Clinical disease activity status (exacerbation versus quiescence) in subjects with relapsing-remitting multiple sclerosis is accurately classified using multivariate serum protein biomarker models
ECTRIMS Online Library. Qureshi F. 09/11/19; 278961; P601
Ferhan Qureshi
Ferhan Qureshi

Abstract: P601

Type: Poster Sessions

Abstract Category: Pathology and pathogenesis of MS - Biomarkers

R. Osan1, F. Qureshi2, M. Becich1, W. Hagstrom3

1Data Science, 2Assay Development, 3Octave Bioscience, Menlo Park, CA, United States

Introduction: Relapsing-Remitting Multiple Sclerosis (RRMS) is a complex and heterogeneous disease, and multiple biological pathways, including inflammation, immune modulation and neurodegeneration are involved in MS pathophysiology. Investigating these pathways, as represented by differential protein biomarker expression in serum, can help inform the development of tools to accurately track disease activity, identify early evidence of relapse, and monitor treatment response.
Objectives: To investigate the performance of multiple protein biomarker models to classify samples from subjects with clinically defined disease activity status (exacerbation vs quiescence) from the Accelerated Cure Project (ACP) repository.
Methods: Serum samples from 124 subjects with RRMS were obtained from the ACP repository. 60 samples represented subjects who were in a clinically defined relapsing stage (exacerbation), and 64 samples represented subjects who were in a clinically defined remitting stage (quiescence). All samples were measured for 1104 proteins, including serum levels of neurofilament light chain (sNfL) using Proximity Extension Assays (PEA) from Olink and for 215 proteins using Luminex immunoassays from Rules Based Medicine (RBM). Protein biomarker expression was quantified and univariate/multivariate machine learning-driven biostatistical techniques were applied to the data. Cross-validation was performed to prevent overfitting and ensure generalizability for predicting the disease status of new samples. Accuracy and area under the curve (AUC) were selected as the key metrics for evaluation of model performance.
Results: Multivariate statistical approaches that included up to 21 biomarkers achieved performance with 0.815 ± 0.043 Accuracy and 0.849 ± 0.035 AUC. Multivariate models outperformed univariate biomarkers including sNfL which was the highest performing univariate marker. Biomarkers that were identified as important features in multivariate classifiers were investigated to establish causation using biological pathway and network models.
Conclusions: Multivariate protein biomarker models representing multiple biological pathways involved in the pathophysiology of MS effectively classified clinical disease activity status (exacerbation versus quiescence) with stronger performance than any single biomarker, thereby warranting further investigation with larger sample numbers and from additional cohorts.
Disclosure: Remus Osan is an employee of Octave Bioscience.
Ferhan Qureshi is an employee of Octave Bioscience.
Michael Becich is an employee of Octave Bioscience.
William Hagstrom is an employee of Octave Bioscience.

By clicking “Accept Terms & all Cookies” or by continuing to browse, you agree to the storing of third-party cookies on your device to enhance your user experience and agree to the user terms and conditions of this learning management system (LMS).

Cookie Settings
Accept Terms & all Cookies