Brain 18F-fluorodeoxyglucose positron emission tomography/computed tomography in COVID-19–associated autoimmune encephalitis: a case report

Abstract

Coronavirus disease 2019 (COVID-19)–associated encephalitis is a rare but severe neurological complication. Conventional diagnostic tools, such as magnetic resonance imaging (MRI) and cerebrospinal fluid (CSF) studies, may yield normal results despite significant functional brain impairment. Brain positron emission tomography/computed tomography (PET/CT) with 18F-fluorodeoxyglucose (18F-FDG) has emerged as a valuable adjunctive imaging modality in such cases. We report the case of a 54-year-old woman with COVID-19 who presented with anosmia, ageusia, and altered behavior. Neurological examination revealed cognitive deficits, including bradyphrenia and mixed aphasia. Routine MRI and CSF analyses were unremarkable. However, brain 18F-FDG PET/CT revealed cortical hypometabolism in frontal and temporal lobes and cerebellar changes, consistent with immune-mediated encephalitis. The patient received immunosuppressive therapy with corticosteroids and intravenous immunoglobulin, leading to gradual clinical improvement and complete recovery at 3 months. This case highlights the diagnostic challenge of COVID-19–associated encephalitis when routine studies are inconclusive. PET/CT functional imaging can detect metabolic brain changes not apparent on structural imaging, supporting immune-mediated pathogenesis. Early recognition and immunotherapy may improve neurological outcomes. To conclude, brain 18F-FDG PET/CT is a useful diagnostic adjunct in COVID-19–related encephalitis with normal MRI and negative CSF, aiding timely diagnosis and treatment monitoring.

Introduction

Encephalitis is a potentially fatal complication of the central nervous system associated with coronavirus disease 2019 (COVID-19) infection. The neuroinvasive potential of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has been demonstrated by the presence of viral particles in endothelial cells of the frontal cortex in histological studies, suggesting direct viral impact on neural tissue [1]. Clinically, neurologic manifestations have been reported in more than half of patients hospitalized with COVID-19, including nonspecific symptoms such as headache, reduced consciousness, agitation, anosmia, and ageusia [2]. More specific neurological manifestations include cerebral infarcts, encephalitis, and white matter lesions detectable by neuroimaging [3].

Brain 18F-fluorodeoxyglucose (18F-FDG) positron emission tomography/computed tomography (PET/CT) has emerged as a valuable functional imaging tool in the evaluation of COVID-19–related neurological complications. 18F-FDG PET/CT studies reveal characteristic patterns of hypometabolism in frontal and temporal regions, as well as cerebellar vermis hypermetabolism in encephalitis cases when conventional magnetic resonance imaging (MRI) and cerebrospinal fluid (CSF) analyses remain normal [4,5]. These findings support immune-mediated mechanisms of neurological dysfunction in SARS-CoV-2 infection, complementing clinical and laboratory criteria for diagnosis [6,7].

In this report, we present a case of possible COVID-19–associated autoimmune encephalitis in which brain 18F-FDG PET/CT was instrumental in diagnosis, highlighting its role as an adjunct diagnostic modality in the context of a normal MRI and negative CSF studies.

Case Report

A 54-year-old woman with no previous medical or psychiatric history presented to the emergency department and reported mild rhinorrhea 4 days before the onset of anosmia and ageusia. As she was previously asymptomatic, these symptoms began 5 days prior to admission and coincided with a positive home rapid antigen test for SARS-CoV-2. Approximately 36 hours before admission, she developed neurological symptoms, including behavioral changes characterized by incoherent speech, impaired ability to manage everyday tasks, cognitive impairment, and insomnia. At presentation, the patient denied fever, cough, dyspnea, chest pain, or other respiratory symptoms. Vital signs were stable, with blood pressure of 141/70 mmHg, heart rate of 96 beats/min, oxygen saturation of 99% on room air, and tympanic temperature of 36.5 °C. At hospital presentation, nasopharyngeal SARS-CoV-2 antigen was positive. There were no known COVID-19 high risk exposures.

