Abstract
Background and Objectives
Myelitis is a common manifestation of aquaporin-4 immunoglobulin G–seropositive (AQP4-IgG+) neuromyelitis optica spectrum disorder (NMOSD), with longitudinally extensive transverse myelitis (LETM) as a characteristic MRI feature. Bright spotty lesions (BSLs) and a novel “double-contoured LETM” sign have also been increasingly observed, but their prevalence and diagnostic utility are unclear. This study aimed to characterize spinal cord MRI findings and evaluate the diagnostic performance of established and novel MRI signs in AQP4-IgG+ NMOSD myelitis.
Methods
We retrospectively reviewed spinal cord MRI scans from adults with AQP4-IgG+ NMOSD myelitis at a university-based hospital across acute (<30 days), subacute (1–3 months), and remission (>3 months) stages. MRI features were descriptively analyzed. Acute scans were assessed for specific MRI signs, including the H sign, double-contoured LETM sign, BSLs, and sagittal ring enhancement and compared with other causes of myelopathy. Sensitivity, specificity, and likelihood ratios were calculated.
Results
Among 217 patients with AQP4-IgG+ NMOSD (median age at onset 44.5 years, 95.9% women), 170 (78.3%) experienced myelitis. A total of 331 NMOSD spinal MRI scans were reviewed. The comparison group included 72 scans, most commonly from patients with spinal cord infarction (39) and idiopathic transverse myelitis (14). Restricting the analysis to first-ever myelitis (53 scans), LETM was observed in 88.7%, and sensitivities of BSLs, the double-contoured LETM sign, and sagittal ring enhancement were 64.0% (95% CI 50.1%–75.9%), 47.2% (95% CI 34.4%–60.3%), and 60.0% (95% CI 46.2%–72.4%), respectively, with specificities of 98.6% (95% CI 92.2%–99.7%), 98.6% (95% CI 92.4%–99.8%), and 96.4% (95% CI 87.9%–99.0%). Combining these signs increased specificity further but reduced sensitivity to approximately 30%. When acute first-ever and recurrent myelitis were analyzed together (161 scans), the prevalence of these signs decreased to 56.5%, 19.9%, and 44.1%, respectively. During remission, complete radiologic resolution was observed in 12.4% of 129 scans, whereas spinal cord atrophy was present in 56.6%.
Discussion
BSLs, the double-contoured LETM sign, and sagittal ring enhancement demonstrate high specificity and may aid in differentiating NMOSD from other causes of longitudinally extensive myelopathy. The double-contoured LETM sign represents a novel imaging feature, particularly in first-ever acute myelitis. However, the comparison cohort was enriched with noninflammatory myelopathies, which may limit generalizability. Further validation in independent cohorts is warranted.
Introduction
Neuromyelitis optica spectrum disorder (NMOSD) is a rare but severe autoimmune inflammatory disease of the CNS characterized primarily by astrocytic injury, with secondary demyelination and neuronal damage.1,2 Clinically, NMOSD presents with a range of core syndromes, including optic neuritis, acute transverse myelitis, area postrema syndrome, and brainstem, diencephalic, or cerebral involvement.3 Accurate differentiation of NMOSD from other inflammatory demyelinating disorders, particularly myelin oligodendrocyte glycoprotein antibody–associated disease (MOGAD), is critical because treatment strategies and long-term outcomes differ substantially.4 This distinction is especially important in Asian populations, where NMOSD is relatively more prevalent.5,6
Acute transverse myelitis is among the most common initial manifestations of NMOSD, occurring in approximately 40% of patients and frequently resulting in significant disability.6,7 A key imaging hallmark is longitudinally extensive transverse myelitis (LETM). LETM has high specificity to NMOSD compared with multiple sclerosis (MS) and is included in the 2015 diagnostic criteria.8,9 However, the presence of a longitudinally extensive spinal cord lesion (LESCL) is not always specific for NMOSD. Similar imaging patterns can also be seen in MOGAD, myelitis from systemic autoimmune diseases, and infectious conditions.10,11 In addition, vascular disorders, including spinal cord infarction and spinal dural arteriovenous fistula, may mimic LETM and represent important noninflammatory differentials.12,13
Beyond LETM, several MRI biomarkers have been proposed. Bright spotty lesions (BSLs) demonstrate high specificity with moderate sensitivity for NMOSD compared with MS and MOGAD.14-18 Marked central canal T2 hyperintensity has also been suggested as a distinguishing feature between NMOSD or MOGAD and MS.19 Contrast-enhancement patterns in NMOSD myelitis are variable, but irregular or rim enhancement is frequently observed.20-22 Based on the experience at the MS and Related Disorders clinic, we identified a novel imaging feature, the double-contoured LETM sign, characterized by 2 distinct T2 signal intensities with a brighter anterior patchy component within a background of less intense T2 hyperintensity.
Rapid diagnosis is crucial for prompt management in NMOSD because MRI is usually available before AQP4 antibody results. Highly specific MRI signs may therefore aid early diagnosis in patients presenting with myelitis. In this study, we systematically characterized spinal cord MRI findings in AQP4-IgG–seropositive (AQP4-IgG+) NMOSD across different stages of myelitis and evaluated the diagnostic performance of established and novel imaging markers for differentiating causes of LESCLs.
Methods
Study Design
We retrospectively reviewed the clinical and spinal cord MRI characteristics of patients at Siriraj Hospital, a university-based hospital, and Siriraj Piyamaharajkarun Hospital, a tertiary-care hospital, between January 1994 and December 2025. Inclusion criteria were (1) age 18 years or older, (2) had at least 1 myelitis attack and an MRI available for review, and (3) AQP4-IgG positive by cell-based assay.8
The study objectives comprised 2 parts:
First, we aimed to systematically characterize spinal cord lesions in AQP4-IgG+ NMOSD patients with myelitis across different clinical stages: acute (MRI performed within 30 days after attack onset), subacute (MRI performed within 1–3 months after onset of the attack), and remission (MRI performed more than 3 months after the attack). We described the characteristics and patterns of T2-weighted abnormalities and contrast-enhanced T1-weighted abnormalities.
Second, we aimed to evaluate the diagnostic performance of the established and the novel MRI signs, including BSLs, double-contoured LETM, H sign, and sagittal ring enhancement patterns. We compared the prevalence of these signs between patients with AQP4-IgG+ NMOSD and a non-NMOSD comparison group and calculated the sensitivity, specificity, and likelihood ratios.
