Integrated microRNA–protein plasma biomarkers improve discrimination of multiple system atrophy from Parkinson’s disease
TL;DR
Findings show that a non-invasive plasma multi-analyte panel can differentiate MSA from PD with clinically meaningful accuracy, and support prospective validation in larger cohorts.
Abstract
Multiple system atrophy (MSA) is a fatal, rapidly progressive atypical parkinsonism that responds poorly to Parkinson’s disease (PD) medications, and diagnostic precision remains a critical unmet need. Earlier, accurate diagnosis would give patients realistic prognostic expectations and enable timely clinical-trial enrollment. We sought plasma biomarkers discriminating MSA from PD. Plasma small RNA sequencing and extracellular vesicle proteomics in discovery cohorts, filtered by the Biomedical Oriented Logistic Dantzig (BOLD) selector, nominated two microRNAs (hsa-miR-520a-5p, hsa-miR-22-3p) and two proteins (LCAT, kallistatin). Candidates were quantified in independent cohorts by XENO-Q qPCR and sandwich ELISA, and multivariate logistic regression models including all pairwise interactions were trained ( N = 65) and validated in an independently recruited testing cohort (MSA N = 22; PD N = 16). The combined microRNA–protein model achieved test-set AUC 0.813 (95% CI 0.654–0.971), with sensitivity 0.818 and specificity 0.750, outperforming microRNA-only (AUC 0.599) and protein-only (AUC 0.747) models and indicating complementary rather than redundant information. The markers converge on pathways of glycerophospholipid metabolism, cholesterol homeostasis, myelin integrity, and inflammatory regulation, consistent with the oligodendrocyte pathology central to MSA. These exploratory findings show that a non-invasive plasma multi-analyte panel can differentiate MSA from PD with clinically meaningful accuracy, and support prospective validation in larger cohorts.