Neuromyelitis optica spectrum disorder and autoimmune glial fibrillary acidic protein astrocytopathy overlap syndrome mimicking a pontine mass: The utility of brainstem biopsy

Abstract

Objective

Diffuse intrinsic pontine gliomas (DIPG) are high‐grade tumors with a dismal prognosis and are classically diagnosed by radiologic features. DIPG is a critical differential consideration for a pediatric patient presenting with an infiltrative brainstem mass. However, inflammatory and infectious etiologies must also be considered, especially in individuals with atypical radiographic features. In a carefully selected clinical scenario, biopsy can be employed to quickly diagnose and direct treatment for patients with brainstem masses.

Results

This 10‐year‐old girl presented with acute onset of dysarthria, ataxia, left‐sided weakness, hypertonicity, and dysmetria. Magnetic resonance imaging revealed an infiltrative pontine lesion with atypical features for that of DIPG or specific inflammatory disease. Due to rapid clinical deterioration, stereotactic brainstem biopsy was performed for diagnostic clarity and showed inflammation but no malignant cells. She was then treated for a presumed antibody‐mediated autoimmune etiology with evaluation later revealing neuromyelitis optica spectrum disorder (NMOSD) and autoimmune glial fibrillary acidic protein (GFAP) astrocytopathy overlap syndrome.

Interpretation

We present a novel example of pediatric NMOSD and autoimmune GFAP astrocytopathy overlap syndrome originally presenting as an infiltrative pontine mass. Our report highlights the safety and utility of brainstem biopsy for brainstem masses atypical for DIPG.

Introduction

Diffuse intrinsic pontine glioma (DIPG) is an aggressive tumor comprising 10%–15% of childhood brain tumors. ¹ Although DIPG has become further defined by the 2021 World Health Organization classification by molecular features, ² classically it is a clinical and radiologic diagnosis. Due to the risks of brainstem injury, biopsy was historically reserved for pontine lesions with imaging features not typical for DIPG ³ but is now increasingly offered to provide targets for therapies and clinical trials.

The radiologic features of pediatric DIPG can be heterogeneous, ⁴ and criteria for “atypical” radiologic features of DIPG are not standardized. ⁵ Importantly, lack of histopathologic and molecular characterization of atypically appearing DIPG may lead to misdiagnosis with significant prognostic and therapeutic implications. ⁶ Thus, we describe a pediatric patient presenting with an infiltrative brainstem mass with an imaging differential including DIPG with atypical features wherein brainstem biopsy quickly ruled out malignancy and redirected clinical suspicion toward an alternative autoimmune etiology. This allowed for prompt initiation of immunotherapy, resulting in an excellent clinical outcome for an eventual diagnosis of neuromyelitis optica spectrum disorder (NMOSD) and autoimmune glial fibrillary acidic protein (GFAP) astrocytopathy overlap syndrome.

Patient Presentation

This 10‐year‐old girl with no significant past medical history presented with 5 days of emotional lability with frequent crying episodes, fatigue, and unsteady gait without any preceding illness. Her initial examination was notable for dysarthria, left arm pronator drift and impaired rapid alternating movements, sustained left ankle clonus, dysmetria of the left arm and left leg, and ataxic gait. Magnetic resonance imaging (MRI) of the brain with and without contrast was notable for a T2 hyperintense lesion infiltrating the near entire pons with extension to the right middle cerebellar peduncle without contrast enhancement or diffusion restriction (Figure 1A–C). Magnetic resonance spectroscopy (MRS) detected mild elevation of choline and a decrease in N‐acetylaspartate (NA) peaks (Figure 1D,E). She also had nonspecific T2 hyperintense lesions in her cerebral white matter. The radiologic differential favored an inflammatory pontine lesion, though the location and infiltrative appearance were atypical for this. A neoplasm, particularly DIPG, could not be excluded based on radiographic appearance. The patient immigrated from the Democratic Republic of Congo 8 years prior to presentation, so bacterial and parasitic infection leading to a pontine encephalitis remained on the differential, although this was considered less likely given the lack of temporally relevant exposures.

Figure 1.

Magnetic resonance imaging showing an infiltrative T2 hyperintense mass (A) without associated enhancement on postcontrast T1 sequences (B) or diffusion restriction on ADC (apparent diffusion coefficient) sequences (C). Magnetic resonance spectroscopy (voxel target illustrated in (D)) showed slight elevation in choline (Ch) peak and a mild decrease in N‐acetylaspartate (NA) peak (E).

