Advertisement

Identification of specific causes of myelopathy in a large cohort of patients initially diagnosed with transverse myelitis

Open AccessPublished:September 28, 2022DOI:https://doi.org/10.1016/j.jns.2022.120425

      Highlights

      • Inflammatory and non-inflammatory myelopathies are under-recognized when the diagnosis of “transverse myelitis” is given.
      • Demographics and temporal profile of symptoms are useful in guiding differential diagnosis in myelopathies.
      • Spinal cord infarction is an important under-recognized diagnosis among myelopathies.
      • An etiologic-based diagnosis of myelopathies facilitates appropriate management and treatment.

      Abstract

      Background and objectives

      Identifying the etiologic diagnosis in patients presenting with myelopathy is essential in order to guide appropriate treatment and follow-up. We set out to examine the etiologic diagnosis after comprehensive clinical evaluation and diagnostic work-up in a large cohort of patients referred to our specialized myelopathy clinic, and to explore the demographic profiles and symptomatic evolution of specific etiologic diagnoses.

      Methods

      In this retrospective study of patients referred to the Johns Hopkins Myelitis and Myelopathy Center between 2006 and 2021 for evaluation of “transverse myelitis”,
      the final etiologic diagnosis determined after comprehensive evaluation in each patient was reviewed and validated. Demographic characteristics and temporal profile of symptom evolution were recorded.

      Results

      Of 1193 included patients, 772 (65%) were determined to have an inflammatory myelopathy and 421 (35%) were determined to have a non-inflammatory myelopathy. Multiple sclerosis/clinically isolated syndrome (n = 221, 29%) and idiopathic myelitis (n = 149, 19%) were the most frequent inflammatory diagnoses, while spinal cord infarction (n = 197, 47%) and structural causes of myelopathy (n = 108, 26%) were the most frequent non-inflammatory diagnoses. Compared to patients with inflammatory myelopathies, patients with non-inflammatory myelopathies were more likely to be older, male and experience chronic symptom evolution (p < 0.001 for all). Hyperacute symptom evolution was most frequent in patients with spinal cord infarction (74%), while chronic symptom evolution was most frequent in patients with structural causes of myelopathy (81%), arteriovenous fistula or arteriovenous malformation (81%), myelopathy associated with rheumatologic disorder (71%), and sarcoidosis-associated myelopathy (61%).

      Conclusions

      Patients initially diagnosed with “transverse myelitis” are eventually found to have a more specific inflammatory or even non-inflammatory cause, potentially resulting in inappropriate treatment and follow-up. Demographic characteristics and temporal profile of symptom evolution may help inform a differential diagnosis in these patients. Etiological diagnosis of myelopathies would provide better therapeutic decisions.

      Graphical abstract

      Keywords

      1. Introduction

      Advanced diagnostics such as greater resolution magnetic resonance imaging (MRI) and auto-antibodies associated with specific myelopathic disorders have improved diagnostic accuracy in patients presenting with myelopathy in recent years. However, evidence suggests that the diagnosis of myelopathies in routine clinical practice remains imprecise [
      • Abbatemarco J.R.
      • Galli J.R.
      • Sweeney M.L.
      • et al.
      Modern look at transverse myelitis and inflammatory myelopathy.
      ], as many patients initially diagnosed with “idiopathic transverse myelitis” or “transverse myelitis” are eventually found to have a more specific inflammatory or even non-inflammatory cause after specialist evaluation [
      • Zalewski N.L.
      • Flanagan E.P.
      • Keegan B.M.
      Evaluation of idiopathic transverse myelitis revealing specific myelopathy diagnoses.
      ,
      • Barreras P.
      • Fitzgerald K.C.
      • Mealy M.A.
      • et al.
      Clinical biomarkers differentiate myelitis from vascular and other causes of myelopathy.
      ]. In this retrospective study of patients referred to a single-center specialized myelopathy clinic with a diagnosis of “transverse myelitis”, we examined the final etiologic diagnoses determined after comprehensive clinical assessment and diagnostic work-up. We also evaluated whether demographic characteristics and temporal profile of symptom onset differed according to etiology of myelopathy.

      2. Methods

      2.1 Study design and participants

      Medical records of patients referred with a diagnosis of “transverse myelitis” to the Johns Hopkins Myelitis and Myelopathy Center (JHMMC) between 2006 and 2021 were retrospectively reviewed. Medical records of all patients were reviewed for completeness of information, availability of MRI images or MRI report, and final physician-assigned diagnosis. All diagnoses were reviewed and validated by two of three physicians (OCM, PB and CAP) with expertise in the diagnosis and management of inflammatory and non-inflammatory myelopathies. Exclusion criteria were: 1) a final non-myelopathy diagnosis; 2) missing clinical information or neuroimaging results, precluding validation of the diagnosis; 3) non-completion of recommended investigations; or 4) clinical uncertainty regarding classification of the myelopathy as either inflammatory or non-inflammatory.

      2.2 Demographic and clinical information

      Race was self-reported. Temporal profile from symptom onset to nadir neurological dysfunction was categorized as hyperacute (<6 h); acute (6 to 48 h); subacute (2 to 21 days); or chronic (>21 days) [
      • Barreras P.
      • Fitzgerald K.C.
      • Mealy M.A.
      • et al.
      Clinical biomarkers differentiate myelitis from vascular and other causes of myelopathy.
      ].

