Automatic segmentation of the spinal cord and intramedullary multiple sclerosis lesions with convolutional neural networks
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. 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. Scans of 1,042 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 1,943 volumes. 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. When compared against manual segmentation, our CNN-based approach showed a median Dice of 95% vs. 88% for PropSeg, 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. The proposed framework is open-source and readily available in the Spinal Cord Toolbox.
Keywords
Cite
@article{arxiv.1805.06349,
title = {Automatic segmentation of the spinal cord and intramedullary multiple sclerosis lesions with convolutional neural networks},
author = {Charley Gros and Benjamin De Leener and Atef Badji and Josefina Maranzano and Dominique Eden and Sara M. Dupont and Jason Talbott and Ren Zhuoquiong and Yaou Liu and Tobias Granberg and Russell Ouellette and Yasuhiko Tachibana and Masaaki Hori and Kouhei Kamiya and Lydia Chougar and Leszek Stawiarz and Jan Hillert and Elise Bannier and Anne Kerbrat and Gilles Edan and Pierre Labauge and Virginie Callot and Jean Pelletier and Bertrand Audoin and Henitsoa Rasoanandrianina and Jean-Christophe Brisset and Paola Valsasina and Maria A. Rocca and Massimo Filippi and Rohit Bakshi and Shahamat Tauhid and Ferran Prados and Marios Yiannakas and Hugh Kearney and Olga Ciccarelli and Seth Smith and Constantina Andrada Treaba and Caterina Mainero and Jennifer Lefeuvre and Daniel S. Reich and Govind Nair and Vincent Auclair and Donald G. McLaren and Allan R. Martin and Michael G. Fehlings and Shahabeddin Vahdat and Ali Khatibi and Julien Doyon and Timothy Shepherd and Erik Charlson and Sridar Narayanan and Julien Cohen-Adad},
journal= {arXiv preprint arXiv:1805.06349},
year = {2018}
}
Comments
38 pages, 7 figures, 2 tables