Emerging Insights of Management of Venous Thromboembolism in Patients With Cancer

3.1. Combined Effects of Virchow’s Triad

Virchow’s triad, comprising endothelial injury, venous stasis, and hypercoagulability, describes the core pathophysiological framework of thrombosis in cancer. ³¹ The thrombotic burden arises from the malignancy itself, as well as its treatments, and associated clinical complications.^13-15

Endothelial injury and dysfunction reflect both tumor invasion and hypoxia-related activation, and are prominently augmented by iatrogenic influences such as surgical trauma and the placement of CVC. ³² Catheter-related infection may further amplify inflammatory activation and prothrombotic conditions. ¹⁵

Venous stasis results from tumor-related vascular compression and reduced mobility in advanced disease, and is frequently exacerbated by postoperative immobilization, prolonged hospitalization, and catheter-related flow disturbance. ³³

Hypercoagulability is driven by tumor-derived procoagulant factors and a systemic inflammatory state, and is further shaped by systemic cancer therapies, particularly chemotherapy. ³⁴

Thus, CAT reflects the convergence of disease-specific, treatment-related, and care-related factors within the framework of Virchow’s triad.

3.2. Tumor Cell Expression of Procoagulants Leading to Hypercoagulability

Tumor cells release large amounts of procoagulant substances, leading to a hypercoagulable state. Tumor cells can release extracellular vehicles (EVs), which are lipid bilayer-bound particles that cannot replicate and do not contain a functional nucleus. ³⁵ In the resting state, phosphatidylserine (PS) is located on the inner leaflet of the EVs’ lipid membranes but translocates to the outer membranes upon stimulation. ³⁶ The exposed PS provides a negatively charged membrane surface for coagulation and offers more accessible sites for the assembly of the prothrombinase complex and tenase complex, thereby promoting thrombin generation and blood clotting.^37,38 Tumor-derived EVs carry a significant amount of intracellular procoagulant material upon release, which is closely related to systemic coagulation. ³⁹ In clinical settings, PS-positive EVs have been detected in the circulation of patients with cancer, ⁴⁰ and observational data indicate that EV profiles and inflammatory profiles are dynamic during anticoagulant treatment, supporting the clinical relevance of PS-positive EVs in CAT.^40,41 Tissue factor (TF), a transmembrane receptor and primary initiator of coagulation, can be expressed by tumor cells or spontaneously released as TF-positive microparticles into the bloodstream, both of which contribute to thrombosis. ⁴² The expression of the TF gene is regulated through various signaling pathways, transcription factors, and microRNAs (miRNAs). ⁴³ TF expression is often upregulated in tumors with specific oncogene (e.g., K-ras, c-Met) and tumor suppressor gene (e.g., p53, PTEN) mutations, as well as in tumor cells undergoing epithelial-mesenchymal transition (EMT) or hypoxia. ⁴⁴ Mutations in oncogenes and tumor suppressor genes activate the mitogen-activated protein kinase/phosphatidylinositol-3 kinase signaling pathway, subsequently inducing full-length tissue factor (flTF) expression. ⁴⁵ TF-mediated coagulation is inhibited by tissue factor pathway inhibitor (TFPI). ⁴⁶ Under hypoxic conditions created by tumor tissues, hypoxia-inducible factor hypoxia-inducible factor-1α (HIF-1α) accumulates and binds to the hypoxia-responsive element (HRE) region of the TFPI promoter, suppressing TFPI transcription while simultaneously activating pro-survival genes such as vascular endothelial growth factor (VEGF), thereby upregulating TF.^47,48 Epigenetically, abnormal downregulation or overexpression of miRNAs such as miR-19 and miR-520g can upregulate TF by suppressing flTF levels.^49,50

Studies have shown elevated plasma levels of von Willebrand factor (vWF) in patients with various cancers. ⁵¹ This reflects endothelial activation, with increased vWF release and surface presentation that promotes platelet adhesion and immune cell recruitment. ⁵² Tumor-derived factors may further amplify this process by promoting vWF secretion and regulating its expression within the tumor microenvironment. High plasma vWF levels have been linked to VTE risk in cancer patients and tend to cluster with more advanced disease and higher clinical risk strata.^53,54

