MPN is a clonal disorder originating from HSCs, marked by the overproliferation of one or more myeloid lines in the bone marrow (BM) (39). In a previous European collaborative study on low-dose aspirin, the cumulative incidence of fatal and nonfatal thrombosis in patients with polycythemia vera (PV) was 5.5% (40). The Gruppo Italiano Studio Paliticemia conducted a study involving 1,213 patients and reported that 19% of them experienced thrombotic events (41). Patients with essential thrombocythemia (ET) experience thrombotic complications at a rate of 2-4% per patient each year, a frequency that parallels the 2.33 events per person-year observed in primary myelofibrosis (PMF) (42). Both a prior history of thrombosis and the Jak2V617F mutation serve as important predictors of thrombotic risk, with established value in clinical risk assessment (36,43). However, there is still a lack of biomarkers for predicting the risk of thrombosis in an individal, especially in younger age groups. The high incidence, complex management and high mortality of thrombus make it a worthy clinical concern in patients with MPN (44). It is necessary to identify new risk factors for thrombosis.

Multiple studies have examined the procoagulant role of NETs among patients with MPN. Using a murine MPN model driven by the Jak2V617F mutation, Wolach et al (13) observed increased NET formation and a predisposition to spontaneous thrombosis in lung tissues. Ruxolitinib, which targets JAK2 and is already in clinical use, was found to abrogate NET production and attenuate thrombosis through suppression of JAK-STAT signaling. Additionally, PAD4 was found to be essential for Jak2V617F-induced NET production and thrombotic events in vivo. In summary, that study suggested that Jak2V617F may promote thrombosis in MPN by enhancing NET formation. According to previous research, essential thrombocythemia-associated thrombosis arises in part from hyperactivation of platelets and white blood cells, a state that is itself closely associated with whether the patient harbors a JAK2 mutation. Among individuals harboring the Jak2V617F mutation or having experienced prior thrombotic events, the levels of TF, P-selectin and vWF were all higher than in those without a history of thrombosis (45,46). Suppressing the JAK-STAT pathway reduces the release of pro-adhesive and procoagulant P-selectin and vWF, as well as the pro-inflammatory cytokine IL-6 (47). Therefore, it can be hypothesized that JAK2 mutations activate platelets through NETs, promote the elevation of TF and vWF, and induce high expression of P-selectin in platelets, thereby increasing thrombotic events in patients with MPN. Ruxolitinib suppresses NET-induced thrombosis by restricting the secretion of IL-6. This hypothesis is presented in Fig. 4. MPO-DNA is a specific marker of NETs that can be measured in plasma. Guy et al (48) found that the MPO-DNA concentration in patients with MPN exhibiting thrombosis was significantly higher than that in the control group, supporting the pathogenic role of NET formation in thrombosis. Notably, these patients with MPN had a history of thrombosis, particularly portal vein thrombosis. However, no differences were found when comparing patients with MPN with thrombosis with those without thrombosis. According to Marin Oyarzún et al (49), only nucleosome levels were elevated. However, nucleosomes lack specificity as NET biomarkers, as they can also be generated through other cell death processes such as apoptosis or necrosis.

Meanwhile, the aforementioned studies have yielded some conflicting findings. In a study by Guy et al (48), 52 patients with newly diagnosed MPN and 54 healthy controls (free of prior thrombotic bleeding episodes) were examined. Neutrophils isolated from the MPN group exhibited a greater propensity for NET formation compared with those obtained from healthy subjects. However, two separate studies conducted by Wolach et al (13) and Marin Oyarzún et al (49) failed to demonstrate that the presence of Jak2V617F in unstimulated neutrophils leads to increased NETosis. The reasons for these discrepancies may be attributed to differences in patient inclusion criteria, treatment status, sample size and MPN subtype heterogeneity across the studies. First, regarding treatment-related confounding, in the two studies by Wolach et al (13) and Marin Oyarzún et al (49), the majority of enrolled patients were undergoing cytoreductive regimens along with JAK inhibitor therapy. These interventions exert myelosuppressive effects and increase the risk of opportunistic infections, leading to neutropenia and altered neutrophil functional status, thereby affecting NET release. Second, regarding cohort characteristics, according to Guy et al (48), half of the patients enrolled in their study had experienced prior thrombotic events. Relative to the data obtained from the other two aforementioned studies, this observation highlights the association between NET release and thrombosis. Third, regarding sample size limitations, the sample sizes of the aforementioned studies were relatively limited (each study included <70 cases), which may be insufficient to detect true intergroup differences, and limits the statistical power of multivariable and subgroup analyses. Fourth, regarding MPN subtype heterogeneity, MPN encompasses three main subtypes, namely PV, ET and PMF, each with distinct disease characteristics. Patients with PV frequently present with erythrocytosis and elevated blood viscosity; patients with ET are characterized by megakaryocyte hyperplasia and thrombocytosis; and patients with PMF exhibit more pronounced BM fibrosis and an inflammatory cytokine storm. These subtype differences may lead to heterogeneity in baseline NETosis levels and their contribution to thrombotic risk. However, due to the limited sample sizes of the aforementioned studies, no subtype-stratified analyses were performed. Differences in the proportions of MPN subtypes included across studies may also represent an important source of result variability.

