Back

Diagnostic Utility of Protein Biomarkers for Distinguishing Acute Ischemic Stroke and Transient Ischemic Attack: A Meta-Analysis

Lin, J. Y.-T.

2025-07-10 neurology
10.1101/2025.07.09.25331200 medRxiv
Show abstract

With ischemic strokes being one of the most pervasive medical issues globally, identifying methods for their accurate diagnosis is becoming increasingly crucial. Two common ischemic stroke subtypes are acute ischemic stroke (AIS) and transient ischemic attack (TIA). Although these two subtypes exhibit similar symptoms, AIS is far more severe than TIA. If AIS is misdiagnosed for TIA, this can lead to inadequate treatment and worse long-term outcomes for patients. Currently, stroke diagnosis relies heavily on patient medical history and imaging techniques, causing high rates of misdiagnosis. However, identifying protein biomarker concentrations associated with AIS and TIA is a promising new method for stroke diagnosis. In this meta-analysis, ten protein biomarkers were analyzed to determine whether or not they would serve as effective diagnostic tools for AIS and TIA. After collecting the mean concentration of each biomarker from 18,160 AIS patients and 3,410 TIA patients, means were compared between AIS and TIA patient groups, determining which biomarkers had a statistically significant difference in concentration between the two stroke subtypes. Biomarkers that yielded statistically significant results were S100B, sNfl, copeptin, IL-6, and MMP-9. MMP-9 yielded the highest difference in concentration between AIS and TIA patients (536.41 {+/-} 134.24 ng/mL in AIS patients vs. 0.11 {+/-} 0.03 ng/mL in TIA patients). A limitation to this study was the smaller sample size of TIA patients included. While this study establishes a baseline for promising protein biomarkers that can be utilized to differentiate AIS and TIA, future studies may want to further investigate each individual biomarker, establishing clear biomarker concentration ranges for AIS and TIA.

Matching journals

The top 4 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.