The previously healthy 13.8-year-old girl was admitted at another hospital with an acute right homonym hemianopsia, increasing sleepiness, anomia, and perturbed consciousness following a febrile upper respiratory tract infection. Despite several therapeutic trials (dexamethason, aciclovir, rocephin intravenous and predison per os) the clinical condition worsened. MRI examinations within the first month after onset of symptoms showed a small lesion in left frontal WM with contrast enhancement, another small lesion within right frontal WM, and a most prominent lesion with widespread edema in the left hemisphere involving parietal, temporal, and occipital lobe and demonstrating a temporary cockade-like contrast enhancement.

On admission to our department 1 month later, clinical examination revealed dysarthria, akalulia, sensoric aphasia, and alexia. Further consecutive MRI examinations showed a regression of the pronounced left parieto-occipital (LPO) WM lesion as depicted in Fig. 1a,b (2 months after onset of symptoms) versus Fig. 1c,d (one month later). Cerebro-spinal fluid investigations revealed a pleozytosis of 30 cells/μl, oligoclonal IgG, and DNA of human herpes virus type 6 (HHV-6) which lead to an intravenous therapy with foscarnet. In a second spinal tap and serum sample after 4 weeks, no HHV-6 DNA was detectable anymore [17]. Over the following 3 months, all symptoms improved significantly except for the hemianopsia which persists up to this day.

Fig. 1 MRI of a 13-year-old child with Balo’s concentric sclerosis at (a,b) 2 months and (c,d) 3 months after onset of symptoms. a Axial FLAIR image demonstrating the concentric pattern of the two left-hemispheric lesions and b T1-weighted image depicting the VOI selected for proton MRS. c Axial FLAIR and d T1-weighted image showing regression of the left parieto-occipital lesion

---

Seven months later, a first relapse occurred with a complex-partial seizure, abulia, and hypokinesia with subsequent remittance. MRI of the brain showed a new contrast-enhancing, concentric lesion in the right frontal premotor region. The known LPO WM lesion appeared smaller. Foscarnet and methylprednisolone were administered intravenously. A maintenance therapy with immunoglobulin (2 g/kg intravenous) was added for 8 months.

The second relapse occurred 3.4 years after the onset with a left hemiparesis, which was again treated with intravenous methylprednisolone. A new concentric lesion was identified within the right suprathalamic region. On T1-weighted images, the LPO WM lesion had transformed to a hypointense, chronic lesion. The remission has so far been incomplete leaving the patient with a left spastic hemiparesis. She now receives an immunomodulatory therapy with α-interferon. Follow-up MRI did not show any new lesion but a considerable atrophy in the LPO WM region.

A total of seven localized proton MRS studies were performed 1, 2, and 4 months as well as 3.4, 3.8, 4.8, and 5.8 years after the onset of symptoms. The institutional review board approved the study and the parents gave informed consent before each examination. The first three studies were conducted at 2 T (Magnetom Vision; Siemens Medical Solutions, Erlangen, Germany), the latter four at 3 T (Magnetom Trio). Proton MR spectra (64 accumulations) were acquired with use of a STEAM localization sequence with repetition time (TR)/echo time (TE)/mixing time (TM) = 6000/20/10 ms as described [18, 19]. The 4.85 ml (2 T) and 4.1 ml (3 T) volume-of-interest (VOI) was placed within the LPO WM covering the structural lesion but excluding the perifocal edema if possible (indicated in Fig. 1b). For the last MRS examination the VOI was reduced to 2.7 ml to account for tissue loss. Absolute concentrations of N-acetylaspartate and N-acetylaspartylglutamate (tNAA), creatine and phosphocreatine (tCr), choline-containing compounds (Cho), myo-inositol (Ins), and lactate (Lac) were determined by LCModel fitting a linear combination of model spectra [20] and compared to age-matched controls (n = 14) taken from Pouwels et al. [21].