      2.3 Diagnostic validation

      Diagnostic validation was in line with methodology described in detail by our group elsewhere [
      • Barreras P.
      • Fitzgerald K.C.
      • Mealy M.A.
      • et al.
      Clinical biomarkers differentiate myelitis from vascular and other causes of myelopathy.
      ], using diagnostic criteria where available [
      • Thompson A.J.
      • Banwell B.L.
      • Barkhof F.
      • et al.
      Diagnosis of multiple sclerosis: 2017 revisions of the McDonald criteria.
      ,
      • Tobin W.O.
      • Guo Y.
      • Krecke K.N.
      • et al.
      Diagnostic criteria for chronic lymphocytic inflammation with pontine perivascular enhancement responsive to steroids (CLIPPERS).
      ,
      • Zalewski N.L.
      • Rabinstein A.A.
      • Krecke K.N.
      • et al.
      Characteristics of spontaneous spinal cord infarction and proposed diagnostic criteria.
      ,
      • Wingerchuk D.M.
      • Banwell B.
      • Bennett J.L.
      • et al.
      International consensus diagnostic criteria for neuromyelitis optica spectrum disorders.
      ,
      • Murphy O.C.
      • Messacar K.
      • Benson L.
      • et al.
      Acute flaccid myelitis: cause, diagnosis, and management.
      ], and otherwise based on interpretation of the clinical features, laboratory and/or neuroimaging findings. Neuromyelitis optica spectrum disorder (NMOSD) included patients with positive aquaporin-4-IgG (AQP4-IgG), or patients with negative serological testing who met clinical diagnostic criteria [
      • Wingerchuk D.M.
      • Banwell B.
      • Bennett J.L.
      • et al.
      International consensus diagnostic criteria for neuromyelitis optica spectrum disorders.
      ]. MOG-IgG testing became available in our institution from 2017, and from this time forward patients with positive MOG-IgG and a consistent clinical syndrome were classified as MOG antibody-associated disease (MOGAD). Infection-associated myelitis was diagnosed when the myelitis emerged in the setting of a documented neurotropic CNS infection, while para/post-infectious myelitis occurred in association with a systemic infection. Idiopathic myelitis and sarcoidosis-associated myelitis were diagnosed in accordance with prior work published by our group [
      • Murphy O.C.
      • Mukharesh L.
      • Salazar-Camelo A.
      • Pardo C.A.
      • Newsome S.D.
      Early factors associated with later conversion to multiple sclerosis in patients presenting with isolated myelitis.
      ,
      • Murphy O.C.
      • Salazar-Camelo A.
      • Jimenez J.A.
      • et al.
      Clinical and MRI phenotypes of sarcoidosis-associated myelopathy.
      ]. Arteriovenous fistulas (AVF) and arteriovenous malformations (AVM) were diagnosed with digital subtraction angiography [
      • Murphy O.C.
      • Hedjoudje A.
      • Salazar-Camelo A.
      • Pardo C.A.
      • Gailloud P.
      Clinical characteristics, misdiagnosis and outcomes of patients with low-flow spinal arteriovenous fistulas.
      ]. Myelopathy associated with connective tissue disease or other rheumatologic disorders (“rheumatologic myelopathy”) was diagnosed according to systemic and/or laboratory features diagnostic of an underlying rheumatological disorder, and a myelopathy consistent with that disorder. Structural myelopathy was diagnosed based on clinical and imaging evidence of spine and/or discal abnormalities impacting the spinal cord. Metabolic myelopathy was diagnosed based on the clinical features supported by relevant laboratory findings. Tumor was diagnosed based on histopathological confirmation from biopsy or resection of intra-axial or extra-axial lesions affecting the cord (or imaging characteristics alone for meningioma).

      2.4 Statistical analyses

      Statistical analyses were performed using Stata version 16 (StataCorp, College Station, TX, USA). The chi-squared, Wilcoxon rank-sum, and one-way analysis of variance (ANOVA) tests were used as appropriate, and p < 0.05 was considered as significant.

      2.5 Ethical approval

      Johns Hopkins University Institutional Review Board approval was obtained for this study, with requirements for patient consent waived.

      3. Results

      3.1 Etiologic diagnoses of the cohort

      Of 1620 patients referred with a working diagnosis of transverse myelitis, 427 patients were excluded and 1193 patients were included. After thorough clinical evaluation and work-up, the myelopathy was categorized as inflammatory in 772 patients (65%) and non-inflammatory in 421 patients (35%, Fig. 1). Only 149 patients (12% of the whole cohort) were given a final diagnosis of idiopathic myelitis after thorough work-up.
      Fig. 1
      Fig. 1Overview of the study population, indicating final etiologic diagnoses after comprehensive clinical evaluation and diagnostic work-up.
      The final diagnosis assigned to each patient after comprehensive work-up is illustrated here. MS = multiple sclerosis, CIS = clinically isolated syndrome, NMOSD = neuromyelitis optica spectrum disorder, AVF = arteriovenous fistula, AVM = arteriovenous malformation.
      Multiple sclerosis/clinically isolated syndrome (MS/CIS, n = 221, 29%), idiopathic myelitis (n = 149, 19%) and NMOSD (n = 132, 17%) were the most frequent inflammatory diagnoses. NMOSD included 101 patients seropositive for AQP4-IgG (70%) and 31 seronegative NMOSD (21%). After 2017, MOGAD was diagnosed in 13 patients out of 384 patients seen after test became available (3.4%). Rheumatologic myelopathy was diagnosed in association with systemic lupus erythematosus (n = 15), Sjogren's syndrome (n = 9), Behcet's disease (n = 2), and mixed connective tissue disease or other (n = 8). Infection-associated myelitis (outlined in Table 2) occurred with enterovirus infection (e.g., acute flaccid myelitis), human immunodeficiency virus (HIV), human T-lymphotropic virus 1 (HTLV-1), mycoplasma pneumoniae, varicella zoster virus (VZV), arboviral (West Nile virus, Jameson Canyon virus), and tick-borne (borreliosis, ehrlichiosis) infections, among others. Inflammatory myelopathies classified as ‘other’ included hyper eosinophilic and hyper-IgE-associated myelitis (“atopic myelitis”), medication-induced (checkpoint inhibitor or anti-TNF therapies), paraneoplastic myelopathy (glial fibrillary acidic protein (GFAP) astrocytopathy, or no known antibody identified), chronic inflammatory pseudotumorous lesions of unknown etiology, post-vaccination myelopathies and chronic lymphocytic inflammation with pontine perivascular enhancement responsive to steroids (CLIPPERS).
      Spinal cord infarction (47%), structural causes of myelopathy (26%), and AVF/AVMs (14%) were the most frequent non-inflammatory diagnoses. Structural causes of myelopathy included cervical spondylosis, disc herniation, spinal cord syrinx, Chiari malformation with cord compression, dorsal arachnoid web, and spinal cord herniation. Metabolic myelopathies included vitamin B12 deficiency, copper deficiency, and mitochondrial disorders. Non-inflammatory myelopathies classified as ‘other’ included radiation myelopathy, hereditary spastic paraparesis, symmetric motor and/or sensory tractopathy of uncertain etiology, and suspected tumors without histologic confirmation.

