THIEME

Original Article 533

AQP4-IgG NMOSD, MOGAD, and double-seronegative

NMOSD: is it possible to depict the antibody subtype
using magnetic resonance imaging?
NMOSD AQP4-IgG, MOGAD e NMOSD duplo
soronegativo: é possível caracterizar o tipo de anticorpo
pela ressonância magnética?
Diego Cardoso Fragoso1 Luana Michelli Oliveira de Paula Salles2
Samira Luisa Apóstolos Pereira2 Dagoberto Callegaro2 Douglas Kazutoshi Sato3
Carolina de Medeiros Rimkus1

1 Universidade de São Paulo, Faculdade de Medicina, Departamento Address for correspondence Diego Cardoso Fragoso
de Radiologia, São Paulo SP, Brazil. (email: [email protected]).
2 Universidade de São Paulo, Faculdade de Medicina, Departamento
de Neurologia, São Paulo SP, Brazil.
3 Pontifícia Universidade Católica do Rio Grande do Sul, Instituto do
Cérebro do Rio Grande do Sul (InsCer), Porto Alegre RS, Brazil.

Arq. Neuropsiquiatr. 2023;81:533–543.

Abstract Background There is clinical and radiological overlap among demyelinating diseases.
However, their pathophysiological mechanisms are different and carry distinct prog-
noses and treatment demands.
Objective To investigate magnetic resonance imaging (MRI) features of patients with
myelin-oligodendrocyte glycoprotein associated disease (MOGAD), antibody against
aquaporin-4(AQP-4)-immunoglobulin G-positive neuromyelitis optica spectrum disor-
der (AQP4-IgG NMOSD), and double-seronegative patients.
Methods A cross-sectional retrospective study was performed to analyze the topog-
raphy and morphology of central nervous system (CNS) lesions. Two neuroradiologists
consensually analyzed the brain, orbit, and spinal cord images.
Results In total, 68 patients were enrolled in the study (25 with AQP4-IgG-positive
NMOSD, 28 with MOGAD, and 15 double-seronegative patients). There were differ-
Keywords ences in clinical presentation among the groups. The MOGAD group had less brain
► Neuromyelitis Optica involvement (39.2%) than the NMOSD group (p ¼ 0.002), mostly in the
► Myelin- subcortical/juxtacortical, the midbrain, the middle cerebellar peduncle, and the
Oligodendrocyte cerebellum. Double-seronegative patients had more brain involvement (80%) with
Glycoprotein larger and tumefactive lesion morphology. In addition, double-seronegative patients
► Magnetic Resonance showed the longest optic neuritis (p ¼ 0.006), which was more prevalent in the
Imaging intracranial optic nerve compartment. AQP4-IgG-positive NMOSD optic neuritis had

received DOI https://doi.org/ © 2023. The Author(s).

November 2, 2022 10.1055/s-0043-1768669. This is an open access article published by Thieme under the terms of the
received in its final form ISSN 0004-282X. Creative Commons Attribution 4.0 International License, permitting copying
January 15, 2023 and reproduction so long as the original work is given appropriate credit
accepted (https://creativecommons.org/licenses/by/4.0/).
February 6, 2023 Thieme Revinter Publicações Ltda., Rua do Matoso 170, Rio de
Janeiro, RJ, CEP 20270-135, Brazil
534 MRI of demyelinating antibody diseases Fragoso et al.

a predominant optic-chiasm location, and brain lesions mainly affected hypothalamic

regions and the postrema area (MOGAD versus AQP4-IgG-positive NMOSD, p¼ 0 .013).
Furthermore, this group had more spinal cord lesions (78.3%), and bright spotty lesions
were a paramount finding to differentiate it from MOGAD (p ¼ 0.003).
Conclusion The pooled analysis of lesion topography, morphology, and signal
intensity provides critical information to help clinicians form a timely differential
diagnosis.

Resumo Antecedentes Há sobreposição clínica e radiológica entre as doenças desmielinizan-

tes. No entanto, seus mecanismos fisiopatológicos são diferentes e apresentam
prognósticos e demandas de tratamento distintos.
Objetivo Investigar as características de imagens de ressonância magnética (RM)
dos pacientes com doença associada à glicoproteína de oligodendrócito de
mielina (MOGAD), a doenças do espectro da neuromielite óptica positivas para
antiaquaporina-4 imunoglobulina G (AQP4-IgG NMOSD), e pacientes duplamente
soronegativos.
Métodos Estudo retrospectivo e transversal para analisar as características e fre-
quência das lesões do sistema nervoso central (SNC). Dois neurorradiologistas
avaliaram consensualmente as imagens do cérebro, das órbitas e da medula espinhal.
Resultados Ao todo, foram incluídos 68 pacientes (25 com AQP4-IgG NMOSD, 28
com MOGAD e 15 duplo-soronegativos). Há diferenças na apresentação clínica entre os
grupos. O grupo MOGAD demonstrou menor frequência de comprometimento do
cérebro (39.2%) comparado com o AQP4-IgG NMOSD (p ¼ 0.002), com predomínio da
distribuição das lesões nas regiões subcortical/justacortical, mesencéfalo, pedúnculos
cerebelares médios e cerebelo. O grupo duplo-soronegativo demonstrou maior
frequência de comprometimento do cérebro (80%), com lesões de maiores dimensões
e com morfologia tumefeita, além de neurite óptica com maior extensão (p ¼ 0.006). O
grupo AQP4-IgG NMOSD demonstrou neurite óptica com predomínio na região óptico-
quiasmática e as lesões encefálicas acometeram predominantemente as regiões
Palavras-chave hipotalâmica e área postrema (MOGAD versus AQP4-IgG NMOSD p ¼ 0.013). Além
► Neuromielite Óptica disso, foram observadas mais lesões na medula espinhal (78.3%) e a presença da “bright
► Glicoproteína spotty lesion” foi um achado primordial para a sua diferenciação com os pacientes
Mielina- MOGAD (p ¼ 0.003).
Oligodendrócito Conclusão A análise pormenorizada das características das lesões por RM dos
► Imageamento por pacientes com doenças desmielinizantes imunomediadas fornece informações
Ressonância fundamentais que auxiliam os médicos no diagnóstico diferencial em um momento
Magnética oportuno.