Patient Selection
Patients diagnosed with AQP4-IgG+ NMOSD were identified from the NMOSD database. For comparison, we included other etiologies that frequently present with the LESCLs. The comparison group was identified from the hospital databases, which included patients with immune-mediated myelopathy such as MOGAD,23 myelitis associated with systemic autoimmune diseases (Behçet disease and sarcoidosis),12 and idiopathic transverse myelitis. Patients with idiopathic transverse myelitis fulfilled the proposed diagnostic criteria and had at least one documented instance of negative testing for both AQP4-IgG and MOG-IgG by cell-based assays.24 Noninflammatory conditions associated with long spinal cord lesions and acute or subacute onset included spinal cord infarction and intravascular large B-cell lymphoma because these conditions frequently mimic acute NMOSD myelitis and are occasionally encountered in clinical practice.13 We also included patients with subacute combined degeneration because of nitrous oxide toxicity and vitamin B12 deficiency, given their similar imaging features and the occasional presentation of acute myelopathy in patients with nitrous oxide use. Patients with MS were not included because MS classically presents with short spinal cord lesions and rarely with LETM, while the clinical course typically evolves over a longer period before a definitive diagnosis can be established according to the 2024 revision of the McDonald criteria.25 Patients with AQP4-IgG–seronegative NMOSD were also excluded because they may represent a different disease entity or have an uncertain diagnosis. In addition, patients without available MRI studies for review were excluded.
Clinical Data Collection
Clinical data from patients with AQP4-IgG+ NMOSD were collected from medical records. Demographic data included current age, age at disease onset, disease duration since diagnosis, and sex. Disease-related data included clinical syndromes, total number of attacks, number of myelitis attacks, and current immunotherapy.
To categorize the clinical stage, details of each myelitis attack, that is, date of attack onset, clinical symptoms, and type and date of treatment, were included.
MRI Acquisition and Assessment
All MRI studies were performed using 1.5-Tesla or 3.0-Tesla MRI scanners. Typical MRI parameters are presented in eTable 1. T2-weighted and contrast-enhanced T1-weighted sequences were primarily reviewed.
MRI scans were categorized into 3 groups: acute, subacute, and remission scans. We collected all scans obtained during the acute and subacute phases of each myelitis attack. For acute and subacute scans, we assessed the presence of spinal cord lesions, axial lesion involvement, presence of LESCLs, sagittal T2 lesion patterns, specific MRI signs (H sign, BSLs, and double-contoured LETM), contrast enhancement, and enhancement patterns. For remission scans, we included only the first scan obtained during remission for each patient and assessed the presence of residual T2 lesions, residual enhancement, and spinal cord atrophy.
The H sign was defined as a prominent hyperintense T2 lesion on axial images involving the entire spinal cord gray matter, resembling the letter “H.”26 BSLs were defined as very bright hyperintense T2 spots within myelitis lesions, visible on axial view, significantly brighter than the surrounding lesion and comparable in intensity to CSF.18 The double-contoured LETM sign was defined as: (1) a longitudinally extensive spinal cord lesion spanning 3 or more vertebral segments, (2) involvement of most of the anteroposterior diameter of the spinal cord, and (3) 2 distinct signal intensities on T2-weighted imaging, consisting of a brighter linear patchy component located anteriorly and shading into a relatively dimmer component comprising the remainder of the lesion, extending along the rostrocaudal axis on sagittal images. Spinal cord atrophy was visually graded as a decrease in spinal cord volume relative to nonlesional areas. All imaging signs were defined a priori based on prior reports and clinical observations.
For the non-NMOSD comparison group, MRI scans obtained within 30 days of clinical onset (acute phase) were included, except for 2 patients with subacute combined degeneration because of vitamin B12 deficiency. The presence of the aforementioned MRI signs was also assessed.
MRI from NMOSD and non-NMOSD comparison groups received independent blinded assessment by 2 neuroradiologists (C.N. and S.B.) for the MRI signs. Any disagreements between the 2 reviewers were resolved by adjudication from a third reviewer (J.J.).
Statistical Analysis
Spinal cord lesion characteristics and their prevalence were summarized using descriptive statistics. Continuous variables were reported as medians with interquartile ranges (IQRs), and categorical variables were reported as counts and percentages. The data were incomplete for some patients, and therefore, denominators varied and were reported accordingly. Differences between groups were analyzed using Student t test, χ2 test, or Fisher exact test, as appropriate. A 2-sided p value < 0.05 was considered statistically significant.
Inter-rater agreement between the 2 reviewers was evaluated using Cohen kappa coefficient. Assuming an anticipated kappa coefficient of 0.70, a minimum acceptable kappa coefficient of 0.55, and a range of expected sign prevalences, a sample size of approximately 100 cases was estimated to be sufficient for the inter-rater reliability analysis. Accordingly, the first 100 consecutive eligible cases were assessed for inter-rater agreement. Sensitivity was calculated as the proportion of true positives divided by the sum of true positives and false negatives, while specificity was calculated as the proportion of true negatives divided by the sum of true negatives and false positives. The reference standard for true-positive cases was seropositivity for AQP4-IgG antibodies by a fixed cell-based assay. Positive and negative likelihood ratios were calculated as sensitivity divided by (1 − specificity) and (1 − sensitivity) divided by specificity, respectively. The 95% CIs for sensitivity, specificity, and likelihood ratios were also calculated. All statistical analyses were performed using R software, version 4.5.2.
Standard Protocol Approvals, Registrations, and Patient Consents
This study was approved by the Siriraj Institutional Board Review (COA No. Si 026/2026). All patients provided written informed consent for inclusion of their clinical data in the MS and Related Disorders Clinic registry for research purposes.
Results
Patient Characteristics
A total of 217 patients with AQP4-IgG+ NMOSD with 331 available spinal MRIs were included. Women comprised 95.9% (208/217), and the median age at onset was 44.5 years (IQR, 27.1–51.4). Myelitis was the most frequent clinical presentation (170/217, 78.3%), with a history of recurrent attacks in most patients (118/170, 69.4%; median number of attacks, 2 [IQR, 1–4]). Most patients were receiving maintenance immunosuppressive therapy, most commonly azathioprine (79/217, 36.4%) and anti-CD20 monoclonal antibodies (66/217, 30.4%) (Figure 1 and Table 1).
Figure 1. Flowchart Summary of Data Review Protocol.
AQP4-IgG+ NMOSD = aquaporin-4 immunoglobulin G–seropositive neuromyelitis optica spectrum disorder; MOGAD = myelin oligodendrocyte glycoprotein antibody–associated disease.
Table 1.