Neurological examination revealed temporospatial disorientation. Cranial nerve function was preserved as muscle strength and sensory modalities with no signs of dysmetria. Cognitive assessment showed bradyphrenia, hypoprosexia, anterograde amnesia, psychomotor agitation, emotional lability, mixed aphasia, impaired working memory, delayed recall, abstraction, acalculia, and mild ideomotor apraxia.

Initial diagnostic workup included a positive nasopharyngeal SARS-CoV-2 polymerase chain reaction (PCR) with a cycle threshold value of 24, consistent with a high viral load at presentation. CSF analysis revealed glucose of 75 mg/dL, proteins at 31.5 mg/dL, and a leukocyte count of 1/mm³ comprising 100% mononuclear cells. There were no erythrocytes, and Gram staining and cultures were negative. The patient’s FilmArray PCR panel (BioFire Diagnostics) for meningitis/encephalitis (including herpes simplex virus, varicella-zoster virus, enterovirus, human herpesvirus 6, Streptococcus pneumoniae, and Neisseria meningitidis) and PCR for SARS-CoV-2 in CSF were both negative. Thyroid function was normal (thyroid stimulating hormone, 0.89 µIU/mL; free thyroxine, 1.31 ng/dL), metabolic and toxicology studies (including renal and liver function, electrolytes, glucose, vitamin B12, folate, and urine drug screening) were unremarkable, and an autoimmune/paraneoplastic antibody panel (including onconeuronal blot, N-methyl-ᴅ-aspartate receptor, leucine-rich glioma-inactivated 1/contactin-associated protein-like 2, α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor, gamma-aminobutyric acid type B receptor, dipeptidyl-peptidase-like protein 6) was negative. Although we did not perform dedicated whole‑body imaging (e.g., chest-abdomen CT or whole‑body FDG‑PET/CT) to screen for underlying malignancy, the absence of systemic symptoms, unremarkable routine laboratory tests, and normal brain MRI made paraneoplastic etiology clinically unlikely in this case.

Brain MRI with intravenous contrast demonstrated normal structures in the posterior fossa, brainstem, and supratentorial white matter without lesions or mass effect. Electroencephalography showed background alpha activity around 8 Hz with low amplitude and reduced reactivity, without clear theta/delta slowing or epileptiform discharges. Brain 18F-FDG PET/CT (Figures 1 and 2) revealed bilateral cortical hypometabolism in the superior medial frontal and mesial temporal lobes, mild hypometabolism in the midbrain, and bilateral cerebellar cortical hypometabolism with preservation of vermis metabolism. Other brain regions, including parietal and occipital lobes, showed normal metabolic activity. Semiquantitative Z-score analysis using a normal brain database confirmed marked right and left superior frontal hypometabolism (standard deviation [SD] of −5.1 and −4.1, respectively) and mild right mesial temporal hypometabolism (SD, −3.3), supporting a network-level encephalitic process.

Figure 1. Sagittal slices of the brain 18F-fluorodeoxyglucose positron emission tomography/computed tomography.

The scans show marked hypometabolism in the bilateral superior medial frontal cortex and mesial temporal regions with mild hypometabolism in the midbrain area. The cerebellar hemispheres have reduced uptake, while the vermis is preserved, suggesting regional functional impairment associated with coronavirus disease 2019-related encephalitis.

Figure 2. Transverse positron emission tomography/computed tomography slices of the brain reveal bilateral cortical hypometabolism involving the frontal and temporal lobes and cerebellar cortex.

The occipital lobes, posterior cingulate gyrus, precuneus, striatum, and thalamus demonstrate preserved metabolic activity, emphasizing the selectivity of neuronal dysfunction in this case.