      3.2 Demographic and clinical characteristics of patients with inflammatory versus non-inflammatory myelopathies

      Compared to patients with an inflammatory myelopathy, patients with a non-inflammatory myelopathy were more likely to be older, male, and experience hyperacute or chronic symptom onset (p < 0.001 for all, Table 1, Fig. 2).
      Table 1Demographic characteristics and temporal profile of symptom evolution patients with myelopathy, according to specific etiologic diagnosis.
      All inflamm-atory (n = 772)MS/CIS (n = 221)Idio-pathic myelitis (n = 149)NMOSD (n = 132)Sarcoidosis-associated myelopathy (n = 83)Infection associated myelitis (n = 70)Rheum-atologic myelopathy (n = 34)Para/Postinfec-tious myelitis (n = 32)p-value
      Across etiologic diagnoses of inflammatory myelopathy.
      All non-inflamm-atory (n = 421)Spinal cord infarct (n = 197)Structural (n = 108)AVF/ AVM (n = 59)p-value
      across etiologic diagnoses of non-inflammatory myelopathy.
      Age, median years (IQR)44 (30–54)44 (35–52)47 (31–56)43 (30–56)48 (39–56)25 (5–56)46 (36–51)23 (12–38)<0.001
      One-way analysis of variance (ANOVA).
      52 (39–63)52 (31–61)50.5 (44.5–62)61 (50–70)<0.001
      One-way analysis of variance (ANOVA).
      Female (%)471 (61)140 (64)86 (58)104 (79)42 (51)24 (34)32 (94)15 (47)<0.001
      chi-squared test.
      195 (46)107 (54)47 (44)13 (22)<0.001
      across etiologic diagnoses of non-inflammatory myelopathy.
      Race
       White (%)564 (73)182 (82)120 (81)74 (56)44 (53)52 (74)23 (68)26 (81)<0.001
      chi-squared test.
      331 (79)156 (79)84 (78)42 (71)0.005
      chi-squared test.
       Black (%)155 (22)27 (12)15 (10)49 (37)36 (43)10 (14)8 (24)2 (6)42 (10)16 (8)18 (17)4 (7)
       Other (%)53 (7)12 (5)14 (9)9 (7)3 (4)8 (11)3 (9)4 (13)48 (11)25 (13)6 (6)13 (22)
      Temporal profile
       Hyperacute (%)16 (2)3 (1)1 (1)7 (5)1 (1)1 (1)1 (3)2 (6)<0.001
      chi-squared test.
      156 (37)145 (74)6 (6)3 (5)<0.001
      chi-squared test.
       Acute (%)162 (21)23 (10)38 (26)41 (31)2 (2)30 (43)4 (12)13 (41)56 (13)40 (20)7 (6)3 (5)
       Subacute (%)349 (45)107 (48)83 (56)67 (51)29 (35)22 (31)5 (15)16 (50)27 (6)8 (4)7 (6)5 (8)
       Chronic (%)245 (32)88 (40)27 (18)17 (13)51 (61)17 (24)24 (71)1 (3)182 (43)4 (2)88 (81)48 (81)
      Specific etiologic diagnoses with n > 30 were included here.
      IQR = interquartile ratio, MS = multiple sclerosis, CIS = clinically isolated syndrome, NMOSD = neuromyelitis optica spectrum disorder, AVM = arteriovenous malformation, AVF = arteriovenous fistula.
      a Across etiologic diagnoses of inflammatory myelopathy.
      b across etiologic diagnoses of non-inflammatory myelopathy.
      c One-way analysis of variance (ANOVA).
      d chi-squared test.
      Fig. 2
      Fig. 2Distribution of age of patients with inflammatory and non-inflammatory myelopathies. Histograms are presented demonstrating (A) the age distribution of patients with inflammatory compared to non-inflammatory myelopathies, and (B) the age distribution of the most common specific etiologic diagnoses. MS = multiple sclerosis, CIS = clinically isolated syndrome, NMOSD = neuromyelitis optica spectrum disorder, AVF = arteriovenous fistula, AVM = arteriovenous malformation.

      3.3 Demographic and clinical characteristics of patients with inflammatory myelopathies

      Among patients with inflammatory myelopathies, age differed across etiologic diagnoses (p < 0.001, Table 1, Fig. 2), driven by the younger age of patients with para/post-infectious myelopathy compared to other etiologic groups. Patients diagnosed with MS/CIS, NMOSD and rheumatologic myelopathy were more frequently female than patients with sarcoidosis-associated myelopathy of infection-associated myelitis. The highest proportion of patients reporting Black race occurred with NMOSD (37%) and sarcoidosis-associated myelitis (43%). Among inflammatory myelopathies, temporal profile of symptom onset was most frequently subacute (45%). A chronic temporal profile was most frequently reported in patients with neurosarcoidosis and myelopathy associated with an underlying rheumatologic disorder, while an acute profile was frequently seen in infection-associated myelopathies (Table 1, Table 2, Fig. 3).
      Table 2Age and temporal profile of symptom evolution in patients with infection-associated myelitis.
      AFM

      (n = 23)
      Myco-plasma (n = 5)VZV

      (n = 10)
      HIV

      (n = 7)
      HTLV-1 (n = 6)Other viral (n = 14)Other bacterial (n = 5)p-value
      Age, median years (IQR)5 (3–6)17.5 (14.5–19)65 (56–73)43.5 (37–45)54.5 (37–56)25 (1–36)59 (35–61)<0.001
      Across etiologic diagnoses of inflammatory myelopathy.
      *
      Temporal profile
       Hyperacute (%)0 (0)0 (0)0 (0)0 (0)0 (0)1 (7)0 (0)<0.001
      across etiologic diagnoses of non-inflammatory myelopathy.
       Acute (%)18 (78)4 (80)1 (10)0 (0)0 (0)6 (43)1 (20)
       Subacute (%)5 (22)1 (20)8 (80)1 (14)0 (0)4 (29)4 (80)
       Chronic (%)0 (0)0 (0)1 (10)6 (86)6 (100)3 (21)0 (0)
      IQR = interquartile ratio, AFM = acute flaccid myelitis, HIV = human immunodeficiency virus, HTLV-1 = human T-lymphotropic virus type VZV = varicella zoster virus; HIV: human immunodeficiency virus; HTLV-1:human T-lymphotropic virus.
      Other viral includes West Nile, Jameson Canyon, Parvovirus, EBV, HSV, Coxsackie.
      Other bacterial includes Lyme and Ehrlichiosis.
      a Across etiologic diagnoses of inflammatory myelopathy.
      b across etiologic diagnoses of non-inflammatory myelopathy.
      Fig. 3
      Fig. 3Distribution of temporal profile of symptom evolution in patients with inflammatory and non-inflammatory myelopathies. Histograms are presented demonstrating (A) the temporal profile of symptom evolution with inflammatory compared to non-inflammatory myelopathies, and (B) the temporal profile of symptom evolution in the most common specific etiologic diagnoses. MS = multiple sclerosis, CIS = clinically isolated syndrome, NMOSD = neuromyelitis optica spectrum disorder, AVF = arteriovenous fistula, AVM = arteriovenous malformation.