INTRODUCTION
ependymal surface and in the circumventricular organs2 and
There have been significant advances in recent decades in the blood—brain barrier.3
understanding demyelinating diseases, notably with the The NMOSD diagnostic criteria comprise two different
discovery of autoantibodies directed against antigens located groups regardless of AQP4-IgG positivity.4 According to Sato
in the central nervous system (CNS). In 2004, an antibody et al., 21.1% of AQP4-IgG-negative NMOSD patients are
against aquaporin-4 immunoglobulin G (AQP4-IgG) was positive for a different antibody, myelin oligodendrocyte
identified, culminating in the description of a new entity glycoprotein-IgG (MOG-IgG).5 Interestingly, although this
called AQP4-IgG-positive neuromyelitis optica spectrum protein was initially thought to be involved in the patho-
disorder (AQP4-IgG NMOSD).1 This antigen, related to water physiogenesis of multiple sclerosis (MS), its role in a new
homeostasis, has a peculiar distribution within the CNS, is entity known as myelin oligodendrocyte glycoprotein asso-
located in the astrocyte endfeet, and is abundant along the ciated-disease (MOGAD) is now recognized.6

Arquivos de Neuro-Psiquiatria Vol. 81 No. 6/2023 © 2023. The Author(s).

 MRI of demyelinating antibody diseases Fragoso et al. 535

The MOGAD spectrum is broad and varies according to protocol comprised fluid-attenuated inversion recovery
age, with acute disseminated encephalomyelitis being more (FLAIR) and axial T1- and T2-weighted imaging (WI) fast
common in children and recurrent optic neuritis (ON) and spin—echo (FSE) sequences. Sagittal 3D-FLAIR sequences
longitudinally extensive transverse myelitis (LETM) being were obtained in 61 out of 68 exams. The other 7 exams
more common in adults.7,8 had 3-mm-thick axial 2D-FLAIR sequences. The orbit proto-
There is clinical and radiological overlap between MS, col comprised axial and coronal T1WI and T2WI FSE and a
MOGAD, AQP4-IgG NMOSD, and antibody-negative coronal T2 short-tau inversion recovery (STIR) sequence. The
NMOSD.9–11 However, the different pathophysiological spinal cord protocol comprised axial and sagittal T1WI and
mechanisms of these diseases carry different prognoses T2WI FSE and a sagittal T2 STIR sequence. Finally, T1WI FSE
and treatment demands, which makes their precise distinc- postcontrast imaging of the brain, spinal cord, and orbit was
tion imperative.12 Overall, antibody-mediated demyelinat- acquired. Magnetic resonance imaging was acquired by the
ing diseases, especially MOGAD, have a low prevalence, even time of the patients’ first visit at our hospital, including 36
lower than that of MS.13 Unfortunately, the autoantibody test during an acute attack (< 30 days) and 32 in the chronic
for these diseases is costly and not universally available. phase.
Together, these factors contribute to antibody analysis scar- Two neuroradiologists with 6 (DCF) and 13 (CMR) years of
city in some places, especially in developing countries. experience consensually assessed all imaging data. The read-
Magnetic resonance imaging (MRI) might support the ers were blinded to the clinical diagnosis and serological
differential diagnosis between MOGAD and AQP4-IgG data. Readers followed a standard assessment to rate the
NMOSD, being especially important for the diagnosis of brain, optic nerve, and spinal cord lesions as per the Flow-
patients who fulfil the latest NMOSD criteria but are negative chart (►Supplementary Material). Afterward, a neuroradi-
for AQP4-IgG and MOG-IgG (to simplify terminology, this ologist (DCF) aligned all FLAIR images into a standard
disease will be referred to as double-seronegative).4,6 There template. Then, lesion masks were manually drawn for
are some overlapping and peculiar MRI characteristics patients with brain lesions using MRIcroGL. Finally, the
among these entities, but a consensus about typical imaging lesions of each group were overlaid, allowing the visualiza-
findings, mostly in the brain compartment, is still lacking. tion of their distribution (►Figure 1).
Here, we aimed to evaluate the brain, optic nerve, and spinal Categorical variables are presented as absolute and relative
cord imaging features of a cohort of AQP4-IgG NMOSD, frequencies and were compared using the Chi-squared (χ2) test.
MOGAD, and double-seronegative patients to identify possi- Continuous variables are presented as the mean and standard
ble discriminators among them. deviation (SD) and were normally distributed and compared by
analysis of variance (ANOVA). Post hoc tests by the Tukey method
were used to identify the source of significant differences when
METHODS
appropriate. A p-value < 0.05 was considered significant. Statis-
This is a cross-sectional retrospective study conducted be- tical analyses were performed using IBM SPSS Statistics for MAC,
tween 2009 and 2018. Initially, patients aged  18 years old version 26 (IBM Corp., Armonk, NY, USA).
at the time of the study with a presumed demyelinating
substrate who demonstrated some “red flags” that favored a
RESULTS
diagnosis other than MS were selected, as described by
Wingerchuk et al.4 According to the tested serology, these Twenty-five AQP4-IgG NMOSD, 28 MOGAD, and 15 double-
patients were divided into three groups: those who tested seronegative patients were included. Most patients were in
positive for MOG-IgG were included in the MOGAD group; an acute or chronic phase with exacerbation (< 30 days from
those who tested positive for AQP4-IgG and met the Wing- clinical attack), named acute disease status. There were no
erchuk criteria for NMOSD were included in the AQP4-IgG differences in either demographics or acute versus chronic
NMOSD group; and those who met the Wingerchuk criteria disease status among the groups (►Table 1).
but tested negative for MOG and AQP4 autoantibodies,
excluding an alternative diagnosis (clinical and/or laborato- Brain
ry), were included in the double-seronegative group. Demo- The present study classified patients into those who had no
graphic data, age at disease onset, and time between disease evidence of brain involvement and those who had brain
onset and MRI analysis were extracted from electronic lesions. Eleven out of 28 (39.2%) MOGAD patients, 18 out
medical records. The present study was approved by the of 25 (72%) AQP4-IgG NMOSD positive, and 12 out of 15 (80%)
local ethics committee (CAPPESQ46571015500000068), and double-seronegative patients had supratentorial lesions.
all patients provided written informed consent. Compared with MOGAD and AQP4-IgG NMOSD patients,
AQP4-IgG and MOG-IgG were detected using a cell-based double-seronegative patients had more supratentorial
assay (CBA) in live HEK-293 cells, as previously reported.5,14 lesions (p ¼ 0.002). Multiple comparisons showed statisti-
The samples were collected during regular visits in our cally significant differences between MOGAD and AQP4-IgG
outpatient clinic and immediately centrifuged and stored NMOSD patients (p ¼ 0.004) and between MOGAD and dou-
at - 80°C. ble-seronegative patients (p ¼ 0.016).
According to service availability, MRI was acquired using There were no differences among MOGAD, AQP4-IgG
1.5 and 3.0 T scanners with a different head coil. The brain NMOSD, and double-seronegative patients related to overall

