Screening of serum biomarkers for coronary artery calcification using DIA quantitative proteomics and construction of...

Using DIA quantitative proteomics technology to screen for coronary artery calcification (CAC)-associated differentially expressed proteins, this study explores the value of SMOC1, HSP90B1, and OPTN as potential serum biomarkers for CAC and constructs an individualized risk prediction model.

A two-stage case-control design was employed, enrolling 320 patients divided into a CAC group (160 cases) and a non-calcified control group (160 cases). Sixty cases were stratified and randomly selected for the discovery cohort (30 cases each), with the remaining 260 cases forming the validation cohort. The discovery cohort samples were randomly divided into three biologically replicated subgroups. Serum samples were pooled in equal volumes for DIA proteomics analysis to screen differentially expressed proteins, followed by bioinformatics analysis. Candidate proteins in the validation cohort were measured via ELISA. Single-factor and multivariate logistic regression analyses were conducted using clinical data to construct a nomogram prediction model. Model performance was evaluated using ROC curves, calibration curves, and decision curves.

A total of 39 differentially expressed proteins were identified (18 upregulated and 21 downregulated). GO enrichment analysis revealed that the differentially expressed proteins were enriched in protein folding, immune receptor signaling pathways, and apoptosis process regulation. KEGG enrichment analysis showed enrichment in T cell receptor signaling pathways, Notch signaling pathways, and calcium reabsorption pathways. Based on the significance of differential expression, biological relevance, and literature evidence, SMOC1 (upregulated), HSP90B1 (downregulated), and OPTN (downregulated) were identified as candidate proteins. ELISA results from the validation set showed that the expression trends of these three proteins were fully consistent with the proteomics findings (P < 0.001). Multivariate logistic regression analysis identified age, serum uric acid, alkaline phosphatase, fasting blood glucose, SMOC1, HSP90B1, and OPTN as independent predictors of CAC (P < 0.05). The AUC of the logistic regression model constructed using the above indicators for predicting CAC was 0.894 (95% CI: 0.855-0.933), which was significantly higher than that of the baseline model containing only clinical indicators (AUC = 0.845; DeLong test, P = 0.0013); The Net Rating Improvement (NRI) was 0.908 (95% CI: 0.707-1.123), and the Integrated Discrimination Improvement (IDI) was 0.128 (95% CI: 0.090-0.167). Calibration curves demonstrated good agreement between predicted and observed probabilities, while decision curves indicated positive clinical net benefit within the 0.1-0.8 probability threshold range, significantly outperforming traditional clinical models.

SMOC1, HSP90B1, and OPTN are potential serum biomarkers for coronary artery calcification (CAC). A nomogram model incorporating these three biomarkers and clinical indicators performs well in predicting CAC.