Tumor cells can also secrete Cancer Procoagulant, which directly activates coagulation factor X independently of factor VII and is unaffected by TFPI. ⁵⁵ Podoplanin, a transmembrane protein, is abnormally elevated in many tumors. It significantly induces platelet aggregation and promotes coagulation by activating platelets through the platelet receptor C-type lectin-like receptor 2 (CLEC-2).^56,57

3.3. Induction of Host Immune and Inflammatory Responses

Tumor development is accompanied by recruitment and activation of neutrophils within the tumor microenvironment, with release of inflammatory mediators that further amplify local immune cell accumulation. ⁵⁸ Activated neutrophils can express tissue factor on the endothelial surface and release neutrophil elastase, thereby promoting thrombin generation, impairing TFPI activity, and facilitating the formation of neutrophil extracellular traps (NETs). Tumor cells can further directly induce NETs formation by neutrophils.^59,60

Experimental studies from the Wagner laboratory have demonstrated that cancer creates a systemic milieu that primes neutrophils for NETs formation and that NETs release directly contributes to CAT in murine models. ⁶¹ Consistent with this, inhibition of NETs formation in mouse thrombosis models reduces thrombus burden, supporting a causal role for NETs in tumor-associated hypercoagulability. ⁶² In the intravascular environment, NETs form extracellular chromatin networks that provide a structural scaffold facilitating platelet adhesion and retention of circulating coagulation factors, thereby promoting thrombus stabilization and local propagation. Beyond this physical scaffold function, NETs also serve as a biochemical platform that concentrates procoagulant molecules within the thrombus microenvironment. Exposure of extracellular DNA and polyphosphates on NETs enhances platelet activation and thrombin generation, while neutrophil-derived proteases can impair endogenous anticoagulant pathways, further shifting the hemostatic balance toward hypercoagulability.^63,64 These effects of NETs occur within established platelet and coagulation pathways, in which vWF-mediated platelet tethering and adhesion contribute to platelet recruitment and thrombus propagation. Work from the Dubois laboratory has shown that tumor-derived microparticles bearing tissue factor can localize to thrombotic sites and interact with intravascular scaffolds, linking tumor-derived procoagulant activity with thrombus propagation associated with NETs. ⁶⁵ Together, these findings establish NETs as a central structural and biochemical amplifier within the tumor-associated thrombotic microenvironment, integrating immune cell activation with thrombus growth dependent on platelet and coagulation pathways. ⁶⁶ Consistent with these mechanisms, elevated NETs levels in CAT patients are associated with an increased risk of thrombosis. ⁶⁷

Immune checkpoint blockade of the PD-1/PD-L1 and CTLA-4 pathways may amplify immune-driven inflammatory activation and thereby promote prothrombotic pathways relevant to CAT. ⁶⁸ Consistent with this, ICI therapy has been linked to a systemic pro-inflammatory state with elevated cytokine levels; activated T cells may promote hypercoagulability by inducing tissue factor expression in monocytes/macrophages. Emerging translational data also suggest that patients who subsequently develop VTE during ICI treatment may exhibit higher baseline inflammatory signatures, including increased myeloid-derived suppressor cells and CXCL8, which may facilitate NETs formation and thereby promote a prothrombotic state.^68,69

3.4. Platelet Activation

Tumor cells secrete platelet agonists such as adenosine diphosphate and thrombin to directly activate platelets. They can also activate platelets through interactions between membrane proteins (e.g., podoplanin) on the tumor cell surface and their specific receptors on platelets. ⁷⁰ Mucinous adenocarcinomas secrete abnormally glycosylated mucins into the bloodstream, which interact with leukocyte L-selectin and platelet P-selectin, leading to platelet activation and aggregation. ⁷¹ Collectively, these interconnected mechanisms contributing to tumor-induced thrombosis are illustrated in Figure 1.OPEN IN VIEWER

4.1. Clinical History and Physical Examination

For cancer patients suspected of having thrombosis, rapid clinical assessment is critical. Medical history should clearly identify the onset time, location, and severity of symptoms, and identify potential precipitating factors such as hospitalization, surgery, chemotherapy, radiotherapy, or central venous catheter placement. Physical examination must check for signs of cancer-associated VTE and bleeding. A comprehensive assessment of comorbidities and medications that may affect bleeding risk or interact with anticoagulant drugs is equally important. According to guideline recommendations, attention should also be given to comorbidities and concomitant therapies that may influence thrombotic or bleeding risk, such as atrial fibrillation, hepatic or renal insufficiency, thrombocytopenia, and anticoagulant or anticancer drugs that may interact with anticoagulants. ⁷²

