Abstracts

Rationale: Several studies have demonstrated an association between impaired verbal memory performance in epilepsy patients and interictal 18F-FDG-PET hypometabolism findings. We aimed to explore the real-world clinical utility of 18F-FDG-PET in predicting pre- and post-operative memory deficits in patients with mesial temporal lobe epilepsy (MTLE). Methods: We recruited 71 patients who underwent a Spencer’s-type anterior temporal lobe resection (ATLR) for drug-resistant MTLE. The hypometabolic region in each patient was identified with reference to 20 healthy controls (p <0.005). Partial correlations were used to determine the relationships between the overall extent of interictal hypometabolism and pre-operative performance on the Rey Auditory Verbal Learning test and verbal Paired Associate Learning. The added value of 18F-FDG-PET in predicting memory outcomes over and above well-established predictors, such as the resection side in relation to hemispheric dominance and pre-operative verbal memory scores, was explored using hierarchical logistic regression. Results: There was no association between the overall extent of 18F-FDG-PET hypometabolism and the pre-operative verbal memory scores (p=0.2). Patients undergoing left ATLR, and those with higher pre-operative verbal memory scores were at a greater risk of post-operative memory decline (p<0.05). The pre-operative 18F-FDG-PET metabolic patterns and the amount of hypometabolism resected had no added predictive value on memory outcomes, nor did the total temporal lobe resection volume estimated from the pre- and post-operative MRIs (p=0.4). Conclusions: The findings of our study demonstrate the limited real-world clinical utility of interictal 18F-FDG-PET in predicting pre- and postoperative memory function in patients undergoing an ATLR for drug-resistant MTLE. Funding: No funding