Prophylactic anticoagulation in isolated tumor thrombus: a single-center retrospective cohort study

Abstract

Background

The optimal management of isolated tumor thrombus (TT) remains unknown. Although therapeutic anticoagulation is indicated for concurrent bland thrombosis or history of venous thromboembolism, current guidelines do not specifically address isolated TT, and recent studies have not shown consistent benefit of anticoagulation in this setting.

Objectives

This study described the characteristics of patients with isolated TT managed at Gustave Roussy Comprehensive Cancer Center for over 5 years and explored outcomes, including overall survival, thrombotic events, and bleeding, across different anticoagulation strategies.

Methods

This retrospective single-center cohort study included consecutive adults with imaging-confirmed isolated TT managed at Gustave Roussy from November, 2018 to June, 2024. Patients receiving therapeutic anticoagulation for any indication were excluded. The primary outcome was overall survival. Secondary outcomes were incident thromboembolic events and bleeding.

Results

Among 1262 screened records, 29 patients with isolated TT met eligibility criteria (mean age 55.6 ± 15.7 years; 75.9% male; 93.1% had stage IV cancer). Ten received prophylactic-dose anticoagulation, and 19 received no anticoagulation. Restricted mean overall survival was 50.7 months (95% CI, 43.3-58.06) in the prophylactic anticoagulation group vs 20.8 months (95% CI, 10.8-30.9) in the no-anticoagulation group. New bland thrombosis and bleeding events were infrequent and did not differ significantly between groups.

Conclusion

In this small retrospective cohort, longer survival was observed in patients receiving prophylactic-dose anticoagulation. This finding may reflect residual confounding and selection of patients with more favorable prognosis rather than a true therapeutic effect and should therefore be interpreted with considerable caution.

Essentials

• •Isolated tumor thrombus (TT) is a rare entity with uncertain management.
• •We compared outcomes in patients who received prophylactic anticoagulation or no anticoagulation.
• •Longer survival was observed in the prophylactic anticoagulation group, likely due to residual confounding.
• •New bland thrombus and bleeding did not differ significantly.

1. Introduction

Intravascular tumoral thrombus (TT) is defined as the intraluminal extension of tumor cells into adjacent vasculature, corresponding to direct vascular invasion by certain malignancies. It is most commonly described in renal cell carcinoma (RCC) and hepatocellular carcinoma (HCC), but may also occur in other advanced solid tumors, with frequent involvement of the inferior vena cava (IVC), renal and visceral vessels, and, less commonly, pulmonary vessels [1,2]. TT worsens prognosis, complicates treatment decisions, and is associated with an increased risk of cancer-associated venous thromboembolism (VTE) [[3], [4], [5]]. Management is primarily directed at controlling the underlying malignancy; surgical thrombectomy may be considered in selected cases when technically feasible, and in HCC with portal vein TT, locoregional treatment with yttrium-90 radioembolization represents an additional option [1,2].

Distinguishing TT from bland (fibrin-rich) thrombus remains challenging and relies primarily on expert radiologic assessment using contrast-enhanced computed tomography (CT) or magnetic resonance imaging (MRI). Fluorodeoxyglucose positron emission tomography (FDG-PET) may provide complementary information in selected metabolically active tumor [6].

This distinction has direct therapeutic implications. Anticoagulation is indicated when concomitant bland thrombus is identified; however, its benefit-risk balance in isolated TT remains uncertain. Current guidelines for the management of cancer-associated VTE do not specifically address this scenario [[7], [8], [9], [10]], and available observational data suggest that anticoagulation may not consistently improve outcomes in patients with TT, while potentially exposing them to significant bleeding risk [4,[11], [12], [13], [14]].

At our institution, complex thrombotic cases in patients with cancer are discussed within a dedicated Thrombosis Multidisciplinary Team (MDT) meeting. In this retrospective, exploratory, single-center cohort study, we aimed to characterize the clinical features and treatment patterns of patients with isolated TT discussed at this MDT over 5.5 years and to explore key outcomes, including overall survival, incident thromboembolic events, and bleeding, according to the anticoagulation strategy used.