Although the aforementioned studies show minor discrepancies in their results, they all consistently demonstrate that NETs promote thrombosis in MPN. Future larger-scale and more precisely stratified studies are needed to further strengthen this preliminary evidence (Table I).

Thrombosis incidence in AML ranges from 2.09 to 8.60% across studies, with variations attributed to diagnostic methods and patient populations (50-52). While disease aggressiveness correlates with VTE risk and poorer outcomes (53), thrombocytopenia in patients with AML has limited their inclusion in thrombosis trials, resulting in scarce management guidelines. This evidence gap underscores the need for an improved mechanistic understanding of thrombosis in AML.

According to previous studies, NET-TF interplay serves as a key driver of thrombus formation. Hell et al (54) conducted a study on 28 patients with malignant diseases accompanied by overt DIC (ISTH score >5), including 15 patients with AML. The authors defined plasma DNA and nucleosomes as parameters related to NETs formation, and measured TF activity in microvesicles (MV-TF) as well as other conventional coagulation indicators. The results showed that NETs-related parameters were closely associated with changes in MV-TF and conventional coagulation parameters. Doubling in plasma DNA levels was linked to a 7.6% decrease in fibrinogen, a 41% increase in D-dimer, a 15.3% reduction in platelets and a 3.9% shortening in prothrombin time. A 10% increase in nucleosomes was associated with a 3.1% decrease in fibrinogen, a 112.7% increase in D-dimer, a 5.0% reduction in platelets and a 1.0% shortening in prothrombin time. Furthermore, upon administration of chemotherapy to patients with AML, NETs parameters and MV-TF activity significantly decreased (nucleosomes and MV-TF activity returned to normal levels after 1 week and 1 month, respectively), and coagulation indicators improved. According to these findings, NETs and TF-carrying MVs are interconnected, and this association may explain why thrombosis frequently complicates AML. Ostafin et al (55) conducted a study on 45 children with leukemia (including 33 patients with ALL, 8 patients with AML and 4 patients with T-cell ALL). During the progression of acute leukemia in children, unlike in adults, neutrophils demonstrate a compromised ability to extrude NETs. Consequently, infectious complications become more probable, whereas thrombotic risk is not similarly elevated. The currently available research is limited, and therefore larger multicenter studies with bigger sample sizes are needed for further validation (Table I).

The progression of AML is associated with inflammatory signaling in the BM (56). Various effector cells drive inflammation, including malignant clonal cells and stromal cells. High levels of pro-inflammatory cytokines, including IL-1β, TNF, IL-6, CXCL8, IFN-γ and granulocyte-macrophage colony-stimulating factor (GM-CSF), are generated by leukemia cells (57-59). Enhanced innate signaling in clonal hematopoietic stem and progenitor cells (HSPCs) leads to elevated secretion of IL-1β (60,61). Through this process, routine hematopoiesis becomes impaired, and the stability of the stem cell niche is undermined, thereby favoring disease progression. Meanwhile, inflammatory cytokines trigger the classic suicidal NETosis pathway in neutrophils, promoting NADPH oxidase-derived ROS-dependent NET release, thereby further contributing to thrombus formation. This working hypothesis is presented in Fig. 5.

