Purpose: The aim of this 18 F-FDG PET study was to determine the diagnostic value of the cortex/striatum metabolic ratio in a large cohort of patients suffering from autoimmune encephalitis (AE) and to search for correlations with the course of the disease. Methods:We retrospectively collected clinical and paraclinical data of patients with AE, including brain 18 F-FDG PET/CT. Whole-brain statistical analysis was performed using SPM8 software after activity parametrization to the striatum in comparison to healthy subjects. The discriminative performance of this metabolic ratio was evaluated in patients with AE using receiver operating characteristic curves against 44 healthy subjects and a control group of 688 patients with MCI. Relationship between cortex/striatum metabolic ratios and clinical/paraclinical data was assessed using univariate and multivariate analysis.Results: Fifty-six patients with AE were included. In comparison to healthy subjects, voxel-based statistical analysis identified one large cluster (p-cluster<0.05, FWE corrected) of widespread decreased cortex/striatum ratio in patients with AE. The mean metabolic ratio was significantly lower for AE patients (1.16 ± 0.13) than for healthy subjects (1.39 ± 0.08; p < 0.001) and than for MCI patients (1.32 ± 0.11; p < 0.001). A ratio threshold of 1.23 allowed to detect AE patients with a sensitivity of 71 %, and a specificity of 82% against MCI patients and 98% against healthy subjects. A lower cortex/striatum metabolic ratio had a trend towards shorter delay before 18 F-FDG PET/CT (p = 0.07) in multivariate analysis. Conclusion:The decrease in the cortex/striatal metabolic ratio has a good early diagnostic performance for the differentiation of AE patients from controls.

No abstract

No abstract

Purpose: The aim was to determine the diagnostic value of the cortex/striatum metabolic ratio in 18F-FDG PET in a large cohort of patients suffering from dysimmune encephalitis (DE) and to search for clinical correlations with the course of the disease.Methods: We retrospectively collected complete clinical and paraclinical data of DE patients, including brain 18F-FDG PET/CT. Whole-brain statistical analysis was performed using SPM8 software after activity parametrization to the striatum and in comparison to healthy subjects. Conventional discriminant analysis between the DE group and controls was performed using cluster metabolic ratios. A correlation analysis between cluster metabolic ratios and clinical/paraclinical data was assessed.Results: Seventy-three patients with DE were included. In comparison to 44 controls, voxel-based statistical analysis identified one large cluster (p-voxel < 0.001 uncorrected; p-cluster<0.05, FWE corrected) of widespread decreased cortical metabolism relative to the striatum in DE patients. The mean parametrized cluster metabolic value was significantly lower for DE patients (1.06 ± 0.13) than for the control group (1.46 ± 0.08; p < 0.001). This cluster metabolic ratio correctly classified 97.4% of the individuals between patients with DE and healthy controls. Correlation analyses showed that a low cluster metabolic ratio was associated with higher risk of death (p = 0.04), the absence of autoantibodies (p = 0.05), and an increased delay between onset of symptoms and diagnostic (p = 0.01).Conclusion: The decrease in the cortex/striatal metabolic ratio has a good diagnostic performance for the differentiation of DE patients from controls and seems to provide prognostic information on the clinical course.