      3.4 Demographic and clinical characteristics of patients with non-inflammatory myelopathies

      Among patients with non-inflammatory myelopathies, age differed across etiologic diagnoses (p < 0.001), driven by the older ages of patients with AVF/AVM compared to patients with spinal cord infarction and structural myelopathies (Table 1). A bimodal age distribution was suggested in patients with spinal cord infarctions (Fig. 2). Patients with AVM/AVF were more frequently male than patients with spinal cord infarction or structural myelopathy. Spinal cord infarction was most frequently hyperacute in onset, while all other non-inflammatory etiologies were most frequently chronic (p < 0.001, Table 1, Fig. 3).

      4. Discussion

      In this study of a very large cohort of patients referred to a single-center specialized myelopathy clinic with a diagnosis of “transverse myelitis”, we have demonstrated that comprehensive clinical assessment and diagnostic testing revealed a more specific inflammatory or non-inflammatory diagnosis in the majority of patients. Furthermore, we have shown that basic clinical information including demographic characteristics and temporal profile of symptom evolution is useful when considering the differential diagnosis of a patient presenting with myelopathy.
      The most important finding of our study was the identification of a specific underlying inflammatory or non-inflammatory diagnosis in most patients. This finding was consistent with a smaller reported cohort from a similar specialized center [
      • Zalewski N.L.
      • Flanagan E.P.
      • Keegan B.M.
      Evaluation of idiopathic transverse myelitis revealing specific myelopathy diagnoses.
      ]. A specific etiologic cause of myelopathy is important in guiding clinical management. Failing to diagnose MS/CIS at presentation may represent a lost opportunity for early treatment to prevent accumulation of long-term disability [
      • Cerqueira J.J.
      • Compston D.A.S.
      • Geraldes R.
      • et al.
      Time matters in multiple sclerosis: can early treatment and long-term follow-up ensure everyone benefits from the latest advances in multiple sclerosis?.
      ,
      • O’Connor P.
      • Goodman A.
      • Kappos L.
      • et al.
      Long-term safety and effectiveness of natalizumab redosing and treatment in the STRATA MS study.
      ,
      • Goodin D.S.
      • Reder A.T.
      • Ebers G.C.
      • et al.
      Survival in MS: a randomized cohort study 21 years after the start of the pivotal IFNβ-1b trial.
      ]. Similarly, NMOSD and neurosarcoidosis represent disabling yet treatable inflammatory disorders requiring specific management and follow-up. Furthermore, treatment of non-inflammatory myelopathies is entirely different to inflammatory myelopathies, and appropriate management can only be facilitated by accurate diagnosis in these cases. Spinal cord infarction is an important under-recognized diagnosis for which proposed diagnostic criteria have only been recently suggested [
      • Wingerchuk D.M.
      • Banwell B.
      • Bennett J.L.
      • et al.
      International consensus diagnostic criteria for neuromyelitis optica spectrum disorders.
      ]. Identification of a spinal cord stroke may offer opportunities for intervention addressing underlying mechanisms in individual patients, [
      • Ullman N.
      • Gregg L.
      • Becker D.
      • Pardo C.
      • Gailloud P.
      Anterior disco-osteo-arterial conflict as a cause of intersegmental arterial flow impairment and spinal cord ischemia.
      ,
      • Murphy O.C.
      • Gailloud P.
      • Newsome S.D.
      Spinal claudication secondary to anterior disco-osteo-arterial conflict and mimicking stiff person syndrome.
      ,
      • Gailloud P.
      • Ponti A.
      • Gregg L.
      • Pardo C.A.
      • Fasel J.H.D.
      Focal compression of the upper left thoracic intersegmental arteries as a potential cause of spinal cord ischemia.
      ] and at a group level facilitate research into mechanisms, therapies and recurrence risks. Furthermore, differentiation of spinal cord stroke from an inflammatory myelopathy may avoid potential harm in individual patients from immune-therapies which may have a pro-coagulable effect (e.g. intravenous immunoglobulin) or serious potential infectious complications (e.g. cyclophosphamide). Patients with structural causes of myelopathy or AVF/AVM represent other groups in whom misdiagnosis may result in lost opportunities for intervention, and irreversible disability [
      • Murphy O.C.
      • Hedjoudje A.
      • Salazar-Camelo A.
      • Pardo C.A.
      • Gailloud P.
      Clinical characteristics, misdiagnosis and outcomes of patients with low-flow spinal arteriovenous fistulas.
      ,
      • Flanagan E.P.
      • Krecke K.N.
      • Marsh R.W.
      • Giannini C.
      • Keegan B.M.
      • Weinshenker B.G.
      Specific pattern of gadolinium enhancement in spondylotic myelopathy.
      ]. Recognition of AVF/AVM can also impact management decisions in the acute setting, as precipitous decline in neurological status has been reported in around 40% of patients with AVF/AVM who were erroneously treated with high-dose corticosteroids – perhaps due to haemodynamic effects [
      • Murphy O.C.
      • Hedjoudje A.
      • Salazar-Camelo A.
      • Pardo C.A.
      • Gailloud P.
      Clinical characteristics, misdiagnosis and outcomes of patients with low-flow spinal arteriovenous fistulas.
      ].
      We have demonstrated that in a large cohort of patients with myelopathies, demographic characteristics and temporal profile of symptom onset differ considerably across etiologic diagnoses. Age is a factor which is rarely included in diagnostic criteria, but is an important consideration in clinical reasoning [
      • Cunha B.A.
      The master clinician’s approach to diagnostic reasoning.
      ]. In most disorders, age distributions in our cohort mirrored findings from the broader literature, e.g. older age distribution of NMOSD and neurosarcoidosis compared to other inflammatory disorders [
      • Arkema E.V.
      • Cozier Y.C.
      Epidemiology of sarcoidosis: current findings and future directions.
      ,
      • Fritz D.
      • van de Beek D.
      • Brouwer M.C.
      Clinical features, treatment and outcome in neurosarcoidosis: systematic review and meta-analysis.
      ,
      • Pandit L.
      • Asgari N.
      • Apiwattanakul M.
      • et al.
      Demographic and clinical features of neuromyelitis optica: a review.
      ]. Our finding regarding a possible bimodal age distribution of spontaneous spinal cord strokes is novel, and to our knowledge has not been previously reported in the literature. [
      • Robertson C.E.
      • Brown R.D.
      • Wijdicks E.F.M.
      • Rabinstein A.A.
      Recovery after spinal cord infarcts: long-term outcome in 115 patients.
      ,
      • Cheng M.
      • Lyu R.
      • Chang Y.
      • et al.
      Spinal cord infarction in chinese patients.
      ,
      • Naess H.
      • Romi F.
      Comparing patients with spinal cord infarction and cerebral infarction: clinical characteristics, and short-term outcome.
      ] This may be because pediatric cases have not been included in various observational studies, or suggest that spinal cord stroke is under-recognized in children. Indeed, the literature on spontaneous spinal cord strokes in children are essentially limited to case reports [
      • Nagata K.
      • Tanaka Y.
      • Kanai H.
      • Oshima Y.
      Acute complete paraplegia of 8-year-old girl caused by spinal cord infarction following minor trauma complicated with longitudinal signal change of spinal cord.
      ,
      • Spencer S.P.
      • Brock T.D.
      • Matthews R.R.
      • Stevens W.K.
      Three unique presentations of atraumatic spinal cord infarction in the pediatric emergency department.
      ,
      • Reisner A.
      • Gary M.F.
      • Chern J.J.
      • Grattan-Smith J.D.
      Spinal cord infarction following minor trauma in children: fibrocartilaginous embolism as a putative cause.
      ]. In adults with spontaneous spinal cord stroke, typical cardiovascular risk factors (such as hypertension) are frequently present [
      • Barreras P.
      • Fitzgerald K.C.
      • Mealy M.A.
      • et al.
      Clinical biomarkers differentiate myelitis from vascular and other causes of myelopathy.
      ,
      • Zalewski N.L.
      • Rabinstein A.A.
      • Krecke K.N.
      • et al.
      Characteristics of spontaneous spinal cord infarction and proposed diagnostic criteria.
      ], suggesting that etiologic mechanisms may be similar to cerebral strokes (such as atheroma and cardio-embolism). Additionally, structural issues such as degenerative spinal disease may result in arterial compression in adults [
      • Gailloud P.
      • Ponti A.
      • Gregg L.
      • Pardo C.A.
      • Fasel J.H.D.
      Focal compression of the upper left thoracic intersegmental arteries as a potential cause of spinal cord ischemia.
      ]. On the other hand, the specific mechanisms of spontaneous spinal cord strokes in young people are not well understood. Trauma may be an important etiologic factor, for example vertebral artery dissection or fibrocartilaginous embolism should be considered in a young person presenting with spontaneous spinal cord infarction following intense exercise, contact sports or load-bearing activity (such as weightlifting) [
      • Reisner A.
      • Gary M.F.
      • Chern J.J.
      • Grattan-Smith J.D.
      Spinal cord infarction following minor trauma in children: fibrocartilaginous embolism as a putative cause.
      ]. Other putative etiological factors in young people may include congenital heart disease, thrombophilic disorders, arteriopathy, or genetic disorders leading to vasculitis (e.g., deficiency of adenosine deaminase 2) [
      • Ganhão S.
      • Loureiro G.B.
      • Oliveira D.R.
