Arterial spin labeling performs comparably to 2-[F]fluoro-2-deoxy-D-glucose positron emission tomography for presurgical evaluation in pediatric lesional epilepsy

Objective: This study was undertaken to test whether arterial spin labeling (ASL) performs comparably to 2-[18F]fluoro-2-deoxy-D-glucose positron emission tomography (FDG-PET), the mainstay functional imaging technique, in pediatric lesional epilepsy, while avoiding radiotracer exposure and additional sedation. Methods: We retrospectively included children with epilepsy due to focal cortical dysplasia, low-grade epilepsy-associated tumors, or hippocampal sclerosis who underwent standardized magnetic resonance imaging (MRI; including single-delay ASL) and FDG-PET during presurgical evaluation. Lesions, perilesional perfusion, and metabolic abnormalities were segmented and coregistered. Spatial overlap was quantified using DICE scores to compare functional modalities with each other (perfusion-to-metabolism: DICEP-to-M), with the lesion (metabolism-to-lesion: DICEM-to-L; perfusion-to-lesion: DICEP-to-L), and, in seizure-free children, with the resection cavity (lesion-, metabolism-, perfusion-to-resection cavity: DICEL-/M-/P-to-Post). We also assessed the temporal stability of perilesional ASL abnormalities and the presence of remote ipsilateral/contralateral abnormalities. Equivalence testing used the Wilcoxon signed-rank equivalence test with FDG-PET as reference; Cohen κ quantified agreement for remote abnormalities. Results: Fifteen children were included; median ages at FDG-PET and ASL were 7.7 and 7.5 years; 53% required sedation. Median perilesional volumes were 11 339 mm3 (FDG-PET) and 10 791 mm3 (ASL); both were larger under sedation (p < .001). Perilesional volumes were equivalent (p = .037). Median DICEM-to-L and DICEP-to-L were .3 and .4; equivalence was confirmed (p < .001). Median DICEP-to-M was .7, indicating strong ASL-FDG-PET concordance. In seizure-free children following surgery, DICEM-to-Post and DICEP-to-Post were both .6 and equivalent (p = .01). ASL findings were stable over time (DICE = .27-.75; n = 4 with repeat ASL). Remote ipsilateral abnormalities were common (ASL 73%, FDG-PET 67%; κ = .53), with poor contralateral agreement (κ = .12). Significance: ASL yielded perilesional findings equivalent to FDG-PET and showed comparable overlap with the resection cavity in seizure-free children. As a radiation-free technique embedded into routine MRI, ASL reduces logistics and avoids an additional sedation session. These findings support ASL as a practical alternative to FDG-PET for presurgical workup, especially when FDG-PET access is limited.