Background
Expanding the number of biomarkers is imperative for studying the etiology and improving venous thromboembolism prediction. In this study, we aimed to identify promising biomarkers or targeted therapies to improve the detection accuracy of early-stage deep vein thrombosis (DVT) or reduce complications.
Methods
Quantibody Human Cytokine Antibody Array 440 (QAH-CAA-440) was used to screen novel serum-based biomarkers for DVT/non-lower extremity DVT (NDVT). Differentially expressed proteins in DVT were analyzed using bioinformatics methods and validated using a customized array. Diagnostic accuracy was calculated using receiver operating characteristics, and machine learning was applied to establish a biomarker model for evaluating the identified targets. Twelve targets were selected for validation.
Results
Cytokine profiling was conducted using a QAH-CAA-440 (RayBiotech, USA) quantimeter array. Cross-tabulation analysis with Venn diagrams identified common differential factors, leading to the selection of 12 cytokines for validation based on their clinical significance. These 12 biomarkers were consistent with the results of previous array analysis: FGF-6 (AUC = 0.956), Galectin-3 (AUC = 0.942), EDA-A2 (AUC = 0.933), CHI3L1 (AUC = 0.911), IL-1 F9 (AUC = 0.898), Dkk-4 (AUC = 0.88), IG-H3 (AUC = 0.876), IGFBP (AUC = 0.858), Gas-1 (AUC = 0.858), Layilin (AUC = 0.849), ULBP-2 (AUC = 0.813)and FGF-9 (AUC = 0.773). These cytokines are expected to serve as biomarkers, targets, or therapeutic targets to differentiate DVT from NDVT.
Conclusions
EDA-A2, FGF-6, Dkk-4, IL-1 F9, Galentin-3, Layilin, Big-h3, CHI3L1, ULBP-2, Gas-1, IGFBP-5, and FGF-9 are promising targets for DVT diagnosis and treatment.