2. Methods

This retrospective, single-center cohort study was conducted at Gustave Roussy, a large Comprehensive Cancer Center in France. The study was approved by the local Ethics Committee (registration number: 2024-436) and conducted in accordance with the Strengthening Reporting of Observational Studies in Epidemiology guidelines [15].

All consecutive cases discussed at our Thrombosis MDT meetings from November, 2018 to June, 2024 were screened. Adult patients aged ≥18 years were eligible if they had: (i) a confirmed diagnosis of TT by contrast-enhanced CT or MRI or FDG-PET; and (ii) active malignancy at the time of TT diagnosis. Patients receiving therapeutic-dose anticoagulation for any indication, including concomitant bland thrombus, were excluded.

Isolated TT was defined as an intravascular soft-tissue component meeting all of the following characteristics: (i) contiguous with the primary tumor or metastatic lesion; (ii) associated venous expansion; (iii) internal enhancement (≥30 Hounsfield units on CT and/or post–contrast enhancement on MRI) and/or intrathrombus hypermetabolism on FDG-PET and (iv) absence of imaging features typical of bland thrombus (including lack of vessel wall enhancement, absent perivascular fat stranding, and metabolic inactivity). Imaging was reviewed by expert oncologic radiologists as part of MDT assessment.

Clinical data were extracted from electronic medical records into a secure, anonymized database. Collected variables included demographics, Eastern Cooperative Oncology Group (ECOG) performance status (PS), cardiovascular factors, renal function, cancer characteristics (cancer type and stage), TT characteristics (imaging modalities, location and extent), MDT recommendations, anticancer treatments, anticoagulation strategy (anticoagulant type, dose, and duration), and outcomes of interest.

Patients were categorized into 2 groups: (i) prophylactic-dose anticoagulation, defined as standard low–molecular-weight heparin dosing per MDT recommendation; or (ii) no anticoagulation. The prophylactic regimen typically consisted of enoxaparin 4000 IU anti-Xa or tinzaparin 4500 IU anti-Xa once daily subcutaneously, or a weight-adjusted equivalent (eg, enoxaparin 1 mg/kg for patients with body weight >100 kg), with dose adjustment for renal function when creatinine clearance was <30 mL/min. Prophylactic anticoagulation was continued until incident thromboembolic event, death, or last follow-up. Treatment allocation was determined by the MDT based on individualized assessment of potential benefits and risks, considering tumor characteristics, ECOG PS, bleeding risk, renal function, platelet count, planned anticancer therapy, and patient preferences.

The primary outcome was overall survival (OS). Secondary outcomes included incident thromboembolic events, defined as objectively confirmed VTE or arterial thrombosis occurring after TT diagnosis, any bleeding, and major bleeding according to ISTH criteria (ie, fatal bleeding, symptomatic bleeding in a critical area or organ, bleeding causing a fall in hemoglobin of at least 2 g/dL, or bleeding requiring transfusion of at least 2 units of red blood cells).

Continuous variables were expressed as mean ± SD or median (IQR) and compared using Student’s t-test or the Mann–Whitney U-test, as appropriate. Categorical variables were presented as numbers and percentages and compared using Fisher’s exact test or the chi-squared test, as appropriate. OS was defined as the time from TT diagnosis to death from any cause or last follow-up. Survival curves were estimated using the Simon–Makuch method to account for the time-varying nature of prophylactic anticoagulation exposure. Between-group differences in survival were analyzed using a time-dependent Cox proportional hazards model. Restricted mean survival time was calculated to provide an absolute measure of OS over a common follow-up horizon. For secondary outcomes, the cumulative incidences of thrombosis and bleeding were estimated using a competing risk model, with death considered as a competing event. All statistical analyses were performed using RStudio (version 2025.05.0-496, Posit Software). All tests were two-sided, and a P <.05 was considered statistically significant. Given the exploratory nature of this study, no adjustment for multiple comparisons was performed.

