Serum MGP and ECP were commonly elevated in both first-onset and recurrent ischemic stroke compared with healthy controls, and two separate combinatorial models discriminated each disease state from healthy controls with high accuracy (test-set AUC 0.993 and 0.959, respectively).
Key Findings
Results
DIA proteomics identified two core serum proteins, matrix Gla protein (MGP) and eosinophil cationic protein (ECP), that were significantly elevated in both first-onset and recurrent ischemic stroke compared with healthy controls.
Discovery phase included 10 patients with first-onset IS, 10 with recurrent IS, and 10 healthy controls analyzed by Data-Independent Acquisition (DIA) proteomics.
Differentially expressed proteins were defined by fold change ≥1.5 or ≤0.67.
Both MGP and ECP were significantly elevated in first-onset and recurrent IS groups versus healthy controls (all p < 0.05).
The two proteins did not differ significantly from each other between the first-onset and recurrent IS patient groups.
Results
A logistic regression model incorporating MGP and ECP discriminated first-onset IS from healthy controls with a test-set AUC of 0.993.
Validation phase included 49 patients with first-onset IS and 50 healthy controls.
Multivariable Firth penalized logistic regression was used, adjusted for age and sex.
Data were randomly split into training and test sets at a 2:1 ratio for ROC analysis.
Results were confirmed by repeated 10-fold cross-validation.
Results
A logistic regression model incorporating MGP and ECP discriminated recurrent IS from healthy controls with a test-set AUC of 0.959.
Validation phase included 30 patients with recurrent IS and 50 healthy controls.
Multivariable Firth penalized logistic regression was used, adjusted for age and sex.
Data were randomly split into training and test sets at a 2:1 ratio for ROC analysis.
Results were confirmed by repeated 10-fold cross-validation.
Results
Enrichment analysis of MGP and ECP interaction partners revealed involvement in neutrophil extracellular trap formation, inflammatory responses, and immune regulation.
Pathway enrichment analysis was performed on protein interaction partners of MGP and ECP.
Three core pathological processes were identified: atherosclerosis and vascular calcification, chronic inflammation, and hypercoagulable state.
These processes collectively informed the proposed mechanistic relevance of the biomarker panel to ischemic stroke pathophysiology.
Methods
The study was conducted in a Chinese cold-region (high-latitude) population, with a two-phase design to discover and validate proteomic biomarkers.
The discovery phase used 10 first-onset IS, 10 recurrent IS, and 10 healthy controls.
The validation phase expanded to 49 first-onset IS, 30 recurrent IS, and 50 healthy controls.
The authors specifically note the relevance of findings to 'high-latitude cold-region populations,' suggesting geographic and environmental context may be relevant.
Firth penalized logistic regression was chosen, likely to address potential small-sample bias in the regression modeling.
What This Means
This research suggests that two blood proteins — matrix Gla protein (MGP) and eosinophil cationic protein (ECP) — are elevated in people who have had an ischemic stroke (a stroke caused by a blood clot), whether it was their first stroke or a repeat event. The study was conducted in northeastern China, a cold, high-latitude region, and used a sophisticated method called DIA proteomics to screen thousands of proteins in blood samples, eventually narrowing down to these two candidates and confirming them in a larger group of patients.
When these two proteins were combined into statistical models, the models were remarkably accurate at distinguishing stroke patients from healthy individuals — achieving accuracy scores (AUC) of 0.993 for first-time stroke and 0.959 for recurrent stroke in test samples. Importantly, the levels of MGP and ECP did not differ significantly between first-time and recurrent stroke patients, suggesting they reflect a common biological process shared across both disease states rather than marking the difference between them. The proteins are linked to processes including vascular calcification, inflammation, and blood clotting — all known contributors to stroke.
This research suggests that measuring MGP and ECP in blood could potentially help identify ischemic stroke with high accuracy, which may be valuable for diagnosis and risk assessment. However, the study is relatively small and conducted in a specific geographic and ethnic population, so further research in larger and more diverse groups is needed before these biomarkers could be considered for clinical use.
Wang N, Guo Z, Jin M, Wu N, Zheng Z, Sun W, et al.. (2026). Serum MGP and ECP are commonly elevated in first-onset and recurrent ischemic stroke: a two-phase DIA proteomic study in a Chinese cold-region cohort.. Frontiers in neurology. https://doi.org/10.3389/fneur.2026.1885191