Дифференцирование патологии острого рассеянного энцефаломиэлита от множественного склероза у детей с использованием диффузных магнитно-резонансных биомаркеров.

Introduction

Multiple sclerosis (MS) is a chronic inflammatory demyelinating disease of the central nervous system (CNS) with a presumed autoimmune origin. Onset in children and adolescents represents 5% or greater of all MS cases. 1 At present, an early MS diagnosis in children is challenging due to important phenotypic overlap with acute disseminated encephalomyelitis (ADEM), particularly in relapsing cases, and the lack of disease-specific biomarkers. 2 ADEM is a frequent demyelinating disorder of childhood, occurring in up to 30% of cases of CNS demyelination. 3 , 4 Most children spontaneously remit, but a large percentage (up to 30%) experience relapse. 5 , 6 ADEM and pediatric-onset MS (POMS) share overlapping clinical features, laboratory findings, and magnetic resonance (MR) imaging (MRI). 7 Conventional MRI scans and published criteria do not reliably discriminate between the conditions at the time of initial presentation due to low sensitivity or specificity. 8 - 11

ADEM and MS share several overlapping neuroimaging features, such as the location and character of white matter lesions and the presence of black holes, which make it difficult to differentiate one disease from the other with certainty. 8 - 10 Despite limitations, classic MS can still exhibit conventional MRI lesional differences from classic ADEM. ADEM lesions are characterized by bilateral, diffuse, large, fluffy, poorly marginated T2/fluid-attenuated inversion recovery-hyperintense lesions exhibiting patchy heterogeneous enhancement with a particular predisposition (notably in myelin oligodendrocyte glycoprotein [MOG] antibody-associated disease [MOGAD] cases) to the cerebellar peduncle and conus medullaris, demonstrating greater gray matter involvement, a more inflammatory appearance (e.g., fat stranding and perineural sheath enhancement), and a higher likelihood of bilateral optic neuritis, in contrast to typical MS, where the lesions are typically more ovoid, smaller, ring enhancing, and well demarcated, situated periventricular and perpendicular to the long axis of the corpus callosum. 7 , 12

Although the presence of MOG-IgG positivity would favor ADEM, antibody levels can fluctuate, 13 and low positive titer has been reported in patients with MS. 4 , 11 Physical examination does not increase diagnostic accuracy. As a result, a final diagnosis can be delayed by several years, increasing the chance of suboptimal treatment. 14

Pathologically, ADEM typically demonstrates a unique pattern of inflammation including diffuse perivenous and confluent demyelination and is associated with MOG-IgG in 50% of cases. 4 MS is characterized mostly by intracortical demyelinating lesions that can develop to slowly increasing demyelinated plaques with oligodendrocyte loss, which is uncommon in ADEM. 15 In analyses of inflammatory cellular infiltration, lesions of patients with MOGAD had CD4+ T cell predominance, 15 which differs from the CD8+ T cell predominance observed in MS lesions. 16 , 17 Furthermore, axonal injury is rare in ADEM, even in severe cases, whereas axonal loss is prominent in the first year of MS onset. 18 , 19

Diffusion tensor imaging (DTI) has been utilized to measure white matter damage in adult and pediatric patients with MS. 20 - 23 In the conventional DTI analysis, a single tensor is utilized to represent the averaged effects of multiple tensors within different microstructural compartments. The method conflates multiple sources of biological variance. Diffusion basis spectrum imaging (DBSI) models the diffusion-weighted MR signals as a linear combination of discrete anisotropic diffusion tensors (representing axonal fiber tracts) and a spectrum of isotropic diffusion tensors (reflecting inflammatory cell infiltration and vasogenic edema, as well as intrinsic cells and extracellular space), overcoming the deficiencies of DTI. 24 - 26

DBSI quantifies inflammation, axonal loss, and vasogenic edema by measuring restricted and nonrestricted isotropic diffusion tensor fractions (DBSI-RF and DBSI-NRF). Specifically, DBSI has the capability to detect axonal integrity through DBSI axial diffusivity (DBSI-AD), demyelination via DBSI radial diffusivity (DBSI-RD), integrity of residual axons to indicate ongoing functional impairment through DBSI fractional anisotropy (DBSI-FA), and apparent axonal density via fiber fraction (DBSI-FF). 26 , 27 DBSI-FF has been shown to be an excellent tool to measure axonal loss in animal models. 28 - 30

In prior work we have shown that DTI can show differences between children with ADEM and those with a similar presentation who go on to have a diagnosis of MS. 31 Recent clinical studies in non-MS and non-ADEM populations report improved sensitivity and specificity to the underlying pathologies of DBSI when compared with traditional DTI. 24 , 27 , 32 DBSI has not been applied in pediatric demyelinating diseases. In this study, we examine white matter tract pathologies in ADEM and pediatric MS using DBSI. The objective is to establish if diffusion metrics may enhance diagnostic procedures and provide insights into neuropathological differences between ADEM and MS.

