The role of brain MRI in mitochondrial neurogastrointestinal encephalomyopathy

Abstract

Leukoencephalopathy is a hallmark of mitochondrial neurogastrointestinal encephalomyopathy (MNGIE) a devastating disorder characterized by ptosis, ophthalmoparesis, gastrointestinal dysfunction and polyneuropathy. To characterize MNGIE-associated leukoencephalopathy and to correlate it with clinical, biochemical and molecular data, four MNGIE patients with heterogeneous clinical phenotypes (enteropathic arthritis, exercise intolerance, CIDP-like phenotype and typical presentation) were studied by magnetic resonance imaging (MRI) and magnetic resonance spectroscopy (MRS). Diffusion weighted imaging (DWI) with apparent diffusion coefficient (ADC) maps were also obtained. In two patients we also investigated the role of brain MRI in monitoring the evolution of leukoencephalopathy by performing follow-up imaging studies at an interval of one and two years. The extension and distribution of leukoencephalopathy were not clearly linked with age, phenotype or disease severity, and did not seem to be related to TYMP mutations, enzyme activity or pyrimidine levels. In the studied patients MRS revealed reduced N-acetyl-aspartate and increased choline signals. Although DWI appeared normal in all patients but one, ADC maps always showed moderate increased diffusivity. Leukoencephalopathy worsened over a two-year period in two patients, regardless of the clinical course, indicating a lack of correlation between clinical phenotype, size and progression of white matter abnormalities during this period. Brain MRI should be considered a very useful tool to diagnose both classical and atypical MNGIE. Serial MRIs in untreated and treated MNGIE patients will help to establish whether the leukoencephalopathy is a reversible condition or not.

Keywords: DWI; MNGIE; MRI; MRS; mitochondrial neurogastrointestinal encephalomyopathy; posterior reversible encephalopathy syndrome.

Figure 1

FLAIR images of semioval centres in the four patients (Pt) investigated. A) Pt 1, B) Pt 2, C) Pt 3, D) Pt 4. Intensity, morphology and distribution of the signal alteration significantly differ among the four patients, although always maintaining a common pattern. Patient 4, with a typical clinical picture at presentation, had the most widespread white matter T2 hyperintensity involving not only periventricular regions, but also extensively the subcortical areas. Deep pyramidal and subcortical U-fibers appear grossly spared.

Figure 2

FLAIR imaging of semioval centres acquired in Patient 3 at the time of the first study (A), 15 (B) and 28 (C) months later. Hyperintensity progressively spread centrifugally towards frontal subcortical white matter (B) and became more evident in the third study with the appearance of focal bright areas within the deep part of precentral gyri.

Figure 3

Images of caudate nuclei and lateral ventricle bodies acquired in Patient 2. A) FLAIR. B) EPI T2-w b0. C) DWI b1000. D) ADC map. Two regions of interest show the technique used to compare T2 hyperintense areas (ROI 1) with NAWM (ROI 2). While the ADC map effectively reveals the increased diffusivity in ROI 1 corresponding to T2 hyperintensity, b1000 DWI shows a comparable signal in the two areas because of a balance between T2 shine through hyperintensity and signal loss due to increased diffusivity as seen in the ventricles (T2 washout phenomenon). This finding, together with increased water signal in MRS studies, could confirm that, at least partially, T2 WM hyperintensity is due to vasogenic interstitial oedema.

Figure 4

4 MR spectroscopy colour maps, obtained by means of multivoxel studies with 135 ms TE in Patient 3 at the time of the third study (A,C) and in a control subject (B,D). Red areas indicate higher N-acetyl-aspartate (NAA) or N-acetyl-aspartate/choline ratio levels, green areas indicate lower levels. Colour scale legend is shown on the left in Figures 4A,B. The colour map showed lower levels of NAA distribution in T2 hyperintense supraventricular areas in a MNGIE affected patient (A) compared to a normal subject (B) In normal patients NAA tended to be uniformly higher in the semioval centres far from cortical grey matter. The choline/NAA ratio was significantly higher in T2 hyperintense areas (red and yellow) compared to subcortical regions in MNGIE-affected patients (C). This pattern was not found in normal patients where the ratio was stable and comparatively lower (D).