Demographic Characteristics of AQP4-IgG–Seropositive NMOSD Included in the Study
| Characteristics | All patients (n = 217) | Patients with myelitis (n = 170) |
| Age; y, median (IQR) | 57.5 (41.5–63.7) | 57.5 (43.3–64.8) |
| Age at onset; y, median (IQR) | 44.5 (27.1–51.4) | 44.5 (26.9–51.5) |
| Disease duration; y, median (IQR) | 13.0 (5.2–17.7) | 13.0 (5.6–20.6) |
| Female, n (%) | 208 (95.9) | 164 (96.4) |
| Clinical syndrome since diagnosis, n (%) | ||
| Myelitis | 170 (78.3) | 170 (100) |
| Optic neuritis | 133 (61.3) | 100 (58.8) |
| Area postrema syndrome | 36 (16.6) | 21 (12.4) |
| Brainstem syndrome | 58 (26.7) | 40 (23.5) |
| Cerebral syndrome | 16 (7.4) | 13 (7.6) |
| Diencephalic syndrome | 7 (3.2) | 5 (2.9) |
| No. of attacks, median (IQR) | 3 (2–7) | 4 (2–8) |
| No. of myelitis attacks, median (IQR) | 2 (1–4) | 2 (1–4) |
| Recurrent myelitis, n (%) | 118 (69.4) | 118 (69.4) |
| Current immunotherapy, n (%) | ||
| Azathioprine | 79 (36.4) | 62 (36.5) |
| Mycophenolate mofetil | 37 (17.1) | 26 (15.3) |
| Rituximab | 66 (30.4) | 52 (30.6) |
| Satralizumab | 1 (0.5) | 1 (0.6) |
Abbreviations: AQP4-IgG = aquaporin-4 immunoglobulin G; IQR = interquartile range; NMOSD = neuromyelitis optica spectrum disorder.
MRI Features of Acute Myelitis
A total of 161 acute-phase MRI examinations were analyzed (median imaging interval, 8.0 days [IQR, 5.0–14.0]). The LESCLs were present in 146/161 (90.7%) scans. Lesions most frequently spanned through the cervical and thoracic regions (74/161, 46.0%), followed by isolated thoracic (39/161, 24.2%) and isolated cervical involvement (34/161, 21.2%). Conus medullaris involvement was uncommon (13/161, 7.9%), and most were continuous caudal extension from cervical or thoracic lesions (11/161, 6.8%). On axial imaging, involvement of more than two-thirds of the cord cross-sectional area was observed in 123/147 (83.7%) (Table 2).
Table 2.
Prevalence of Spinal Cord MRI Abnormalities in AQP4-IgG–Seropositive NMOSD Myelitis
| MRI characteristics | Acute scan (n = 161) | Subacute scan (n = 41) | ||
| No. of patients n (%) | Median timing of MRI (IQR) day | No. of patients n (%) | Median timing of MRI (IQR) day | |
| T2-weighted abnormalities | ||||
| Sagittal view | ||||
| LESCLs (≥3 vertebral segments) | 146 (90.7) | 8.0 (5.0–13.8) | 31 (75.6) | 56.0 (50.0–70.5) |
| Short lesion | 15 (9.3) | 10.0 (3.5–17.5) | 10 (24.4) | 51.0 (46.5–62.8) |
| Location | ||||
| Isolated cervical | 34 (21.1) | 11.0 (5.0–16.0) | 10 (24.4) | 65.0 (51.0–75.5) |
| Isolated thoracic | 39 (24.2) | 8.0 (6.0–12.0) | 10 (24.4) | 54.5 (52.0–62.8) |
| Isolated conus medullaris | 2 (1.2) | 17.0 (15.5–18.5) | 1 (2.4) | 46.0c |
| Cervical through thoracic | 74 (46.0) | 7.0 (4.0–13.0) | 18 (43.9) | 53.5 (43.5–63.0) |
| Thoracic through conus medullaris | 3 (1.9) | 9.0 (6.5–9.0) | 2 (4.9) | 52.0 (45.0–59.0) |
| Cervical through conus medullaris | 8 (5.0) | 5.5 (4.5–8.5) | 1 (2.4) | 52.0c |
| T2-weighted abnormalities pattern | ||||
| Patchy/heterogeneous | 103 (64.0) | 8.0 (5.0–15.0) | 36 (87.8) | 55.0 (50.0–67.0) |
| Homogeneous | 26 (16.1) | 8.0 (4.0–12.5) | 3 (7.3) | 71.0 (59.5–73.5) |
| Double-contoured LETM | 32 (19.9) | 6.5 (3.0–11.8) | 2 (4.9) | 42.5 (42.3–42.8) |
| Cord enlargement | 101 (62.7) | 6.0 (4.0–11.0) | 13 (31.7) | 46.0 (42.0–54.0) |
| Axial viewa | ||||
| Location | ||||
| >2/3 cross-sectional area | 123 (83.7) | 8.0 (4.5–14.0) | 25 (65.8) | 59.0 (48.0–72.0) |
| Central | 10 (6.8) | 11.0 (2.8–16.8) | 10 (26.3) | 53.5 (50.5–56.8) |
| Anterior | 0 (0) | — | 1 (2.6) | 46.0c |
| Posterior | 11 (7.5) | 7.0 (4.5–8.5) | 2 (5.3) | 58.0 (55.0–61.0) |
| Hemicord | 3 (2.0) | 5.0 (4.5–11.0) | 0 (0) | — |
| Bright spotty lesion (BSL) | 83 (56.5) | 8.0 (5.0–14.0) | 15 (39.5) | 54.0 (44.5–74.0) |
| BSL location | ||||
| Isolated central | 40 (27.2) | 7.0 (4.8–14.3) | 8 (21.1) | 53.0 (41.8–72.3) |
| Isolated peripheral | 19 (12.9) | 8.0 (6.0–12.5) | 3 (7.9) | 72.0 (62.0–74.5) |
| Central and peripheral | 24 (16.3) | 9.0 (5.0–14.3) | 4 (10.5) | 51.0 (45.3–62.3) |
| H sign | 11 (7.5) | 6.0 (4.0–7.0) | 2 (5.3) | 38.0 (36.0–40.0) |
| Contrast-enhanced T1-weighted abnormalitiesb | ||||
| Contrast enhancement | 115 (75.7) | 7.0 (4.0–13.5) | 20 (52.6) | 53.0 (42.3–67.3) |
| Enhancement pattern | ||||
| Sagittal ring | 67 (44.1) | 6.0 (4.0–13.5) | 4 (10.5) | 44.5 (42.8–67.3) |
| Patchy/heterogeneous | 42 (27.6) | 7.0 (5.3–12.8) | 12 (31.6) | 52.0 (41.0–57.8) |
| Homogeneous | 2 (1.3) | 6.3 (6.3–6.8) | 2 (5.3) | 76.0 (73.5–78.5) |
| Nodular | 5 (3.3) | 14.0 (5.0–15.0) | 2 (5.3) | 68.0 (66.0–70.0) |
Abbreviation: AQP4-IgG = aquaporin-4 immunoglobulin G; IQR = interquartile range; LESCL = longitudinally extensive spinal cord lesion; LETM = longitudinally extensive transverse myelitis; NMOSD = neuromyelitis optica spectrum disorder.
Total number: acute scan (147) and subacute scan (38).