      • Dos-Reis-Maia R.
      • Aguiar F.
      • Quental R.
      • Moura C.
      • Barreira J.L.
      • Rodrigues M.
      • Brito I.
      Two cases of ADA2 deficiency presenting as childhood polyarteritis nodosa: novel ADA2 variant, atypical CNS manifestations, and literature review.
      ] among others. Exploring the mechanisms of spontaneous spinal cord stroke represents a key area for future research.
      Our findings also suggest that temporal profile of symptom evolution is useful in guiding an initial differential diagnosis in a patient with myelopathy. For example, spinal cord infarction should be considered in the differential of any hyperacute myelopathic syndrome, while disorders in which a chronic presentation is likely include sarcoidosis-associated myelopathy, rheumatologic myelopathy, structural myelopathies, and AVM/AVF.
      Importantly, other factors not studied here that have been identified as informative in differentiating etiologies of myelopathy include MRI characteristics and CSF pleocytosis [
      • Barreras P.
      • Fitzgerald K.C.
      • Mealy M.A.
      • et al.
      Clinical biomarkers differentiate myelitis from vascular and other causes of myelopathy.
      ]. The absence of an MRI T2-lesion can be informative finding, for example MRI may appear normal in the hyperacute assessment of patients with spinal cord infarction (with lesions becoming apparent on later interval imaging) [
      • Zalewski N.L.
      • Rabinstein A.A.
      • Krecke K.N.
      • et al.
      Characteristics of spontaneous spinal cord infarction and proposed diagnostic criteria.
      ]; myelitis in MOGAD may be MRI-negative [
      • Sechi E.
      • Krecke K.N.
      • Pittock S.J.
      • et al.
      Frequency and characteristics of MRI-negative myelitis associated with MOG autoantibodies.
      ]; and a T2-lesion may not be evident in patients with low-flow spinal dural AVF (although other features such as CSF flow voids may be present) [
      • El Mekabaty A.
      • Pardo C.A.
      • Gailloud P.
      The yield of initial conventional MRI in 115 cases of angiographically confirmed spinal vascular malformations.
      ]. Length of T2 lesion is a key MRI characteristic that can help differentiate specific inflammatory myelopathies, with longitudinally extensive lesions extending over 3 vertebral segments being more characteristic of NMOSD, sarcoidosis-associated myelitis, MOGAD, or idiopathic transverse myelitis, than of MS. [
      • Murphy O.C.
      • Mukharesh L.
      • Salazar-Camelo A.
      • Pardo C.A.
      • Newsome S.D.
      Early factors associated with later conversion to multiple sclerosis in patients presenting with isolated myelitis.
      ,
      • Murphy O.C.
      • Salazar-Camelo A.
      • Jimenez J.A.
      • et al.
      Clinical and MRI phenotypes of sarcoidosis-associated myelopathy.
      ,
      • Asnafi S.
      • Morris P.P.
      • Sechi E.
      • et al.
      The frequency of longitudinally extensive transverse myelitis in MS: a population-based study.
      ,
      • Sepúlveda M.
      • Blanco Y.
      • Rovira A.
      • et al.
      Analysis of prognostic factors associated with longitudinally extensive transverse myelitis.
      ,
      • Flanagan E.P.
      • Kaufmann T.J.
      • Krecke K.N.
      • et al.
      Discriminating long myelitis of neuromyelitis optica from sarcoidosis.
      ,
      • Dubey D.
      • Pittock S.J.
      • Krecke K.N.
      • et al.
      Clinical, radiologic, and prognostic features of myelitis associated with myelin oligodendrocyte glycoprotein autoantibody.
      ] Axial location of T2-lesion is also important, with central lesions being characteristic of AVF, and gray matter delineation being a key feature of AFM, for example [
      • Barreras P.
      • Fitzgerald K.C.
      • Mealy M.A.
      • et al.
      Clinical biomarkers differentiate myelitis from vascular and other causes of myelopathy.
      ,
      • Murphy O.C.
      • Messacar K.
      • Benson L.
      • et al.
      Acute flaccid myelitis: cause, diagnosis, and management.
      ]. Enhancement patterns that may be characteristic of specific etiologic diagnosis include the ‘trident sign’ or dorsal subpial enhancement in sarcoidosis-associated myelitis [
      • Flanagan E.P.
      • Kaufmann T.J.
      • Krecke K.N.
      • et al.
      Discriminating long myelitis of neuromyelitis optica from sarcoidosis.
      ,
      • Zalewski N.L.
      • Krecke K.N.
      • Weinshenker B.G.
      • Aksamit A.J.
      • Conway B.L.
      • McKeon A.
      • Flanagan E.P.
      Central canal enhancement and the trident sign in spinal cord sarcoidosis.
      ], ‘pancake-like’ transverse enhancement in spondylotic myelopathy [
      • Flanagan E.P.
      • Krecke K.N.
      • Marsh R.W.
      • Giannini C.
      • Keegan B.M.
      • Weinshenker B.G.
      Specific pattern of gadolinium enhancement in spondylotic myelopathy.
      ], ring or partial ring enhancement in AQP4-IgG seropositive NMOSD [
      • Flanagan E.P.
      • Kaufmann T.J.
      • Krecke K.N.
      • et al.
      Discriminating long myelitis of neuromyelitis optica from sarcoidosis.
      ], or the ‘missing piece’ sign in spinal dural AVF [
      • Zalewski N.L.
      • Rabinstein A.A.
      • Brinjikji W.
      • Kaufmann T.J.
      • Nasr D.
      • Ruff M.W.
      • Flanagan E.P.
      Unique gadolinium enhancement pattern in spinal dural arteriovenous fistulas.
      ]. Indeed, education regarding typical enhancement patterns of specific myelopathies can aid physicians in making the correct diagnosis [
      • Mustafa R.
      • Passe T.J.
      • Lopez-Chiriboga A.S.
      • et al.
      Utility of MRI enhancement pattern in myelopathies with longitudinally extensive T2 lesions.
      ]. Additional advanced imaging studies should also be considered where clinically-appropriate. FDG-PET CT of the body is a useful technique to identify evidence for underlying systemic sarcoidosis when sarcoidosis-associated myelitis is suspected, [
      • Fritz D.
      • van de Beek D.
      • Brouwer M.C.
      • Booij J.
      Whole-body 18F-FDG PET-CT in the diagnosis of neurosarcoidosis.
      ] while rigorous spinal digital subtraction angiography with careful interpretation is considered the gold standard for diagnosis of AVF [
      • Barreras P.
      • Heck D.
      • Greenberg B.
      • Wolinsky J.
      • Pardo C.A.
      • Gailloud P.
      Analysis of 30 spinal angiograms falsely reported as normal in 18 patients with subsequently documented spinal vascular malformations.
      ]. In addition, CSF parameters during the acute phase of symptoms such as pleocytosis and the presence of oligoclonal bands (OCBs) would favor a diagnosis of inflammatory myelopathy, while a completely normal CSF or absence of pleocytosis or OCBs would decrease the likelihood of diagnosis of inflammatory or demyelinating myelopathies [
      • Barreras P.
      • Fitzgerald K.C.
      • Mealy M.A.
      • et al.
      Clinical biomarkers differentiate myelitis from vascular and other causes of myelopathy.
      ]. The presence of CSF-restricted OCBs in combination of MRI features such as short-length lesions or multi-focal lesions may also provide support for the diagnosis of demyelinating myelopathies and/or serve as a predictor for the potential conversion to MS in patients presenting with isolated myelitis [
      • Murphy O.C.
      • Mukharesh L.
      • Salazar-Camelo A.
      • Pardo C.A.
      • Newsome S.D.
      Early factors associated with later conversion to multiple sclerosis in patients presenting with isolated myelitis.
      ].
      The current study has a number of limitations. Our myelopathy center is a specialized service that may have introduced bias towards diagnostically challenging cases. Infection-associated myelitis may be underdiagnosed in this cohort (and perhaps misclassified as idiopathic myelitis), since patients were assessed in an outpatient setting weeks to months after symptom onset, at which point causative pathogens are unlikely to be detected. MOG-IgG testing was not commercially available in our region until 2017, so patients seen prior to this with MOG-IgG associated disease may have been misclassified as idiopathic myelitis or seronegative NMOSD. This situation and the fact MOG-IgG antibody titers may decline after the acute phase and re-testing in some patients may explain the relative lower frequency of MOGAG in our cohort compared to recent studies [
      • Kim K.H.
      • Kim S.H.
      • Hyun J.W.
      • Kim Y.
      • Park H.
      • Kim H.J.
      Seroprevalence of anti-myelin oligodendrocyte glycoprotein antibodies in adults with myelitis.
      ]. We did not include MRI and CSF findings in this study, as these paraclinical features of myelopathies have already been analyzed in detail in other work from our group [
      • Barreras P.
      • Fitzgerald K.C.
      • Mealy M.A.
      • et al.
      Clinical biomarkers differentiate myelitis from vascular and other causes of myelopathy.
      ], and the focus of the current study was limited to misdiagnosis and the identification of demographic and clinical clues to specific etiologic diagnoses. All patients in the current study underwent a battery of routine diagnostic tests including MRI and serum biomarkers, advanced diagnostics such as angiography to detect AVF/AVM and FDG-PET CT to detect systemic sarcoidosis were necessarily limited to patients with suggestive clinical or paraclinical features. Additionally, sensitive and specific tests for certain myelopathic disorders (e.g. spinal cord infarction) are still lacking, meaning that misclassification of diagnoses in some patients cannot be excluded. Finally, it is important to note that this study focused on cases diagnosed or misdiagnosed as ‘transverse myelitis’, and should not be considered an incidence or prevalence study for specific etiologic diagnoses.

