Brain MRI and motor function in leukodystrophies.

There is a need to develop clinical outcome measures for future treatment trials, particularly for pediatric studies and for studies of rare disorders for which such outcome metrics are neither currently well-defined nor validated. Patient-reported outcome measures should reflect outcomes that are meaningful to patient-perceived quality of life and sensitive to patient-identified change over time. Physician-reported outcomemeasures are typically quantifiable and consistently applied across multiple study sites by different investigators, which requires that the metrics be well-defined and that investigator training be rigorous. While clinically relevant outcomes are considered the “gold standard,” and are often the mandated primary outcome of federally funded research, many slowly progressive disorders do not demonstrate clinically apparent change over the time course of a clinical trial. Thus, there is a need for biomarkers, such as neuroimaging measures, that may be more sensitive to change over the typical 2to 5-year period of a trial. Key to the applicability of such surrogate measures, however, is that they ultimately correlate with changes in neurologic function. In this issue of Neurology®, Steenweg et al. studied 28 patients with MRI-defined hypomyelinating leukodystrophies. The authors measured T2 relaxation time, magnetization transfer ratio (MTR), fractional anisotropy (FA), mean, axial, and radial diffusivities, and brain metabolites in the splenium, parietal deep white matter, and corticospinal tracts in the centrum semiovale. Steenweg et al. then looked for correlations with motor function as measured by the gross motor function classification system (GMFCS). As expected, virtually all MRI measures differed between patients and healthy controls. Within patients, radial diffusivity (but not other diffusion tensor imaging measures) correlated best with GMFCS scores, although the N-acetylaspartate/ creatine ratio was also independently correlated. These findings suggest that in patients with hypomyelinating leukodystrophies, loss of myelin integrity (and thus increased diffusivity radially) contributes to greater motor dysfunction than does axonal damage (as measured by FA). However, FA values do change in the context of demyelination as well as with axonal loss or axonal transection, and thus the fact that FA did not correlate with GMFCS scores is somewhat puzzling. The reduction in N-acetylaspartate implicated neuronal dysfunction or loss, either directly by reduction inmotor neurons, or more globally as a measure of neuronal degeneration. It would have been of interest to also evaluate magnetization transfer imaging (MTR) in patients with hypomyelinating leukodystrophies. MTR is sensitive to change in myelin and axonal content, and increases inMTR are thought to reflect remyelination in acquired inflammation, such as is seen in multiple sclerosis plaques. MTR is a more difficult sequence to analyze, and interscanner differences make evaluation across sites challenging, which is probably why MTR was not performed by the authors. Overall, it is promising that diffusion tensor imaging and magnetic resonance spectroscopy, and possibly MTR, can be used as biomarkers capable of quantifying white matter integrity and that such metrics may correlate with neurologic dysfunction in patients with leukodystrophy. Longitudinal studies are now required to evaluate these MRI metrics as correlates of disease progression, and future studies will explore their role in disease monitoring during clinical trials. The MRI findings in patients with hypomyelinating leukodystrophies are similar to the role of MRI in monitoring patients with PLP1 (proteolipid protein 1) mutations. In these patients who have this quintessential hypomyelinating leukodystrophy, diffusion tensor imaging revealed abnormal radial and parallel diffusivity (indicative of loss of myelin integrity) across all patients, compared to controls, with more severely affected patients also demonstrating reduced FA (indicative of axonal damage). A limitation of the work of Steenweg et al. is the crosssectional nature of their study. As mentioned above, whether neuroimaging will be more sensitive than clinical tools in evaluating patients over time and their response to various therapies requires serially acquired, highquality imaging. It is of interest that radial diffusivity and Nacetylaspartate/creatine explain only about 50% of the variance in GMFCS scores. Therefore, it is possible that