Assessment of White Matter Damage in Subacute Sclerosing Panencephalitis Using Quantitative Diffusion Tensor MR Imaging

Materials and Methods

We describe 21 children with SSPE (14 boys and 7 girls) with a mean age of 8.6 years (range, 5–14 years). All patients were consistent right-handers. The diagnosis of SSPE was based on typical clinical presentation, EEG pattern, and elevated CSF anti-measles antibody titer.⁴ The characteristic EEG pattern was suppression burst episodes, in which high-amplitude slow and sharp waves of 3–5 seconds recur at intervals of 5 to 8 seconds on a slow background in all these cases. Passive particulate agglutination test (Serodia-Measle; Fujirebio, Tokyo, Japan) for the detection and titration of antibodies to measles virus was used. CSF antibody titer ≥1:128 was considered positive for measles virus infection. The age at measles infection, which could be ascertained in 14 children, ranged from 6 months to 2 years in 13 children and at age 4 years in the remaining child. There was a mean latent period of 7.6 years (range, 4–11 years) for the onset of clinical symptoms in patients with a history of measles. There was a mean gap of 5.8 months (range, 2–12 months) between the onset of symptoms and the imaging in all these cases.

The usual clinical presentation was myoclonus, altered behavior, and cognitive decline. In addition, 12 of the 21 children had extrapyramidal symptoms and difficulty in walking. Jabbour et al⁴ classified these children with SSPE into the following 4 stages on the basis of clinical status: stage I, cerebral signs (mental, behavioral); stage II, convulsive motor signs (myoclonus, incoordination, choreoathetosis, and tremors); stage III, coma, opisthotonus, decerebrate rigidity, and no responsiveness to any stimulus; stage IV, mutism, loss of cerebral cortex function, less frequent myoclonus, and diminished hypertonia. We used the Jabbour classification for the clinical staging and found that all the 21 children were in stage II. Most of children had such cognitive decline that they could not give information beyond their names or identify simple objects like a pen or watch. Detailed neuropsychologic evaluation could not be performed in these patients. We also performed MR imaging on 10 age- and sex-matched right-handed healthy controls (age range, 5–12 years, 8 boys and 2 girls) to compare age-related changes in brain parenchyma in children with SSPE. None of the controls had any neurologic or unstable medical disease nor were they on any medication. The institutional research ethics committee approved the study. Informed consent was obtained from the parents of patients and controls for MR imaging examination.

All MR imaging was performed on a 1.5T MR imaging system (Signa, GE Healthcare, Milwaukee, Wis) equipped with an actively shielded whole-body magnetic field gradient set with a maximal strength of 33 mT/m and a quadrature birdcage receive and transmit radio-frequency head coil. The conventional MR imaging protocol included a T2-weighted fast spin-echo sequence with TR/TE/echo train length/number of excitations, 6000 /200 ms/16/2; a T1-weighted spin-echo sequence with TR/TE/number of excitations, 800/14 ms/1; and a fluid-attenuated inversion recovery (FLAIR) sequence with TR/TE/ TI, 9000/89/2200 ms. Gadodiamide (Gd-DTPA-BMA, Omniscan, Amersham Health, Oslo, Norway), 0.2 mL/kg of body weight, was injected intravenously in these patients. DTI data were acquired by using a single-shot echo-planar dual spin-echo sequence with ramp sampling. The diffusion weighting b-factor was set to 1000 s/mm², TR ∼ 8 seconds, TE ∼100 ms. A total of 36 axial sections was acquired with an image matrix of 256 × 256 (following zero-filling). The diffusion tensor encoding used was the balanced rotationally invariant dodecahedral scheme with 10 uniformly distributed directions over the unit hemisphere.¹⁵ To enhance the signal-to-noise ratio and reduce phase fluctuations, the scanner software repeated (number of excitations, 8) and temporally averaged the magnitude-constructed images. All imaging was performed in the axial plane and had identical geometric parameters: field of view, 240 × 240 mm²; section thickness, 3 mm; section gap, 0; and number of sections, 36.

DTI Data Processing

The DTI data were processed as described in detail elsewhere.¹⁶ Briefly, following image cropping and distortion corrections, the data were interpolated to attain isotropic voxels and decoded to obtain the tensor field for each voxel. The tensor field data were then diagonalized by using the analytic diagonalization method¹⁵ to obtain the eigenvalues (λ₁, λ₂, and λ₃) and the 3 orthonormal eigenvectors. The tensor field data and eigenvalues were used to compute the DTI metrics such as the mean diffusivity (MD) and fractional anisotropy (FA) for each voxel.

