To examine the associations between serum protein biomarker profiles and patient-reported disability in people with Multiple Sclerosis (pwMS).
In this multicenter cross-sectional study we included 431 adults (mean age 49 years, 81% women, 94% non-Hispanic White) with MS between 2017 and 2020. For exposure, we included 19 serum protein biomarkers potentially associated with MS inflammatory disease activity and 7 key clinical factors (age at sample collection, sex, race/ethnicity, disease subtype, disease duration, disease-modifying treatment, and time interval between sample collection and closest PRO assessment). Using 4 machine learning approaches (Least Absolute Shrinkage and Selection Operator [LASSO] regression, Random Forest, XGBoost, and Support-Vector Machines, we examined model performance in predicting Patient Determined Disease Steps (PDDS) and Patient-Reported Outcomes Measurement Information System (PROMIS) physical function as outcomes.
Using binary outcomes, models comprising both routine clinical factors and the 19 proteins as features consistently outperformed base models (containing clinical features alone) in predicting severe (PDDS≥4, PROMIS<35) versus mild/moderate (PDDS<4, PROMIS≥35) disability for all machine learning approaches, with LASSO displaying the best area under the curve (AUCPDDS=0.91, AUCPROMIS=0.90). Using continuous outcomes, LASSO models with combined clinical and 19 proteins as features (R2PDDS=0.31, R2PROMIS=0.35) again outperformed base models. The four LASSO models with combined clinical and protein features shared 2 clinical features (disease subtype, disease duration) and 4 protein biomarkers (CDCP1, IL-12B, NEFL, PRTG).
Serum protein biomarker profiles have potential clinical utility beyond clinical profile or single protein in predicting real-world MS disability status.