Leucoencephalopathy with brainstem and spinal cord involvement and high lactate: quantitative magnetic resonance imaging.

Leucoencephalopathy with brainstem and spinal cord involvement and elevated lactate is a white matter disorder caused by DARS2 mutations. The pathology is unknown. We observed striking discrepancies between improvement on longitudinal conventional magnetic resonance images and clinical deterioration and between large areas of high signal on diffusion-weighted imaging and small areas with low apparent diffusion coefficient values. These observations prompted a longitudinal and quantitative magnetic resonance imaging study. We investigated eight patients (two males, mean age 27 years). Maps of T(2) relaxation times, fractional anisotropy, apparent diffusion coefficients, signal on diffusion-weighted imaging, and axial and radial diffusivities were generated. Brain metabolites, obtained by chemical shift imaging, were quantified. Data analysis focused on: (i) white matter with low apparent diffusion coefficient; (ii) white matter with high T(2) values; (iii) white matter with intermediate T(2) values; and (iv) normal-appearing white matter. The areas were compared with similarly located areas in eight matched controls. In five patients, T(2)-weighted images, spectroscopy, apparent diffusion coefficient maps and diffusion-weighted imaging maps were compared with those obtained 5-7 years ago. In white matter with low apparent diffusion coefficient, axial and radial diffusivities were decreased and fractional anisotropy was high. T(2) values were intermediate. These areas with truly restricted diffusion were small and often observed at the periphery of areas with high T(2) values. In the white matter with high and intermediate T(2) values, apparent diffusion coefficients and axial and radial diffusivities were increased and fractional anisotropy decreased. The signal on diffusion-weighted imaging was highest in white matter with high T(2) values, an effect of T(2) shinethrough. Chemical shift imaging in both white matter types showed increased lactate, increased myo-inositol and decreased N-acetylaspartate, most pronounced in white matter with high T(2) values. Normal-appearing white matter was comparable with white matter of control subjects. Over time, mild decreases in T(2) signal intensities, signal on diffusion-weighted imaging and in extent of the low apparent diffusion coefficient areas were seen. In conclusion, the disease process in leucoencephalopathy with brainstem and spinal cord involvement and elevated lactate is extremely slow. We hypothesize that diffusion restriction is the first stage of the disease caused by intramyelinic water accumulation, followed by slow shift and then loss of the surplus of water. On conventional T(2) images this leads to improvement. We hypothesize that it is loss of water rather than structural restoration that causes the change in T(2) signal intensity, which would be in better agreement with the slow clinical deterioration.