scores.

There is level 3 evidence (from seven case control studies: Choe et al. 2017; D’Souza et al. 2017; Ellingson et al. 2008a; Facon et al. 2005; Kim et al. 2015; Shanmuganathan et al. 2008; Shanmuganathan et al. 2017) that DTI may be an effective tool to measure microstructure abnormalities in individuals with an SCI compared to healthy controls.

There is level 3 evidence (from one case control study: Shanmuganathan et al. 2017) that axial diffusivity may be positively correlated with the presence of hemorrhagic contusion, International Standard of Neurological Classification for SCI motor scores, and spinal cord independence measure III at both baseline and follow-up.

There is level 3 evidence (from one case control study: Shanmuganathan et al. 2017) that axial diffusivity, radial diffusivity, and mean diffusivity may be a significant predictor of International Standard of Neurological Classification for SCI motor scores, but not spinal cord independence measure III scores, for both individuals with and without SCI.

There is conflicting level 3 evidence against (from one case control study (Ellingson et al. 2008a) and one observational study (Wang et al. 2016)) and level 3 evidence for (from one case control study: Koskinen et al. 2013) that DTI grading may have prognostic value in determining motor and sensory scores in individuals with SCI.