Diagnosis based on clinical presentation and imaging findings was consistent with immune-mediated encephalitis according to Graus et al.’s criteria [6] in the setting of COVID-19 infection. Symptomatic treatment with trazodone 100 mg orally at night and quetiapine 150 mg/day (divided doses) was initiated for insomnia and behavioral symptoms. The patient received antiviral therapy with remdesivir, consisting of a 10-day course with a loading dose of 200 mg on day 1 followed by 100 mg once daily for 9 additional days. However, immunotherapy was initiated shortly after admission due to a lack of improvement. This included methylprednisolone 250 mg daily for 3 days, followed by intravenous immunoglobulin (IVIG) at 0.4 g/kg/day for 5 consecutive days. The patient showed gradual clinical improvement in both behavioral symptoms and memory function, with complete resolution observed at 3 months of follow-up.

This case report was conducted in accordance with ethical standards for research and publication. The authors confirm that all procedures involving human participants were performed in compliance with the ethical standards of the institutional research committee and with the 1964 Helsinki declaration and its later amendments. Informed consent was obtained from the patient for the publication of this case report and any accompanying images. The patient has understood the purpose of this publication and has given permission for the information to be shared while maintaining confidentiality and anonymity.

Discussion

Neurological symptoms, including anosmia and ageusia, have been reported in more than half of hospitalized COVID-19 patients [2,7]. Although anosmia and ageusia began only a few days before hospital admission in our patient, the actual SARS-CoV-2 infection may have occurred earlier, making a para- or postinfectious late immune-mediated encephalitic process more plausible in this case. Considering the brief episode of rhinorrhea reported by the patient 4 days before anosmia and ageusia, and the typical incubation period of SARS‑CoV‑2, it is plausible that the initial infection occurred at least several days before the onset of symptoms. If a subclinical infection is assumed to have started approximately one week before rhinorrhea, the total interval from viral exposure to the development of encephalitic symptoms (which appeared 36 hours before admission) may reasonably exceed 2 weeks. This estimated 2 to 3‑week window is consistent with a para‑ or postinfectious adaptive immune‑mediated process, in line with previous reports of COVID‑19–associated immune encephalitis [3,5]. In the context of COVID-19, neurological manifestations range from nonspecific symptoms to encephalitis and other severe syndromes, and our case exemplifies this spectrum by presenting an immune-mediated encephalitic picture with normal MRI and CSF.

Although COVID-19–related delirium is common, several features in this case argue against purely toxic metabolic encephalopathy or isolated delirium: the subacute onset of multidomain cognitive dysfunction (memory, language, executive functions), the absence of metabolic, toxic, or structural abnormalities, and the presence of a reproducible pattern of frontal-temporal and cerebellar hypometabolism on 18F-FDG PET/CT. Furthermore, the progressive and sustained improvement after immunotherapy is more consistent with an inflammatory/autoimmune encephalitic process than with transient delirium. No repeat CSF analysis or follow-up 18F-FDG PET/CT was performed in this case, which limits our ability to correlate longitudinal metabolic changes with clinical evolution. Future reports that include serial imaging and CSF studies will be important to better characterize disease dynamics.

Delorme et al. [5] described COVID-19 patients with acute cognitive disorders and a similar pattern of frontal hypometabolism and cerebellar vermis involvement on 18F-FDG PET/CT, with clinical improvement after immunotherapy. Brain 18F-FDG PET/CT has shown characteristic but nonspecific patterns in autoimmune encephalitis, most commonly involving frontotemporal cortical hypometabolism, with frequent cerebellar involvement, even when MRI is normal. Similar configurations of frontotemporal and cerebellar hypometabolism have been reported in COVID-19–related immune-mediated encephalitis and other autoimmune encephalitis syndromes, and the pattern observed in our patient is therefore consistent with previously described cases, although it is not pathognomonic.