Total number: acute scan (152) and subacute scan (38).
Because the characteristic was observed in only one patient, only the timing of the MRI is presented.
On sagittal T2-weighted imaging, patchy or heterogeneous hyperintensity was the predominant pattern (103/161, 64.0%), whereas homogeneous hyperintensity was less common (26/161, 16.1%). The double-contoured LETM pattern was observed in 32/161 scans (19.9%) (Figures 2A and 3). Cord enlargement was present in 102/161 (62.7%). The BSLs were identified in 83/147 scans (56.5%), more frequently in central than peripheral locations (64/147 [43.5%] vs 43/147 [29.3%]), with 24/147 (16.3%), demonstrating combined central and peripheral BSLs (Figure 2B). The H sign was observed in 11/147 scans (7.5%).
Figure 2. Characteristics of MRI Abnormalities in Acute AQP4-IgG+ NMOSD Myelitis.
Image A illustrated patterns of T2-weighted abnormalities. Note the area postrema involvement in figure Aa (star), and the double-contoured T2 lesion in figure Ab, shown as an anterior brighter linear T2 hyperintensity line (thin arrow) and faded T2 hyperintensity at the back (thick arrow). Image B illustrated the BSLs, and image C illustrated contrast enhancement patterns. AQP4-IgG+ NMOSD = aquaporin-4 immunoglobulin G–seropositive neuromyelitis optica spectrum disorder; BSL = bright spotty lesion.
Figure 3. Characteristics of the Double-Contoured LETM Sign in Acute AQP4-IgG–Seropositive NMOSD Myelitis.
Images A.a-E.a illustrate the double-contoured LETM sign; images A.b-E.b show the corresponding axial views. The double-contoured LETM sign showed a brighter signal in the anterior portion (thin arrow) compared with the remainder of the T2 hyperintense lesion (thick arrow). AQP4-IgG+ = aquaporin-4 immunoglobulin G seropositive; LETM = longitudinally extensive transverse myelitis; NMOSD = neuromyelitis optica spectrum disorder.
Contrast enhancement was present in 115/152 scans (75.7%). Sagittal ring enhancement was the most frequent enhancement pattern (67/152, 44.1%), followed by patchy or heterogeneous enhancement (42/152, 27.6%). Homogeneous (2/152, 1.3%) and nodular enhancement (5/152, 3.3%) were infrequent (Figure 2C). Ring-enhancing lesions demonstrated thick, irregular rim enhancement, most pronounced along the posterior margin, with incomplete rims in most cases. A small subset of lesions showed nodular enhancing caps at the superior and inferior margins of the lesion (eFigure 1).
MRI Features in Subacute and Remission Phase
In the 41 subacute-phase scans, the median imaging interval was 55.0 days (IQR, 48.0–70.0); the prevalence of LESCLs decreased to 31/41 (75.6%), with a relative increase in short-segment lesions (10/41, 24.4%). Compared with the acute phase, cord enlargement (13/41, 31.7%) and contrast enhancement (20/38, 52.6%) were less frequent. BSLs were identified in 15/38 examinations (39.5%). The double-contoured LETM sign and the H sign were uncommon (2/41 [4.9%] and 2/38 [5.3%], respectively) (Table 2).
In the remission phase (129 examinations; median imaging interval, 600.0 days (IQR, 222.0–1,595.0), complete resolution of T2 abnormalities was observed in 16 scans (12.4%), with the median imaging interval of 590.5 days (IQR, 205.8–887.0). Spinal cord atrophy was the predominant chronic imaging abnormality, present in 73/129 (56.6%) with a median number of prior myelitis attacks of 3 (IQR 1–4), and a median imaging interval of 734.0 days (IQR 265.0-1,841.0). Residual contrast enhancement was infrequently observed in 13/113 (10.1%); the median interval was 141.0 days (IQR, 109.0–204.0), and most were faint patchy enhancement (eTable 2).
Diagnostic Performance of Imaging Signs in Acute NMOSD Myelitis
A total of 72 non-NMOSD cases were included for comparison: inflammatory myelopathies (idiopathic transverse myelitis [n = 14], MOGAD [n = 7], neuro-Behçet disease [n = 3], and neurosarcoidosis [n = 3]) and noninflammatory myelopathies (spinal cord infarction [n = 39], intravascular large B-cell lymphoma [n = 3], subacute combined degeneration due to nitrous oxide intoxication [n = 1], and vitamin B12 deficiency [n = 2]). Analysis of inter-rater reliability showed high agreement rates for the double-contoured LETM sign, the presence of enhancement, and the sagittal ring enhancement (Cohen kappa [95% CI]: 0.71 [0.54–0.88], 0.80 [0.67–0.93], and 0.83 [0.71–0.95], respectively) and moderate agreement rates for the BSLs and the H sign (Cohen kappa [95% CI]: 0.57 [0.41–0.73] and 0.63 [0.42–0.85], respectively) (eTable 3).
Compared with other causes of myelopathy, LESCLs were more frequent in NMOSD than other myelopathies (146/161 [90.7%] vs 50/72 [69.4%], p < 0.001). Several imaging features were highly specific for acute myelitis from AQP4-IgG+ NMOSD, including BSLs (83/147 [56.5%] vs 1/69 [1.4%], p < 0.001), the double-contoured LETM sign (32/161 [19.9%] vs 1/72 [1.4%], p < 0.001), and sagittal ring enhancement (67/152 [44.1%] vs 2/57 [3.5%], p < 0.001). Cord enlargement was also more frequent in NMOSD (101/161 [62.7%] vs 34/72 [47.2%], p = 0.027), whereas the H sign did not differ significantly (11/147 [7.5%] vs 8/69 [11.6%], p = 0.320) (Table 3 and eTable 4).
Table 3.
Prevalence of Abnormal Signs in T2-Weighted and Contrast-Enhanced T1-Weighted MRI Abnormalities in Acute AQP4-IgG–Seropositive NMOSD Myelitis (<30 d After Onset) and Other Diseases
| Characteristics | AQP4-IgG+ NMOSD (n = 161) | Other myelopathya (n = 72) | p Valueb |
| T2-weighted characteristics: n (%) | |||
| LESCLs | 146 (90.7) | 50 (69.4) | <0.001 |
| Short lesion (<3 vertebral segments) | 15 (9.3) | 19 (26.4) | <0.001b |
| H signc | 11 (7.5) | 8 (11.6) | 0.320 |
| Bright spotty lesion (BSL)c | 83 (56.5) | 1 (1.4) | <0.001b |
| Double-contour LETM | 32 (19.9) | 1 (1.4) | <0.001b |
| Cord enlargement | 101 (62.7) | 34 (47.2) | 0.027b |
| Enhancement: n (%)d | |||
| Presenced of enhancement | 115 (75.7) | 33 (57.9) | 0.012b |
| Sagittal ring | 67 (44.1) | 2 (3.5) | <0.001b |
| Patchy/heterogeneous | 42 (27.6) | 27 (47.4) | 0.007b |
| Homogeneous | 2 (1.3) | 0 (0) | >0.999 |
| Nodular | 5 (3.3) | 4 (7.0) | 0.260 |
Abbreviations: AQP4-IgG+ NMOSD = aquaporin-4 immunoglobulin G–seropositive neuromyelitis optica spectrum disorder; LESCL = longitudinally extensive spinal cord lesion; LETM = longitudinally extensive transverse myelitis; MOGAD = myelin oligodendrocyte glycoprotein antibody–associated disease.