Arquivos de Neuro-Psiquiatria Vol. 81 No. 6/2023 © 2023. The Author(s).

536 MRI of demyelinating antibody diseases Fragoso et al.

Figure 1 Lesion map depicts the location, morphology, and distribution of AQP4-IgG-positive-NMOSD, double seronegative, and MOGAD
groups, including all patients that had brain lesions.

Table 1 Demographic and disease-related characteristics

Variable Demyelinating disease Test

MOGAD (28) AQP4-IgG NMOSD Double seronegative
(25) (15)
Sex Male 11 11 5 Chi-squared
0.557
Female 17 14 10
Age (mean  SD) 32.3  2.1 38.4  4.1 28.2  3.0 ANOVA
0.218
Disease status Active 18 14 7 Chi-squared
0.846
Inactive 10 11 8

Active attack to MRI 7.8  2.6 17.2  2.3 19.1  3.3 ANOVA
(mean  SD) 0.01
Time elapse from last attack 9.5  2.4 46  18.3 16.3  11.2 ANOVA
to MRIτ (mean  SD) 0.172

Abbreviations: ANOVA, analysis of variance; AQP4-IgG NMOSD, antibody against aquaporin-4-immunoglobulin G-positive neuromyelitis optica
spectrum disorder; MOGAD, myelin-oligodendrocyte glycoprotein associated disease; MRI, magnetic resonance imaging; SD, standard deviation.
Notes: days; τmonths.

subependymal lesions when we considered patients with or AQP4-IgG NMOSD patients had this lesion subtype
brain lesions. Nonetheless, the subependymal “extensive” (p ¼ 0.018). There were significant differences between
subtype showed significant differences among the groups AQP4-IgG NMOSD and double-seronegative patients
(p ¼ 0.010); it was present in 9.1% of MOGAD patients, 16.7% (p ¼ 0.024) and between double-seronegative and MOGAD
of AQP4-IgG NMOSD patients, and 58.3% of double-seroneg- patients (p ¼ 0.048).
ative patients. There were significant differences between There were no differences in “focal” or “extensive” corti-
MOGAD and double-seronegative patients (p ¼ 0.017) and cospinal tract lesions among the groups. Notwithstanding,
between AQP4-IgG NMOSD and double-seronegative regardless of corticospinal lesion subtype, 50% of double-
patients (p ¼ 0.025). seronegative patients, 18.2% of MOGAD patients, and 11.1%
There were no basal ganglia lesions in MOGAD or AQP4- of AQP4-IgG NMOSD patients presented lesions (p ¼ 0.004),
IgG NMOSD patients. Although two double-seronegative with significant differences between MOGAD and double-
patients (16.7%) demonstrated basal ganglia lesions, there seronegative patients (p ¼ 0.005) and between AQP4-IgG
were no differences among the groups (p ¼ 0.081). NMOSD and double-seronegative patients (p ¼ 0.002).
Corpus callosum lesions were not different among the The overall infratentorial compartment lesions did not
groups. However, 25% of double-seronegative patients had reach any differences among the groups. Twelve of 18 AQP4-
the “extensive” corpus callosum lesion subtype. No MOGAD IgG NMOSD patients (67%), 1 of 11 MOGAD patients (9.1%),

Arquivos de Neuro-Psiquiatria Vol. 81 No. 6/2023 © 2023. The Author(s).