3. Results and Discussion

Among 1262 medical records reviewed, 39 patients with TT were identified. After excluding 10 patients who required therapeutic anticoagulation for other indications, 29 patients were included in the final analysis (Figure 1). Baseline characteristics stratified by anticoagulation strategy are summarized in Table 1. Mean age was 55.6 ± 15.7 years, and 22 patients (75.9%) were male. PS was generally preserved, with 96.6% of patients having an ECOG PS ≤2. The most common malignancies were lung cancer, sarcoma, and melanoma (each n = 5, 17.2%), followed by RCC (n = 4, 13.8%), other urologic malignancies (n = 4, 13.8%), pancreatic cancer (n = 2, 6.9%), thyroid cancer (n = 2, 6.9%), and HCC (n = 1, 3.4%). Most patients (93.1%) had stage IV disease. The median interval between cancer diagnosis and TT diagnosis was 10.2 months (IQR, 3.4-47.4). The most frequent TT locations were the IVC territory and visceral vessels (each n = 9, 31.0%), followed by the pulmonary vessels (n = 6, 20.7%) and the superior vena cava territory (n = 5, 17.2%). Representative imaging cases are provided in Supplementary Figures 1-3. At the time of TT diagnosis, 25 patients (86.2%) were receiving active anticancer treatment, and 4 patients (13.8%) were receiving palliative or supportive care alone.

Figure 1.

Study flowchart.

Table 1.

Baseline characteristics of the 29 patients with isolated tumor thrombosis included in the study.

Characteristics	Overall population N = 29	Prophylactic-dose anticoagulation N = 10	No anticoagulation N = 19	P-value
Age, y	55.62 (15.7)	54.20 (16.42)	56.37 (15.71)	.731
Male sex	22 (75.9)	8 (80)	14 (73.7)	1.000
BMI, kg/m2	25.75 (7.5)	26.5 (2.12)	25 (12.72)	.716
eGFR, mL/min	92.08 (25.23)	90.73 (26.7)	92.79 (25.15)	.839
ECOG performance status score				.751
0	14 (48.3)	6 (60)	8 (42.1)
1	10 (34.5)	3 (30)	7 (36.8)
2	4 (13.8)	1 (10)	3 (15.8)
3	1 (3.4)	0	1 (5.3)
Presence of 2 or more cardiovascular risk factors	14 (48.3)	4 (40)	10 (52.6)	.700
Primary tumor site				.252
Sarcoma	5 (17.2)	2 (20)	3 (15.8)
Lung	5 (17.2)	1 (10)	4 (21.1)
Melanoma	5 (17.2)	0	5 (26.3)
Renal	4 (13.8)	3 (30)	1 (5.3)
Urothelial	4 (13.8)	2 (20)	2 (10.5)
Pancreatic	2 (6.9)	0	2 (10.5)
Thyroid	2 (6.9)	1 (10.0)	1 (5.3)
Adrenal	1 (3.4)	1 (10.0)	0
Hepatic	1 (3.4)	0	1 (5.3)
Stage of cancer				.460
Stage I-II	0	0	0
Stage III	1 (3.4)	0	1 (5.3)
Stage IV	27 (93.1)	10 (100)	18 (94.7)
Cancer treatment status				1.0
Active	25 (86.2)	9 (90)	16 (84.2)
Palliative/supportive care	4 (13.8)	1 (10)	3 (15.8)
Time between cancer diagnosis and tumoral thrombus diagnosis, mo	10.2 (3.4-47.4)	4.5 (2.5-28)	27.4 (4.5-49.9)	.211
Site of tumoral thrombus				.552
Superior vena cava territory (jugular, superior vena cava, innominate trunk)	5 (17.2)	1 (10)	4 (21.1)
Inferior vena cava territory (inferior vena cava, iliac vein)	9 (31.0)	4 (40)	5 (26.3)
Visceral vessel (hepatic artery, portal, renal, mesenteric, hepatic or splenic vein)	9 (31.0)	4 (40)	5 (26.3)
Pulmonary vessel (pulmonary artery, pulmonary vein.)	6 (20.7)	1 (10)	5 (26.3)
Duration of thromboprophylaxis, mo	–	31.7 (19.1-49.3)	–

Data presented as n (%) or mean (SD) or median (IQR), as appropriate.

BMI, body mass index; ECOG, Eastern cooperative oncology group; eGFR, estimated glomerular filtration rate; TNM, tumor, node, metastasis.

