Plasma proteomics improves thrombosis prediction in patients with cancer and identifies targetable IL-17-driven...

Thrombosis remains a major cause of morbidity and mortality in patients with cancer. Existing risk models fail to reliably predict venous thromboembolism (VTE), underscoring the need for more accurate predictive models. In this study, we conducted a high-throughput proteomic analysis of 1105 plasma proteins in peripheral blood samples from patients with newly diagnosed lung or gastric cancer who were prospectively monitored for VTE development. Using a Bayesian probabilistic machine learning approach, we developed a predictive model incorporating 11 protein biomarkers and five clinical parameters (age, sex, history of VTE, body mass index, and hemoglobin), which outperformed the standardly used Khorana prediction score [c statistic 0.84 (0.79 to 0.90) as compared with 0.36 (0.27 to 0.45)]. Orthogonal validation in an external placebo cohort from a phase 3 trial confirmed the model's predictive power. Further investigation into the mechanistic role of CD200 receptor 1 (CD200R1), an immune checkpoint receptor known to limit leukocyte inflammatory response that contributed strongly to the model, showed that reduced concentrations in plasma correlated with higher D-dimer concentrations and thrombosis risk. CD200R1-deficient mice were characterized by features of a prothrombotic state, with elevated thrombin-antithrombin complexes, increased interleukin-17A (IL-17A), and endothelial inflammation. Administration of anti-IL-17A antibodies to CD200R1-deficient mice normalized thrombin-antithrombin complexes in vivo, and a meta-analysis of human COVID-19 studies showed reduced pulmonary thromboembolism in those on anti-IL-17A antibodies. These findings highlight the utility of plasma proteomics to improve prediction of thrombosis in patients with cancer and to identify unanticipated mechanistic insights and therapeutic targets in thrombo-inflammatory disease.