 MRI of demyelinating antibody diseases Fragoso et al. 537

and 4 of 12 double-seronegative patients (33.3%) had lesions The other spinal cord variables showed no differences and
in the area postrema (p ¼ 0.016), with a significant difference are summarized in ►Table 4.
between MOGAD and AQP4-IgG NMOSD patients (p ¼ 0.013).
The lesion morphology analysis showed significant differ- Optic nerve
ences in the “tumefactive” subtype (p ¼ 0.018). Three of 12 Thirteen patients did not undergo orbital MRI (5 MOGAD, 5
double-seronegative patients (25%) had “tumefactive” AQP4-IgG NMOSD, and 3 double-seronegative).
lesions, and no “tumefactive” lesions were observed in any The presence of ON did not reach significant differences
patients in the other groups, with a significant difference regardless of laterality among the groups. The MOGAD group
between MOGAD and double-seronegative patients showed shorter ON and fewer affected segments (p ¼ 0.006),
(p ¼ 0.048) and between AQP4-IgG NMOSD and double- with a significant difference between MOGAD and AQP4-IgG
seronegative patients (p ¼ 0.024). All other variables are NMOSD patients (p ¼ 0.010) and between MOGAD and dou-
included in ►Table 2. ble seronegative patients (p ¼ 0.040).
Brain MRI analysis of a subgroup of patients with active or Double-seronegative and AQP4-IgG NMOSD patients had
inactive disease depicted additional information (►Table 3). more lesions in the optic-chiasm and optic-tract segments
In active disease, double-seronegative patients presented than MOGAD patients, with a tendency toward significant
more tumefactive lesions than the other groups, more differences in the optic-chiasm (p ¼ 0.053) and optic-tract
deep gray matter lesions than MOGAD patients, and more (p ¼ 0.063) segments among these groups. Nonetheless,
corticospinal cord lesions than AQP4-IgG NMOSD patients. In multiple comparisons analyses depicted a significant differ-
inactive disease, double-seronegative patients had more ence in the optic-chiasm segment between MOGAD and
infiltrative lesions than MOGAD patients. AQP4-IgG NMOSD patients (p ¼ 0.048). The other optic nerve
Lesion distribution maps allowed the visualization of variables showed no differences and are summarized
specific exciting characteristics. MOGAD and double-sero- in ►Table 5.
negative patients showed more subcortical and juxtacortical
lesions than AQP4-IgG NMOSD patients. Double-seronega-
DISCUSSION
tive patients showed larger lesions in the cerebral hemi-
spheres than patients in the other two groups, which was Current knowledge of autoantibody-mediated demyelinat-
statistically supported by the presence of more lesions with ing diseases has allowed the demonstration of the presumed
tumefactive morphology. Similarly, more lesions were ob- initial pathophysiologic mechanism of each known nosolog-
served in the basal ganglia and along the corticospinal tract ical entity. Antibodies against AQP4, predominantly located
in the double-seronegative group than in the other groups. in astrocyte endfeet, disrupt local homeostasis, leading to
AQP4-IgG NMOSD patients showed more linear lesions along astrocyte destruction and ultimately promoting secondary
the supratentorial ependymal surface, with further evidence demyelination, which are critical findings of NMOSD.12,15 In
of optic-chiasmatic and hypothalamic lesions. More lesions contrast, MOGAD is a primarily inflammatory demyelinating
in the midbrain, middle cerebellar peduncles, and cerebellar disease, as their antibodies target a protein expressed on the
hemispheres were observed in the MOGAD group than in the outer myelin surface in the CNS.15 Although pathophysio-
other groups. At the same time, medulla oblongata im- logical mechanisms are different, patients from both groups
pairment, particularly in the area postrema, was more can have overlapping imaging findings. The possibility of an
manifest in AQP4-IgG NMOSD patients. undiscovered secondary factor common to both entities
could contribute to understanding this intersection.16
Spinal cord We investigated the MRI hallmarks of MOGAD, AQP4-IgG
Four patients did not undergo spinal MRI (1 MOGAD, 2 AQP4- NMOSD, and double-seronegative patients. In line with
IgG NMOSD, and 1 double-seronegative). previous reports,13,17 MOGAD patients had a higher predi-
Ten out of 27 MOGAD patients (37%), 18 out of 23 AQP4- lection for ON (68.2%), followed by spinal cord (38.5%) and
IgG NMOSD patients (78.3%), and 7 out of 14 double-sero- brain (37%) lesions, while the leading clinical phenotype of
negative patients (50%) demonstrated spinal cord lesions AQP4-IgG NMOSD patients was spinal cord (75%), followed
(p ¼ 0.012), with a significant difference between MOGAD by brain (73.1%) and ON (71.4%). Regarding the double-
and AQP4-IgG NMOSD patients (p ¼ 0.009). seronegative clinical phenotypes, we noted a different pre-
The double-seronegative group showed more affected sentation: ON was the leading phenotype (78.6%), followed
segments per patient than the other groups (p ¼ 0.049), by brain (76.5%) and spinal cord (50%) lesions.
with a significant difference between MOGAD and double- Studies have shown that patients with autoantibody-medi-
seronegative patients (p ¼ 0.044). However, LETM was more ated demyelinating diseases have more brainstem lesions than
prevalent in AQP4-IgG NMOSD patients (94.4%) than in MS patients.16,18,19 We did not include MS patients in our
MOGAD (60%) and double-seronegative (85.7%) patients, cohort, but brainstem lesions were a common and universal
with a tendency toward a difference between MOGAD and finding in the analyzed CNS demyelinating diseases.
AQP4-IgG NMOSD patients (p ¼ 0.055). Topographical lesion analysis demonstrates essential dif-
Bright spotty lesions were the best discriminator between ferences. Lesions in the area postrema are a characteristic
MOGAD and AQP4-IgG NMOSD patients, with a significant finding in the differentiation between AQP4-IgG NMOSD and
difference by multiple comparisons (p ¼ 0.003). MOGAD patients, which is consistent with other

Arquivos de Neuro-Psiquiatria Vol. 81 No. 6/2023 © 2023. The Author(s).