Other myelopathy included spinal cord infarction (n = 39), MOGAD (n = 7), idiopathic transverse myelitis (n = 14), neurosarcoidosis (n = 3), neuro-Behçet disease (n = 3), intravascular large B-cell lymphoma (n = 3), and subacute combined degeneration (n = 3).
statistically significant results.
Total number: AQP4-IgG+ NMOSD (147) and other myelopathy (69).
Total number: AQP4-IgG+ NMOSD (152) and other myelopathy (57).
In first-ever and recurrent acute NMOSD myelitis cases, the BSLs demonstrated the highest sensitivity among NMOSD-specific imaging markers (56.5% [48.4%–64.2%]) and high specificity (98.6% [92.2%–99.7%]). Sagittal ring enhancement and the double-contoured LETM sign demonstrated high specificity (96.5% [88.1–99.0] and 98.6% [92.5%–99.8%], respectively) but lower sensitivity (44.1% [36.4%–52.0%] and 19.9% [14.5–26.7]). Combinations of imaging signs yielded a specificity of 100% (94.9%–100%) but markedly reduced sensitivity (Table 4).
Table 4.
Sensitivity and Specificity of Imaging Signs for Acute AQP4-IgG–Seropositive NMOSD Myelitis (<30 d After Onset)
| Imaging signs | First-ever myelitis (n = 53) | First ever + recurrent myelitis (n = 161) | ||||||
| Sensitivity (%) | Specificity (%) | LR + | LR - | Sensitivity (%) | Specificity (%) | LR + | LR - | |
| H sign | 16.0 (8.3–28.5) | 88.4 (78.8–94.0) | 1.4 (0.6–3.4) | 1.0 (0.8–1.1) | 7.5 (4.2–12.9) | 88.4 (78.8–94.0) | 0.7 (0.3–1.5) | 1.1 (1.0–1.2) |
| Bright spotty lesion (BSL) | 64.0 (50.1–75.9) | 98.6 (92.2–99.7) | 44.2 (6.2–312.5) | 0.4 (0.3–0.5) | 56.5 (48.4–64.2) | 98.6 (92.2–99.7) | 39.0 (5.5–274.1) | 0.4 (0.4–0.5) |
| Double-contoured LETM | 47.2 (34.4–60.3) | 98.6 (92.4–99.8) | 33.5 (4.7–239.4) | 0.5 (0.4–0.7) | 19.9 (14.5–26.7) | 98.6 (92.5–99.8) | 14.3 (2.0–102.7) | 0.8 (0.8–0.9) |
| Sagittal ring enhancement | 60.0 (46.2–72.4) | 96.4 (87.9–99.0) | 16.8 (4.2–66.8) | 0.4 (0.3–0.6) | 44.1 (36.4–52.0) | 96.5 (88.1–99.0) | 12.6 (3.2–49.6) | 0.6 (0.5–0.7) |
| Cord enlargement | 77.4 (64.5–86.5) | 52.1 (40.7–63.3) | 1.6 (1.2–2.21) | 0.4 (0.3–0.8) | 62.7 (55.0–69.8) | 52.8 (41.4–63.9) | 1.3 (1.0–1.7) | 0.7 (0.5–1.0) |
| BSL + double-contoured LETM | 28.3 (18.0–41.6) | 100 (94.9–100) | 42.9 (2.6–701.3)a | 0.7 (0.6–0.8) | 13.2 (8.8–19.3) | 100 (94.9–100) | 19.6 (1.2–319.5)a | 0.9 (0.8–0.9) |
| BSL + sagittal ring enhancement | 39.6 (27.6–53.1) | 100 (94.9–100) | 57.9 (3.6–933.8)a | 0.6 (0.5–0.7) | 29.9 (23.2–37.5) | 100 (94.9–100) | 42.6 (2.7–681.6)a | 0.7 (0.6–0.8) |
| Double-contoured LETM + sagittal ring enhancement | 32.7 (21.5–46.2) | 100 (94.9–100) | 47.6 (2.9–773.1)a | 0.7 (0.6–0.8) | 13.8 (9.3–19.9) | 100 (94.9–100) | 20.4 (1.3–331.8)a | 0.9 (0.8–0.9) |
| BSL + double-contoured LETM + sagittal ring enhancement | 20.8 (12.0–33.5) | 100 (94.9–100) | 31.3 (1.9–518.6)a | 0.8 (0.7–0.9) | 9.4 (5.8–14.9) | 100 (94.9–100) | 14.1 (0.9–231.8)a | 0.9 (0.9–1.0) |
Abbreviation: AQP4-IgG = aquaporin-4 immunoglobulin G; BSL = bright spotty lesion; LETM = longitudinally extensive transverse myelitis; LR+ = positive likelihood ratio; LR− = negative likelihood ratio; NMOSD = neuromyelitis optica spectrum disorder.
Owing to zero false-positive cases in our cohort, a 0.5 continuity correction was applied to all cells to prevent infinite likelihood-ratio estimates.
When the analysis was restricted to first-ever myelitis attacks, higher proportions of the double-contoured LETM sign (p < 0.001) and sagittal ring enhancement (p = 0.013) were observed compared with recurrent attacks (eTable 5). The prevalence of BSLs did not differ between first-ever and recurrent attacks (p = 0.571). Stratifying patients based on myelitis attack history improved the sensitivity of the BSLs, double-contoured LETM sign, and sagittal ring enhancement substantially (BSLs 64.0% [50.1%–75.9%], double-contoured LETM 47.2% [34.4%–60.3%], and sagittal ring enhancement 60.0% [46.2%–72.4%]) without impeding the specificity. Positive likelihood ratios demonstrated that the presence of BSLs, the double-contoured LETM sign, sagittal ring enhancement, or combinations of these MRI signs strongly rule in acute AQP4-IgG+ NMOSD (Table 4).