      5. Conclusions

      We have demonstrated that specific inflammatory and non-inflammatory disorders are often missed in patients diagnosed with “transverse myelitis”. Neurologists should focus on an etiologic-based diagnosis in order to facilitate appropriate treatment and follow-up rather than establishing the generic diagnosis of “transverse myelitis”. History-taking remains a cornerstone of clinical neurology, with demographic characteristics and temporal profile of symptom evolution potentially contributing to diagnostic reasoning in patients with myelopathies.

      Study funding

      This study received no specific funding.

      Author contributions

      OCM contributed to conceptualization, acquisition, analysis and interpretation of the data, drafting and revising the manuscript.
      PB contributed to acquisition and interpretation of the data, figures and critical revision of the manuscript.
      AVB contributed to acquisition and interpretation of the data.
      MM contributed to study conceptualization and design, acquisition of the data and critical revision of the manuscript.
      CAP contributed to study conceptualization and design, analysis and interpretation of the data and critical revision of the manuscript.

      Disclosures

      The authors have no disclosures relevant to the manuscript.

      Acknowledgments

      Thank you to the Bart McLean Fund for Neuroimmunology Research for the funding and support of the Johns Hopkins Myelitis and Myelopathy Center (JHMMC), the Siegel Rare Neuroimmune Association for supporting Dr. Olwen C. Murphy's fellowship through the James T. Lubin Award, and the Foundation for Sarcoidosis Research for supporting Dr. Paula Barreras' fellowship. Thank you to Drs. Laura Muñoz, Maria A. Garcia, Maria I. Reyes, Andrea Salazar-Camelo, Daniel Becker and Philippe Gailloud for their work and contributions to the JHMMC.