538 MRI of demyelinating antibody diseases Fragoso et al.

Table 2 Brain MRI features

Variable Demyelinating disease p-value Multiple comparison p-value

Affected variable No./total No. (%)
MOGAD (28) AQP4-IgG Double
NMOSD seronegative
(25) (15)
Topography Supratentorial Supratentorial 11/28 (39.2) 18/25 (72) 12/15 (80) 0.002 M vs. NMOþ 0.004
lesions
M vs. DN 0.016
Juxtacortical 4/11 (36.4) 3/18 (16.7) 3/12 (25) 0.507
Subcortical 6/11 (54.5) 5/18 (27.8) 6/12 (50) 0.299
Subependymal† 4/11 (36.4) 7/18 (38.8) 9/12 (75) 0.256
1. Extensive 1/11 (9.1) 3/18 (16.7) 7/12 (58.3) 0.010 M vs. DN 0.017
NMOþ vs. DN 0.025
2. Dawson finger 2/11 (18.2) 1/18 (5.5) 2/12 (16.7) 0.534
3. Nonperpendicular 1/11 (9.1) 3/18 (16.7) 2/12 (25) 0.842
focal lesion
Third ventricle 1/11 (9.1) 3/18 (16.7) 3/12 (25) 0.617
Temporal lobe 2/11 (18.2) 1/18 (5.5) 3/12 (25) 0.329
Diencephalic 1/11 (9.1) 3/18 (16.7) 3/12 (25) 0.617
Thalamus 2/11 (18.2) 1/18 (5.5) 4/12 (33.3) 0.147
Basal ganglia 0 0 2/12 (16.7) 0.081
á
Deep gray matter 2/11 (18.2) 1/18 (5.5) 4/12 (33.3) 0.147
Corpus callosum‡ 3/11 (27.3) 2/18 (11.1) 6/12 (50) 0.082
1. Extensive 0 0 3/12 (25) 0.018 M vs. DN 0.048
NMOþ vs. DN 0.024
2. Oval 2/11 (18.2) 0 1/12 (8.3) 0.198
3. Round 1/11 (9.1) 2/18 (11.1) 1/12 (8.3) 0.968
4. Band 0 0 1/12 (8.3) 0.306
Corticospinal tract 2/11 (18.2) 2/18 (11.1) 6/12 (50) 0.004 NMOþ vs. DN 0.003
1. Focal 1/11 (9.1) 1/18 (5.5) 4/12 (33.3) 0.093
2. Extensive 1/11 (9.1) 1/18 (5.5) 2/12 (16.7) 0.620
Infratentorial Infratentorial 8/28 (28.6) 13/25 (52) 7/15 (46.6) 0.094
Area postrema 1/11 (9.1) 12/18 (67) 4/12 (33.3) 0.016 M vs. NMOþ 0.013
Brainstem 8/11 (72.7) 10/18 (55) 7/12 (58.3) 0.657
Periaqueductal 1/11 (9.1) 1/18 (5.5) 2/12 (16.7) 0.620
Fourth ventricle 5/11 (45.4) 3/18 (16.7) 2/12 (16.7) 0.173
Cerebellum 3/11 (27.3) 1/18 (5.5) 2/12 (16.7) 0.283
Cerebellar peduncles 3/11 (27.3) 2/18 (11.1) 1/12 (8.3) 0.393
Morphology Oval well-circumscribed 2/11 (18.2) 1/18 (5.5) 2/12 (16.7) 0.534
Infiltrative 5/11 (45.4) 13/18 (72) 10/12 (83.3) 0.140
Non-specific 4/11 (36.4) 7/18 (39) 4/12 (33.3) 0.956
Cavitated 2/11 (18.2) 1/18 (5.5) 4/12 (33.3) 0.147
Tumefactive§ 0 0 3/12 (25) 0.018 M vs. DN 0.048
NMOþ vs. DN 0.024
Round well-circumscribed 1/11 (9.1) 0 1/12 (8.3) 0.458

Abbreviations: M, MOGAD; DN, double-seronegative; ANOVA, analysis of variance; AQP4-IgG NMOSD, antibody against aquaporin-4-immunoglobulin
G-positive neuromyelitis optica spectrum disorder; MOGAD, myelin-oligodendrocyte glycoprotein associated disease; MRI, magnetic resonance imaging;
NMO, neuromyelitis optica; SD, standard deviation; vs., versus.
Notes: Data presented in % (compared to total cases with brain lesion). The MS-like lesion was defined as an area of focal hyperintensity on a FLAIR or
T2-weighted sequence, round to ovoid, and should be at least 3 mm in their long axis.12 †Extensive lesions (defined as > 2 cm, not perpendicular);
perpendicular round focal lesions (Dawson’s fingers), and nonperpendicular focal lesions; ‡Extensive, Focal oval, round, or band; §Defined as
lesions > 2 cm; áThalamus þ Basal Ganglia. Statistically significant differences are in bold.

Arquivos de Neuro-Psiquiatria Vol. 81 No. 6/2023 © 2023. The Author(s).

 MRI of demyelinating antibody diseases Fragoso et al. 539

Table 3 Subgroup analysis (active x inactive disease)

Disease status Variable p-value Multiple comparison p-value

Active Tumefactive lesion 0.043 DN > M 0.076
DN > NMOþ 0.053
Area postrema 0.024 NMOþ > M 0.019
Subependymal 0.046 DN > M 0.037
Deep gray matter 0.046 DN > M 0.050
Corticospinal tract 0.035 DN > NMOþ 0.029
Inactive Subependymal 0.049 NMOþ > M 0.056
Subependymal extensive 0.000 DN > M 0.000
DN > NMOþ 0.000
Infiltrative lesion 0.042 DN > M 0.035

Abbreviation: M, MOGAD; DN, double-seronegative; NMO, neuromyelitis optica.

Notes: Disease status refers to less (active) or > 30 days (inactive) from clinical attack. Tendency to statically significant differences.

Table 4 Spinal MRI features

Variable Demyelinating Disease p-value Multiple p-value

N Affected variable n / total n (%) comparison
MOGAD (27) AQP4-IgG NMOSD Double
(23) seronegative
(14)
Spinal lesion 10/27 (37) 18/23 (78.3) 7/14 (50) 0.012 M vs. NMOþ 0.009
†
Number of segments 6 [  5.3] 7.6 [  4.7] 12.3 [  5.5] 0.049 M vs. DN 0.044
LETM 6/10 (60) 17/18 (94.4) 6/7 (85.7) 0.068 M vs. NMOþ 0.055
Column segment Cervical 6/10 (60) 17/18 (94.4) 5/7 (71.4) 0.124
Thoracic 7/10 (70) 11/18 (61.1) 7/7 (100) 0.840
Lumbar 4/10 (40) 1/18 (5.5) 3/7 (42.8) 0.857
Enhancement 5/10 (50) 7/18 (38.9) 4/7 (57.1) 0.791
Bright spotty 0 11/18 (61.1) 3/7 (42.8) 0.004 M vs. NMOþ 0.003
H medullar sign 4/10 (40) 1/18 (5.5) 3/7 (42.8) 0.621
Longitudinal line 4/10 (40) 2/18 (11.1) 4/7 (57.1) 0.684