Discussion
In this cohort of AQP4-IgG+ NMOSD, we systematically characterized spinal cord MRI features across different stages of myelitis and evaluated the diagnostic performance of established and novel imaging markers. LETM remained the predominant imaging phenotype, present in more than 90% of acute attacks. Short myelitis lesions occurred in 9.3% of patients, comparable with prior reports of short myelitis as the initial presentation of AQP4-IgG+ NMOSD of 4%–14%.20,27 Several lesion-specific MRI features, including BSLs, the double-contoured LETM sign, and sagittal ring enhancement, demonstrated very high specificity for acute NMOSD myelitis; however, with limited sensitivity. The diagnostic yield of these features was influenced substantially by attack history.
BSLs emerged as the most sensitive NMOSD-specific imaging marker in this cohort, consistent with prior reports describing sensitivities of 26.7%–69.0% and specificities of 89.1%–100%.18 In our study, BSLs demonstrated a sensitivity of 64.0% and a specificity of 98.6% for first-ever myelitis, including the ability to differentiate from spinal cord infarction, one of the most common acute noninflammatory myelopathies. Pooled analysis of BSLs in patients with first and recurrent myelitis showed a slight reduction in sensitivity, which remained relatively high (56.5%). This feature may aid in differentiating new inflammatory lesions from residual changes or pseudo-relapses, which are common in NMOSD (Table 4).
We also described and validated the double-contoured LETM sign as a novel imaging feature of AQP4-IgG+ NMOSD. Although this sign exhibited limited overall sensitivity (19.9%), it demonstrated high specificity (98.6%) and substantially higher sensitivity in first-ever myelitis (47.2%). Accordingly, this sign may be most informative when MRI is obtained at the time of initial disease presentation because T2 resolution is often incomplete in patients with recurrent myelitis and may overlap with new acute lesions, making the sign difficult to distinguish.
The pathophysiologic mechanisms underlying BSLs and the double-contoured LETM sign remain incompletely understood. Pathologic studies in NMOSD have demonstrated severe astrocytic injury mediated by complement activation, with secondary neuronal and axonal damage.28,29 Tissue involvement extends beyond gray matter and the central canal region to include extensive white matter injury, particularly within the posterior and lateral columns. Perivascular inflammatory infiltrates support hematogenous delivery of AQP4-IgG and complement through both anterior and posterior spinal vascular arteries,30 resulting in heterogeneous lesion appearance across central and peripheral regions. The markedly hyperintense foci seen because BSLs may reflect severe tissue destruction or microcystic change, resulting in signal intensities approaching that of CSF on T2-weighted images.14 However, the pattern of spinal cord atrophy after myelitis was often diffuse, suggesting that inflammation was severe throughout the lesion rather than confined to the BSLs. This supports the hypothesis that BSLs may be related to fluid trapping or microcystic changes. The preferential central distribution of BSLs likely corresponds to regions of high AQP4 expression.
The layered appearance of the double-contoured LETM sign may reflect intralesional gradients of tissue injury, ranging from intense inflammation in severely affected regions to less severe inflammation and vasogenic edema in surrounding tissue. The predilection for an anterior hyperintense component on sagittal imaging may reflect preferential involvement of the anterior gray matter, anterior to the central canal region, which is rich in AQP4 expression (Figure 3). This sign may closely resemble the marked central canal hyperintensity or pseudodilatation of the central canal observed in MOGAD.19,23 However, in MOGAD, the sharp sagittal hyperintense line is centrally located and surrounded by more subtle T2 hyperintensity both anteriorly and posteriorly. By contrast, in the double-contoured LETM sign, the anterior component appears as a bright, linear patchy T2 hyperintensity shading into a relatively fainter T2 hyperintensity confined to the posterior aspect of the lesion. In comparison, the relatively lower signal intensity of the anterior component observed in MOGAD and neurosarcoidosis may also reflect less severe inflammation and tissue injury in these disorders (eFigure 2).31,32
In this AQP4-IgG+ NMOSD cohort, the H sign was observed in 7.5% of patients with acute myelitis, consistent with previous reports of approximately 8%, although less frequently than in MOGAD (29%).23 Pathologic studies have suggested that AQP4 expression is higher in the spinal cord gray matter, causing lesions to appear centrally located or predominantly involve the gray matter, which may result in the characteristic H appearance. In addition, the inhomogeneous nature of longitudinally extensive lesions may increase the likelihood of observing the H sign on certain axial sections.33
Sagittal ring enhancement was the predominant contrast-enhancement pattern in acute AQP4-IgG+ NMOSD myelitis. Although enhancement patterns in NMOSD have been less extensively studied, prior series have reported comparable prevalences of rim-enhancing lesions of approximately 31.2%–35.0%.34-36 An autopsy study in a reported case demonstrated central tissue vacuolation and loss of AQP4 immunoreactivity with peripheral gliosis, consistent with prior severe inflammatory activity at the lesion periphery during the acute attack.36 The frequent observation of incomplete and posteriorly accentuated rims in our cohort suggests spatial heterogeneity of inflammatory activity within lesions. Nodular enhancing caps at lesion poles may reflect active lesion propagation along spinal cord tracts. Although ring enhancement can be observed in MS, its reported prevalence is lower than that observed in our NMOSD cohort (6.2%–8.7%).37,38 In first-ever acute NMOSD myelitis, approximately one-third exhibited any 2 of these specific signs and one-fifth exhibited all 3 (Table 4).
Longitudinal assessment demonstrated that radiologic recovery after NMOSD myelitis is incomplete in most patients. Complete resolution of T2 abnormalities was rare, and more than half of patients developed spinal cord atrophy after a median of 3 prior attacks, providing a structural correlate for the poor recovery potential after NMOSD attacks and reinforcing the importance of early diagnosis and effective relapse prevention strategies.39
Several limitations need to be addressed. This retrospective study included MRI examinations acquired using heterogeneous protocols and variable image quality, which may have limited the assessment of certain imaging features in some cases. In addition, multiple scans from different disease stages in the same patient were included, potentially introducing within-subject correlation. However, this approach enabled characterization of imaging phenotypes across different disease stages and clinical contexts, thereby reflecting real-world practice. To mitigate within-subject correlation, we limited inclusion to one MRI per disease stage per patient, except for the acute phase, in which all scans from separate myelitis attacks were included. Stratified analyses restricted to first-ever myelitis were also performed to further address this limitation. Furthermore, the comparison cohort was enriched for a noninflammatory etiology associated with longitudinally extensive lesions, which may limit generalizability to other inflammatory myelopathies. Finally, patients with MS were not included in the analysis. Although MS is one of the most common demyelinating diseases, it rarely presents with LETM, and its clinical course often requires longitudinal clinical and radiologic information before a definitive diagnosis can be established.
In summary, highly specific MRI features, including BSLs, the double-contoured LETM sign, and sagittal ring enhancement, substantially improve diagnostic confidence for AQP4-IgG+ NMOSD in patients presenting with acute myelitis. Although sensitivity remains imperfect, these imaging markers provide clinically meaningful discrimination from major NMOSD mimics, especially when serologic testing is delayed or unavailable. The identification of the double-contoured LETM sign further expands the spectrum of NMOSD-associated imaging phenotypes and appears particularly informative in first-ever myelitis.
