Quantitative targeted proteomics for occult cancer screening in patients with unprovoked venous thromboembolism: results from the prospective PLATO-VTE study

Res Pract Thromb Haemost. 2025 Aug 19;9(6):103018. doi: 10.1016/j.rpth.2025.103018. eCollection 2025 Aug.

Abstract

Background: About 5% of patients with unprovoked venous thromboembolism (VTE) have occult cancer. Despite standard cancer screening, 50% of cancers remain undetected.

Objectives: We used quantitative targeted proteomics to identify novel cancer biomarkers among patients with unprovoked VTE.

Methods: Patients aged ≥40 years with a first unprovoked VTE and without a malignancy in the preceding 5 years were invited to an international prospective cohort study. Plasma samples were collected within 10 days after VTE. The primary outcome was an adjudicated cancer diagnosis during 12-month follow-up. Concentrations of 269 plasma proteins covering coagulation, complement, and cancer-associated pathways were measured using quantitative mass spectrometry-based targeted proteomics. In a nested case-control study, protein profiles of patients with cancer were compared with those of randomly sampled unique control patients (ratio 3:1). Proteins with an unadjusted P value < .05 and fold change ≥15% were combined in a multivariable logistic regression model. To address the variability in the obtained model, the protein selection and model-building approach were replicated in 250 bootstrap samples, and an optimism-adjusted c-statistic was calculated.

Results: Of the 476 included participants, 28 (5.9%) were newly diagnosed with cancer. Plasma samples were available for 24 cases, which were compared with those of 75 control patients. Concentrations of P-selectin, β-2 microglobulin, complement component 7, intracellular adhesion molecule 1, and lumican were higher in cases than in controls, whereas coagulation factor (F)VII, FX, and FXII, β-Ala-His dipeptidase, and kalistatin were lower. The optimism-adjusted c-statistic of the multivariable logistic regression model including these proteins was 0.78 (95% CI, 0.70-0.87).

Conclusion: Ten differentially abundant proteins were identified in patients with occult cancer, suggesting potential of plasma proteomic tests as novel biomarker for occult cancer in patients with unprovoked VTE.

Keywords: early detection of cancer; liquid biopsy; neoplasms; proteomics; venous thromboembolism.