4.2. Laboratory Assessment

In clinical laboratory assessments, conventional coagulation tests (PT/INR/APTT) are widely used to screen for coagulation disorders. However, these tests have limited clinical utility in detecting hypercoagulable states or predicting thrombotic risk in cancer patients. D-dimer, a fibrin degradation product, is the most widely used laboratory biomarker in routine clinical practice to aid the evaluation of suspected VTE in cancer patients. Due to its low specificity, it is primarily used as an exclusionary test, while persistent elevation during follow-up may indicate a higher risk of VTE recurrence. ⁷³ Parameters such as leukocyte count, soluble P-selectin, and platelet count have been investigated in research settings and shown to be associated with VTE risk. These parameters may provide mechanistic insights or contribute to future risk models. However, it is important to note that these associations have not been validated for clinical application, therefore they are not currently recommended for inclusion in the clinical evaluation of VTE.^4,74

4.3. Imaging Evaluation

Ultrasound is the first choice for imaging examinations of venous thrombosis. Different examination techniques have different degrees of accuracy. If the purpose is to diagnose proximal DVT, compression ultrasound (CUS) alone may be the most appropriate technique (with a sensitivity of over 90% and specificity of nearly 100%). ⁷⁵ Conversely, if the aim is to identify distal DVT, duplex or triplex ultrasound may be more appropriate techniques. ⁷⁶ Based on the latest literature, CUS has become the preferred diagnostic method for detecting DVT and has been applied in multiple departments, including hospital emergency rooms and outpatient clinics. ⁷⁵ Study has suggested that emergency physicians can use portable ultrasound machines in the emergency room to perform 3-point CUS examinations of the lower limbs, which can effectively detect possible DVT in a short period of time. ⁷⁷ For pregnant women, patients with severe renal failure, or patients with a history of severe reactions to iodine contrast agents, CTPA is typically contraindicated. In such cases, CUS can provide a bedside, simple, and non-invasive examination to determine the presence of DVT. When assessing for DVT in the abdomen and pelvis or planning preoperatively, CT venography and MR venography are more appropriate. However, the ionizing radiation from CTV and the cost-effectiveness of MRV, along with the specialized expertise required for each modality, must be considered. ⁷⁸ For patients suspected of pulmonary embolism (PE), CT angiography remains the gold standard for diagnosis.

