<  Back to the Polytechnique Montréal portal

Automatic multiclass intramedullary spinal cord tumor segmentation on MRI with deep learning

Andreanne Lemay, Charley Gros, Zhizheng Zhuo, Jie Zhang, Yunyun Duan, Julien Cohen-Adad and Yaou Liu

Article (2021)

Published Version
Terms of Use: Creative Commons Attribution .
Download (2MB)
Cite this document: Lemay, A., Gros, C., Zhuo, Z., Zhang, J., Duan, Y., Cohen-Adad, J. & Liu, Y. (2021). Automatic multiclass intramedullary spinal cord tumor segmentation on MRI with deep learning. NeuroImage - Clinical, 31. doi:10.1016/j.nicl.2021.102766
Show abstract Hide abstract


Spinal cord tumors lead to neurological morbidity and mortality. Being able to obtain morphometric quantification (size, location, growth rate) of the tumor, edema, and cavity can result in improved monitoring and treatment planning. Such quantification requires the segmentation of these structures into three separate classes. However, manual segmentation of three-dimensional structures is time consuming, tedious and prone to intra- and inter-rater variability, motivating the development of automated methods. Here, we tailor a model adapted to the spinal cord tumor segmentation task. Data were obtained from 343 patients using gadolinium-enhanced T1-weighted and T2-weighted MRI scans with cervical, thoracic, and/or lumbar coverage. The dataset includes the three most common intramedullary spinal cord tumor types: astrocytomas, ependymomas, and hemangioblastomas. The proposed approach is a cascaded architecture with U-Net-based models that segments tumors in a two-stage process: locate and label. The model first finds the spinal cord and generates bounding box coordinates. The images are cropped according to this output, leading to a reduced field of view, which mitigates class imbalance. The tumor is then segmented. The segmentation of the tumor, cavity, and edema (as a single class) reached 76.7 ± 1.5% of Dice score and the segmentation of tumors alone reached 61.8 ± 4.0% Dice score. The true positive detection rate was above 87% for tumor, edema, and cavity. To the best of our knowledge, this is the first fully automatic deep learning model for spinal cord tumor segmentation. The multiclass segmentation pipeline is available in the Spinal Cord Toolbox (https://spinalcordtoolbox.com/). It can be run with custom data on a regular computer within seconds.

Uncontrolled Keywords

Deep learning, Automatic segmentation, Spinal cord tumor, MRI, Multiclass, CNN

Open Access document in PolyPublie
Subjects: 1900 Génie biomédical > 1900 Génie biomédical
1900 Génie biomédical > 1901 Technologie biomédicale
2500 Génie électrique et électronique > 2500 Génie électrique et électronique
Department: Département de génie électrique
Institut de génie biomédical
Research Center: NeuroPoly - Laboratoire de Recherche en Neuroimagerie
Funders: CRSNG/NSERC, IVADO, Canada Research Chair in Quantitative Magnetic Resonance Imaging, Canadian Institute of Health Research, Canada Foundation for Innovation, Fonds de Recherche du Québec - Santé, Fonds de Recherche du Québec - Nature et Technologies, Canada First Research Excellence Fund (IVADO and TransMedTech), Courtois NeuroMod project and the Quebec BioImaging Network, Spinal Research and Wings for Life (INSPIRED project), National Science Foundation of China, Beijing Municipal Natural Science Foundation for Distinguished Young Scholars, Special Fund of the Pediatric Medical Coordinated Development Center of Beijing Hospitals Authority
Grant number: EX-2018-4, 950-230815, CIHR FDN-143263, 32454, 34824, 28826, 2015-PR-182754, RGPIN-2019-07244, 5886, 35450, 81870958, 81571631, JQ20035, XTYB201831
Date Deposited: 19 Apr 2022 13:20
Last Modified: 20 Apr 2022 01:20
PolyPublie URL: https://publications.polymtl.ca/9276/
Document issued by the official publisher
Journal Title: NeuroImage - Clinical (vol. 31)
Publisher: Elsevier
Official URL: https://doi.org/10.1016/j.nicl.2021.102766


Total downloads

Downloads per month in the last year

Origin of downloads


Repository Staff Only