Abbreviations: M, MOGAD; DN, double-seronegative; AQP4-IgG NMOSD, antibody against aquaporin-4-immunoglobulin G-positive neuromyelitis
optica spectrum disorder; LETM, longitudinally extensive transverse myelitis; MRI, magnetic resonance imaging; NMO, neuromyelitis optica; SD,
standard deviation; vs., versus.
Note: †Mean deviation within brackets; Tendency to statically significant differences. Statistically significant differences are in bold.

articles,18,20,21 supporting their inclusion in the Cacciaguera NMOSD.21 In the future, multicenter studies are required to
criteria for the differentiation of these diseases.19 confirm this hypothesis.
Compared with AQP4-IgG NMOSD patients, MOGAD The MOGAD patients had fewer supratentorial lesions
patients had fewer brainstem lesions,18 similar to our results. than AQP4-IgG NMOSD and double-seronegative patients,
Analysis of the lesion distribution showed that MOGAD regardless of disease activity. In our cohort, lesions in
patients had more lesions in the midbrain, middle cerebellar MOGAD patients were predominantly peripherally (subcor-
peduncles, and cerebellar hemispheres than their counter- tical and juxtacortical) located, whereas those in AQP4-IgG
parts. Our study includes a small MOGAD sample, which NMOSD patients were more centrally located, mostly in the
might be a limitation when generalizing this finding. Not- hypothalamic-chiasmatic region. Several studies found sig-
withstanding, MOGAD is a rare disease, being less prevalent nificant differences in this affected topographical region
than MS and AQP4-IgG NMOSD, and small samples are also a between MOGAD and AQP4-IgG NMOSD patients, with the
limitation to several other previous MOGAD studies. Never- former showing more lesions in the subcortical/juxtacortical
theless, the characteristics of middle cerebellar peduncle location.19–21 Double-seronegative patients demonstrated a
lesions deserve special attention, as they have been consid- distinct lesion distribution pattern, with lesions situated
ered a discriminating factor between MOGAD and AQP4-IgG both peripherally and centrally.

Arquivos de Neuro-Psiquiatria Vol. 81 No. 6/2023 © 2023. The Author(s).

540 MRI of demyelinating antibody diseases Fragoso et al.

Table 5 Orbital MRI features

Variable Demyelinating disease p-value Multiple p-value

N Affected variable n / total n (%) comparison
MOGAD AQP4-IgG Double
(23) NMOSD seronegative
(20) (12)
Neuritis 16/23 (69.5) 14/20 (70) 10/12 (83.3) 0.660
Unilateral 10/16 (62.5) 7/14 (50) 6/10 (60) 0.087
Bilateral 6/16 (37.5) 7/14 (50) 4/10 (40)
Number of segments 1.7 [  1.4] 3.7 [  2.8] 3.6 [  2.3] 0.006 M vs. NMOþ 0.010
M vs. DN 0.040
Segment Intraorbital 14/16 (87.5) 12/14 (85.7) 9/10 (90) 0.134
Canalicular/prechiasmic 13/16 (81.2) 13/14 (92.8) 10/10 (100) 0.093
Optic-chiasm 2/16 (12.5) 8/14 (57.1) 6/10 (60) 0.053 M vs. NMOþ 0.048
Optic-tract 1/16 (6.2) 7/14 (50) 3/10 (30) 0.063 M vs. NMOþ 0.052
Extension < 50% 10/16 (62.5) 5/14 (35.7) 4/10 (40) 0.472
> 50% 6/16 (37.5) 9/14 (64.3) 7610 (60) 0.720
Perineuritis 4/16 (25) 1/17 (5.9) 2/10 (20) 0.789

Abbreviations: M, MOGAD; DN, double-seronegative; AQP4-IgG NMOSD, antibody against aquaporin-4-immunoglobulin G-positive neuromyelitis
optica spectrum disorder; MRI, magnetic resonance imaging; NMO, neuromyelitis optica; vs., versus.
Notes: Tendency to statically significant differences. Statistically significant differences are in bold.

Subependymal lesions might offer additional substrate hallmark of NMOSD patients.2 However, we demonstrated
information to differentiate these diseases. A recent study that corticospinal lesions also occurred in MOGAD patients
demonstrated that AQP4-IgG NMOSD patients presented more frequently than in AQP4-IgG NMOSD patients. In a
more lesions adjacent to the lateral ventricles than MOGAD study comparing MOGAD and AQP4-IgG NMOSD patients,
patients.16 Furthermore, lesions in the corpus callosum Chen et al. showed that MOGAD patients had approximately
occurred in all groups. However, double-seronegative 2.5 times more lesions in the internal capsule than AQP4-IgG
patients showed more “extensive” subependymal lesions NMOSD patients.21
than MOGAD patients and more “extensive” corpus callosum Compared with MOGAD patients, AQP4-IgG NMOSD
lesions than MOGAD and AQP4-IgG NMOSD patients. These patients had more spinal cord lesions, with the cervical
findings are in line with the morphological lesion analysis segment being most frequently involved, similar to other
and the interpretation of the lesion distribution map, as studies.7,12 Although conus medullaris involvement has
double-seronegative patients had more aggressive charac- been reported as a hallmark of MOGAD patients,23 our cohort
teristics from all groups, with more extensive lesions and a of Brazilian patients did not show differences in this topog-
tumefactive pattern. raphy among the groups, similar to a Chinese study.21
An explanation for the presence of a lower number of Double-seronegative patients exhibited LETM with a
brain lesions in MOGAD patients could be the temporal higher number of affected segments than MOGAD patients.
evolution of their lesions. A recent study showed that brain This finding must be analyzed with caution, given that the
lesions in MOGAD patients tend to resolve more completely number of affected segments changes throughout the course
than those in AQP4-IgG NMOSD patients.22 This finding of MOGAD and NMOSD,24 with MOGAD lesions tending to
should be accounted for in our analysis, as some patients resolve in the chronic phase.22
underwent MRI in the chronic phase of the disease. This Bright spotty lesions were a characteristic finding of
information has additional clinical relevance. In a hypothet- NMOSD patients, especially AQP4-IgG patients. In cases
ical patient with a brain lesion suggestive of a non-MS-type with nonspecific brain lesions or where the analysis of
demyelinating lesion, a sequential MR exam could help to autoantibodies is difficult or unavailable, the detection of a
differentiate antibody-related diseases. bright spotty lesion might have paramount relevance.
Lesions in the basal ganglia and along the corticospinal The optic nerve is the primary target of autoantibody-
tract could also reflect the more severe neurological im- mediated demyelinating diseases.10,18,20 Similar to cerebral
pairment observed in double-seronegative patients. These and spinal cord involvement, double-seronegative patients
findings are divergent in the literature and should be care- have also demonstrated more extensive lesions of the optic
fully analyzed due to the number of patients included.16,20 pathways, with more segments affected and predominantly
Corticospinal tract involvement has been described as a in the intracranial segments. Compared with MOGAD