5.1. Primary Thromboprophylaxis

Cancer patients are at a significantly elevated risk of VTE compared with the general population. Predictors of VTE include patient-related factors such as a prior history of VTE, advanced age, obesity, and comorbidities; tumor-related factors, particularly cancer site (pancreatic, gastric, and biliary tract regions); anticancer treatment such as chemotherapy, and biomarkers, among which D-dimer is a promising candidate for baseline risk stratification.^80-84 The Khorana score is the most widely validated and applied tool to address this heterogeneous risk, stratifying patients into low (0 points), intermediate (1–2 points), and high (≥3 points) VTE risk categories prior to the initiation of systemic anticancer therapy.^82,85 However, external validation studies indicate that the Khorana score may have a limited ability to reliably distinguish patients at higher and lower risk of VTE across diverse cancer populations. In particular, analyses from the Van Es group have shown reduced sensitivity for capturing patients who subsequently develop VTE, with variable performance across major cancer subgroups.^85,86 In addition to the Khorana score, other risk assessment models, including the Vienna CATS and ONKOTEV scores, have been proposed for VTE risk stratification in cancer patients.^87,88 More recently, the COMPASS-CAT score, prospectively developed in common solid tumors and incorporating broader clinical and comorbidity-related factors, has been proposed as an alternative tool that, unlike the Khorana score, can be applied after initiation of anticancer treatment. Subsequent studies suggest moderate discriminatory performance and potential advantages in selected populations, including lung cancer and immunotherapy-treated cohorts, although substantial heterogeneity and limited prospective external validation remain.^89,90 Accumulating evidence suggests that exposure to ICIs is independently associated with an increased risk of venous thromboembolism, even after adjustment for traditional clinical and tumor-related factors. However, existing risk assessment models, including the Khorana score, were developed prior to the widespread use of ICIs and primarily in chemotherapy-treated populations, and their performance in ICI-treated patients appears limited and inconsistent, highlighting the need for dedicated, prospectively validated risk stratification tools and more individualized approaches to thromboprophylaxis in this population.^91,92 Routine primary thromboprophylaxis is not advised for all ambulatory cancer patients owing to variable absolute VTE risk and possible bleeding complications.^82,93 Current clinical guidelines therefore recommend applying validated risk stratification tools to identify ambulatory cancer patients who have an intermediate to high risk of VTE (e.g., Khorana score≥2) and a low bleeding risk, and to consider initiating primary pharmacological thromboprophylaxis in this subgroup.^82,93,94 In this population, prophylaxis with low-molecular-weight heparin (LMWH) or direct oral anticoagulants (DOACs) is recommended and should generally be continued for at least 12 weeks.^94,95 For example, in ambulatory patients with pancreatic cancer receiving systemic chemotherapy, primary pharmacological thromboprophylaxis is recommended with LMWH (grade 1A) or DOACs (rivaroxaban or apixaban; grade 1B). ⁹⁴ In hospitalized cancer patients without VTE, pharmacological thromboprophylaxis is generally recommended in the absence of contraindications, with LMWH preferred.^94,95 However, extended thromboprophylaxis is not routinely recommended; a study by Osataphan et al showed no significant reduction in VTE incidence between the extended (28-42 days) and standard (14 days) prophylaxis group, while the extended group exhibited a higher risk of clinically relevant bleeding. ⁹⁶ For cancer patients undergoing surgery, thromboprophylaxis should be administered both preoperatively and postoperatively according to guidelines, and mechanical methods such as intermittent pneumatic compression (IPC) may be used as adjuncts.^93-95 Overall, personalized risk assessment remains essential, incorporating not only thrombotic and bleeding risks but also factors such as treatment-related thrombocytopenia, renal and hepatic function, and potential drug interactions. ⁹⁷ These prophylactic strategies form an integral component of a clinical management paradigm for VTE in cancer patients, laying the groundwork for subsequent treatment options and ongoing follow-up and reassessment throughout the disease course.

5.2. Anticoagulant Therapy

5.2.1. Comparison of Anticoagulants

More than two decades ago, vitamin K antagonists (VKAs), particularly warfarin, were the mainstay for CAT until the landmark CLOT trial in 2003 showed that LMWH significantly reduced recurrence (9% vs. 17%; HR 0.48; P = 0.002), with no significant difference in major bleeding (6% vs. 4%; P=0.27).^98,99 Warfarin has largely lost patient favor due to the frequent and challenging INR monitoring it requires, as well as its associated complications. ¹⁰⁰ However, daily injections, cost, and injection-related discomfort (e.g., hematoma) diminished patient adherence,^101,102 prompting demand for effective and convenient oral anticoagulants. ¹⁰³

DOACs are receiving growing favor and progressively emerging as first-line treatments, attributed to their ease of oral administration and similarly high effectiveness. Subsequent RCTs comparing DOACs with LMWH are summarized in Table 2. The SELECT-D trial demonstrated lower VTE recurrence with rivaroxaban (4% vs. 11%; HR 0.43; 95%CI, 0.19 to 0.99), yet also higher major bleeding and clinically relevant non-major bleeding (CRNMB) rates. ¹⁰⁴ In HOKUSAI Cancer trial in 2018, edoxaban was noninferior to dalteparin for a composite outcome of recurrent VTE (7.9% vs. 11.3%; HR 0.71; 95%CI, 0.48 to 1.06) or major bleeding (6.9% vs. 4.0%; HR 1.77; 95%CI, 1.03 to 3.04), but showed slightly higher major bleeding. ¹⁰⁵ ADAM-VTE trial and Caravaggio trial in 2020 both favored apixaban: ADAM-VTE trial achieved markedly reduced recurrence (0.7% vs. 6.3%; HR 0.099; 95% CI, 0.37 to 1.07; P < 0.001) with no major bleeding; Caravaggio trial confirmed noninferiority (5.6% vs. 7.9%; HR 0.63; 95%CI: 0.37 to 1.07; P<0.001) with similar safety shown in the lower risk of major bleeding risk (3.8% vs. 4.0%, HR, 0.82; 95%CI, 0.40 to 1.69).^106,107 In CANVAS trial in 2021, physicians and patients could choose any DOACs or LMWH (with warfarin transition option). ¹⁰⁸ In the CASTA-DIVA trial in 2022, although it did not meet the prespecified non-inferiority criterion, rivaroxaban showed numerically lower rates of recurrent VTE and major bleeding than dalteparin, with overall directional results consistent with those of prior randomized trials. ¹⁰⁹ meta-analysis incorporating these trials showed that DOACs significantly reduced VTE recurrence (RR 0.67; 95% CI 0.52–0.85), with no meaningful difference in major bleeding (RR 1.20; 95% CI 0.85–1.70), although CRNMB risk increased (RR 1.63; 95% CI 1.25–2.11). ¹¹⁰

