AUTHOR=Ion-Mărgineanu Adrian , Kocevar Gabriel , Stamile Claudio , Sima Diana M. , Durand-Dubief Françoise , Van Huffel Sabine , Sappey-Marinier Dominique 

TITLE=Machine Learning Approach for Classifying Multiple Sclerosis Courses by Combining Clinical Data with Lesion Loads and Magnetic Resonance Metabolic Features

JOURNAL=Frontiers in Neuroscience

VOLUME=Volume 11 - 2017

YEAR=2017

URL=https://www.frontiersin.org/journals/neuroscience/articles/10.3389/fnins.2017.00398

DOI=10.3389/fnins.2017.00398

ISSN=1662-453X

ABSTRACT=Purpose. The purpose of this study is classifying multiple sclerosis (MS) patients in the four clinical
forms as de□ned by the McDonald criteria using machine learning algorithms trained on clinical data
combined with lesion loads and magnetic resonance metabolic features.
Materials and Methods. Eighty-seven MS patients (12 Clinically Isolated Syndrome (CIS), 30
Relapse Remitting (RR), 17 Primary Progressive (PP) and 28 Secondary Progressive (SP)) and
eighteen healthy controls were included in this study. Longitudinal data available for each MS patient
included clinical (e.g. age, disease duration, Expanded Disability Status Scale), conventional magnetic
resonance imaging and spectroscopic imaging. We extract N-acetyl-aspartate (NAA), Choline (Cho),
and Creatine (Cre) concentrations, and we compute three features for each spectroscopic grid by
averaging metabolite ratios (NAA/Cho, NAA/Cre, Cho/Cre) over good quality voxels. We built
linear mixed-e□ects models to test for statistically signi□cant di□erences between MS forms. We test
nine binary classi□cation tasks on clinical data, lesion loads, and metabolic features, using a leaveone-
patient-out cross-validation method based on 100 random patient-based bootstrap selections. We
compute F1-scores and BAR values after tuning Linear Discriminant Analysis (LDA), Support Vector
Machines with gaussian kernel (SVM-rbf), and Random Forests.
Results. Statistically signi□cant di□erences were found between the disease starting points of each
MS form using four di□erent response variables: Lesion Load, NAA/Cre, NAA/Cho, and Cho/Cre
ratios. Training SVM-rbf on clinical and lesion loads yields F1-scores of 71-72% for CIS vs. RR and
CIS vs. RR+SP, respectively. For RR vs. PP we obtained good classi□cation results (maximum F1-
score of 85%) after training LDA on clinical and metabolic features, while for RR vs. SP we obtained
slightly higher classi□cation results (maximum F1-score of 87%) after training LDA and SVM-rbf on
clinical, lesion loads and metabolic features.
Conclusions. Our results suggest that metabolic features are better at di□erentiating between
relapsing-remitting and primary progressive forms, while lesion loads are better at di□erentiating
between relapsing-remitting and secondary progressive forms. Therefore, combining clinical data
with magnetic resonance lesion loads and metabolic features can improve the discrimination between
relapsing-remitting and progressive forms.