Individual 18F-FDG PET and arterial spin labeling coupling based on simultaneous PET/MRI predicting anti-seizure medication response in temporal lobe epilepsy patients.

AIMS: Despite advances in epilepsy management, reliable prediction of anti-seizure medication (ASM) response remains an unmet need in temporal lobe epilepsy (TLE). Emerging evidence indicates that neurovascular decoupling (NVD), defined as the mismatch between neuronal activity and cerebral hemodynamics, may function as an imaging biomarker that reflects dysfunction within the epileptogenic network. This prospective study uses simultaneous 18 F-FDG PET/MRI to characterize patterns of metabolic-perfusion dissociation and to evaluate their value in predicting ASM outcomes in TLE. METHODS: This prospective cohort enrolled 52 unilateral TLE patients before initiation of standardized ASM regimens. Simultaneous 18 F-FDG PET/MRI was conducted before treatment to measure cerebral glucose metabolism through standardized uptake value (SUV) and hemodynamics through arterial spin labeling-derived cerebral blood flow (CBF). Participants were classified into ASM-responsive (N = 27) and ASM-resistant (N = 25) groups according to seizure recurrence at 12-month follow-up. Multimodal normalization was performed using age-matched healthy controls (N = 23) to generate z-score maps for metabolic (SUV_z) and perfusion (CBF_z) parameters. Neurovascular coupling was quantified using the SUV-CBF ratio at the whole-brain voxel level. Regions of interest (ROIs) were defined based on regions showing significant differences, and predictive performance was assessed with receiver operating characteristic (ROC) curve analysis. RESULTS: The analysis identified distinct neurovascular decoupling gradients in TLE. The SUV-CBF ratio decreased in the temporal lobe and cerebellum, whereas it increased in the frontal and occipital lobes compared with healthy controls. In the ASM-resistant group, the SUV-CBF ratio z-score was lower in the posterior cingulate gyrus and higher in the frontotemporal parietal occipital cortex compared with responders. The posterior cingulate cortex was chosen as the primary ROI. Prediction of ASM response based on SUV, CBF, and the SUV-CBF ratio in the posterior cingulate cortex yielded AUCs of 0.553, 0.676, and 0.836, respectively. CONCLUSIONS: Neurovascular decoupling characteristics derived from 18 F-FDG PET/MR may serve as a promising biomarker for identifying patients at risk of ASM-resistant outcomes.