Statistical parametric mapping analysis of FDG- PET studies in autoimmune encephalitis

Early initiation of therapy in Autoimmune encephalitis (AIE) has good results. Timely diagnosis is therefore crucial but clinical diagnosis can be delayed for want of antibody results or if MRI is non-contributory. Metabolic patterns on F-18 Fluorodeoxyglucose positron emission tomography (FDG- PET) have proved useful to support the clinical diagnosis of AIE. Routinely visual analysis is performed but the confidence of interpretation can be enhanced by the use of more objective computer supported techniques like Statistical Parametric Mapping (SPM). We undertook this analysis to evaluate concordance between visual and SPM analysis of FDG-PET scans in AIE. In this retrospective study, a total of 39 patients (age range of 9-75 years) with clinically suspicion of AIE, undergoing FDG-PET studies were included for analysis. Each scan was interpreted visually. This was followed by SPM analysis on a MATLAB platform. 28 normal controls were included from our normal brain database for SPM analysis. The metabolic pattern (hypermetabolism and hypometabolism) demonstrated on visual analysis and SPM was compared for cortical and subcortical regions. Basal ganglia and mesial temporal lobe hypermetabolism was detected in 17 and 9 patients respectively on SPM and showed almost perfect and fair concordance with visual analysis. Mesial temporal hypermetabolism was better demonstrated on SPM. Hypometabolism involving the frontal, parietal, occipital or temporal cortices showed moderate to substantial concordance between SPM and visual analysis. SPM proved to be better than visual analysis for mesial temporal hypermetabolism as well as for demonstrating cortical hypometabolism in AIE, thereby supporting the utility of computer-supported reading with SPM.