This study aimed to test the hypothesis that DBSI might detect differences in the pathologic changes of normal-appearing white matter (NAWM) present in ADEM and MS.

Materials and Methods

Human subjects

Procedures were approved by the Institutional Review Board of Washington University. We obtained consents from a parent/guardian and assents from children who were older than seven years. We prospectively enrolled 23 children presenting with CNS demyelination and eight with demyelination and encephalopathy fulfilling the 2013 International Pediatric MS Study Group (IPMSSG) criteria 33 for ADEM and 15 who fulfilled IPMSSG criteria for MS at the time of recruitment.

Clinical information

We obtained the following information: age at first demyelination, localization of first event, diagnosis at first DBSI/DTI imaging, disease duration at first DBSI/DTI, disease duration at last DBSI/DTI, total number relapses in the first two years of disease, total duration of follow-up, Expanded Disability Status Scale at last follow-up, and MOG-IgG positivity at any time.

MRI acquisition

Figure 1 summarizes the study design. In addition to standard anatomic sequences, the MR protocol included a 99-direction multiple b-value DBSI protocol performed on 3T Siemens Prisma Scanner. DBSI/DTI scans were obtained following the first demyelination attack, and follow-up scans were performed at varying intervals, as clinically indicated, from a minimum of two months to a maximum of 36 months. Data on diffusion-weighted MRI were acquired in the axial plane, encompassing the entire brain, at a resolution of 2 × 2 × 2 mm 3 . The repetition time and echo time were maintained at 10,000/120 milliseconds. The 99-direction sampling scheme as prescribed by Wang et al. 25 with a maximum b-value of 1500 s/mm 2 was employed for acquiring diffusion-weighted MRI data for DTI/DBSI analyses. DTI/DBSI metric maps were computed using an in-house MATLAB software developed at the Mallinckrodt Institute of Radiology of Washington University in St. Louis. 25 DBSI applies a multitensor model to analyze diffusion-weighted MRI signals. 25 B0 and DTI-FA were used for regions-of-interest analysis on NAWM at midsagittal corpus callosum ( Fig 2 ). We measured DBSI-AD, DBSI-RD, DBSI-FA, DBSI-RF, DBSI-NRF, and DBSI-FF in these regions of interest as well as DTI axial diffusivity, DTI radial diffusivity, and DTI fractional anisotropy. Mean and standard deviations (S.D.s) were used to summarize DBSI measures for each group.

Study design. The color version of this figure is available in the online edition.

Anatomic magnetization-prepared rapid acquisition gradient echo (A) was used as the “template” for registering T2-weighted preliminary study images. (B) B0. After coregistration, region of interest was defined in anatomic images (A and B) and mapped to diffusion magnetic resonance imaging metric maps, for example, diffusion tensor imaging (DTI)-fractional anisotropy (C), DTI-apparent diffusion coefficient (D), diffusion basis spectrum imaging (DBSI)-restricted fraction (E), DBSI-hindered fraction (F), and DBSI nonrestricted fraction (G). The color version of this figure is available in the online edition.

Statistical analysis

Data analysis was performed using SPSS software version 26 (IBM SPSS Statistics, Armonk, NY, USA). Continuous data were expressed as the mean ± S.D. if the distribution was normal and median (minimum-maximum) otherwise. Categorical variables were presented as a number and percentage and compared using the Fisher exact test (independent parameters). Kruskal-Wallis test is used to compare DBSI and DTI measures. Univariate analysis was performed to assess diffusivity measurements in POMS and ADEM over time. Covariance analysis was used to compare age-adjusted measurements among the two groups. Statistical significance was defined as a two-tailed P value of <0.05.

Results

Patient characteristics are summarized in Table 1 . The age at the first demyelination event was nearly half in patients with ADEM compared with the patients with MS ( P = 0.002). There was no significant gender ratio difference between the two groups. Similarly, patients with ADEM had a younger average age at initial imaging than the patients with MS ( P = 0.03). Both groups experienced clinical relapses during follow-up. MOG IgG positivity was present in patients with ADEM (87%), but not in patients with MS.

