Abstracts

Rationale: Glucose transporter 1 deficiency syndrome (GLUT1-DS) is caused by impaired glucose transport across blood-brain barrier. It is diagnosed by low CSF glucose, impaired glucose uptake into erythrocytes, and SLC2A1 gene mutation. Although Neuroimaging findings are considered nonspecific, it has been reported that FDG-PET shows low uptake in thalami and cerebral cortex and relatively high uptake in basal ganglia. However it has not been clarified if the PET findings are useful to diagnose GLUT1-DS among patients with epilepsy and unknown etiology. The purpose of this study is to establish an objective method of PET studies that is helpful in the diagnosis of GLUT1-DS. Methods: We performed FDG-PET in consecutive 7 patients (mean age 12 years, range 3 to 23 years) with GLUT1-DS. PET was compared with those of 45 controls (mean age 10 years, range 2 to 21 years) by using SPM12 software (Wellcome Trust Centre for Neuroimaging, London, UK). The controls had epilepsy of unknown etiology, and MRI and PET were normal on visual inspection. We also measured ratio of mean radioactivity in lenticular nuclei and thalami (LN/Thal ratio) in each patient and control by using regions of interest generated from WFU PickAtlas. To adjust the difference of the ratio by age, we derived linear regression line from the age and the ratios of 45 controls. From the obtained regression equation: y = ax + b (x: age, y: ratio), we calculated age-corrected ratio as: (ratio of each subject) - a*(age of each subject). We determined sensitivity and specificity of the age-corrected ratio for the differential diagnosis of GLUT1-DS and controls.  Results: SPM showed significantly decreased uptake in bilateral thalami and increased uptake in bilateral basal ganglia in patients with GLUT1-DS. The mean of age-corrected LN/Thal ratio was 1.78 ± 0.15 in patients and 1.37 ± 0.08 in controls. In ROC curve analysis, AUC was 0.994 and best cut-off value was 1.54. With the cut-off value of 1.54, sensitivity of age-corrected LN/Thal ratio to detect patients with GLUT1-DS was 100% and specificity was 98%. Positive predictive value of patients with GLUT1-DS was 88% and negative predictive value was 100%. Conclusions: Age-corrected LN/Thal ratio of FDG-PET is highly useful to detect patients with GLUT1-DS among patients with epilepsy of unknown etiology. The unique pattern of glucose metabolism on PET is supposed to be caused by expression of glucose transporters and local glucose metabolism in human brain from fetal period, and abnormal thalamo-cortical networks. Funding: The study was supported by Japan Epilepsy Research Foundation (JERF).