Abstracts

Rationale: [18F]-Fluorodeoxyglucose positron emission tomography ([18F]-FDG-PET) is a common tool for epileptogenic zone identification, which presented an area with hypometabolism. Extratemporal hypometabolism was observed in temporal lobe epilepsy (TLE) patients, and thought as the network disorder. Besides, epilepsy patients usually have declined cognitive function, which might be related to both metabolic and structural changes of the cortexes. However, whether the deep grey matter, thalamus, related to cognitive performances in epilepsy patients had not been identified.

Methods: We retrospectively collected the TLE patients, who received brain MRI and PET for presurgical survey between June 2020 and Feburary 2022 in Veteran Taichung General Hospital. There were 15 right-TLE (7 female, 8 male) and 21 left-TLE (8 female, 13 male) cases enrolled. The PET images were acquired by the GE Healthcare Discovery STE PET/CT system, while MR images were collected by Simens Magnetom 1.5T system. The grey matters were segmented and then registered to the PET images. The whole brain grey matter was chosen as a reference for intensity normalization. The SUV map was spatially smoothed with Gaussian smoothing kernel of 6 mm full width at half maximum. The volumes and the SUVs of each thalamic subnuclei (anterior, lateral, ventral, intralaminar, medial, posterior) were calculated. The asymmetrical index [(L-R)/(L+R)] was used to study the relationship between the metabolic/structural images and five indices of cognition (Full-Scale IQ: FSIQ; Verbal IQ: VIQ; Performance IQ: PIQ; Working Memory Index: WMI; Processing Speed Index: PSI). A P value less than 0.01 was thought as statistically significant.

Results: Five aspects of intelligence quotient did not show obvious difference between left-TLE and right-TLE. In left-TLE, only the asymmetrical index of anterior nucleus volume was related to VIQ. PSI was negatively related to the asymmetrical index of lateral, intralaminar, and the whole thalamic SUV. The longer seizure duration, the smaller left subnuclei volume than the right corresponding areas. In right-TLE, only PSI was correlated with the asymmetrical index of anterior nucleus volume. However, the cognitive function was not related to any SUV asymmetrical index. The SUV of the right thalamic subfields is relatively stronger than the left corresponding site in the cases with longer seizure duration.

Conclusions: Our study suggested thalamus participate in the network of cognitive function of TLE patients, presenting structural and metabolic changes. Seizure duration was related to the thalamic structural changes of left-TLE and metabolic changes of right-TLE. It suggests left-TLE induces ipsilateral thalamus atrophy, while right-TLE relatively enhance the ipsilateral thalamus metabolism. Among all cognitive aspects, PSI is well reflected by the imaging. PSI is maintained in left-TLE when contralateral thalamic SUV is preserved. In contrast, processing speed is kept in right-TLE when contralateral anterior thalamic nuclei volume is relatively larger.

Funding: None