TABLE 1. Clinical Characteristics of POMS and ADEM Clinical Characteristics POMS (n = 15) ADEM (n = 8) P Value Sex M/F, n (%)   6 (40%)/9 (60%)    2 (25%)/6 (75%) 0.471 Average age at first demyelination (years/S.D.) 12.2 (4.3) 6.7 (3.1) 0.002 Average age at first imaging (years/S.D.) 15.5 (14) 11.7 (11) 0.037 Disease duration at first imaging (years/S.D.)   2.7 (2.4)   1.3 (1.9) 0.25 Disease duration at last FU (years/S.D.)   4.4 (1)   4.5 (4.1) 0.96 EDSS at last FU (S.D.)   1.0 (1)   1.1 (1.1) 0.69 MOG-IgG positivity 0   7 (1:40-1:10,000) Treated with DMT (Yes/No) 15/0 5/3 Abbreviations: ADEM = Acute disseminated encephalomyelitis DMT = Disease modifying therapy EDSS = Expanded Disability Status Scale F = Female FU = Follow-up M = Male MOG = Myelin oligodendrocyte glycoprotein POMS = Pediatric-onset multiple sclerosis

Clinical Characteristics of POMS and ADEM

Abbreviations:

ADEM = Acute disseminated encephalomyelitis

DMT = Disease modifying therapy

EDSS = Expanded Disability Status Scale

F = Female

FU = Follow-up

M = Male

MOG = Myelin oligodendrocyte glycoprotein

POMS = Pediatric-onset multiple sclerosis

Analyses of baseline and follow-up DBSI and DTI measurements between the ADEM and POMS groups are summarized in Tables 2 and 3 , respectively. There was no significant difference in DBSI-FA of fibers, DBSI-RD, DBSI-RF, and DBSI-NRF at baseline measurements between the ADEM and POMS groups. There was also no significant difference in baseline and over-time DTI measurements between the ADEM and POMS groups.

TABLE 2. Univariate Analysis of Baseline Measurements in POMS and ADEM DBSI and DTI Measurements ADEM POMS P Value N = 8 N = 15 DBSI (mean)  Axial diffusivity 2.06 2.09 0.73  Fractional anisotropy 0.89 0.87 0.18  Fiber fraction 0.7 0.74 0.23  Radial diffusivity 0.21 0.25 0.14  Nonrestricted fraction 0.13 0.11 0.46  Restricted fraction 0.16 0.15 0.36 DTI (mean)  Axial diffusivity 1.31 1.35 0.54  Fractional anisotropy 0.72 0.72 0.85  Radial diffusivity 0.32 0.32 0.96 Abbreviations: ADEM = Acute disseminated encephalomyelitis DBSI = Diffusion basis spectrum imaging DTI = Diffusion tensor imaging POMS = Pediatric-onset multiple sclerosis

Univariate Analysis of Baseline Measurements in POMS and ADEM

Abbreviations:

ADEM = Acute disseminated encephalomyelitis

DBSI = Diffusion basis spectrum imaging

DTI = Diffusion tensor imaging

POMS = Pediatric-onset multiple sclerosis

TABLE 3. Univariate Analysis of Average Over-Time Measurements in POMS and ADEM DBSI and DTI Measurements ADEM POMS P Value N = 17 N = 31 DBSI (mean)  Axial diffusivity 2.12 2.09 0.78  Fractional anisotropy 0.89 0.87 0.002  Fiber fraction 0.71 0.74 0.046  Radial diffusivity 0.21 0.25 0.016  Nonrestricted fraction 0.13 0.1 0.005  Restricted fraction 0.16 0.17 0.32 DTI (mean)  Axial diffusivity 0.35 0.30 0.15  Fractional anisotropy 0.73 0.73 0.08  Radial diffusivity 0.31 0.3 0.30 Abbreviations: ADEM = Acute disseminated encephalomyelitis DBSI = Diffusion basis spectrum imaging DTI = Diffusion tensor imaging POMS = Pediatric-onset multiple sclerosis

Univariate Analysis of Average Over-Time Measurements in POMS and ADEM

Abbreviations:

ADEM = Acute disseminated encephalomyelitis

DBSI = Diffusion basis spectrum imaging

DTI = Diffusion tensor imaging

POMS = Pediatric-onset multiple sclerosis

In analysis of follow-up DBSI measurements between the ADEM and POMS groups, there were significant differences in DBSI-FA, DBSI-FF, DBSI-RD, and DBSI-NRF over time. DBSI FA was lower in POMS than ADEM ( P = 0.002). Higher DBSI-FF ( P = 0.046) and DBSI-RD ( P = 0.016) but lower DBSI-NRF ( P = 0.005) was noted in POMS. Kruskal-Wallis nonparametric testing was used to assess differences. Given our alpha = 0.05, we concluded that there was a statistically significant difference in the respective averages over time of the DBSI fiber-FA, DBSI-FF, DBSI-fiber radial, and DBSI-NRF measures.