Ten patients (34.5%) received prophylactic-dose anticoagulation, while 19 (65.5%) received no anticoagulation. Baseline characteristics were broadly comparable between groups; however, in this small exploratory cohort, the absence of statistically significant differences did not exclude potentially important clinical imbalances, and residual confounding is likely (Table 1).

During a median follow-up of 25.36 months (IQR, 16.39-46.98), 13 patients died, corresponding to a 36-month–cumulative incidence of death of 51.3 % (95% CI, 21-70), including 1 patient in the prophylactic-dose anticoagulation group and 12 in the no-anticoagulation group (Figure 2). All deaths were related to cancer progression. Restricted mean survival time was longer in patients receiving prophylactic-dose anticoagulation compared with those receiving no anticoagulation (50.7 months [95% CI, 43.3-58.06] vs 20.8 months [95% CI, 10.8-30.9]; P = .018).

Figure 2.

Overall survival according to anticoagulation strategy. Survival curves were estimated using the Simon–Makuch method to account for the time-varying nature of prophylactic anticoagulation.

Three patients (10.3%) developed a new bland thrombus: 1 patient in the prophylactic-dose anticoagulation group and 2 patients in the no-anticoagulation group, at a median of 1.4 months (IQR, 1.4-4.2) after TT diagnosis. For the only patient who developed a new bland thrombus in the prophylactic-dose anticoagulation group, the interval between initiation of prophylactic-dose anticoagulation and the event was 6.9 months. The 36-month cumulative incidence of new bland thrombus was 10% (95% CI, 0.45-37) in the prophylactic-dose anticoagulation group vs 11% (95% CI, 1.7-29) in the no-anticoagulation group (P > .9; Table 2). All 3 patients were subsequently switched to therapeutic-dose anticoagulation.

Table 2.

The 36-month–cumulative incidence of death of 51.3 % (95% CI, 21-70), including 1 patient in the prophylactic-dose anticoagulation group and 12 in the no-anticoagulation group.

Outcome	Prophylactic-dose anticoagulation N = 10	No anticoagulation N = 19	P-value
Cumulative incidence of death, %			.0031a
At 6 mo	0	42.1 (15.0-60.5)
At 12 mo	0	53.9 (24-72.1)
At 24 mo	0	53.9 (24-72.1)
At 36 mo	20 (0-48.4)	69.3 (21.1-88.1)
Restricted mean survival time, mo	50.7 (43.3-58.06)	20.8 (10.8-30.9)	.018b
Cumulative incidence of new bland thrombus, %			>0.9a
At 6 mo	0	11 (1.7-29)
At 12 mo	10 (0.45-37)	11 (1.7-29)
At 24 mo	10 (0.45-37)	11 (1.7-29)
At 36 mo	10 (0.45-37)	11 (1.7-29)
Cumulative incidence of any bleeding, %			.6a
At 6 mo	0	11 (1.7-29)
At 12 mo	11 (0.47-41)	11 (1.7-29)
At 24 mo	11 (0.47-41)	11 (1.7-29)
At 36 mo	26 (2.8-60)	11 (1.7-29)
Cumulative incidence of major bleeding, %			0.2a
At 6 mo	0	0
At 12 mo	11 (0.47-41)	0
At 24 mo	11 (0.47-41)	0
At 36 mo	11 (0.47-41)	0

Values in parentheses are 95% CI.

^aGray’s test.

^bWald test.

Safety outcomes were comparable between groups. Bleeding event occurred in 4 patients with 2 events per group, at a median of 6.2 months (IQR, 0.8-14.6) after TT diagnosis. One major bleeding event, an intestinal hemorrhage at a metastatic site, occurred in the prophylactic-dose anticoagulation group 11.4 months after TT diagnosis.

This cohort study describes a rare but clinically challenging scenario encountered in routine oncologic practice. Major guidelines for the management of cancer-associated VTE [[7], [8], [9], [10]] do not specifically address isolated TT, and data on the role of anticoagulation in this setting remain extremely limited.