Further imaging with an MRI whole spine with and without contrast was unremarkable. The patient then underwent a lumbar puncture, and cerebrospinal fluid (CSF) was notable for pleocytosis (19 white blood cells) with a lymphocytic predominance. There were no CSF unique oligoclonal bands, and the IgG index was mildly elevated at 0.71. To investigate for an autoimmune antibody, the Pediatric Autoimmune Encephalopathy/CNS Disorder Evaluation, Spinal Fluid (PCDEC) and Pediatric Autoimmune Encephalopathy/CNS Disorder Evaluation, Serum (PCDES) panels were sent to Mayo Clinic Laboratories, with an anticipated turnaround time of 2 weeks. Infectious evaluation thus far, including CSF meningitis/encephalitis polymerase chain reaction (PCR) panel, Lyme antibody titers, HIV‐1/2 antibody titers, and HSV‐1/2 PCR, returned negative. Four days after her initial presentation and despite treatment with dexamethasone 2 mg twice daily, the patient's neurological examination declined with left facial weakness and right lateral gaze palsy. Autoimmune antibodies were pending. Escalation to high‐dose intravenous methylprednisolone were considered but risked clinical worsening if the etiology was infectious. Additionally, a positive treatment response to steroids would not necessarily rule out a malignancy or provide sufficient diagnostic clarity to guide further treatment. Due to clinical deterioration and ongoing equipoise regarding whether the lesion was inflammatory, neoplastic, or infectious, she underwent brain biopsy the next day.

The neurosurgical team performed an MRI‐based stereotactic brainstem biopsy through a retrosigmoid burr hole, through the middle cerebellar peduncle. Pathology demonstrated perivascular chronic inflammation with increased microglial cells as well as foci of vasculitis in the brainstem sample (Figure 2). Given the absence of both tumor cells and pathologic evidence of infection on the biopsy, the patient was initiated on immunotherapy for a presumed antibody‐mediated autoimmune process with high‐dose methylprednisolone 30 mg/kg daily for 5 days and five sessions of plasmapheresis. After 1 day of methylprednisolone and just prior to initiating the plasmapheresis, she had worsening bulbar signs and respiratory deterioration and was transferred to the pediatric intensive care unit in anticipation of intubation for impending respiratory failure with rising CO₂. However, she responded quickly to plasmapheresis with improved bulbar, truncal, and appendicular strength as well as alertness and speech. The PCDEC and PCDES panels returned 4 days after biopsy with positive NMO antibody (AQP4‐IgG) via serum fluorescence‐activated cell sorting assay (titer 1:10 000) and positive GFAP antibody (GFAP‐IgG) via CSF immunofluorescence assay and cell‐based assay (titer 1:32). The serum GFAP immunofluorescence assay returned negative. Thus, the patient met clinical diagnostic criteria for NMOSD due to AQP4‐IgG positivity with acute brainstem syndrome as well as a likely diagnosis of GFAP astrocytopathy due to GFAP‐IgG positivity in the CSF with clinical features of encephalitis. She was given further immunotherapy with IVIG 2 g/kg and the first induction dose of rituximab 750 mg/m². Her neurological examination improved throughout admission, and she was discharged to acute inpatient rehabilitation.

Figure 2.

Hematoxylin and eosin–stained section (×200 magnification) showing dense perivascular chronic inflammation and damage to vascular wall (A) and smooth muscle actin immunohistochemical stain (×200 magnification) highlighting focal damage to vessel (B).

One week after discharge, she returned to the hospital due to febrile neutropenia and at the time had complete B‐cell depletion, so the second induction dose of rituximab was deferred. At her 1‐month follow‐up appointment, she had mild ataxic dysarthria, left‐sided dysmetria, and impaired rapid alternating movements but had returned to close to her functional baseline. Posttreatment MRI brain obtained 3 months after her original scan demonstrated decreased diffuse T2 abnormal signal in the midbrain and pons without abnormal enhancement, and her neurological examination had normalized. The neutropenia was thought to be an idiosyncratic response to rituximab, and the decision was made to switch to satralizumab for maintenance immunotherapy.

Discussion

For lesions with classic radiographic features of DIPG, previous studies have supported MRI as sufficient in making the diagnosis and differentiating from alternative space‐occupying lesions. ⁷ , ⁸ Our patient highlights the importance of biopsy in individuals whose imaging features are unusual—in this case, the lack of significant mass effect and atypical findings on MRS. In a study of 12 pediatric patients with DIPG, baseline MRS showed significantly reduced peaks of NA, a marker of neuronal health, and choline, a marker of proliferation. ⁹ The MRS findings in our patient showed elevated choline and only slightly decreased NA peaks, which was unexpected for DIPG. In addition, the rapid onset and precipitous deterioration were unusual.