When corrected for differences in age between the two groups, the previously described significant differences in DBSI measures between POMS and ADEM were lost and no significant difference was found between the two groups. Additionally, the comparison between baseline and over-time measures of DBSI and DTI did not reveal any significant difference within both ADEM and POMS groups.

Discussion

DTI is now one of the most common neuroimaging techniques to quantify the white matter microstructure. 34 - 36 DTI has been widely utilized to measure white matter damage in adult and pediatric patients with MS in recent years. 20 - 23 , 31 Studies reveal reduced FA in T2-hyperintense lesions, and in NAWM in pediatric MS cases. 37 A small retrospective pediatric study with a short follow-up revealed lower FA and AD with higher RD in NAWM-MS compared with control subjects. 21 Another longitudinal study reported low FA and high mean diffusivity in NAWM-MS in contrast to relatively stable FA and mean diffusivity in NAWM-ADEM. 38 In our small previous published preliminary study, 31 we found that DTI metrics at baseline differed by final diagnosis with low FA and high RD in patients with POMS compared with those with ADEM and age-matched control subjects but the results were not significant in adjusted comparisons. DTI has limitations. Previous studies have demonstrated the limited sensitivity and specificity of DTI metrics in the presence of coexisting inflammation in white matter tracts. 24 We have shown that DBSI reliably distinguishes and quantifies coexisting axonal injury/loss, demyelination, and inflammation in CNS white matter tracts. 25 , 26 , 39 , 40

Both ADEM and POMS are characterized by demyelination, axonal damage, edema, and cellular inflammation. Pathologic studies showed that axonal injury is present in ADEM, but the predominant pattern is perivascular/periventricular demyelination, and scattered cortical microglial activation with limited axonal loss even in very severe disease. 19 Conversely, axonal loss is prominent in the first year of MS onset. 18 A recent postmortem analysis of pediatric MS brain tissue revealed reduced axonal density and presence of axonal spheroids positive for amyloid precursor protein. Reduced axonal density was noted in early active demyelinating lesions in pediatric MS cases, which was found to be associated with the severity of the illness in pathology. 41

The main objectives of the study were to investigate differences in DBSI and DTI measures between MS and ADEM to identify potential pathologic distinctions between the two conditions. We investigated if DBSI was able to distinguish and quantify coexisting axonal injury, demyelination, and inflammation in CNS white matter tracts in children with POMS and relapsing ADEM. Our results showed that when comparing DBSI measures between ADEM and MS groups, there were significant differences in DBSI-FA, DBSI-FF, DBSI-RD, and DBSI-NRF over time. Specifically, there was higher DBSI- RD in POMS than ADEM ( P = 0.016) and lower DBSI-FA ( P = 0.002), indicative of worse demyelination in POMS. Higher DBSI-FF ( P = 0.046) but lower DBSI-NRF ( P = 0.005) in patients with POMS suggests less axon injury and edema in POMS. 42

When corrected for differences in age between the two groups, there are no significant differences in DBSI measures between POMS and ADEM, which could be due to changes in brain water during early childhood, which makes it difficult to correct for different ages in small sample size. 43 Accordingly, more extensive studies with similar ages in both groups are needed to be conducted to further investigate these findings. Although DBSI may not add much in differentiating typical ADEM and MS cases when lesions are characteristic of their respective conditions, it may prove useful in atypical cases, such as differentiating tumefactive ADEM from tumefactive MS in older children, where there is greater overlap in conventional MRI findings. 44 , 45

Acknowledgments

This research was supported by the Institute of Clinical and Translational Sciences at Washington University in St. Louis (UL1 TR002345).

Footnotes

Statistical analysis performed by: Jyoti Arora, Washington University School of Medicine, Saint Louis, MO.

CRediT authorship contribution statement

Esra Pehlivan: Writing – review & editing, Visualization, Formal analysis. Martin Kinuthia Mwangi: Writing – original draft, Funding acquisition, Data curation. Vihas Abraham: Writing – review & editing. Urmi Mange: Writing – review & editing, Data curation. Sheng-Kwei Song: Writing – review & editing, Software, Formal analysis. Peng Sun: Writing – review & editing, Formal analysis, Conceptualization. Soe Soe Mar: Writing – review & editing, Supervision, Methodology, Investigation, Funding acquisition, Conceptualization.

Declaration of competing interest

I have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

References