DTI was performed on axial slices of 6 mm thickness at 2 T and 4 mm at 3 T using a diffusion-weighted partial Fourier single-shot STEAM MRI sequence with TR = 6.7 s, TE = 74 ms (2 T) and TR = 16.2 s, TE = 50 ms (3 T), and readout flip angle 10° [22, 23]. The protocol comprised a reference image without diffusion-weighting and diffusion-weighted images at a b value of 1000 s mm⁻² along 24 directions of alternate polarity (icosahedron scheme). Maps of the fractional anisotropy (FA) and mean diffusivity (MD) were calculated from the estimated diffusion tensor and co-registered to anatomical images using FSL 3.2 (FMRIB Software Library, Oxford, U.K.). Regional mean FA and MD values were determined for the same LPO WM lesion as studied by proton MRS (indicated in Fig. 2a). These values were compared to data from age- and gender-matched controls (seven healthy female subjects, age range 13–20 years, mean ± standard deviation (SD) 16.2 ± 2.9 years).

Fig. 2 Thresholded axial fractional anisotropy maps overlaid onto corresponding T1-weighted images at a 1 month, b 2 months, c 4 months, d 3.4 years, e 4.8 years, and f 5.8 years after the onset of symptoms. The ellipse shown in a indicates the location for ROI-analysis of FA and MD values

---

To allow for adequate fiber tractography, DTI in the 6th examination was performed at 2.2 mm isotropic resolution (38 contiguous slices, 3 averages). Estimates of axonal projections were computed by the fiber assignment by continuous tracking (FACT) algorithm using software developed in-house [24, 25]. Tracking terminated when the FA value was lower than 0.15 or the main diffusion directions in consecutive steps differed by more than 40° (empirically optimized thresholds). The geniculo-calcarine tracts were determined by region-to-region tracking using regions-of-interest (ROIs) defined on color-coded maps of the main diffusion direction. In order to study secondary degenerative effects of the corpus callosum (CC), the surface of the CC was reconstructed. To compare CC abnormalities of the patient with an age-matched control, a recently developed CC scheme was applied to delineate the fiber topography of the CC by separating transcallosal projections into prefrontal, premotor (and supplementary motor), primary motor, primary sensory, parietal, temporal, and occipital regions [26].

The initial metabolite pattern of the LPO WM lesion with strikingly reduced tNAA and high Cho and Lac concentrations is in line with active demyelination. Almost identical metabolic disturbances have frequently been reported for demyelinating processes in conjunction with various leukodystrophies, for example in children with adrenoleukodystrophy [27]. In this context, the tNAA reduction may be taken as a sign of neuroaxonal damage. This is because tNAA has been shown to be predominantly localized within neurons and axons and validated as an axon-specific marker of cerebral WM [28–30]. Complementary, the MRS-detectable Cho resonance comprises precursor molecules for membrane synthesis as well as the corresponding degradation products and therefore relates to structural integrity [31]. The initial high and for almost 3 years elevated Cho concentration may be attributed to ongoing active demyelination and breakdown of myelin phospholipids. However, increased membrane turnover has also been associated with inflammatory cellular infiltration which is known to occur in BC and therefore a likely contributor to a high Cho level [7, 32]. The subsequent gradual decrease of Cho probably reflects the cessation of active demyelinating processes which are usually replaced by reactive astrocytosis and gliosis. In fact, the course of the Ins concentration, which has been found in high concentrations in astrocytes [33], is in excellent agreement with this histopathological observation.

The Lac concentration was strongly increased 1 month after disease onset and slowly disappeared over time. In general, this finding may indicate non-oxidative glycolysis during ischemia/hypoxia or disturbances in the respiratory chain. Although active BC lesions resemble aspects found in early hypoxic/ischemic WM as well as in active pattern III MS lesions [7], these are not satisfactory explanations for our patient. Instead, Lac production may be preferentially due to the activation of macrophages and their non-oxidative glucose utilization, which have been found in abundance in BC lesions [6, 34]. The observation of mobile lipids in lesions of adult patients with BC [8, 10] could not be confirmed in our study of a child despite the use of short-echo time proton MRS.

