Abstract Details

Title Constructing a Molecular Model of Disease Severity in Multiple Sclerosis

Topic Multiple Sclerosis

Presentation(s) P3 - Poster Session 3 (5:30 PM-6:30 PM)

Poster/Presentation
Number 15-006

Objective To construct and independently validate a cerebrospinal-fluid (CSF) biomarker-based model that predicts future rates of disability progression.

Background Neurology currently lacks validated models capable of predicting multiple sclerosis (MS) severity and identifying patient-specific biological processes that drive central nervous system (CNS) tissue destruction.

Design/Methods Untreated MS patients were prospectively recruited under IRB approved protocols and signed written informed consent. All patients underwent diagnostic work-up, including collection of CSF in untreated stage and standard clinical follow-up. CSF biomarkers were analyzed blindly, using one of the two versions of DNA-aptamer-based assay that measured 1128 and 1317 proteins, respectively. A novel statistical learning pipeline built using random forest regression models implemented on training datasets (N=185 and N=140, respectively) constructed a model that predicts disease progression, as measured by a new, sensitive MS severity scale (MS-DSS) at both first and last clinical visits. Independent validation cohorts (N=92 and N=72, respectively) assessed the out-of-sample performance of these models. STRING analysis was explored on the retained markers for biological interpretations.

Results Ratios of approximately 100 individual proteins were uncovered across the ensemble of models, with substantial overlap between proteins uncovered in the individual models. Model predictions correlated strongly with measured MS severity scales (MS-DSS, and EDSS-based MS severity scores [MSSS and ARMSS]) as well as with disease progression slopes in the independent validation cohort. Several biological processes showed correlation with speed of disability progression, representing potential therapeutic candidates.

Conclusions

CSF biomarkers can predict future progression rates and identify pathogenic processes in MS patients, allowing for more targeted therapeutic regimes in precision neurology. This research was supported by the Intramural Research Program of the National Institute of Allergy and Infectious Diseases (NIAID) of the National Institutes of Health (NIH).

Authors/Disclosures
PRESENTER	No disclosure on file
Peter Kosa, PhD (NIH/NINDS)	Dr. Kosa has nothing to disclose.
	No disclosure on file
Bibiana Bielekova, MD, FAAN (Neuroimmunological Diseases Section/NIAID/NIH)	Dr. Bielekova has received research support from National Institutes of Allergy and Infectious Diseases. Dr. Bielekova has received intellectual property interests from a discovery or technology relating to health care. Dr. Bielekova has received publishing royalties from a publication relating to health care.

��ɫ��

��ɫ��