In our cohort, patients receiving prophylactic-dose anticoagulation had longer OS compared with those receiving no anticoagulation, whereas rates of new bland thrombus formation and bleeding were similar between groups. These findings should nevertheless be interpreted with considerable caution given the retrospective design, very small sample size, and substantial risk of residual confounding. Treatment allocation was determined by MDT judgment, and patients selected for anticoagulation may have differed from untreated patients in ways not fully captured by collected variables. Unmeasured factors, including disease biology, clinician perception of bleeding risk, anticipated prognosis, and planned anticancer therapy, may all have influenced both treatment allocation and outcomes. In a cohort of only 29 patients, even modest imbalances could have generated an apparently large effect estimate. Importantly, the present study was designed as an exploratory analysis to describe real-world management and outcomes in patients with isolated TT, and to generate hypotheses rather than establish treatment efficacy.

Our findings must also be interpreted in the context of the emerging literature on anticoagulation in patients with TT, which overall has not demonstrated a consistent benefit. In RCC, Kaptein et al. observed a 6-fold higher VTE risk in patients with TT; anticoagulation showed a nonsignificant trend toward VTE reduction (hazard ratio, 0.56; 95% CI, 0.13-2.48), but was associated with higher major bleeding rates (hazard ratio, 3.44; 95% CI, 0.95-12.42), emphasizing the narrow therapeutic margin in this population [4]. In a retrospective cohort of 153 patients with solid tumor-associated TT, 26.8% received anticoagulation with no significant survival difference between the groups [11]. Similarly, in a cohort of 211 patients with TT, of whom 50.2% received anticoagulation (including 21.8% with concurrent VTE at TT diagnosis), rates of thrombosis, bleeding, and mortality were high and comparable between treated and untreated patients [12]. Two recent systematic reviews reached concordant conclusions—while anticoagulation may lower VTE risk in RCC-associated TT, bleeding risk remains high in both groups, and all available studies carry considerable risk of bias [13,14]. These prior data contrast with the signal observed in our cohort, and support a restrained interpretation of our results.

The baseline characteristics of our cohort were broadly consistent with those reported in prior studies enrolling patients with TT across multiple cancer types: majority were male (75.9%), most had metastatic disease (93.1%), and TT most commonly involved the IVC (31%) and visceral vessels (31%) [11,12]. Notably, unlike prior cohorts, we included only patients with isolated TT, systematically excluding those with concomitant bland thrombus or other indications for therapeutic anticoagulation, thereby focusing on the subgroup with the greatest management uncertainty. Furthermore, RCC and HCC accounted for only 13.8% and 3.4% of our cohort, respectively, while they were predominant in the previous studies. These differences in patient selection and tumor distribution may partly explain discrepancies between our findings and prior studies.

The primary goal of anticoagulation in isolated TT should be to prevent thromboembolic complications, rather than thrombus regression or survival improvement, without increasing bleeding risk. We did not observe a clear excess of bleeding with prophylactic-dose anticoagulation; however, the numbers of events were very small, and the study was substantially underpowered to assess safety. Larger retrospective cohorts and systematic reviews consistently identify bleeding as a major competing risk in TT, particularly in advanced malignancy, around invasive procedures, or with tumors involving highly vascular sites. The apparent absence of excess bleeding in our cohort should therefore not be generalized to all patients with isolated TT.

Our study has several strengths. All cases were reviewed within a dedicated Thrombosis MDT, improving consistency in diagnostic adjudication and treatment decision-making. Imaging-based criteria defined isolated TT, and the cohort reflects real-world practice across a broad spectrum of malignancies and anatomical locations at a tertiary comprehensive cancer center.

Nevertheless, limitations must be acknowledged. Beyond the retrospective single-center design and small sample size, misclassification of thrombus type remains possible, as distinguishing isolated TT from mixed thrombus is challenging even with expert review. Important prognostic variables, including treatment response and time-varying disease burden, were not fully accounted for. These limitations reduce internal validity and generalizability and preclude firm therapeutic conclusions.

Taken together, our data remain exploratory and hypothesis-generating. They do not support a causal-survival benefit of prophylactic-dose anticoagulation and should not be used to justify routine use in all patients with isolated TT. Decisions should remain individualized and multidisciplinary, incorporating tumor type, thrombus location, bleeding risk, renal function, platelet count, prognosis, and planned therapy. Future multicenter prospective studies are needed to determine whether any subgroup of patients with isolated TT derives a net clinical benefit from anticoagulation and to better define the balance between thromboembolic prevention and bleeding risk.