In the German Hodgkin Study Group trials (n=5,773), patients with HL showed a 3.3% thrombosis incidence (193 events) (62). Given the significant mortality impact of thrombosis in this curable malignancy, improved risk prediction is clinically crucial. Current tools such as the Khorana score inadequately predict VTE risk in HL (63), highlighting the need for refined risk assessment to guide timely thromboprophylaxis and improve outcomes.

Francischetti et al (21) analyzed 32 cases of cHL, among which, 75% (9 out of 12) of the nodular sclerosis (NS) subtype exhibited NET formation associated with fibrosis. Histones and elastase, two key NET components, can stimulate TF production within the endothelium. Through subsequent interaction with FVIIa, this response may activate the extrinsic arm of the coagulation system (64). Consequently, immunohistochemistry was used to measure TF expression, revealing 100% positivity (12/12) in the NS subtype. Through the combined action of NETs and TF, ECs become activated, which in turn fosters a reciprocal cycle linking coagulation and inflammation. This cycle emits signals that draw lymphocytes and neutrophils to the site (64-67). Hodgkin Reed-Sternberg (HRS) cells in NS cHL express PAR-2, indicating their tumorigenic and inflammatory phenotype. These events plausibly participate in fostering an inflammatory tumor microenvironment, operating via pathways that bear conceptual resemblance to immunothrombosis (15) (Fig. 6). In summary, these data provide the first evidence that NETs promote immunothrombosis and tumor growth in NS cHL. Future studies should explore the potential of targeting NETs and the PAR/protease axis. Such approaches may offer a way to retard the progression of cHL (Table I). In the aforementioned study (21), no NET formation was observed in patient samples of 5 mixed-cellularity, 5 lymphocyte-rich, 1 lymphocyte-depleted and 4 nodular lymphocyte-predominant HL cases. Thus far, the presence or absence of NETs in these HL subtypes has not been addressed by subsequent investigations.

VTE occurs in >10% of patients with MM, driven by disease-specific factors (such as M-protein hyperviscosity, proinflammatory cytokines, newly diagnosed disease, renal impairment, and chromosome 11 abnormalities) (4,68-71) and hemostatic alterations (including elevated FVIII, vWF, fibrinogen, and thrombin generation) (72-74). Despite existing International Myeloma Working Group and European risk stratification guidelines (2014-2015) (75,76), current models cannot reliably distinguish between low-, intermediate- and high-risk patients, and validated predictive biomarkers remain lacking, thus highlighting the need for deeper mechanistic insights to improve VTE risk assessment in MM.

In a previous study by Nielsen et al (77), 38 patients diagnosed with MM, 19 patients with monoclonal gammopathy of undetermined significance (MGUS) and 34 healthy subjects were enrolled. The hypercoagulable state of the patients was assessed through coagulation analysis. The results revealed that both patients with MM and MGUS exhibited increased TG and procoagulant phospholipid activity, and all patients with MM showed an increase in MV-TF. Additionally, plasma levels of cfDNA, employed as a proxy indicator of NET generation, were assessed. Patients with MM exhibited markedly elevated cfDNA concentrations, and, in certain individuals, these values ranged from 5 to 10-fold above those observed in the control group. The authors suggested that the hypercoagulability observed in patients with MM may be attributed to the production of NETs. However, since the study was limited by a small sample size, this hypothesis remains to be validated (Table I).

MM progression is associated with tumor-related tissue inflammation (78). Myeloma cells can autonomously secrete IL-6 and TNF-α, while mesenchymal stromal cells and osteoblasts also secrete IL-6 (79,80). Additionally, immunoglobulin-specific mechanisms contribute to VTE through inflammatory cytokine effects and acquired activated protein C resistance (81). Massive deposition of M-protein on the vascular endothelium triggers inflammatory responses. The hypercoagulable state in MM arises not only from the disease itself but also from therapeutic interventions. The use of immunomodulatory drugs (IMiDs), such as lenalidomide, promotes the release of inflammatory cytokines and induces stress and injury in ECs (82). It can be hypothesized that these inflammatory cytokine effects, in addition to sustaining tumor growth, may partially promote thrombosis by stimulating NET formation. This working hypothesis is presented in Fig. 7.