OPEN IN VIEWER

Table 2.

Comparison Between DOACs and LMWH in CAT From Key RCTs

Trial	DOACs vs. LMWH	Primary endpoint	Duration	Results
				Recurrent VTE	Major bleeding
SELECT-D (2018)	Rivaroxaban vs Dalteparin	Recurrent VTE	6 months	4.0% vs 11.0%; HR 0.43; 95%CI 0.19 to 0.99	6.0% vs 4.0%; HR 1.83; 95%CI 0.68 to 4.96
HOKUSAI Cancer (2018)	Edoxaban vs Dalteparin	Recurrent VTE or Major bleeding	6 to 12 months	7.9% vs 11.3%; HR 0.71; 95%CI 0.48 to 1.06	6.9% vs 4.0%; HR 1.77; 95%CI 1.03 to 3.04
ADAM-VTE (2020)	Apixaban vs Dalteparin	Major bleeding	6 months	0.7% vs 6.3%; HR 0.10; 95%CI 0.01 to 0.78	0.0% vs 1.4%; HR not estimable. p = 0.138
Caravaggio (2020)	Apixaban vs Dalteparin	Recurrent VTE	6 months	5.6% vs 7.9%; HR 0.63; 95%CI 0.37 to 1.07	3.8% vs 4.0%; HR 0.82; 95%CI 0.40 to 1.69
CANVAS (2021)	DOACs vs LMWH (±warfarin)	Recurrent VTE	6 months	6.1% vs 8.8%; 90%CI -6.1 to 0.7	5.2% vs 5.6%; 90%CI -2.5 to 1.7
CASTA-DIVA (2022)	DOACs vs Dalteparin	Recurrent VTE	3 months	6.4% vs 10.1%; SHR 0.75; 95%CI 0.21 to 2.66	1.4% vs 3.7%; SHR 0.36; 95%CI 0.04 to 3.43

Abbreviations: VTE, venous thromboembolism; LMWH, low-molecular-weight heparin; DOACs, Direct oral anticoagulants; HR, hazard ratio; SHR, Subdistribution Hazard Ratio.

Notably, apixaban demonstrated a major GI bleeding rate similar to dalteparin in the Caravaggio trial. ¹⁰⁴ Furthermore, it showed a significantly lower rate of pulmonary embolism recurrence and severe bleeding in another comparison with LMWH. ¹¹¹ An observational study suggests that apixaban may be associated with lower bleeding and fewer hospitalization events than rivaroxaban. ¹¹² A network meta-analysis also suggests that in patients with CrCl >80 mL/min, apixaban is associated with a lower risk of major bleeding than warfarin, rivaroxaban, or dabigatran. ¹¹³ These results highlights the potential advantages of apixaban among the numerous DOACs, though no head-to-head RCTs of DOACs exist, and dabigatran is not included in current recommendations. Currently, international guidelines all generally consider LMWH and DOACs (apixaban, rivaroxaban, and edoxaban) as the preferred treatment options for CAT. In many eligible patients without contraindications, DOACs are preferred owing to their inherent advantages, especially during the short-term (3 to 6 months) maintenance phase after initial therapy, with careful assessment for the patients’ suitability required. In contrast, warfarin is classified as a backup or secondary option in guidelines, to be considered when LMWH or DOACs are not feasible, with decisions based on individual patient circumstances.^{94,95,114,115}

5.3. Clinical Scenarios With Special Considerations