      References

        • Abbatemarco J.R.
        • Galli J.R.
        • Sweeney M.L.
        • et al.
        Modern look at transverse myelitis and inflammatory myelopathy.
        Neurol. Neuroimmunol. Neuroinflamm. 2021; 8e1071https://doi.org/10.1212/NXI.0000000000001071
        • Zalewski N.L.
        • Flanagan E.P.
        • Keegan B.M.
        Evaluation of idiopathic transverse myelitis revealing specific myelopathy diagnoses.
        Neurology. 2018; 90: e96-e102https://doi.org/10.1212/WNL.0000000000004796
        • Barreras P.
        • Fitzgerald K.C.
        • Mealy M.A.
        • et al.
        Clinical biomarkers differentiate myelitis from vascular and other causes of myelopathy.
        Neurology. 2018; 90: e12-e21https://doi.org/10.1212/WNL.0000000000004765
        • Thompson A.J.
        • Banwell B.L.
        • Barkhof F.
        • et al.
        Diagnosis of multiple sclerosis: 2017 revisions of the McDonald criteria.
        Lancet Neurol. 2018; 17: 162-173
        • Tobin W.O.
        • Guo Y.
        • Krecke K.N.
        • et al.
        Diagnostic criteria for chronic lymphocytic inflammation with pontine perivascular enhancement responsive to steroids (CLIPPERS).
        Brain. 2017; 140: 2415-2425https://doi.org/10.1093/brain/awx200
        • Zalewski N.L.
        • Rabinstein A.A.
        • Krecke K.N.
        • et al.
        Characteristics of spontaneous spinal cord infarction and proposed diagnostic criteria.
        JAMA Neurol. 2019; 76: 56-63https://doi.org/10.1001/jamaneurol.2018.2734
        • Wingerchuk D.M.
        • Banwell B.
        • Bennett J.L.
        • et al.
        International consensus diagnostic criteria for neuromyelitis optica spectrum disorders.
        Neurology. 2015; 85: 177-189https://doi.org/10.1212/WNL.0000000000001729
        • Murphy O.C.
        • Messacar K.
        • Benson L.
        • et al.
        Acute flaccid myelitis: cause, diagnosis, and management.
        Lancet. 2021; 397: 334-346https://doi.org/10.1016/S0140-6736(20)32723-9
        • Murphy O.C.
        • Mukharesh L.
        • Salazar-Camelo A.
        • Pardo C.A.
        • Newsome S.D.
        Early factors associated with later conversion to multiple sclerosis in patients presenting with isolated myelitis.
        J. Neurol. Neurosurg. Psychiatry. 2021; 92: 831-838https://doi.org/10.1136/jnnp-2020-325274
        • Murphy O.C.
        • Salazar-Camelo A.
        • Jimenez J.A.
        • et al.
        Clinical and MRI phenotypes of sarcoidosis-associated myelopathy.
        Neurol. Neuroimmunol. Neuroinflamm. 2020; 7e722https://doi.org/10.1212/NXI.0000000000000722
        • Murphy O.C.
        • Hedjoudje A.
        • Salazar-Camelo A.
        • Pardo C.A.
        • Gailloud P.
        Clinical characteristics, misdiagnosis and outcomes of patients with low-flow spinal arteriovenous fistulas.
        J. Neurol. Sci. 2020; 413116863https://doi.org/10.1016/j.jns.2020.116863
        • Cerqueira J.J.
        • Compston D.A.S.
        • Geraldes R.
        • et al.
        Time matters in multiple sclerosis: can early treatment and long-term follow-up ensure everyone benefits from the latest advances in multiple sclerosis?.
        J. Neurol. Neurosurg. Psychiatry. 2018; 89: 844-850https://doi.org/10.1136/jnnp-2017-317509
        • O’Connor P.
        • Goodman A.
        • Kappos L.
        • et al.
        Long-term safety and effectiveness of natalizumab redosing and treatment in the STRATA MS study.
        Neurology. 2014; 83: 78-86https://doi.org/10.1212/WNL.0000000000000541
        • Goodin D.S.
        • Reder A.T.
        • Ebers G.C.
        • et al.
        Survival in MS: a randomized cohort study 21 years after the start of the pivotal IFNβ-1b trial.
        Neurology. 2012; 78: 1315-1322https://doi.org/10.1212/WNL.0b013e3182535cf6
        • Ullman N.
        • Gregg L.
        • Becker D.
        • Pardo C.
        • Gailloud P.
        Anterior disco-osteo-arterial conflict as a cause of intersegmental arterial flow impairment and spinal cord ischemia.
        Neuroradiology. 2016; 58: 1109-1115https://doi.org/10.1007/s00234-016-1739-1
        • Murphy O.C.
        • Gailloud P.
        • Newsome S.D.
        Spinal claudication secondary to anterior disco-osteo-arterial conflict and mimicking stiff person syndrome.
        JAMA Neurol. 2019; https://doi.org/10.1001/jamaneurol.2019.1007
        • Gailloud P.
        • Ponti A.
        • Gregg L.
        • Pardo C.A.
        • Fasel J.H.D.
        Focal compression of the upper left thoracic intersegmental arteries as a potential cause of spinal cord ischemia.
        AJNR Am. J. Neuroradiol. 2014; 35: 1226-1231https://doi.org/10.3174/ajnr.A3833
        • Flanagan E.P.
        • Krecke K.N.
        • Marsh R.W.
        • Giannini C.
        • Keegan B.M.
        • Weinshenker B.G.
        Specific pattern of gadolinium enhancement in spondylotic myelopathy.
        Ann. Neurol. 2014; 76: 54-65https://doi.org/10.1002/ana.24184
        • Cunha B.A.
        The master clinician’s approach to diagnostic reasoning.
        Am. J. Med. 2017; 130: 5-7https://doi.org/10.1016/j.amjmed.2016.07.024
        • Arkema E.V.
        • Cozier Y.C.
        Epidemiology of sarcoidosis: current findings and future directions.
        Ther. Adv. Chronic Dis. 2018; 9: 227-240https://doi.org/10.1177/2040622318790197
        • Fritz D.
        • van de Beek D.
        • Brouwer M.C.
        Clinical features, treatment and outcome in neurosarcoidosis: systematic review and meta-analysis.
        BMC Neurol. 2016; 16: 220https://doi.org/10.1186/s12883-016-0741-x
        • Pandit L.
        • Asgari N.
        • Apiwattanakul M.