Arquivos de Neuro-Psiquiatria Vol. 81 No. 6/2023 © 2023. The Author(s).

 MRI of demyelinating antibody diseases Fragoso et al. 541

patients, AQP4-IgG NMOSD patients also had more frequent NMOSD may be partially explained by differential antigen
lesions in the optic tracts and chiasm, findings already expression in specific areas.2,12,18 Indeed, AQP4 lesions are
supported in the literature.12 In our cohort, perineural frequently encountered in areas with high AQP4 expression
enhancement did not allow differentiation among the dis- (ON, periependymal, periaqueductal, hypothalamic, and spi-
eases, contrary to what was previously reported.24 nal cord).20 Data on MOG expression within the CNS are still
►Figure 2 illustrates the CNS lesion characteristics of our lacking, but the peripheral location near the cortex has been
cohort that are typical and those that might be depicted observed more frequently than their counterparts.19 Double-
("overlap") in any of the antibody-mediated studied diseases. seronegative patients are an undefined entity. One could
Considering all these aspects, we wondered whether MRI argue the possibility that double seronegativity could repre-
topographical and morphological differences among sent a MOGAD or AQP4-IgG NMOSD disease with undetect-
MOGAD, AQP4-IgG NMOSD, and double-seronegative able or negative antibodies due to limited test sensitivities.

Figure 2 Radiological guide ;- characteristics of CNS lesions that are typical ( (hint)) and those that might be depicted (‘overlap’) in any of the
antibody-mediated disease. (a1a) Extensive corpus callosum and (a1b) tumefactive lesions might be indicative of Double-seronegative. (a2)
Middle cerebellar peduncle lesions, suggestive of MOGAD. (a3) Area postrema and (b1) bright spotty lesion, suggestive of AQP4-IgG-
positiveNMOSD. (c1 and c2) Long ON, notably in the intracranial segments, might suggest Double-seronegative and AQP4-IgG-positive-NMOSD.
Arrows in c2d and c2e demonstrated a cavitated left chiasm and tract. Corpus callosum lesion could be found in Double-seronegative (d1),
MOGAD (d2), and AQP4-IgG-positive-NMOSD (d3). Conus medullary lesion could be found in Double-seronegative (e1), MOGAD (e2), and
AQP4-IgGpositive-NMOSD (e3). Perineuritis could be found in Double-seronegative (f1), MOGAD (f2), and AQP4-IgGpositive-NMOSD (f3).

Arquivos de Neuro-Psiquiatria Vol. 81 No. 6/2023 © 2023. The Author(s).

