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)
1
EPM -
École Polytechnique de Montréal
2 Montreal Neurological Institute and Hospital
3 Department of Radiology and Biomedical Imaging [San Francisco]
4 Xuanwu Hospital of Capital
5 Deutsche Telekom Group
6 Department of Computer Science [Liverpool]
7 NIRS - National Institute of Radiological Sciences
8 Juntendo University Hospital [Tokyo]
9 Hôpital Cochin [AP-HP]
10 Karolinska Institutet [Stockholm]
11 Centre Hospitalier Universitaire de Rennes [CHU Rennes] = Rennes University Hospital [Ponchaillou]
12 VisAGeS - Vision, Action et Gestion d'informations en Santé
13 CIC - Centre d'Investigation Clinique [Rennes]
14 Département de neurologie [Montpellier]
15 CRMBM - Centre de résonance magnétique biologique et médicale
16 CEMEREM - Centre d'Exploration Métabolique par Résonance Magnétique [Hôpital de la Timone - APHM]
17 OFSEP - Observatoire Français de la Sclérose En Plaques [Lyon]
18 Vita-Salute San Raffaele University and Center for Translational Genomics and Bioinformatics
19 Brigham and Women's Hospital [Boston]
20 HMS - Harvard Medical School [Boston]
21 NMR Research Unit [London]
22 UCL - University College of London [London]
23 Vanderbilt University [Nashville]
24 Athinoula A. Martinos Center for Biomedical Imaging
25 NIH - National Institutes of Health [Bethesda, MD, USA]
26 Biospective [Montréal]
27 University of Toronto
28 UNF-CRIUGM - Unité de Neuroimagerie Fonctionnelle [Montréal]
29 Stanford University
30 MNI - McConnell Brain Imaging Centre
31 NYU Langone Medical Center - New York University Langone Medical Center
2 Montreal Neurological Institute and Hospital
3 Department of Radiology and Biomedical Imaging [San Francisco]
4 Xuanwu Hospital of Capital
5 Deutsche Telekom Group
6 Department of Computer Science [Liverpool]
7 NIRS - National Institute of Radiological Sciences
8 Juntendo University Hospital [Tokyo]
9 Hôpital Cochin [AP-HP]
10 Karolinska Institutet [Stockholm]
11 Centre Hospitalier Universitaire de Rennes [CHU Rennes] = Rennes University Hospital [Ponchaillou]
12 VisAGeS - Vision, Action et Gestion d'informations en Santé
13 CIC - Centre d'Investigation Clinique [Rennes]
14 Département de neurologie [Montpellier]
15 CRMBM - Centre de résonance magnétique biologique et médicale
16 CEMEREM - Centre d'Exploration Métabolique par Résonance Magnétique [Hôpital de la Timone - APHM]
17 OFSEP - Observatoire Français de la Sclérose En Plaques [Lyon]
18 Vita-Salute San Raffaele University and Center for Translational Genomics and Bioinformatics
19 Brigham and Women's Hospital [Boston]
20 HMS - Harvard Medical School [Boston]
21 NMR Research Unit [London]
22 UCL - University College of London [London]
23 Vanderbilt University [Nashville]
24 Athinoula A. Martinos Center for Biomedical Imaging
25 NIH - National Institutes of Health [Bethesda, MD, USA]
26 Biospective [Montréal]
27 University of Toronto
28 UNF-CRIUGM - Unité de Neuroimagerie Fonctionnelle [Montréal]
29 Stanford University
30 MNI - McConnell Brain Imaging Centre
31 NYU Langone Medical Center - New York University Langone Medical Center
Lydia Chougar
- Fonction : Auteur
- PersonId : 785545
- ORCID : 0000-0001-9306-5687
- IdRef : 232888671
Gilles Edan
- Fonction : Auteur
- PersonId : 838812
Pierre Labauge
- Fonction : Auteur
- PersonId : 759958
- ORCID : 0000-0001-7759-8555
- IdRef : 059860367
Virginie Callot
- Fonction : Auteur
- PersonId : 179488
- IdHAL : virginie-callot
- ORCID : 0000-0003-0850-1742
- IdRef : 06867919X
Jean Pelletier
- Fonction : Auteur
- PersonId : 756935
- ORCID : 0000-0001-9730-7567
- IdRef : 078804868
Bertrand Audoin
- Fonction : Auteur
- PersonId : 756934
- ORCID : 0000-0002-9860-7657
Jean-Christophe Brisset
- Fonction : Auteur
- PersonId : 918238
Maria Rocca
- Fonction : Auteur
- PersonId : 793341
- ORCID : 0000-0003-2358-4320
Massimo Filippi
- Fonction : Auteur
- PersonId : 888099
Daniel Reich
- Fonction : Auteur
- PersonId : 793342
- ORCID : 0000-0002-2628-4334
Julien Cohen-Adad
- Fonction : Auteur correspondant
- PersonId : 758004
- ORCID : 0000-0003-3662-9532
Résumé
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.
Domaines
Ingénierie biomédicale
Origine : Fichiers produits par l'(les) auteur(s)
Loading...