Contrast-Enhancing Lesion Segmentation in Multiple Sclerosis: A Deep Learning Approach Validated in a Multicentric Cohort-Reference-Cited by-同舟云学术

Contrast-Enhancing Lesion Segmentation in Multiple Sclerosis: A Deep Learning Approach Validated in a Multicentric Cohort

Published:2024-08-22 Issue:8 Volume:11 Page:858
ISSN:2306-5354
Container-title:Bioengineering
language:en
Short-container-title:Bioengineering

Author:

Greselin Martina¹²³,Lu Po-Jui¹²³^ORCID,Melie-Garcia Lester¹²³,Ocampo-Pineda Mario¹²³^ORCID,Galbusera Riccardo¹²³^ORCID,Cagol Alessandro¹²³⁴^ORCID,Weigel Matthias¹²³⁵^ORCID,de Oliveira Siebenborn Nina¹²³⁶,Ruberte Esther¹²³⁶^ORCID,Benkert Pascal⁷,Müller Stefanie⁸,Finkener Sebastian⁹^ORCID,Vehoff Jochen⁸,Disanto Giulio¹⁰^ORCID,Findling Oliver⁹^ORCID,Chan Andrew¹¹^ORCID,Salmen Anke¹¹¹²^ORCID,Pot Caroline¹³^ORCID,Bridel Claire¹⁴^ORCID,Zecca Chiara¹⁰¹⁵^ORCID,Derfuss Tobias³,Lieb Johanna M.¹⁶^ORCID,Diepers Michael¹⁷,Wagner Franca¹⁸,Vargas Maria I.¹⁹,Pasquier Renaud Du¹³,Lalive Patrice H.¹⁴,Pravatà Emanuele¹⁵²⁰^ORCID,Weber Johannes²¹,Gobbi Claudio¹⁰¹⁵,Leppert David³^ORCID,Kim Olaf Chan-Hi²¹,Cattin Philippe C.²²^ORCID,Hoepner Robert¹¹^ORCID,Roth Patrick²³,Kappos Ludwig¹²³^ORCID,Kuhle Jens²³,Granziera Cristina¹²³

Affiliation:

1. Translational Imaging in Neurology (ThINk) Basel, Department of Biomedical Engineering, Faculty of Medicine, University Hospital Basel, University of Basel, 4123 Basel, Switzerland

2. Department of Neurology, University Hospital Basel, 4031 Basel, Switzerland

3. Research Center for Clinical Neuroimmunology and Neuroscience Basel (RC2NB), University Hospital Basel, University of Basel, 4031 Basel, Switzerland

4. Department of Health Sciences, University of Genova, 16132 Genova, Italy

5. Division of Radiological Physics, Department of Radiology, University Hospital Basel, 4031 Basel, Switzerland

6. Medical Image Analysis Center (MIAC), 4051 Basel, Switzerland

7. Clinical Trial Unit, Department of Clinical Research, University Hospital Basel, University of Basel, 4031 Basel, Switzerland

8. Department of Neurology, Cantonal Hospital St. Gallen, 9000 St. Gallen, Switzerland

9. Department of Neurology, Cantonal Hospital Aarau, 5001 Aarau, Switzerland

10. Neurology Department, Neurocenter of Southern Switzerland, 6900 Lugano, Switzerland

11. Department of Neurology, Inselspital, Bern University Hospital, University of Bern, 3010 Bern, Switzerland

12. Department of Neurology, St. Josef-Hospital, Ruhr-University Bochum, 44791 Bochum, Germany

13. Service of Neurology, Department of Clinical Neurosciences, Lausanne University Hospital (CHUV), University of Lausanne, 1005 Lausanne, Switzerland

14. Division of Neurology, Department of Clinical Neurosciences, Faculty of Medicine, Geneva University Hospitals, 1205 Geneva, Switzerland

15. Faculty of biomedical Sciences, Università della Svizzera Italiana, 6962 Lugano, Switzerland

16. Division of Diagnostic and Interventional Neuroradiology, Clinic for Radiology and Nuclear Medicine, University Hospital Basel, University of Basel, 4031 Basel, Switzerland

17. Department of Radiology, Cantonal Hospital Aarau, 5001 Aarau, Switzerland

18. Department of Diagnostic and Interventional Neuroradiology, Inselspital, Bern University Hospital, University of Bern, 3010 Bern, Switzerland

19. Department of Radiology, Faculty of Medicine, Geneva University Hospital, 1205 Geneva, Switzerland

20. Department of Neuroradiology, Neurocenter of Southern Switzerland, 6900 Lugano, Switzerland

21. Department of Radiology, Cantonal Hospital St. Gallen, 9000 St. Gallen, Switzerland

22. Center for medical Image Analysis & Navigation, Department of Biomedical Engineering, University of Basel, 4123 Allschwil, Switzerland

23. Department of Neurology, University Hospital of Zurich, University of Zurich, 8091 Zurich, Switzerland

Abstract

The detection of contrast-enhancing lesions (CELs) is fundamental for the diagnosis and monitoring of patients with multiple sclerosis (MS). This task is time-consuming and suffers from high intra- and inter-rater variability in clinical practice. However, only a few studies proposed automatic approaches for CEL detection. This study aimed to develop a deep learning model that automatically detects and segments CELs in clinical Magnetic Resonance Imaging (MRI) scans. A 3D UNet-based network was trained with clinical MRI from the Swiss Multiple Sclerosis Cohort. The dataset comprised 372 scans from 280 MS patients: 162 showed at least one CEL, while 118 showed no CELs. The input dataset consisted of T1-weighted before and after gadolinium injection, and FLuid Attenuated Inversion Recovery images. The sampling strategy was based on a white matter lesion mask to confirm the existence of real contrast-enhancing lesions. To overcome the dataset imbalance, a weighted loss function was implemented. The Dice Score Coefficient and True Positive and False Positive Rates were 0.76, 0.93, and 0.02, respectively. Based on these results, the model developed in this study might well be considered for clinical decision support.

Publisher

MDPI AG

Link

https://www.mdpi.com/2306-5354/11/8/858/pdf

Reference38 articles.

1. Progressive multiple sclerosis: Pathology and pathogenesis;Lassmann;Nat. Rev. Neurol.,2012

2. The pathology of multiple sclerosis is the result of focal inflammatory demyelination with axonal damage;J. Neurol.,2005

3. Gadolinium enhancement increases the sensitivity of MRI in detecting disease activity in multiple sclerosis;Miller;Brain,1993

4. Gadolinium should always be used to assess disease activity in MS—Yes;Granziera;Mult. Scler. J.,2020

5. ECTRIMS/EAN Guideline on the pharmacological treatment of people with multiple sclerosis;Montalban;Mult. Scler. J.,2018