A most remarkable finding of this study is the substantial recovery of tNAA during disease progression. The synthesis of NAA from l-aspartate and acetyl-CoA catalyzed by N-acetyl-transferase takes place in the neuronal mitochondria. A translocase transports it into the cytosolic compartment from where it travels the axons via active axonal transport. As NAA cannot be hydrolyzed in these cells, the molecule participates in an intercompartmental cycle between neurons and oligodendrocytes where it is catabolized by aspartoacylase. It has been postulated that NAA turnover is conducted via molecular water pumps and that it plays a major role in brain osmoregulation [35]. Furthermore, NAA has been suggested to act as a precursor for the neurotransmitter NAAG, as an acetyl donor during myelination, and as a molecule participating in interneuronal and intercellular signaling [35, 36].

According to the aforementioned functions, several factors may contribute to the reversible reduction of tNAA. Firstly, edema, as a prominent feature of the acute lesion, may simply reduce the relative number of axons per VOI and later reverse the effect upon absorption [32]. Secondly, owing to the osmoregulatory role of NAA, the dramatic changes in extracellular water content during the formation and resolution of the edema could directly affect the intracellular NAA concentration. In fact, in animal models without loss of neuron viability, NAA was greatly reduced or even absent presumably solely because of its function as a neuronal osmolyte [37]. Thirdly, the mitochondrial function may be reversibly impaired during inflammation in which case NAA synthesis is resumed to a significant extent at later stages [36]. Fourthly, it could be speculated that the axonal transport is re-established after subsidence of active demyelination in parallel to remyelination. The persisting reduction of tNAA to about 70% of control values can very well be attributed to the residual atrophy, that is permanent neuroaxonal loss, within the LPO WM. This explanation is not only appreciable on MRI but also supported by the structural deficits revealed by DTI.

WM diffusion anisotropy reflects the integrity of both axonal membranes and myelin sheaths. FA and MD values in our patient were significantly altered in the BC lesion and in contrast to the MRS results showed no recovery. Reduced anisotropy and increased diffusivity point to a structural damage within the delicate microarchitecture of axonal membranes, periaxonal spaces, and myelin layers which apparently remained without repair. These results are in line with data from MS patients where a range of studies document similar changes in response to a structural tissue damage, for example see Rovaris et al. [38]. Two DTI studies of adult BC reported initially reduced rather than enhanced water diffusivity in active lesions [15, 39]. The fact that we did not see such effects one month after disease onset may be due to their use of very early time points as close as day one of the disease. Almost immediate reductions of the apparent diffusion coefficient that several days later are followed by a reversal to higher values than normal are a typical observation in ischemic brain tissue after acute stroke [40].

Fiber tractography has so far not been performed in BC. It revealed a significantly impaired optic radiation in the left hemisphere almost 5 years after disease onset. The abrupt fiber truncation at the anterior border of the lesion is in full accordance with the persisting clinical symptom of right homonym hemianopsia. Neuroaxonal damage together with gliotic changes may be responsible not only for the lack of structural integrity and, hence, diffusion anisotropy, but also for impairing normal axonal transport and function. Furthermore, a surface reconstruction of the CC showed a pronounced thinning of fiber regions interconnecting right- and left-hemispheric premotor and supplementary motor areas as well as primary motor cortices [26]. In a previous study, Sydykova et al. reported a correlation between reduced FA values in the CC and a decline in fiber integrity and neurodegeneration in corresponding cortical areas [41]. Therefore, reduced FA values in CC regions covering fibers of cortical premotor and supplementary motor areas (compare Fig. 5b) fit well to the observed premotor lesions in both hemispheres.