FDG-PET/CT in autoimmune encephalitis: metabolic patterns, MRI correlation, and relevant paraclinical findings

Introduction: The diagnosis of autoimmune encephalitis (AE) remains challenging due to the lengthy time required for autoantibody testing and significant clinical overlap with other autoimmune and infectious diseases of the nervous system. In 2016, consensus diagnostic criteria were established to facilitate rapid diagnosis using neurological exam findings, EEG, MRI, and serum and CSF testing. However, studies have shown that FDG-PET/CT may be more sensitive than MRI in detecting abnormalities in AE.

Objectives: To determine how AE affects FDG uptake in the brain, and evaluate possible correlations between FDG-PET uptake and the different types of autoimmune encephalitis. To correlate brain FDG-PET abnormalities with brain MRI, EEG, and CSF used in the initial evaluation and assessment of this patient population.

Methods: Retrospective review identified 49 patients with clinical suspicion for AE who received both brain MRI and FDG-PET/CT imaging. Consensus diagnostic criteria by Graus et al. (2016) were used to identify patients with definite AE and specific subtypes. Each MRI and FDG-PET/CT study was evaluated in consensus by two radiologists. The brain FDG-PET/CT images were spatial and intensity normalized, and a statistical analysis comparing each patient’s image to a database of normals was performed. For each brain region, the Z-score was used to quantitatively assess differences in FDG uptake compared to a database of 43 healthy controls (19 females and 24 males, age range 41- 80 years).

Results: Seventeen patients met consensus criteria for AE and underwent FDG-PET/CT imaging at median 10 weeks of symptoms and median 4 days from MRI. The most common subtypes were limbic encephalitis (n=8) and anti-NMDA receptor encephalitis (n=5). CSF pleocytosis was found in 11/15 (73.3%) patients, while EEG abnormalities were present in 12/15 (80%) patients. FDG-PET/CT and MRI were both consistent with AE in 10/17 (58.8%) patients each. Either MRI or FDG-PET/CT findings supported the diagnosis of AE in 15/17 (88.2%) patients, while only 5/17 (29.4%) patients had supportive findings in both. Equal numbers of patients only had supportive imaging findings in one modality (5/17, 29.4% each). Most AE patients demonstrated mixed hypermetabolism and hypometabolism (6/10, 60%) on FDG-PET/CT, while fewer demonstrated only hypermetabolism or hypometabolism (2/10, 20%, each). Definite AE patients had significant hypermetabolism, given as (Z-score ± standard error, p value), in the temporal lobe (Z=2.425 ± 0.747, p=0.002) while the subtype of anti-NMDA receptor encephalitis patients demonstrated hypometabolism in the parietal (Z=-1.528 ± 0.749, p=0.044) and occipital (Z=-2.069 ± 0.475, p<0.0001) lobes.

Conclusions: Our results suggest that FDG-PET/CT was equal to MRI, and comparable to EEG and CSF studies in patients with criteria-defined AE. FDG-PET/CT most often demonstrated mixed hypometabolism and hypermetabolism in AE patients, while significant hypometabolism was seen in the parietal and occipital lobes of anti-NMDA receptor encephalitis.