The California Cancer Registry analyzed 2,482 cases of ALL from 1993 to 1995, accounting for 4.5% of the cohort (51). Ziegler et al (50) studied 185 cases of ALL and found a thrombosis incidence of 2.09%. Adolescents with ALL face an elevated thrombotic risk compared with other age groups. A meta-analysis comprising 17 prospective studies (1,752 participants) reported that 5.2% of pediatric patients with ALL experienced thrombotic complications (83). VTE occurrence, meanwhile, elevates the likelihood of mortality within the first year by 40% (51). The incomplete understanding of ALL-associated thrombosis mechanisms and the absence of standardized thromboprophylaxis underscore the need for personalized prevention strategies tailored to patient-specific risk profiles.

Kim et al (84) examined 29 adult patients with acute leukemia (mean age, 50 years, among whom, 7 had ALL). The authors found increased levels of both factor XIIa and extracellular trap markers, specifically histone-DNA complex and cell-free dsDNA, and noted a strong correlation between these measurements. These results offer a previously unrecognized mechanistic explanation for the fact that adult patients with ALL are prone to thrombotic events. NETs may activate factor XII, thus triggering the intrinsic clotting cascade and resulting in a prothrombotic state along with increased thrombotic risk. To investigate the link between coagulation parameters and NET formation in pediatric ALL, Kumar et al (85) performed a research study where the authors measured NET-related markers, including circulating cfDNA, nucleosomes and citrullinated H3Cit, as well as endogenous thrombin potential (ETP) in 17 patients and a control group. The results showed that, with the exception of cfDNA, the concentrations of H3Cit and nucleosomes in plasma remained comparable to those observed in the control group at all the measured time points during treatment. Subsequent treatment of plasma samples with DNase I only revealed a significant reduction in ETP. cfDNA is considered to elevate the likelihood of thrombotic events. The above study also faced limitations due to its relatively small sample size, which restricted the ability to assess the associations among thrombin generation, NETs and clinical thrombosis (Table I). Therefore, future research should expand the sample size of patients with ALL and conduct multicenter studies to validate the association between NET markers and thrombosis. A concurrent in-depth investigation of the differences in NET formation and thrombotic risk mechanisms between pediatric and adult ALL would be of great importance.

Notably, four common issues emerge from the literature on the aforementioned diseases (1), including i) sample size: The majority of studies are single-center exploratory analyses (n<50), lacking multicenter, prospective, large-cohort validation (n>200), as well as prospective stratification by treatment stage (newly diagnosed, post-chemotherapy, or post-transplantation) and disease subtype; ii) NET marker standardization: cfDNA and nucleosomes are easily measurable but lack specificity; MPO-DNA and H3Cit exhibit high specificity, yet their detection methods (ELISA, immunofluorescence, and flow cytometry) have not been standardized. At the same time, as NET research has advanced, the understanding of traditional indicators has evolved. Free plasma DNA and nucleosomes have been discussed as non-exclusive indicators of NET formation, since they can likewise derive from other modes of cell death, namely cancer-associated or treatment-induced apoptosis and necrosis (49). To the best of our knowledge, no study has to date performed a parallel comparison of the predictive efficacy of cfDNA, nucleosomes, MPO-DNA and H3Cit for thrombotic events within the same cohort, nor has a unified positive threshold been established. In future research, a combined detection strategy is recommended, and the specificity of plasma markers should be verified whenever possible by immunocytochemistry or flow cytometry to guarantee the robustness of the findings; iii) therapeutic confounding: The majority of studies have not adequately controlled for or performed stratified analyses on the direct effects of cytoreductive therapy, JAK inhibitors, IMiDs, asparaginase or GM-CSF on NETosis. An ideal study design should involve dynamic sample collection before treatment (baseline), during treatment (acute phase) and after treatment (remission phase) to distinguish the contributions of the disease itself vs. therapeutic interventions to NET formation; and iv) lack of causal validation: Existing studies are predominantly correlational in nature and lack causal chain validation (such as using PAD4 inhibitors, DNase I or PAD4-knockout mice in disease models to demonstrate that NETs lead to thrombosis). Furthermore, in vitro experiments often employ non-physiological strong stimuli such as PMA to induce NETosis, which may differ mechanistically from NETosis induced by tumor-related stimuli in vivo (characterized by low-concentration, chronic stimulation).