Neuroimaging spectrum of myelin oligodendrocyte glycoprotein antibody-associated disease with brain involvement: description of various cerebral syndromes

Myelin oligodendrocyte glycoprotein (MOG), a surface antigen located on oligodendrocyte membranes and the outermost layer of myelin sheaths, has been recognized over the past decade as a key target in immune-mediated inflammation of the central nervous system (CNS) [1]. MOG antibody-associated disease (MOGAD) has emerged as a leading cause of acquired demyelinating syndromes [2‐8], manifesting as acute disseminated encephalomyelitis (ADEM), neuromyelitis optica spectrum disorder (NMOSD) without anti-aquaporin-4 antibodies, optic neuritis, and transverse myelitis in both pediatric and adult populations [4, 5]. The clinical and radiological features of MOGAD are increasingly recognized as distinct from those of multiple sclerosis and aquaporin-4 antibody-positive demyelinating disease [5, 9‐12]. Early recognition of MOGAD is essential, as the condition typically responds well to corticosteroid therapy, whereas treatment delays may result in significant long-term neurological impairment.

Recently, the International MOGAD Panel introduced new diagnostic criteria for MOGAD [13]. These criteria emphasize the presence of MOG-IgG (immunoglobulin G) antibodies and magnetic resonance imaging (MRI) findings, including optic neuritis, transverse myelitis, and brain, brainstem, or cerebellar syndromes. According to the guideline, MRI plays a central role in interpreting MOG antibody test results. Furthermore, radiologic suspicion of MOGAD can influence the decision to initiate MOG-IgG testing, especially considering its variable availability and delayed turnaround time. Since universal screening for MOG-IgG is not recommended in all cases of CNS demyelination [13], it is crucial to accurately define and describe the imaging features associated with MOGAD.

Unlike optic neuritis and transverse myelitis, MOGAD involving the brain and brainstem exhibits considerable heterogeneity in both clinical presentation and MRI findings. Brain involvement in MOGAD has traditionally been associated with pediatric ADEM or “ADEM-like” syndromes. However, over the past decade, the recognized radiological spectrum of MOGAD has broadened, with emerging patterns that do not align with the classical definition of ADEM. Even the most recent diagnostic criteria do not fully capture this range of brain imaging presentations, raising concerns about potential underdiagnosis using current guidelines [14]. This issue is particularly relevant in children, as brain involvement is more frequently observed in the pediatric population compared to adults [15], and approximately 5–10% of pediatric MOGAD cases are reported to present with atypical features [16].

Furthermore, while previous studies have occasionally described radiologic findings in MOGAD with brain involvement—such as ADEM, cortical encephalitis, and other manifestations [4, 6, 9, 17, 18]—no prior research has systematically focused on MRI characteristics or conducted a comprehensive evaluation of neuroradiological features specific to pediatric MOGAD with brain involvement. Therefore, the objective of this study was to provide a detailed overview of the radiologic spectrum of pediatric MOGAD presenting with various cerebral syndromes, to enhance understanding of this disease subgroup.

A total of 55 patients were included in the study (25 boys and 30 girls; median age, 6 years; range, 15 months to 17 years) (Table 1). Initial clinical symptoms included seizures (n = 20), gait instability or ataxia (n = 34), psychiatric manifestations (n = 15), dysphasia (n = 19), movement disorders (n = 18), brainstem dysfunction (n = 8), focal weakness (n = 29), memory dysfunction (n = 2), and altered consciousness (n = 24). Fever exceeding 38 °C was present in 36 patients (65.5%), and headache was reported in 16 (29.0%). A detailed breakdown of symptoms by MRI subtype is provided in the supplementary table. The median interval between symptom onset and initial MRI was 2 days. CSF profiles at presentation were available for 52 patients. The median CSF protein concentration was 41.5 mg/dL (range, 19–122), with elevated levels noted in 26 patients (50%). The median CSF white blood cell count was 20 cells/mm³ (range, 0–556), with elevated counts in 41 patients (78.9%). Aquaporin-4 antibody testing was performed in 36 patients, all of whom were seronegative. N-Methyl-d-aspartate receptor antibody testing was conducted in 15 patients, with all results negative.