To facilitate region of interest (ROI) placement for quantitative analysis, we displayed the DTI-derived maps and overlaid them on images with different contrasts in the 3 orthogonal planes for a visual inspection. Regional FA and MD values were obtained by placing rectangular ROIs on the genu, midbody, and splenium of the corpus callosum at the midsagittal plane in patients as well as in healthy controls. To obtain FA and MD values in the posterior limb of internal capsule, and in the frontal, parietal, and occipital lobes, we placed elliptic and rectangular ROIs at an axial plane through the third ventricle (Fig 1C). Elliptic ROIs were placed at the level of midbrain to obtain FA and MD values from the temporal lobe white matter. ROIs were placed on the same location in both patients with normal as well as abnormal conventional imaging. The ROI placement was based on the description of involvement of various regions on histopathology⁵ and imaging.^11,17 The normal-appearing and hyperintense white matter regions of 4 cerebral lobes of the patients with SSPE on T2-weighted/FLAIR imaging were compared with those of the healthy controls. Patients with hyperintensity on T2-weighted/FLAIR imaging in brain parenchyma were considered as patients with abnormal findings on imaging. To account for the different geometric configuration of both right and left hemispheres, we compared the mean regional FA and MD values of the right hemisphere with those in the left hemisphere of healthy controls. The size of the ROI varied from 2 × 2 to 6 × 6 pixels. Areas of the genu, midbody, and splenium of the corpus callosum were measured by counting the number of pixels in each of the segments on DTI-reconstructed T2-weighted midsagittal images. The shape of the ROI for area measurements was either elliptic or rectangular or drawn freehand, depending on the sampling area within the corpus callosum. All these measurements were performed on both patients and healthy controls.

• Download figure
• Open in new tab
• Download powerpoint

Fig 1.

A 6-year-old healthy control. T2-weighted (A), FA (B), and color-coded FA images fused with apparent diffusion coefficient (ADC) maps (C) through the lateral ventricles show normal distribution of white matter. The rectangular and elliptic ROIs are placed (C) on the frontal, parietal, and occipital cerebral lobes and the posterior limb of the internal capsule for quantification of FA and MD values. The cutoff value for the color-coded FA for display is kept at 0.2 (C), above which the color-coded regions reflect the white matter only (red [right–left)], green [anteroposterior], and blue [superior–inferior)]).

Results

Qualitative Analysis

On conventional MR imaging, 10 of the 21 patients showed abnormalities. Lesions were hyperintense on T2-weighted/FLAIR images and iso- to hypointense on T1-weighted images and were distributed in the periventricular and/or subcortical white matter. In the remaining 11 patients, no abnormalities were detected on conventional MR imaging. In 9 of the 10 patients with abnormal findings on conventional imaging, lesions were distributed in the centrum semiovale (n = 7), frontal white matter (n = 7), parietooccipital white matter (n = 9), temporal white matter (n = 5), splenium (n = 2), and thalamus (n = 1). In the remaining patient, gross cerebral atrophy was present. There was no abnormal parenchymal or meningeal enhancement on postcontrast T1-weighted imaging in any of the patients. On the basis of the imaging findings, the patients were grouped into patients with normal (Fig 2A) and abnormal (Fig 3A) findings on conventional imaging for the purpose of quantitative DTI analysis. Four of 11 patients with normal findings on imaging showed motor dysfunction on clinical examination, whereas in 8 of 10 patients with abnormal findings on imaging, motor functions were impaired. These patients had gait abnormalities, spasticity, brisk deep tendon reflexes, and extensor plantar responses.

• Download figure
• Open in new tab
• Download powerpoint

Fig 2.

A 12-year-old boy with clinical findings of SSPE shows normal appearance on the T2-weighted image (A). FA map (B) shows bilateral significantly low FA values in the white matter (right frontal white matter, 0.15; left frontal white matter, 0.18; right parietal white matter, 0.16; left parietal white matter, 0.15; right occipital white matter, 0.17; left occipital white matter, 0.12). Color-coded FA fused with the ADC map (C) shows the abnormality more clearly.

• Download figure
• Open in new tab
• Download powerpoint

Fig 3.

A 7-year-old boy with SSPE has hyperintensities on the T2-weighted (A) image (arrow) in the right frontal and parietooccipital region. The FA map (B) shows widespread bilateral abnormal white matter (right frontal white matter, 0.06; left frontal white matter, 0.14; right parietal white matter, 0.09; left parietal white matter, 0.13; right occipital white matter, 0.12; left occipital white matter, 0.17) and thinning of the genu and splenium of the corpus callosum. Color-coded FA fused with the ADC map (C) shows the abnormality more clearly.

