The Detection of Invisible Abnormal Metabolism in the FDG-PET Images of Patients With Anti-LGI1 Encephalitis by Machine Learning

Cortical metabolic characteristics of anti-leucine-rich glioma-inactivated 1 antibody encephalitis based on 18F-FDG PET

Purpose

A general glucose metabolism pattern is observed in patients with anti-leucine-rich glioma-inactivated 1 (LGI1) antibody encephalitis; however, it is unclear whether further subregional metabolic differences exist. Therefore, the present study aimed to conduct an in-depth exploration of the features of glucose metabolism within specific brain areas using 18F-fluorodeoxyglucose positron emission tomography (18F-FDG PET).

Materials and methods

This retrospective study enrolled thirteen patients confirmed with LGI1 antibody encephalitis who were admitted to Beijing Tiantan Hospital from June 2021 to September 2022. All patients underwent 18F-FDG PET before initiating clinical treatment. Changes in glucose metabolism in specific brain areas were analyzed using Cortex ID software. The laterality of 18F-FDG uptake was assessed, and differences in specific brain areas were compared using paired t-tests.

Results

Significant metabolic changes in at least one brain region in 11 out of 13 patients (84.6%) were revealed by semi-quantitative analysis (z-score > 2). A bilateral decrease in the 18F-FDG metabolic pattern was revealed in almost all brain regions of interest; in contrast, a hypermetabolic pattern was observed in the medial temporal region, with mean z-scores of 1.75 ± 3.27 and 2.36 ± 5.90 on the left and right sides, respectively (p = 0.497). In the prefrontal and temporal lobes, 18F-FDG metabolism was significantly lower in the lateral region than in the medial region on both sides. For the cingulate cortex, significant hypometabolism was also observed in the posterior part compared to the anterior counterpart on both the left (z-score: −1.20 ± 1.93 vs. −0.42 ± 1.18, respectively; p = 0.047) and right (z-score: −1.56 ± 1.96 vs. −0.33 ± 1.63, respectively; p = 0.001) sides. However, a significant difference in regional metabolism was observed only on the left side (p = 0.041).

Conclusion

An asymmetric 18F-FDG metabolic pattern exists in patients with anti-LGI1 encephalitis. Meanwhile, varied regional metabolic differences were revealed bilaterally in specific cerebral areas, which could be associated with the clinical manifestations.

Positron emission tomography in autoimmune encephalitis: Clinical implications and future directions

¹⁸ F-fluoro-deoxyglucose position emission tomography (¹⁸ F-FDG-PET) has been proven as a sensitive and reliable tool for diagnosis of autoimmune encephalitis (AE). More attention was paid to this kind of imaging because of the shortage of MRI, EEG, and CSF findings. FDG-PET has been assessed in a few small studies and case reports showing apparent abnormalities in cases where MRI does not. Here, we summarized the patterns (specific or not) in AE with different antibodies detected and the clinical outlook for the wide application of FDG-PET considering some limitations. Specific patterns based on antibody subtypes and clinical symptoms were critical for identifying suspicious AE, the most common of which was the anteroposterior gradient in anti- N -methyl- d -aspartate receptor (NMDAR) encephalitis and the medial temporal lobe hypermetabolism in limbic encephalitis. And the dynamic changes of metabolic presentations in different phases provided us the potential to inspect the evolution of AE and predict the functional outcomes. Except for the visual assessment, quantitative analysis was recently reported in some voxel-based studies of regions of interest, which suggested some clues of the future evaluation of metabolic abnormalities. Large prospective studies need to be conducted controlling the time from symptom onset to examination with the same standard of FDG-PET scanning.