Acknowledgment
The authors thank all participants taking part in the study and research associates of the Siriraj Neuroimmunology Center: Ms. Onpawee Sangsai, M.Sc., Ms. Jirawan Bodkhum, M.Sc., and Ms. Kamonchanok Aueaphatthanawong, B.S., who helped recruit participants and performed basic clinical screening at the Siriraj MS and Related Disorders Clinic.
Glossary
- AQP4-IgG+
aquaporin-4 immunoglobulin G seropositive
- BSL
bright spotty lesions
- IQR
interquartile range
- LESCL
longitudinally extensive spinal cord lesion
- LETM
longitudinally extensive transverse myelitis
- MOGAD
myelin oligodendrocyte glycoprotein antibody–associated disease
- MS
multiple sclerosis
- NMOSD
neuromyelitis optica spectrum disorder
Footnotes
Editorial, page e218604
Author Contributions
N. Sukhonpanich: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data. C. Ngamsombat: drafting/revision of the manuscript for content, including medical writing for content; analysis or interpretation of data. S. Buathong: drafting/revision of the manuscript for content, including medical writing for content; analysis or interpretation of data. N. Thaneerat: major role in the acquisition of data. D. Tharacheewin: major role in the acquisition of data. S. Siritho: drafting/revision of the manuscript for content, including medical writing for content. T. Ongphichetmetha: drafting/revision of the manuscript for content, including medical writing for content. N. Prayoonwiwat: drafting/revision of the manuscript for content, including medical writing for content. N. Rattanathamsakul: drafting/revision of the manuscript for content, including medical writing for content. J. Jitprapaikulsan: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data.
Study Funding
The authors report no targeted funding.
Disclosure
The authors report no relevant disclosures. Go to Neurology.org/N for full disclosures.
References
- 1.Lennon VA, Wingerchuk DM, Kryzer TJ, et al. A serum autoantibody marker of neuromyelitis optica: distinction from multiple sclerosis. Lancet. 2004;364(9451):2106-2112. doi: 10.1016/S0140-6736(04)17551-X [DOI] [PubMed] [Google Scholar]
- 2.Kalinowska-Łyszczarz A, Guo Y, Lucchinetti CF. Update on pathology of central nervous system inflammatory demyelinating diseases. Neurol Neurochir Pol. 2022;56(3):201-209. doi: 10.5603/PJNNS.a2022.0046 [DOI] [PubMed] [Google Scholar]
- 3.Wingerchuk DM, Lucchinetti CF. Neuromyelitis optica spectrum disorder. N Engl J Med. 2022;387(7):631-639. doi: 10.1056/NEJMra1904655 [DOI] [PubMed] [Google Scholar]
- 4.Cacciaguerra L, Flanagan EP. Updates in NMOSD and MOGAD diagnosis and treatment: a tale of two central nervous system autoimmune inflammatory disorders. Neurol Clin. 2024;42(1):77-114. doi: 10.1016/j.ncl.2023.06.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Hor JY, Asgari N, Nakashima I, et al. Epidemiology of neuromyelitis optica spectrum disorder and its prevalence and incidence worldwide. Front Neurol. 2020;11:501. doi: 10.3389/fneur.2020.00501 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Tisavipat N, Lapanakoakiat S, Siengwattana P, et al. A quarter-century report on neuromyelitis optica spectrum disorder in Thailand: a single-center tertiary care cohort. Mult Scler Relat Disord. 2022;63:103907. doi: 10.1016/j.msard.2022.103907 [DOI] [PubMed] [Google Scholar]
- 7.Ou Yang WY, Lin CJ, Yu KW, Guo YC, Liao YC. Neuromyelitis optica spectrum disorder (NMOSD): recent advances and insights from Taiwan. J Formos Med Assoc. 2026;125(Suppl 1):S55-S64. doi: 10.1016/j.jfma.2025.10.019 [DOI] [PubMed] [Google Scholar]
- 8.Wingerchuk DM, Banwell B, Bennett JL, et al. International consensus diagnostic criteria for neuromyelitis optica spectrum disorders. Neurology. 2015;85(2):177-189. doi: 10.1212/WNL.0000000000001729 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Kim HJ, Paul F, Lana-Peixoto MA, et al. MRI characteristics of neuromyelitis optica spectrum disorder: an international update. Neurology. 2015;84(11):1165-1173. doi: 10.1212/WNL.0000000000001367 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Pekcevik Y, Mitchell CH, Mealy MA, et al. Differentiating neuromyelitis optica from other causes of longitudinally extensive transverse myelitis on spinal magnetic resonance imaging. Mult Scler. 2016;22(3):302-311. doi: 10.1177/1352458515591069 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Anamnart C, Jitprapaikulsan J. Longitudinally extensive transverse myelitis: don't forget syphilis. Pract Neurol. 2022;22(1):75-76. doi: 10.1136/practneurol-2021-003148 [DOI] [PubMed] [Google Scholar]
- 12.Alkabie S, Casserly CS, Morrow SA, Racosta JM. Identifying specific myelopathy etiologies in the evaluation of suspected myelitis: a retrospective analysis. J Neurol Sci. 2023;450:120677. doi: 10.1016/j.jns.2023.120677 [DOI] [PubMed] [Google Scholar]
- 13.Hsu JL, Cheng MY, Liao MF, et al. A comparison between spinal cord infarction and neuromyelitis optica spectrum disorders: clinical and MRI studies. Sci Rep. 2019;9(1):7435. doi: 10.1038/s41598-019-43606-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Yonezu T, Ito S, Mori M, et al. “Bright spotty lesions” on spinal magnetic resonance imaging differentiate neuromyelitis optica from multiple sclerosis. Mult Scler. 2014;20(3):331-337. doi: 10.1177/1352458513495581 [DOI] [PubMed] [Google Scholar]
- 15.Hyun JW, Kim SH, Jeong IH, Lee SH, Kim HJ. Bright spotty lesions on the spinal cord: an additional MRI indicator of neuromyelitis optica spectrum disorder?. J Neurol Neurosurg Psychiatry. 2015;86(11):1280-1282. doi: 10.1136/jnnp-2014-309761 [DOI] [PubMed] [Google Scholar]
- 16.Rabasté S, Cobo-Calvo A, Nistiriuc-Muntean V, Vukusic S, Marignier R, Cotton F, OFSEP, NOMADMUS Study Group. Diagnostic value of bright spotty lesions on MRI after a first episode of acute myelopathy. J Neuroradiol. 2021;48(1):28-36. doi: 10.1016/j.neurad.2020.04.006 [DOI] [PubMed] [Google Scholar]