Discussion

On the basis of the observations in the present study, it appears that patients with clinical stage II disease showing normal findings on conventional MR imaging have widespread microstructural changes in the white matter. Our observations of abnormal DTI measures in stage II SSPE can be explained by pathologic changes, like astrogliosis, neuronal loss and degeneration, demyelination, neurofibrillary tangles, and infiltration of inflammatory cells.^5–9 These observations are in contradiction to the conventional MR imaging findings, which are usually normal in the early stage.² Early imaging study findings may appear normal, with the sequential involvement of posterior parts of cerebral hemisphere as the disease progresses.¹⁷ With time, high-signal-intensity changes in the deep white matter and severe cerebral atrophy occur on T2-weighted images.¹⁷

In the present study, patients with normal findings on conventional imaging (in white matter of all 4 cerebral lobes) showed significantly lower FA values than healthy controls. However, we observed significantly increased MD values in the white matter of the frontal, parietal, and occipital lobes. These observations suggest a net loss and disorganization of the structural barriers to molecular diffusion of water in the NAWM and may be explained by documented pathologic characteristics of SSPE, like demyelination, infiltration of inflammatory cells, and cytolysis of the oligodendrocytes secondary to infection by measles virus.^5,6,18 In addition, our observations are in line with the previous in vivo proton MR spectroscopy studies showing high choline/Creatine(Cr) and myoinositol/Cr ratios with normal N-acetylaspartate in the NAWM regions on conventional imaging.^3,12 However, significantly decreased FA with no significant change in MD in the temporal region probably reflects the early stage of the disease, which is associated with accumulation of inflammatory cells, including macrophages, and the presence of a myelin breakdown product that could potentially restrict water molecular motion, resulting in pseudonormalization of MD.¹⁹

In the present study, we observed significantly reduced FA in the periventricular white matter in patients with abnormal imaging findings compared with both patients with normal conventional imaging findings and healthy controls. Similar results have been obtained on previous DTI studies in patients with other demyelinating inflammatory disease, such as multiple sclerosis, that showed significantly reduced FA values with increased MD values in the lesions compared with NAWM.²⁰

The corpus callosum is associated with a different aspect of human behavior, cognition, and normal aging.²¹ The rostrum and genu and midbody and splenium of the corpus callosum are associated with the parcellation units comprising frontal lobe structures and sensorimotor, midtemporal, and occipital units, respectively. In this study, regional FA values in the patients in both groups were significantly reduced in the genu, midbody, and splenium of the corpus callosum compared with those in healthy controls, suggestive of microstructural damage in the interhemispheric fibers from the hemispheric lobes that pass from the genu, midbody, and splenium of corpus callosum. The involvement of the corpus callosum was further substantiated by callosal cross-sectional-area reduction observed in the genu, midbody, and splenium. The involvement of a callosal structural component could be either due to the interruption of some interhemispheric fibers either by lesion-related secondary damage in white matter tracts of corpus callosum or by primary involvement of oligodendrocytes by the mutant viral antigen. Anlar et al¹¹ reported thinning of the corpus callosum and suggested that its involvement is secondary to periventricular white matter abnormality rather then primary involvement with the disease process. We speculate that the involvement of the corpus callosum is secondary to wallerian degeneration of the fibers crossing the corpus callosum because there was no focal lesion in the corpus callosum in most of the patients in this study.

Cognitive decline is the common clinical feature in patients with SSPE.⁵ For example, occipital-callosal fiber tracts, passing through the splenium, play an important role in normal vision and cognitive functions.²² The significant decrease in the corpus callosum area in patients compared with controls along with decrease in FA confirms its involvement and may be partly responsible for the cognitive decline; however, this hypothesis could not be supported from the results of the present study because detailed cognitive assessment was not performed in these patients.

Pyramidal signs and symptoms usually develop in later stages of SSPE.⁵ Significantly decreased FA was observed in the posterior limb of the internal capsule in patients with normal as well as abnormal findings on imaging compared with healthy controls. However, motor dysfunction was present in only 4 patients with normal findings and in 8 patients with abnormal findings on imaging. Decreased FA in the absence of motor dysfunction suggests that it may be an early indicator of structural abnormality of these motor fibers in SSPE.

Although the current study is limited by the absence of patients with other clinical stages and lack of long-term follow-up, DTI-derived metrics obtained regionally helped in reducing the discrepancy between imaging and clinical findings. Demonstration of involvement of the parietooccipital lobe, with or without involvement of the splenium on DTI, especially in patients with normal findings on conventional imaging, may be used as an early sign of SSPE in young children with a history of childhood fever, skin rashes, and poor school performance along with behavioral changes.

Conclusion

Quantitative DTI can detect the changes in the white matter in the patients with SSPE earlier than conventional MR imaging and may be helpful in treatment planning for these patients.