        • et al.
        Demographic and clinical features of neuromyelitis optica: a review.
        Mult. Scler. 2015; 21: 845-853https://doi.org/10.1177/1352458515572406
        • Robertson C.E.
        • Brown R.D.
        • Wijdicks E.F.M.
        • Rabinstein A.A.
        Recovery after spinal cord infarcts: long-term outcome in 115 patients.
        Neurology. 2012; 78: 114-121https://doi.org/10.1212/WNL.0b013e31823efc93
        • Cheng M.
        • Lyu R.
        • Chang Y.
        • et al.
        Spinal cord infarction in chinese patients.
        CED. 2008; 26: 502-508https://doi.org/10.1159/000155988
        • Naess H.
        • Romi F.
        Comparing patients with spinal cord infarction and cerebral infarction: clinical characteristics, and short-term outcome.
        Vasc. Health Risk Manag. 2011; 7: 497-502https://doi.org/10.2147/VHRM.S22950
        • Nagata K.
        • Tanaka Y.
        • Kanai H.
        • Oshima Y.
        Acute complete paraplegia of 8-year-old girl caused by spinal cord infarction following minor trauma complicated with longitudinal signal change of spinal cord.
        Eur. Spine J. 2017; 26: 1432-1435https://doi.org/10.1007/s00586-017-4995-9
        • Spencer S.P.
        • Brock T.D.
        • Matthews R.R.
        • Stevens W.K.
        Three unique presentations of atraumatic spinal cord infarction in the pediatric emergency department.
        Pediatr. Emerg. Care. 2014; 30: 354-357https://doi.org/10.1097/PEC.0000000000000130
        • Reisner A.
        • Gary M.F.
        • Chern J.J.
        • Grattan-Smith J.D.
        Spinal cord infarction following minor trauma in children: fibrocartilaginous embolism as a putative cause.
        J. Neurosurg. Pediatr. 2013; 11: 445-450https://doi.org/10.3171/2013.1.PEDS12382
        • Ganhão S.
        • Loureiro G.B.
        • Oliveira D.R.
        • Dos-Reis-Maia R.
        • Aguiar F.
        • Quental R.
        • Moura C.
        • Barreira J.L.
        • Rodrigues M.
        • Brito I.
        Two cases of ADA2 deficiency presenting as childhood polyarteritis nodosa: novel ADA2 variant, atypical CNS manifestations, and literature review.
        Clin. Rheumatol. 2020; 12: 3853-3860https://doi.org/10.1007/s10067-020-05210-4
        • Sechi E.
        • Krecke K.N.
        • Pittock S.J.
        • et al.
        Frequency and characteristics of MRI-negative myelitis associated with MOG autoantibodies.
        Mult. Scler. 2021; 27: 303-308https://doi.org/10.1177/1352458520907900
        • El Mekabaty A.
        • Pardo C.A.
        • Gailloud P.
        The yield of initial conventional MRI in 115 cases of angiographically confirmed spinal vascular malformations.
        J. Neurol. 2017; 264: 733-739https://doi.org/10.1007/s00415-017-8419-x
        • Asnafi S.
        • Morris P.P.
        • Sechi E.
        • et al.
        The frequency of longitudinally extensive transverse myelitis in MS: a population-based study.
        Mult. Scler. Relat. Disord. 2019; 37101487https://doi.org/10.1016/j.msard.2019.101487
        • Sepúlveda M.
        • Blanco Y.
        • Rovira A.
        • et al.
        Analysis of prognostic factors associated with longitudinally extensive transverse myelitis.
        Mult. Scler. 2013; 19: 742-748https://doi.org/10.1177/1352458512461968
        • Flanagan E.P.
        • Kaufmann T.J.
        • Krecke K.N.
        • et al.
        Discriminating long myelitis of neuromyelitis optica from sarcoidosis.
        Ann. Neurol. 2016; 79: 437-447https://doi.org/10.1002/ana.24582
        • Dubey D.
        • Pittock S.J.
        • Krecke K.N.
        • et al.
        Clinical, radiologic, and prognostic features of myelitis associated with myelin oligodendrocyte glycoprotein autoantibody.
        JAMA Neurol. 2019; 76: 301-309https://doi.org/10.1001/jamaneurol.2018.4053
        • Zalewski N.L.
        • Krecke K.N.
        • Weinshenker B.G.
        • Aksamit A.J.
        • Conway B.L.
        • McKeon A.
        • Flanagan E.P.
        Central canal enhancement and the trident sign in spinal cord sarcoidosis.
        Neurology. 2016; 87: 743-744https://doi.org/10.1212/WNL.0000000000002992
        • Zalewski N.L.
        • Rabinstein A.A.
        • Brinjikji W.
        • Kaufmann T.J.
        • Nasr D.
        • Ruff M.W.
        • Flanagan E.P.
        Unique gadolinium enhancement pattern in spinal dural arteriovenous fistulas.
        JAMA Neurol. 2018; 75: 1542-1545https://doi.org/10.1001/jamaneurol.2018.2605
        • Mustafa R.
        • Passe T.J.
        • Lopez-Chiriboga A.S.
        • et al.
        Utility of MRI enhancement pattern in myelopathies with longitudinally extensive T2 lesions.
        Neurol. Clin. Pract. 2021; 11: e601-e611https://doi.org/10.1212/CPJ.0000000000001036
        • Fritz D.
        • van de Beek D.
        • Brouwer M.C.
        • Booij J.
        Whole-body 18F-FDG PET-CT in the diagnosis of neurosarcoidosis.
        Mayo Clin. Proc. 2020; 95: 1082-1084https://doi.org/10.1016/j.mayocp.2020.01.032
        • Barreras P.
        • Heck D.
        • Greenberg B.
        • Wolinsky J.
        • Pardo C.A.
        • Gailloud P.
        Analysis of 30 spinal angiograms falsely reported as normal in 18 patients with subsequently documented spinal vascular malformations.
        AJNR Am. J. Neuroradiol. 2017; 38: 1814-1819https://doi.org/10.3174/ajnr.A5275
        • Kim K.H.
        • Kim S.H.
        • Hyun J.W.
        • Kim Y.
        • Park H.
        • Kim H.J.
        Seroprevalence of anti-myelin oligodendrocyte glycoprotein antibodies in adults with myelitis.
        Ann. Clin. Transl. Neurol. 2022 Aug 6; https://doi.org/10.1002/acn3.51642. Epub ahead of print