Diffusion MRI abnormalities localise the epileptogenic zone in paediatric drug-resistant epilepsy
Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
Objectives
Seizure freedom follows surgery in 67% of children with drug-resistant epilepsy, but imaging biomarkers could improve patient selection, epileptogenic zone localisation, and outcome prediction. We assessed whether diffusion MRI abnormalities predicted seizure freedom when resected; aligned with stereo-EEG seizure-onset and interictal zones; distinguished patients with/without an identified onset zone; and localised the epileptogenic zone in MRI-negative cases.
Methods
We retrospectively studied children undergoing focal resection (n=110) or stereo-EEG (n=61) between 2015–2023, alongside healthy controls aged 6–18 (n=69), using identical 3T MRI diffusion protocols. Age- and sex-adjusted normative models generated mean diffusivity (MD) and fractional anisotropy (FA) voxelwise z-score maps to identify dominant abnormalities. In resection patients, abnormality–resection overlap was related to ≥12-month Engel outcome. In stereo-EEG patients, z-scores were compared across seizure-onset, interictal, and uninvolved grey-matter contacts. In lesion-negative patients, abnormalities were compared with PET hypometabolism and stereo-EEG findings.
Results
Greater resection of MD abnormalities predicted seizure freedom, especially with overlap >0 (mean differences 17%, 23%; P perm =0.0086, <0.001), and higher odds of Engel I (OR 1.31/10%; AUC 0.723; Youden threshold 36%). FA showed similar associations. Seizure-onset contacts had higher MD and lower FA z-scores than uninvolved grey matter (+0.65, −0.18; P perm <0.001, =0.001), with interictal contacts intermediate. Abnormalities were more widespread and contralateral in patients without an identified onset zone. In lesion-negative patients, abnormalities co-localised with PET (20/28 fully, 6/28 partially) and stereo-EEG (17/24 fully, 3/24 partially).
Interpretation
Diffusion MRI abnormalities aligned with invasive electrophysiology, predicted seizure freedom when resected, and co-localized with PET/stereo-EEG, supporting an important role in presurgical evaluation.
