Surgical Resection of Diffusion-weighted MRI Abnormalities Results in Seizure Freedom
Abstract number :
2.303
Submission category :
5. Neuro Imaging / 5A. Structural Imaging
Year :
2024
Submission ID :
705
Source :
www.aesnet.org
Presentation date :
12/8/2024 12:00:00 AM
Published date :
Authors :
Presenting Author: Jonathan Horsley, BSc – Newcastle University
Gerard Hall, PhD – Newcastle University
Csaba Kozma, BSc – Newcastle University
Callum Simpson, BSc – Newcastle University
Rhys Thomas, MD – Newcastle University
Yujiang Wang, PhD – Newcastle University
Jane de Tisi, BSc – University College London
Anna Miserocchi, MD – University College London
Andrew McEvoy, MD – University College London
Sjoerd Vos, PhD – University of Western Australia
Gavin Winston, BM BCh, MA, PhD, EMBA, FRCP, CSCN (EEG) – Queen's University
John Duncan, MD – University College London
Peter Taylor, PhD – Newcastle University
Rationale: Following surgical resection, many people with epilepsy continue to have seizures. This suggests that new approaches or data are needed to successfully identify and remove the epileptogenic zone. Localisation using diffusion-weighted MRI (dMRI) has shown early promise, but this has not been comprehensively investigated, nor is it currently used for this purpose clinically. Here, we investigate whether surgical resection of voxel-wise abnormalities, derived from diffusion-weighted MRI, results in seizure freedom.
Methods: We investigated 211 people with epilepsy, all of whom had pre-op dMRI and both pre- and post-op T1w MRI. Additionally, 97 healthy controls were used as a normative baseline. Diffusion tensor maps were obtained and registered to the same MNI-152 space. We calculated normative diffusion tensor maps using the mean and standard deviation in each voxel from control scans. For each voxel, we calculated deviations from the normative maps in individual subjects. We then applied probabilistic threshold-free clustering to boost the signal-to-noise ratio, thresholded abnormalities and identified abnormal clusters. Resection masks were derived from T1w MRIs, and the spatial location of the resected tissue was compared to the abnormal clusters. We hypothesised that surgical resection of the largest abnormal cluster would lead to seizure freedom after surgery.
Results: ILAE 1 or 2 at 12-months post-surgery was classed as seizure free and the largest cluster was classed as resected if at least half of it overlapped with the resection mask. The best results were obtained using axial diffusivity (AD) abnormalities. Resection of at least half of the largest cluster of abnormalities increased likelihood of seizure freedom ten-fold (odds ratio: 10.2; 95% CI: [2.4 , 43.6]). Of the 39 individuals with the largest cluster resected, 37 of them were seizure free (positive predictive value = 95%). Of the 63 individuals who were not seizure free, 61 of them did not have the largest abnormal cluster resected (specificity = 97%).
Conclusions: The targeting of dMRI abnormalities in surgery may lead to considerable improvements in seizure freedom rate. However, many subjects that did not have the largest abnormal cluster resected, potentially suggesting that the resection had sufficient impact on the epileptogenic network, or that alternative mechanisms exist.
Funding: P.N.T. and Y.W. are both supported by UKRI Future Leaders Fellowships (MR/T04294X/1, MR/V026569/1). J.J.H. is supported by the Centre for Doctoral Training in Cloud Computing for Big Data (EP/L015358/1). JD, JdT are supported by the NIHR UCLH/UCL Biomedical Research Centre. G.P.W. was supported by the MRC (G0802012, MR/M00841X/1).
Neuro Imaging