Radiologic phenotypes in the MOGAD cohort were classified as follows: ADEM in 39 patients, cortical encephalitis in three, aseptic meningitis in six, tumefactive demyelinating lesions in three, cerebellitis/brainstem encephalitis in two, other patterns in two patients (one leukodystrophy-like and one multiple sclerosis-like).

For patients with the ADEM phenotype, the mean widespread score was 2.7 (median 2, range 1–5). In seven cases, only multifocal supratentorial white matter lesions were observed, corresponding to a widespread score of 1. Spinal MRI was performed in 22 of the 39 patients (56%), and myelitis was identified in 10 of these 22 patients (45%). Except for one patient, nine cases showed longitudinally extensive lesions involving the central gray matter on axial images (Fig. 3). Conus involvement was identified in two out of the nine patients; in one patient, evaluation of the conus was not possible due to incomplete MRI coverage. A summary of the MRI distribution of affected regions in MOGAD patients with the ADEM phenotype is shown in Table 2. Typical ADEM patterns (Fig. 1) were observed in 15 of 39 patients (38%), whereas the remaining 24 patients (62%) exhibited at least one atypical feature for ADEM (Fig. 2). Among these 24 cases, 10 (42%) had one atypical feature, nine (38%) had two atypical features, and five (21%) presented with three atypical features. The most frequently observed atypical finding was the predominance of subcortical white matter lesions (17 of 39 patients, 44%), followed by the presence of small lesions (12 of 39 patients, 31%). Ten patients (26%) had both small and predominantly subcortical white matter lesions. These findings were compared with those reported by Baumann et al. [24], as summarized in Table 3.

Bild vergrößern

A cortical encephalitis pattern was identified in three patients, and an aseptic meningitis pattern was seen in six patients (Figs. 4 and 5, respectively). Tumefactive demyelinating lesions were observed in three cases. These lesions were large (> 2 cm) and unifocal, resembling either neoplastic processes or localized abscesses (Fig. 6). The lesions were located in the periventricular white matter, thalamus, and subcortical white matter/cortex, respectively. Unlike typical tumors, these lesions did not demonstrate mass effect. A cerebellitis/brainstem encephalitis pattern was found in two patients (Fig. 7). One patient exhibited non-enhancing lesions in the cerebellar white matter and midbrain, while the other had poorly defined cerebellar cortical lesions with curvilinear enhancement. Among the miscellaneous patterns, one patient showed a leukodystrophy-like presentation (Fig. 8a), and another exhibited an MS-like pattern (Fig. 8b–d).

Bild vergrößern

Bild vergrößern

Bild vergrößern

Bild vergrößern

Bild vergrößern

In children, MOGAD commonly presents as isolated or recurrent optic neuritis, isolated or combined myelitis, or ADEM. Since then, various atypical forms of brain involvement have been identified, and it is now recognized that, in contrast to optic neuritis or myelitis, brain involvement demonstrates a heterogeneous pattern that may include cortical encephalitis, aseptic meningitis, cerebellitis, and brainstem encephalitis. This study analyzed the radiologic features of anti-MOGAD presenting with cerebral syndromes involving the brain and brainstem in pediatric patients from a single center. While earlier studies have intermittently described imaging findings of MOGAD in presentations such as ADEM, cortical encephalitis, and others [4, 6, 17, 18, 25], we consider the present study to be valuable in offering a comprehensive review of neuroradiological features of MOGAD in children and in providing a broader radiologic overview of the disease.

