Segmentation of spinal rootlets across MRI contrasts with RootletSeg.
Authors
Affiliations (13)
Affiliations (13)
- Department of Biomedical Engineering, FEEC, Brno University of Technology, Brno, Czechia.
- NeuroPoly Lab, Institute of Biomedical Engineering, Polytechnique Montreal, Montreal, QC, Canada.
- Max Planck Research Group Pain Perception, Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany.
- Aix Marseille Univ, CNRS, CRMBM, Marseille, France.
- APHM, CHU Timone, Pôle d'Imagerie Médicale, CEMEREM, Marseille, France.
- Mila - Quebec AI Institute, Montreal, QC, Canada.
- Functional Neuroimaging Unit, CRIUGM, Université de Montréal, Montreal, QC, Canada.
- Centre de Recherche du CHU Sainte-Justine, Université de Montréal, Montreal, QC, Canada.
- NeuroPoly Lab, Institute of Biomedical Engineering, Polytechnique Montreal, Montreal, QC, Canada. [email protected].
- Department of Neurosurgery, Faculty of Medicine and Dentistry, Palacký University Olomouc, Olomouc, Czechia. [email protected].
- Department of Neurology, Faculty of Medicine and Dentistry, Palacký University Olomouc, Olomouc, Czechia. [email protected].
- Spinal Cord Injury Center, Balgrist University Hospital, University of Zurich, Zurich, Switzerland. [email protected].
- Faculty of Medicine and Dentistry, Palacký University Olomouc, Hněvotínská 976/3, Nová Ulice, 779 00, Olomouc, Czechia. [email protected].
Abstract
Segmentation of spinal nerve rootlets is relevant for spinal level estimation, lesion classification, neuromodulation therapy, and group-level analyses. The aim of this study was to develop a deep learning method for the automatic segmentation of C2-T1 dorsal and ventral spinal nerve rootlets on various MRI scans. The study included MRI scans from two open-access and one private dataset, consisting of 3D isotropic 3T turbo spin echo T2-weighted (T2w) and 7T MP2RAGE (T1-weighted [T1w] INV1 and INV2, and UNIT1) MRI scans. A deep learning model, RootletSeg, was developed on 93 MRI scans from 50 healthy adults (mean age, 28.70 years ± 6.53 [SD]; 28 [56%] males, 22 [44%] females) and achieved a mean ± SD Dice score of 0.67 ± 0.09 for T1w-INV2, 0.65 ± 0.11 for UNIT1, 0.64 ± 0.08 for T2w, and 0.62 ± 0.10 for T1w-INV1 contrasts. RootletSeg accurately segmented C2-T1 spinal rootlets across MRI contrasts, enabling the determination of spinal levels directly from MRI scans. The method is open-source and can be used for a variety of downstream analyses.