Skip to Main content Skip to Navigation
Journal articles

Automatic segmentation of the spinal cord and intramedullary multiple sclerosis lesions with convolutional neural networks

Charley Gros 1 Benjamin de Leener 1 Atef Badji 1 Josefina Maranzano 2 Dominique Eden 1 Sara Dupont 1, 3 Jason Talbott 3 Ren Zhuoquiong 4 Yaou Liu 4 Tobias Granberg 5 Russell Ouellette 6 Yasuhiko Tachibana 7 Masaaki Hori 8 Kouhei Kamiya 8 Lydia Chougar 8, 9 Leszek Stawiarz 10 Jan Hillert 10 Elise Bannier 11, 12 Anne Kerbrat 11, 12 Gilles Edan 13, 12 Pierre Labauge 14 Virginie Callot 15, 16 Jean Pelletier 15, 16 Bertrand Audoin 15, 16 Henitsoa Rasoanandrianina 15, 16 Jean-Christophe Brisset 17 Paola Valsasina 18 Maria Rocca 18 Massimo Filippi 18 Rohit Bakshi 19 Shahamat Tauhid 20 Ferran Prados 21 Marios Yiannakas 22 Hugh Kearney 22 Olga Ciccarelli 22 Seth Smith 23 Constantina Andrada Treaba 24 Caterina Mainero 24 Jennifer Lefeuvre 25 Daniel Reich 25 Govind Nair 25 Vincent Auclair 26 Donald Mclaren 26 Allan Martin 27 Michael Fehlings 27 Shahabeddin Vahdat 28, 29 Ali Khatibi 28, 30 Julien Doyon 30, 28 Timothy Shepherd 31 Erik Charlson 31 Sridar Narayanan 30 Julien Cohen-Adad 1, *
* Corresponding author
12 VisAGeS - Vision, Action et Gestion d'informations en Santé
INSERM - Institut National de la Santé et de la Recherche Médicale : U1228, Inria Rennes – Bretagne Atlantique , IRISA-D5 - SIGNAUX ET IMAGES NUMÉRIQUES, ROBOTIQUE
21 NMR Research Unit [London]
Institute of Neurology [London]
Abstract : The spinal cord is frequently affected by atrophy and/or lesions in multiple sclerosis (MS) patients. Segmentation of the spinal cord and lesions from MRI data provides measures of damage, which are key criteria for the diagnosis, prognosis, and longitudinal monitoring in MS. Automating this operation eliminates inter-rater variability and increases the efficiency of large-throughput analysis pipelines. Robust and reliable segmentation across multi-site spinal cord data is challenging because of the large variability related to acquisition parameters and image artifacts. In particular, a precise delineation of lesions is hindered by a broad heterogeneity of lesion contrast, size, location, and shape. The goal of this study was to develop a fully-automatic framework - robust to variability in both image parameters and clinical condition - for segmentation of the spinal cord and intramedullary MS lesions from conventional MRI data of MS and non-MS cases. Scans of 1042 subjects (459 healthy controls, 471 MS patients, and 112 with other spinal pathologies) were included in this multi-site study (n = 30). Data spanned three contrasts (T1-, T2-, and T2∗-weighted) for a total of 1943 vol and featured large heterogeneity in terms of resolution, orientation, coverage, and clinical conditions. The proposed cord and lesion automatic segmentation approach is based on a sequence of two Convolutional Neural Networks (CNNs). To deal with the very small proportion of spinal cord and/or lesion voxels compared to the rest of the volume, a first CNN with 2D dilated convolutions detects the spinal cord centerline, followed by a second CNN with 3D convolutions that segments the spinal cord and/or lesions. CNNs were trained independently with the Dice loss. When compared against manual segmentation, our CNN-based approach showed a median Dice of 95% vs. 88% for PropSeg (p ≤ 0.05), a state-of-the-art spinal cord segmentation method. Regarding lesion segmentation on MS data, our framework provided a Dice of 60%, a relative volume difference of -15%, and a lesion-wise detection sensitivity and precision of 83% and 77%, respectively. In this study, we introduce a robust method to segment the spinal cord and intramedullary MS lesions on a variety of MRI contrasts. The proposed framework is open-source and readily available in the Spinal Cord Toolbox.
Document type :
Journal articles
Complete list of metadata

Cited literature [106 references]  Display  Hide  Download
Contributor : Elise Bannier Connect in order to contact the contributor
Submitted on : Saturday, November 23, 2019 - 4:24:41 PM
Last modification on : Monday, October 11, 2021 - 1:23:40 PM
Long-term archiving on: : Monday, February 24, 2020 - 1:21:04 PM


Files produced by the author(s)



Charley Gros, Benjamin de Leener, Atef Badji, Josefina Maranzano, Dominique Eden, et al.. Automatic segmentation of the spinal cord and intramedullary multiple sclerosis lesions with convolutional neural networks. NeuroImage, Elsevier, 2019, 184, pp.901 - 915. ⟨10.1016/j.neuroimage.2018.09.081⟩. ⟨hal-01934566⟩



Record views


Files downloads