Older reviews reported permanent morbidity and mortality from brainstem stereotactic biopsy as high as 9% and 4%, respectively. ¹⁰ Several more recent studies using modern surgical techniques have reported much lower morbidity and mortality rates, demonstrating the safety of stereotactic biopsies in both the pediatric and adult populations. ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ In one large meta‐analysis of 735 patients who underwent stereotactic biopsy for pediatric brainstem tumors, only 0.6% experienced permanent morbidity and 0.6% experienced mortality. ¹⁰

There is a clear trend toward biopsy even in classically appearing DIPG, for a variety of reasons. Increasingly, molecular diagnostics are being leveraged for targeted therapies, given that standard treatment with radiation does not confer a significant survival benefit. ¹ , ¹⁶ Similar to our patient, although very unusual and with clinical suspicion also present, biopsies of brainstem masses can rapidly confirm nonmalignant inflammatory lesions, which require significantly different treatment strategies. In a study of 106 children with brainstem masses, stereotactic biopsy allowed for diagnosis of tuberculosis, toxoplasmosis, or encephalitis. ¹⁷

Immune‐mediated and inflammatory isolated brainstem lesions can be challenging to accurately diagnose given a large spectrum of clinical, laboratory, and radiologic features as well as significant lag time of antibody‐based testing. Of note, our patient had a CSF pleocytosis, which is described in many inflammatory etiologies but also reported in DIPG and other pediatric brain tumors. ¹⁸ , ¹⁹ Time to treatment can be critical in conditions like NMOSD, where time to rituximab initiation has been associated with disability scores. ²⁰ NMOSD is a distinct inflammatory CNS disorder associated with aquaporin‐4 immunoglobulin G (AQP4‐IgG) antibodies that bind to astrocytes and commonly presents with clinical characteristics of optic neuritis, myelitis, area postrema syndrome, brainstem syndrome, diencephalic syndrome, or brain lesions. ²¹ Without a known positive AQP4‐IgG result, diagnostic criteria for NMOSD requires at least two core clinical characteristics, ²¹ and our patient fulfilled only one of brainstem syndrome, limiting our diagnostic certainty prior to antibody testing returning. In our patient, the decision to initiate plasmapheresis prior to antibody confirmation enabled her to avoid intubation and even worse clinical deterioration.

Interestingly, our patient's clinical presentation was also consistent with autoimmune GFAP astrocytopathy, as her encephalopathy was thought to be out of proportion to her brainstem lesion. GFAP astrocytopathy is a more recently identified CNS autoimmune disorder associated with CSF antibodies against the astrocytic cytoplasmic intermediate filament protein, typically resulting in a meningoencephalomyelitis. ²² This disease entity is better characterized in the adult population; however, case series of pediatric patients with GFAP‐IgG reported up to 93.75% also presenting with meningoencephalitis. ²³ , ²⁴ , ²⁵ There is a range of brain lesions associated with GFAP astrocytopathy, with linear perivascular enhancement in the basal ganglia and thalamus as the most commonly described. ²⁶ Although our patient lacked these typical radiologic features, her GFAP‐IgG positivity in the CSF is highly specific for autoimmune GFAP astrocytopathy, with reports of only 0.9% of GFAP‐IgG positivity by cell‐based assay in CSF controls. ²⁷

Patients with GFAP astrocytopathy commonly have concurrent autoantibodies, but simultaneous AQP4‐IgG and GFAP‐IgG positivity is reported less frequently. ²² In a large case series of adult patients with GFAP‐IgG, 40% had one or more coexisting serum or CSF antibodies, with 10% being AQP4‐IgG. ²⁷ Overlap syndromes with clinical features of both NMOSD and GFAP astrocytopathy have been rarely reported. However, a case series of 10 adult patients with GFAP astrocytopathy overlap syndromes identified five patients with AQP4‐IgG who fulfilled diagnostic criteria for NMOSD at time of presentation, ²⁸ suggesting these distinct immune‐mediated processes could occur simultaneously. One pediatric patient with GFAP astrocytopathy has been reported to have low‐titer AQP4‐IgG in the CSF, although the clinical picture was inconsistent with NMOSD. ²⁵ Thus, we believe our patient had a true overlap syndrome with both encephalitis and a brainstem lesion, which represents a novel characteristic of this case.

Although multiple studies report a high specificity and sensitivity with detected GFAP‐IgG in the CSF, ²² , ²⁷ , ²⁹ , ³⁰ there is no consensus on clinical diagnostic criteria for GFAP astrocytopathy, which can make overlap syndromes difficult to diagnose at time of presentation. Ultimately, the heterogenous nature of this condition was reflected in our patient as she presented with atypical imaging, symptoms, and a timeline that did not fit one condition. Even though tissue biopsy is rarely indicated in suspected antibody‐mediated immune conditions, the lack of diagnostic clarity and confounding clinical features in our patient necessitated timely brainstem biopsy to rule out malignancy and initiate immunomodulatory treatment.

Conclusions

In appropriate clinical scenarios, stereotactic biopsy has a favorable risk profile, provides diagnostic value, and can alter the treatment strategy and clinical outcome for pediatric patients with brainstem masses. Our patient demonstrates how biopsy played a critical role in properly guiding treatment in a clinically decompensating individual.

Author Contributions

Vivien X. Xie: Investigation; writing—original draft; writing—review and editing. Gilbert Vezina: Writing—review and editing. John S. Myseros: Writing—review and editing. Lakshmi Ramachandran Nair: Writing—review and editing. Elizabeth M. Wells: Conceptualization; writing—review and editing. Benjamin I. Siegel: Conceptualization; writing—review and editing; supervision.

Conflict of Interest

The authors declare no conflicts of interest.