In this study, ADEM was identified as the most frequent radiologic presentation of MOGAD with cerebral syndrome, accounting for 71% (39 out of 55) of the cases. Previous reports have also indicated that ADEM constitutes the largest subgroup (40–50%) in pediatric MOGAD and represents 70–80% of cases with cerebral involvement [4, 9, 26]. Characteristic MRI features of ADEM include diffuse, poorly defined, large (> 1–2 cm) lesions predominantly affecting the cerebral white matter [23]. Baumann et al. compared MOG-positive and MOG-negative ADEM and found that MOG-positive ADEM exhibited a more consistent pattern of typical ADEM, marked by large, bilateral, and widespread lesions. In contrast, MOG-positive ADEM in our cohort exhibited a higher frequency of atypical features, with small lesions being more prevalent (P = 0.005) compared to Baumann’s findings. Regarding lesion distribution, white matter involvement was primarily located in the subcortical regions in 44% of our MOG-ADEM cases. This pattern of preferential subcortical over periventricular or deep white matter involvement was also observed by Konuskan et al. in their MOG-ADEM cohort [27]. Consequently, in our study, approximately two-thirds of the cases (24 of 39, 62%) presented with atypical ADEM patterns, and about one-quarter (10 cases, 26%) displayed small, predominantly subcortical white matter lesions. Similarly, Dong et al. reported that at least one atypical ADEM feature was observed in 78.8% of their MOG-positive ADEM patients [28]. Therefore, we suggest that a large multicenter study is necessary to determine the true prevalence of typical versus atypical ADEM patterns in pediatric MOGAD. Nevertheless, a potential selection bias exists in our cohort. Patients with typical ADEM, who tend to recover quickly, may have been less likely to be referred to our institution, which is a national referral center. As a result, patients with more severe or atypical presentations were more likely to be included, possibly contributing to the relatively low proportion of “typical” ADEM in our study.

Although ADEM has traditionally been considered the predominant phenotype of MOGAD [4, 6], approximately 20% of pediatric MOGAD cases have been reported to present with a non-ADEM encephalitic phenotype [4]. In our study, 29% (16 out of 55) of MOGAD cases with cerebral syndromes did not meet the criteria for ADEM. Among these, aseptic meningitis was the second most common radiologic pattern following the ADEM phenotype, observed in six of the 55 patients (11%). This pattern is characterized on imaging by localized leptomeningeal enhancement or sulcal hyperintensity on FLAIR sequences, without associated brain parenchymal abnormalities. Consequently, such findings may be missed on pre-contrast MRI if post-contrast sequences are not performed. We suspect that some previously reported MOGAD cases with normal MRI findings may have actually represented aseptic meningitis [4, 9]. The relatively high frequency of aseptic meningitis in our cohort is concerning, as this subtype may not be identified under the 2023 MOGAD diagnostic criteria [14]. Based on our findings, we propose that aseptic meningitis should be recognized as a principal cerebral syndrome associated with MOGAD. Furthermore, our results support the routine inclusion of contrast-enhanced sequences in MRI protocol for the evaluation of suspected MOGAD.

Cerebral cortical encephalitis is another well-established non-ADEM phenotype of MOGAD [15]. Initially described in the seminal work by Ogawa et al., it is now recognized as one of the main non-ADEM cerebral syndromes [18]. Budhram et al. conducted a literature review of patients with serum MOG antibodies, encephalitis, and seizures and reported that MRI in these cases frequently showed unilateral cortical involvement. They introduced the acronym FLAMES (unilateral FLAIR-hyperintense lesions in anti-MOG-associated encephalitis with seizures) to describe this presentation [29]. Similarly, Wegener-Panzer et al. described 10 pediatric cases of encephalitis associated with MOG antibodies, in which MRI abnormalities were mainly confined to the cortex and deep gray matter, without cerebral white matter involvement [7]. This pattern was also identified in a small subset of our patients (three cases, 5%). Therefore, MOG-related encephalitis should be included in the differential diagnosis of cortical encephalitis. Recognition of this condition is essential, as it can easily be mistaken for viral encephalitis [16]. Another imaging pattern that may lead to diagnostic confusion is the tumefactive demyelination type, observed in three cases (5%). It typically presents as a solitary T2-hyperintense lesion resembling a tumor or abscess. Minimal or absent mass effect and evolution on follow-up MRIs may help distinguish it from neoplastic or infectious processes. Additionally, cerebellitis associated with MOGAD has been documented [30]. According to the 2023 MOGAD diagnostic criteria, such cases fall under the core clinical category of brainstem or cerebellar symptoms. In our study, one patient had lesions in both cerebellar white matter and the brainstem, while another presented with isolated cortical cerebellitis.