542 MRI of demyelinating antibody diseases Fragoso et al.

However, the different clinical presentation compared with 3 Papadopoulos MC, Verkman AS. Aquaporin water channels in the
the other groups may be partially explained by a different nervous system. Nat Rev Neurosci 2013;14(04):265–277
4 Wingerchuk DM, Banwell B, Bennett JL, et al; International Panel
inflammatory cascade or undiscovered autoantibodies.5 A
for NMO Diagnosis. International consensus diagnostic criteria for
recent international randomized, double-blinded, placebo-
neuromyelitis optica spectrum disorders. Neurology 2015;85
controlled study (phase II and III) demonstrated different (02):177–189
therapeutic responses among MOGAD, AQP4-IgG NMOSD, 5 Sato DK, Callegaro D, Lana-Peixoto MA, et al. Distinction between
and double-seronegative patients using specific immunopa- MOG antibody-positive and AQP4 antibody-positive NMO spec-
thogenic targets, supporting a distinct mechanism among trum disorders. Neurology 2014;82(06):474–481
6 Jarius S, Paul F, Aktas O, et al. MOG encephalomyelitis: interna-
them.25 Double-seronegative patients showed more exten-
tional recommendations on diagnosis and antibody testing. J
sive brain, ON, and spinal cord lesions.
Neuroinflammation 2018;15(01):134
Prospective and histopathological studies will be needed 7 Carandini T, Sacchi L, Bovis F, et al. Distinct patterns of MRI lesions
to investigate whether this disease group’s distinct substrate in MOG antibody disease and AQP4 NMOSD: a systematic review
promotes more extensive damage to the CNS, with an and meta-analysis. Mult Scler Relat Disord 2021;54:103118
intersection pattern between MOGAD and AQP4-IgG 8 Fadda G, Armangue T, Hacohen Y, Chitnis T, Banwell B. Paediatric
multiple sclerosis and antibody-associated demyelination: clini-
NMOSD.
cal, imaging, and biological considerations for diagnosis and care.
Our study has some limitations. First, it is a cross- Lancet Neurol 2021;20(02):136–149
sectional retrospective study with a small sample, which 9 Spadaro M, Gerdes LA, Krumbholz M, et al. Autoantibodies to
may not reflect the precise lesion distribution and mor- MOG in a distinct subgroup of adult multiple sclerosis. Neurol
phology. These diseases are not prevalent, with the same Neuroimmunol Neuroinflamm 2016;3(05):e257
limitation occurring in most studies. Thus, prospective and 10 Jarius S, Ruprecht K, Kleiter I, et al; in cooperation with the
Neuromyelitis Optica Study Group (NEMOS) MOG-IgG in NMO
multicenter studies are encouraged to reduce potential
and related disorders: a multicenter study of 50 patients. Part 2:
selection biases and to respond with greater statistical Epidemiology, clinical presentation, radiological and laboratory
power to these questions. Second, patients were at differ- features, treatment responses, and long-term outcome. J Neuro-
ent stages of the disease. Some MOGAD patients might inflammation 2016;13(01):280
have demonstrated resolution of CNS lesions more often 11 Schmidt FA, Chien C, Kuchling J, et al. Differences in Advanced
Magnetic Resonance Imaging in MOG-IgG and AQP4-IgG Sero-
than AQP4-IgG NMOSD patients. Therefore, follow-up MRI
positive Neuromyelitis Optica Spectrum Disorders: A Compara-
exams, including moments of acuteness and intercrisis,
tive Study. Front Neurol 2020;11:499910
may provide an additional underpinning to reinforce our 12 Lana-Peixoto MA, Talim N. Neuromyelitis Optica Spectrum Dis-
findings. This is one of the most critical hints that could order and Anti-MOG Syndromes. Biomedicines 2019;7(02):42
help neurologists form the correct diagnosis in clinical 13 Dos Passos GR, Oliveira LM, da Costa BK, et al. MOG-IgG-Associ-
practice. ated Optic Neuritis, Encephalitis, and Myelitis: Lessons Learned
From Neuromyelitis Optica Spectrum Disorder. Front Neurol
Our study provides further evidence that MOGAD,
2018;9:217
AQP4-IgG NMOSD, and double-seronegative patients 14 Oliveira LM, Apóstolos-Pereira SL, Pitombeira MS, Bruel Torretta
have distinct MRI features. The pooled analysis of lesion PH, Callegaro D, Sato DK. Persistent MOG-IgG positivity is a
topography, morphology, and signal intensity provides predictor of recurrence in MOG-IgG-associated optic neuritis,
essential information to help clinicians make an early encephalitis and myelitis. Mult Scler 2019;25(14):1907–1914
differential diagnosis. 15 Mader S, Kümpfel T, Meinl E. Novel insights into pathophysiology
and therapeutic possibilities reveal further differences between
AQP4-IgG- and MOG-IgG-associated diseases. Curr Opin Neurol
Authors’ Contributions
2020;33(03):362–371
LMOPS, DCF, CMR: substantial contributions to the design 16 Jurynczyk M, Geraldes R, Probert F, et al. Distinct brain imaging
and development of the study; DCF, LMOPS, SLQP: sub- characteristics of autoantibody-mediated CNS conditions and
stantial contributions in the collection, analysis, and multiple sclerosis. Brain 2017;140(03):617–627
interpretation of data; DCF, DC, DKS, CMR: substantial 17 Marignier R, Hacohen Y, Cobo-Calvo A, et al. Myelin-oligodendro-
cyte glycoprotein antibody-associated disease. Lancet Neurol
contributions in the writing of the article, and in its
2021;20(09):762–772
critical revision; All authors: substantial contributions 18 Cobo-Calvo A, Ruiz A, Maillart E, et al; OFSEP and NOMADMUS
in the approval of the final version. Study Group. Clinical spectrum and prognostic value of CNS MOG
autoimmunity in adults: The MOGADOR study. Neurology 2018;
Conflict of Interest 90(21):e1858–e1869
The authors have no conflict of interest to declare. 19 Cai M-T, Zheng Y, Shen C-H, et al. Evaluation of brain and spinal
cord lesion distribution criteria at disease onset in distinguishing
NMOSD from MS and MOG antibody-associated disorder. Mult
References Scler 2021;27(06):871–882
1 Lennon VA, Wingerchuk DM, Kryzer TJ, et al. A serum autoanti- 20 Salama S, Khan M, Shanechi A, Levy M, Izbudak I. MRI differences
body marker of neuromyelitis optica: distinction from multiple between MOG antibody disease and AQP4 NMOSD. Mult Scler
sclerosis. Lancet 2004;364(9451):2106–2112 2020;26(14):1854–1865
2 Dutra BG, da Rocha AJ, Nunes RH, Maia ACM. Neuromyelitis Optica 21 Chen C, Liu C, Fang L, et al. Different magnetic resonance imaging
Spectrum Disorders: Spectrum of MR Imaging Findings and Their features between MOG antibody- and AQP4 antibody-mediated
Differential Diagnosis. Radiographics 2018;38(01):169–193 disease: A Chinese cohort study. J Neurol Sci 2019;405:116430

Arquivos de Neuro-Psiquiatria Vol. 81 No. 6/2023 © 2023. The Author(s).

 MRI of demyelinating antibody diseases Fragoso et al. 543

22 Sechi E, Krecke KN, Messina SA, et al. Comparison of MRI Lesion 24 Shor N, Deschamps R, Cobo Calvo A, et al; NOMADMUS study
Evolution in Different Central Nervous System Demyelinating group. MRI characteristics of MOG-Ab associated disease in
Disorders. Neurology 2021;97(11):e1097–e1109 adults: An update. Rev Neurol (Paris) 2021;177(1-2):39–50
23 Dubey D, Pittock SJ, Krecke KN, et al. Clinical, Radiologic, and 25 Ringelstein M, Ayzenberg I, Lindenblatt G, et al; Neuromyelitis
Prognostic Features of Myelitis Associated With Myelin Oligoden- Optica Study Group (NEMOS) Interleukin-6 Receptor Blockade in
drocyte Glycoprotein Autoantibody. JAMA Neurol 2019;76(03): Treatment-Refractory MOG-IgG-Associated Disease and Neuro-
301–309 myelitis Optica Spectrum Disorders. Neurol Neuroimmunol Neu-
roinflamm 2021;9(01):e1100

Arquivos de Neuro-Psiquiatria Vol. 81 No. 6/2023 © 2023. The Author(s).