Clinical Proteomics: Serum Panel Flags Artery Calcification
- Aug 3
- 4 min read
Coronary artery calcification usually shows up on a CT calcium scan: radiation, a scanner slot, and a cost most primary-care patients never absorb. A blood test would change that math. Clinical proteomics offers one route. Cui and Liu screened serum by data-independent acquisition mass spectrometry and pulled a three-protein signature out of a crowded plasma background. Their two-stage case-control design, 60 subjects for discovery and 260 for validation, flags SMOC1, HSP90B1, and OPTN as markers that separate calcified from non-calcified arteries.
Key Takeaways
A three-protein serum panel found by clinical proteomics predicted coronary artery calcification at an AUC near 0.89, ahead of a clinical-risk-factor model.
Data-independent acquisition mass spectrometry surfaced low-abundance calcification markers from serum without a targeted antibody assay.
Adding SMOC1, HSP90B1, and OPTN to routine chemistry sharpened risk discrimination over standard predictors.
The panel is single-center and needs prospective, multi-site validation before it could triage who gets a CT scan.
How Clinical Proteomics Narrowed the Serum Proteome
Serum is an awkward matrix for protein profiling; a few proteins like albumin dominate the signal and bury anything clinically useful. Working from 60 subjects, the authors quantified 2,676 proteins by DIA and found 39 that shifted between groups. Regularized regression and multivariate logistic modeling trimmed the list to three, retested by proteome analysis in a separate 260-subject cohort. Reaching low-abundance markers in blood is an old problem; Geyer and colleagues (2016) set out much of the plasma proteome profiling approach that work like this still depends on.
Key Findings
Discovery MS quantified 2,676 serum proteins, of which 39 were differentially expressed (18 up, 21 down).
SMOC1 rose in calcified patients while HSP90B1 and OPTN fell, all reproduced in the 260-subject validation cohort.
The combined model hit AUC 0.894 (95% CI 0.855 to 0.933) versus 0.845 for clinical factors alone, DeLong P = 0.0013.
Age, serum uric acid, alkaline phosphatase, fasting glucose, and the three proteins stayed independent predictors at P < 0.05.
Figure 1. Volcano plot of the discovery serum proteome, plotting log2 fold change against negative log10 P-value. Red points are 18 upregulated and green points 21 downregulated proteins in coronary artery calcification versus controls; 2,637 gray points were unchanged. Adapted from Cui and Liu (2026).
What the Three Proteins Point To
SMOC1 is a matricellular protein tied to mineralization, so its rise in calcified patients fits the biology rather than fighting it. HSP90B1, an endoplasmic-reticulum chaperone, and OPTN, an autophagy adaptor, both dropped, hinting at stressed protein handling in the vessel wall. The mechanistic story is thin, though; these are serum associations, not proof that any of the three drives calcification.
A Serum Readout Next to CT Scoring
The appeal is obvious: one blood draw against a calcium scan that needs a scanner and a radiation dose. An AUC of 0.89 in a matched case-control set tends to soften in real screening populations, where prevalence and comorbidity run messier. I would read this as a promising triage signal, not a replacement for imaging yet.
From Discovery Screen to a Usable Assay
For biomarker teams, the useful lesson is how much a small, well-chosen panel adds on top of routine chemistry once the discovery step is clean. In client work at Dalton we see a serum signature earn its place only when depletion and acquisition are controlled tightly enough that a two-fold change is real and not a plate artifact, which is why our serum proteomics services pair discovery DIA with same-instrument validation. Studies built like this one make a sensible entry point for anyone scoping a multi-omics biomarker discovery program from screen to validated test.
Frequently Asked Questions
What is clinical proteomics? Clinical proteomics is the large-scale measurement of proteins in patient samples such as serum or plasma, usually by mass spectrometry, to find disease markers. It reads out hundreds to thousands of proteins from a single blood draw. The aim is a signature that tracks diagnosis, risk, or treatment response.
Can clinical proteomics detect coronary artery calcification from blood? In this study a three-protein serum panel (SMOC1, HSP90B1, OPTN) told calcified from non-calcified patients apart at an AUC near 0.89. That does not replace a CT calcium scan, but it suggests blood could help decide who needs imaging. Prospective validation comes next.
How many samples does a serum proteomics biomarker study need? This one used 60 samples for discovery and 260 for validation, a common two-stage split. Discovery stays small because deep mass spectrometry is the costly part, while validation needs enough cases and controls to hold statistically. Powering the validation arm well matters more than a huge discovery set.
Conclusion
A compact serum panel that beats standard risk factors is believable here; the DIA workflow and the validation cohort are solid enough to trust the direction of the finding. What isn't settled is whether it holds in older, multi-ethnic, or asymptomatic groups, since both authors drew from one center. For a CRO client, this reads as a validation-ready hypothesis, not a finished assay.
Related Reading
See how we run these analyses in one lab: Dalton's multi-omics CRO services.
Citation
Cui, R., & Liu, X. (2026). Screening of serum biomarkers for coronary artery calcification using DIA quantitative proteomics and construction of a regression model. Frontiers in Cardiovascular Medicine, 13, 1824102. https://doi.org/10.3389/fcvm.2026.1824102
Note
This blog post summarizes findings from the above-cited research. Figures are adapted from the original publication. For full details, please refer to the source article.
By Seungjun Yeo, CEO at Dalton Bioanalytics. Specializing in multi-omics mass spectrometry for drug discovery and biomarker research.