Cancer-associated thrombosis Mechanisms, prediction, and treatment

This thesis explores the mechanisms, prediction, and treatment of cancer-associated thrombosis (CAT). 
In Part I, we investigated biological mechanisms contributing to venous thromboembolism (VTE) in patients with cancer. In a review, we described germline and somatic mutations associated with increased VTE risk. A transcriptomic analysis in pancreatic cancer suggested a potential role of interleukin-33 in VTE development, whereas a protein profiling approach did not identify candidate protein predictors.
In Part II, we focused on risk prediction. A new CAT risk score was validated in a large Danish cohort, classifying more patients with VTE as high risk compared with the guideline-recommended Khorana score, while maintaining similar discriminatory performance. An individual patient data meta-analysis of >2,200 patients with cancer identified predictors of recurrent VTE during anticoagulation (age, breast cancer, metastatic disease, direct oral anticoagulant use, and index deep-vein thrombosis). However, predictive accuracy was modest, underscoring the limitations of models relying only on baseline predictors at the start of anticoagulation.
In Part III, we evaluated treatment of recurrent VTE despite anticoagulation. A post-hoc analysis of the Hokusai VTE Cancer trial revealed heterogeneous management strategies and high rates of recurrence, bleeding, and mortality. In a prospective observational study, a low-molecular-weight heparin dose escalation strategy did not clearly reduce second recurrences.
Overall, this thesis underscores the complexity of cancer-associated thrombosis, the limitations of current prediction tools, and the poor outcomes of recurrent VTE, highlighting the need for improved biomarkers and treatment strategies.