A leukodystrophy-like pattern, considered a rare presentation, was identified in one case in our study. This phenotype has been recently reported in pediatric MOGAD and appears to occur more frequently in younger children [21]. In line with this, our case involved a 3-year-old girl. Another distinct pattern observed in our study was a multiple sclerosis-like presentation, seen in just one patient. This was characterized by small, well-defined white matter lesions oriented perpendicular to the ventricles. MOG antibodies are rarely linked to multiple sclerosis in the pediatric population [4], which may explain why only one such case was identified in our cohort. In the study by Baumann M et al., multiple sclerosis-like patterns were absent in the MOG-positive ADEM group, while two cases (14%) were observed in the MOG-negative ADEM group [24]. The patient with the multiple sclerosis-like pattern in our study was lost to follow-up, so we could not determine whether this represented true multiple sclerosis. It is also noteworthy that no cases of limbic encephalitis were detected in our cohort. Although MOGAD can present as limbic encephalitis [31], bilateral medial temporal lobe involvement is considered atypical for MOGAD [25]. Furthermore, limbic encephalitis generally presents at an older age compared to other MOGAD subtypes such as ADEM and cortical encephalitis [31].

Finally, our study highlights a significant relationship between patient age and the radiological presentation of MOGAD with cerebral involvement. Younger children are more likely to exhibit the ADEM pattern, while older patients tend to present with atypical manifestations in children older than 7 years. Imaging remains critical for differentiating pediatric demyelinating diseases, which, besides MOGAD, include pediatric multiple sclerosis, aquaporin-4-antibody-positive NMOSD, and other less clearly defined categories. Radiologists must also consider a broad differential diagnosis when evaluating suspected MOGAD, including infectious encephalitis, CNS vasculitis, CNS hemophagocytic lymphohistiocytosis, brain tumors, and neurometabolic disorders [32, 33].

This study has several significant limitations. First, we did not include a comparison group of children who met the diagnostic criteria but were negative for MOG antibodies. However, the objective of this study was to characterize the diverse radiologic features of MOGAD, rather than to compare findings with MOG-negative cases. Second, follow-up imaging was not analyzed. In our cohort, some patients experienced relapses with different radiologic patterns. Since incorporating follow-up data would add complexity, we consider this a topic better suited for future research. Third, optic neuritis was not evaluated in the study. Although optic neuritis is a common manifestation of pediatric MOGAD and may co-occur with cerebral syndromes, it was not evaluated here due to the lack of dedicated imaging sequences for the optic nerves in most cases. Fourth, given the retrospective design of this study, certain clinical data, such as electroencephalography findings, were incomplete or unavailable. Fifth, the possibility of selection bias must be acknowledged. As a major national tertiary center, our institution is more likely to have received patients with severe symptoms, while patients with milder clinical and radiologic presentations may have been underrepresented. Therefore, we do not believe this study captures the full radiologic spectrum of pediatric MOGAD, and a large-scale multicenter study is warranted.

In summary, pediatric MOGAD involving the brain presents with a variety of imaging patterns. While the ADEM phenotype was most frequently observed, it commonly displayed atypical features, particularly small, subcortical white matter lesions. We also described imaging characteristics of less commonly reported phenotypes, including cerebral cortical encephalitis, aseptic meningitis, tumefactive demyelinating lesions, and cerebellitis. Together, these variants constitute a considerable portion of brain-involved MOGAD cases and occur more frequently in older children. Recognizing the broad imaging spectrum of MOG-associated brain disease is essential for achieving an accurate diagnosis.