Our discussion of this COVID-19–associated encephalitis case highlights the complex interplay of mechanisms and the diagnostic role of advanced imaging. SARS-CoV-2’s neuroinvasive potential has been supported by detection of viral particles in brain endothelial cells [1], but encephalitis often appears more related to immune-mediated injury than direct viral infection. A cytokine storm marked by interleukin-6, tumor necrosis factor alpha, and additional pro-inflammatory mediators can damage the blood-brain barrier, recruit glial cells, and lead to encephalopathy without direct viral presence in the brain parenchyma [8]. This inflammatory cascade underlies the clinical and imaging findings in our patient, who showed no SARS-CoV-2 in CSF but had typical encephalitic symptoms. In our patient, neurological symptoms developed several days after the onset of anosmia and ageusia, which were preceded by a brief episode of mild rhinorrhea. This temporal pattern, together with the negative SARS-CoV-2 PCR in CSF and the complete clinical response to immunotherapy, is more consistent with an immune-mediated encephalitic process than with direct viral neuroinvasion.

A comprehensive differential diagnosis was considered. The patient’s thyroid function tests were normal, and Hashimoto encephalopathy can therefore be reasonably excluded. Metabolic and toxic causes of encephalopathy were also ruled out. An extensive autoimmune and paraneoplastic antibody panel was negative. This case fulfills Graus’s criteria for possible autoimmune encephalitis [6], based on multidomain cognitive and behavioral symptoms, compatible neuropsychiatric features, supportive 18F-FDG PET/CT abnormalities, and reasonable exclusion of alternative diagnoses, despite negative neuronal antibodies. Neurosyphilis was considered in the differential diagnosis, but there were no historical risk factors, no systemic or dermatological features suggestive of syphilis, and CSF protein and cell counts were within normal or near‑normal limits, making this diagnosis unlikely. Syphilis serology was not obtained and represents a minor limitation of the work‑up. Although the clinical context and combination of laboratory, neuroimaging, and therapeutic response strongly support COVID-19–associated immune-mediated encephalitis, other autoimmune or inflammatory etiologies remain a theoretical possibility, and should be kept in mind when interpreting a single case observation. In this context, our patient’s positive SARS-CoV-2 test was interpreted within a broad differential framework to minimize diagnostic anchoring and over attribution of findings to COVID-19.

Brain 18F-FDG PET/CT functional imaging proved critical for revealing cortical hypometabolism of the frontal and temporal lobes and cerebellar involvement despite normal MRI and CSF analyses. Similar PET hypometabolic patterns have been reported in COVID-19–associated encephalitis and long COVID cognitive impairment, supporting network-based functional disruption likely due to immune-mediated mechanisms [4,5,9]. PET imaging thus emerges as a valuable complementary tool to diagnose and monitor these neurological manifestations. We consider the 18F-FDG PET/CT findings as supportive functional evidence in the setting of normal structural MRI.

In terms of management, our therapeutic strategy combined antiviral treatment with remdesivir and immunotherapy with corticosteroids and IVIG. Remdesivir was administered for 10 days in line with institutional practice for hospitalized COVID-19 cases, while methylprednisolone was given at a dose of 250 mg/day for 3 days, followed by IVIG 0.4 g/kg/day for 5 days in line with regimens used in autoimmune encephalitis [10]. This lower than standard dose of methylprednisolone was deliberately chosen to balance the need for immunomodulation against the risk of enhancing viral replication in the setting of active SARS-CoV-2 infection. The optimal corticosteroid dose in severe COVID-19 remains uncertain and higher doses have not consistently demonstrated superior clinical benefit over standard regimens in randomized trials and meta-analyses [11]. The favorable clinical evolution and complete recovery at 3 months support the effectiveness and safety of this combined antiviral and immunomodulatory approach in this setting. Clinical improvement further supports immune-mediated pathogenesis rather than direct viral cytopathy. Emerging reports of autoimmune encephalitis cases post-COVID-19 reinforce this hypothesis and highlight the importance of early recognition and immunotherapy [12,13].