Predicting venous thromboembolism in patients with pancreatic cancer.

The association between venous thromboembolism (VTE), which comprises pulmonary
embolism and deep venous thrombosis, and cancer is well-established. The
short-term risk of VTE in cancer patients receiving chemotherapy is around 6%.2
Cancer-associated VTE is an important complication being the second cause of
death after cancer itself. The incidence of VTE varies across patients with
specific types of cancer, with the highest rates observed in patients with
pancreatic cancer, especially in those with locally advanced or metastatic
disease.
Cancer patients are at high risk for VTE because of a cancer-induced
hypercoagulable state, often in combination with other predisposing factors
such as surgery, central lines, infection, immobilization and chemotherapy.
Extracellular vesicles (EV) are small, cell-derived vesicles released from
various types of cells, including platelets, monocytes, leukocytes,
erythrocytes, and endothelial cells. EVs can be detected in blood, but also in
urine, saliva, seminal fluid, and tear fluid. High concentrations of EVs have
been described in cardiovascular disease, thrombotic conditions, and in
patients with different types of cancer. As EVs share at least some
characteristics of the parent cell, such as the exposure of cell-specific
antigens, EVs are considered as being active players in processes such as
cellular communication, angiogenesis, coagulation and invasiveness. Cancer
cells are able to release these 200-1,500 nm sized membrane-enclosed vesicles,
and EVs are thought to contribute to the hypercoagulable state in at least some
types of cancer patients. The procoagulant activity of EVs in cancer can be
attributed to the exposure of coagulant tissue factor (TF).
Recently, the role of human TF-exposing EVs (TF+ EVs) in the
pathogenesis of cancer associated VTE was confirmed in mouse models. In mice
with ortotopic patient-derived xenographic tumors, human (tumor-derived) TF was
found to be present in venous blood clots. Furthermore, these blood clots in
tumor-bearing mice were larger than in mice without tumors, plasma level of
human TF correlated with tumor volume, and clot size in tumor-bearing mice was
smaller when TF was inhibited by anti-human TF.
In cancer patients with VTE, significantly higher levels of TF+ EVs
occur compared to cancer patients without thrombosis. While increased levels of
TF+ EVs are present in various cancer types, patients with pancreatic cancer
may have the highest level of TF+ EVs.
To measure the TF+ EV dependent coagulation in cancer patients, we
previously developed the fibrin generation test (FGT). Briefly, in the FGT the
clotting time of platelet-poor plasma is measured in the absence (control,
saline) or presence of anti-human factor VIIa. A delay of at least 13% in the
clotting time in the presence of anti-human factor VIIa is considered as
evidence that TF+ EVs are present and that a patient is at risk of developing
VTE. Subsequently, we evaluated the FGT as independent predictor of VTE in
cancer patients in a prospective cohort of 648 cancer patients. The study
showed that a *high* FGT result, i.e. at least 13% delay by anti-factor VIIa,
is associated with a two-fold increased risk for VTE (hazard ratio, HR 2.0; 95%
confidence interval, CI, 1.1-3.6). The association was stronger in patients
with pancreatic cancer (HR 4.1; 95% CI, 0.91-19) than in those with other tumor
types (HR 1.5; 95% CI, 0.72-3.1).
We recently improved the test by optimizing the procedure of blood
sample collection. By using plastic instead of glass tubes for sample
collection, and by preparing EV-containing but platelet-depleted plasma (two
centrifugation steps to remove platelets) instead of EV-containing
platelet-poor plasma (a single centrifugation step to remove platelets), the
problem of contact activation leading to unwanted clotting and unrelated to the
presence of TF+ EVs, could be strongly reduced (manuscript in preparation).
Because in cancer-associated thrombosis, coagulation is triggered by relatively
low amounts or numbers of TF+ EVs, this TF-induced clotting is mediated by
factor XIa and not by factor XIIa. By titrating polybrene, an inhibitor of
contact activation, we were able to find a concentration of polybrene that
efficiently blocks factor XIIa, thus blocks contact activation, whereas the
same concentration of polybrene hardly affects factor XIa activity. With these
aforementioned improvements of the test, we expect to minimize false positive
test results and to significantly improve the predictive performance of the
FGT.
In this current study we sought to (1) evaluate the performance of the improved
FGT in predicting VTE in patients with pancreatic cancer, (2) assess the
reproducibility of the FGT, (3) and compare to FGT performance to another TF+
EV activity assay, the TF-dependent factor Xa generation test, and (4) assess
other biomarkers potentially predictive of VTE.

Primary Objective:
- To identify pancreatic cancer patients at high risk of VTE by measuring EV
procoagulant activity with the improved FGT.

Secondary Objective(s):
- To evaluate the reproducibility of the improved FGT by measuring it in
duplicate.
- To identify pancreatic cancer patients at high risk of VTE by measuring the
EV procoagulant activity with the TF-dependent factor Xa generation test.
- To assess several other independent (bio)markers potentially predictive of
VTE in patients with pancreatic cancer, and to assess their association with
the FGT

This is an exploratory observational prospective cohort study. We will collect
blood samples at baseline from patients with newly diagnosed pancreatic cancer.
The follow-up duration will be 6 months from start of chemotherapy.

As this is an observatory study the risks of participation are limited to
complications of vena puncture. Subjects will have no direct benefits or risks
by participating in the study.