- 17.Hyun JW, Lee HL, Park J, et al. Brighter spotty lesions on spinal MRI help differentiate AQP4 antibody-positive NMOSD from MOGAD. Mult Scler. 2022;28(6):989-992. doi: 10.1177/13524585211060326 [DOI] [PubMed] [Google Scholar]
- 18.Salama S, Levy M. Bright spotty lesions as an imaging marker for neuromyelitis optica spectrum disorder. Mult Scler. 2022;28(11):1663-1666. doi: 10.1177/1352458521994259 [DOI] [PubMed] [Google Scholar]
- 19.Webb LM, Cacciaguerra L, Krecke KN, et al. Marked central canal T2-hyperintensity in MOGAD myelitis and comparison to NMOSD and MS. J Neurol Sci. 2023;450:120687. doi: 10.1016/j.jns.2023.120687 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Dumrikarnlert C, Siritho S, Chulapimphan P, Ngamsombat C, Satukijchai C, Prayoonwiwat N. The characteristics of spinal imaging in different types of demyelinating diseases. J Neurol Sci. 2017;372:138-143. doi: 10.1016/j.jns.2016.11.035 [DOI] [PubMed] [Google Scholar]
- 21.Dutra BG, da Rocha AJ, Nunes RH, Maia ACMJ. Neuromyelitis optica spectrum disorders: spectrum of MR imaging findings and their differential diagnosis. Radiographics. 2018;38(1):169-193. doi: 10.1148/rg.2018170141 [DOI] [PubMed] [Google Scholar]
- 22.Ciccarelli O, Cohen JA, Reingold SC, Weinshenker BG, International Conference on Spinal Cord Involvement and Imaging in Multiple Sclerosis and Neuromyelitis Optica Spectrum Disorders. Spinal cord involvement in multiple sclerosis and neuromyelitis optica spectrum disorders. Lancet Neurol. 2019;18(2):185-197. doi: 10.1016/S1474-4422(18)30460-5 [DOI] [PubMed] [Google Scholar]
- 23.Dubey D, Pittock SJ, Krecke KN, et al. Clinical, radiologic, and prognostic features of myelitis associated with Myelin Oligodendrocyte Glycoprotein autoantibody. JAMA Neurol. 2019;76(3):301-309. doi: 10.1001/jamaneurol.2018.4053 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Transverse Myelitis Consortium Working Group. Proposed diagnostic criteria and nosology of acute transverse myelitis. Neurology. 2002;59:499-505. doi: 10.1212/wnl.59.4.499 [DOI] [PubMed] [Google Scholar]
- 25.Montalban X, Lebrun-Frénay C, Oh J, et al. Diagnosis of multiple sclerosis: 2024 revisions of the McDonald criteria. Lancet Neurol. 2025;24(10):850-865. doi: 10.1016/S1474-4422(25)00270-4 [DOI] [PubMed] [Google Scholar]
- 26.Banwell B, Bennett JL, Marignier R, et al. Diagnosis of myelin oligodendrocyte glycoprotein antibody-associated disease: International MOGAD Panel proposed criteria. Lancet Neurol. 2023;22(3):268-282. doi: 10.1016/S1474-4422(22)00431-8 [DOI] [PubMed] [Google Scholar]
- 27.Flanagan EP, Weinshenker BG, Krecke KN, et al. Short myelitis lesions in aquaporin-4-IgG-positive neuromyelitis optica spectrum disorders. JAMA Neurol. 2015;72(1):81-87. doi: 10.1001/jamaneurol.2014.2137 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Duan T, Smith AJ, Verkman AS. Complement-dependent bystander injury to neurons in AQP4-IgG seropositive neuromyelitis optica. J Neuroinflammation. 2018;15(1):294. doi: 10.1186/s12974-018-1333-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Uzawa A, Oertel FC, Mori M, Paul F, Kuwabara S. NMOSD and MOGAD: an evolving disease spectrum. Nat Rev Neurol. 2024;20(10):602-619. doi: 10.1038/s41582-024-01014-1 [DOI] [PubMed] [Google Scholar]
- 30.Hayashida S, Masaki K, Yonekawa T, et al. Early and extensive spinal white matter involvement in neuromyelitis optica. Brain Pathol. 2017;27(3):249-265. doi: 10.1111/bpa.12386 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Takai Y, Misu T, Fujihara K, Aoki M. Pathology of myelin oligodendrocyte glycoprotein antibody-associated disease: a comparison with multiple sclerosis and aquaporin 4 antibody-positive neuromyelitis optica spectrum disorders. Front Neurol. 2023;14:1209749. doi: 10.3389/fneur.2023.1209749 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Westendorp WF, van de Beek D, Brouwer MC. Neurosarcoidosis: a review from diagnosis and future perspectives. EClinicalMedicine. 2025;90:103653. doi: 10.1016/j.eclinm.2025.103653 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Matiello M, Schaefer-Klein J, Sun D, Weinshenker BG. Aquaporin 4 expression and tissue susceptibility to neuromyelitis Optica. JAMA Neurol. 2013;70(9):1118-1125. doi: 10.1001/jamaneurol.2013.3124 [DOI] [PubMed] [Google Scholar]
- 34.Yokote H, Nose Y, Ishibashi S, et al. Spinal cord ring enhancement in patients with neuromyelitis optica. Acta Neurol Scand. 2015;132(1):37-41. doi: 10.1111/ane.12354 [DOI] [PubMed] [Google Scholar]
- 35.Flanagan EP, Kaufmann TJ, Krecke KN, et al. Discriminating long myelitis of neuromyelitis optica from sarcoidosis. Ann Neurol. 2016;79(3):437-447. doi: 10.1002/ana.24582 [DOI] [PubMed] [Google Scholar]
- 36.Zalewski NL, Morris PP, Weinshenker BG, et al. Ring-enhancing spinal cord lesions in neuromyelitis optica spectrum disorders. J Neurol Neurosurg Psychiatry. 2017;88(3):218-225. doi: 10.1136/jnnp-2016-314738 [DOI] [PubMed] [Google Scholar]
- 37.Klawiter EC, Benzinger T, Roy A, Naismith RT, Parks BJ, Cross AH. Spinal cord ring enhancement in multiple sclerosis. Arch Neurol. 2010;67(11):1395-1398. doi: 10.1001/archneurol.2010.271 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Lopaisankrit T, Thammaroj J. Brain and spinal cord MRI findings in Thai multiple sclerosis patients. J Imaging. 2023;9(2):27. doi: 10.3390/jimaging9020027 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Tunç A, Koc ER, Carnero Contentti E, et al. Predictors of relapse and disability in NMOSD, MOGAD, and double-seronegative demyelinating syndromes: an international multicenter retrospective cohort study. J Neurol. 2025;272(10):680. doi: 10.1007/s00415-025-13425-1 [DOI] [PubMed] [Google Scholar]