DISCUSSION:
The major focus of this study was to review the role of anticoagulation
in the treatment of tumor thrombus. Our study showed that a minority
(20%) of patients with tumor thrombus at a large tertiary center
received anticoagulation with the purpose of preventing bland thrombus
formation and/or embolization. Furthermore, there was no significant
difference in the rate of bland thrombus formation and/or embolization
or in overall survival in pediatric patients with intravascular
extension of a solid tumor or tumor thrombus treated with or without
anticoagulation. Patients who received anticoagulation experienced a
statistically significant increase in bleeding events.
Tumor thrombus is overall an uncommon clinical entity. However, its
presence can complicate the management plan. Hence, early and accurate
detection is helpful in surgical planning and in some cases, tumor
staging 9,10. Imaging is critical for early detection
and differentiation from bland thrombus 11. In adults,
renal cell carcinoma, hepatocellular carcinoma, and adrenocortical
carcinoma are the malignancies most frequently associated with tumor
thrombus12; while in children, Wilms tumor has a
tendency for vascular invasion 3,13. In Wilms tumor,
20-35% patients have renal vein involvement, while extension into the
IVC is reported in 4-10% of patients 4,13. There are
isolated case reports reporting vascular invasion and tumor thrombus in
patients with Ewing sarcoma, osteosarcoma, neuroblastoma, and
hepatoblastoma 14–17. In our study, the most common
malignancy associated with tumor thrombus was hepatoblastoma, followed
by osteosarcoma and Wilms tumor.
Most of the patients were asymptomatic and tumor thrombus was detected
on routine imaging, similar to the observation in previous
studies5,13. However, two of our patients presented
with PE and one with lower extremity pain and swelling. PE has been
previously reported as a presenting sign of primary or relapsed
osteosarcoma associated with tumor thrombus 18,19.
Yeduluri et al speculated that this could be attributable to the
predilection of metastatic spread by vascular invasion in osteosarcoma20. Our patient with Ewing sarcoma had tumor
thrombus-related PE, similar to a report by Dotson et al describing a
patient with Ewing sarcoma who developed systemic emboli and myocardial
infarction from tumor thrombus located in the left atrium21.
Although chemotherapy, surgery, and/or radiation form the backbone of
management of primary tumor and potentially, tumor thrombus, there is a
risk of tumor rupture or spread with potential complications such as PE
or systemic emboli of bland thrombus components. This risk is
particularly high in cases where tumor thrombus is located in the IVC or
extending to the heart. Hence considering anticoagulation, either by
prophylactic or therapeutic dosing, is not unreasonable. However, the
use of anticoagulation for patients with tumor thrombus is inconsistent.
While all Wilms tumor patients with intracardiac extension of the tumor
received anticoagulation post-resection for a minimum of 3 months in a
study by Cox et al 22, anticoagulation was not
attempted in some other similar studies 4,13,23 and
the results were comparable in terms of survival and response to
treatment. In a review assessing OS patients with cardiovascular
involvement, the authors reported one patient was on anticoagulation
before the diagnosis of tumor thrombus was correctly established, but
otherwise almost all patients received chemotherapy and/or surgery with
no adjuvant anticoagulation with good outcomes 20. In
a study of patients with renal cell carcinoma, the most common tumor
associated with tumor thrombus in adults, presenting with PE, 5 of the 7
patients received anticoagulation with either unfractionated heparin or
enoxaparin; anticoagulation was held in the remaining two patients due
to bleeding 24. The authors’ rationale was that
patients with malignancy are already at a high risk for thrombosis and
in the presence of an IVC tumor thrombus, there is additional
disturbance of flow through the IVC, further increasing their risk of
bland thrombosis or thromboembolism. In our patient cohort, we noted a
similar approach by the providers in that most of patients did not
receive anticoagulation, but it was initiated in 20% of the patients
for management of signs or symptoms and/or to prevent further
complications from progression of bland components of the tumor thrombus
or thromboembolism. However, we want to emphasize that although
anticoagulation could be helpful, the backbone of treatment of tumor
thrombus is surgical with or without chemotherapy and radiation (when
indicated) for the underlying malignancy 3. This is
supported by findings in our patient cohort. Our patient with relapsed
osteosarcoma continued to have worsening tumor thrombus despite
therapeutic anticoagulation until thrombectomy was performed, and
despite initial improvement, had another episode of PE in the absence of
cancer-directed therapy. In contrast, our patient with Ewing sarcoma
received anticoagulation in conjunction with chemotherapy and surgery
and had a complete resolution of the tumor thrombus. The underlying
characteristics of tumor thrombus are different than bland thrombus
making it more likely to resolve with chemotherapy and respond minimally
to anticoagulation.
The benefits of anticoagulation must be weighed against the risks of
bleeding. Malignancy itself places patients at a high risk of bleeding
and thrombosis. In our dataset, 40% patients on anticoagulation
developed bleeding, while this percentage was only 5% for the group not
on anticoagulation. This difference in bleeding was statistically
significant, which is not surprising given the use of anticoagulation in
one group, but further underscores the importance of considering
anticoagulation on a case-by-case basis.
There are some limitations of this study. The design is retrospective
with subjects being identified by reported imaging findings. It is
possible that some patients may not have been included if tumor thrombus
was diagnosed intra-operatively but not appreciated on imaging, leading
to underreporting of the finding in our study. This could possibly
explain why there was a lower proportion of Wilms tumor thrombus in our
study compared to other pediatric reports. Pathological confirmation was
not made for all the suspected tumor thrombus cases, raising the
possibility that some tumor thrombi may have a component of bland
thrombus and hence, may have responded in part to anticoagulation. As a
single center study, our sample size is small, limiting the statistical
power to draw conclusions. Hence, further larger multi-center studies
are to needed to confirm these findings.
In conclusion, this study reports the largest case series of pediatric
tumor thrombus treated with or without anticoagulation in combination
with cancer-directed therapy. An important advantage of our study over
previous reports is that we focused on the role of anticoagulation in
lessening the potential risk of bland thrombus formation and/or
embolization in this cohort of patients since data is limited on this
subject. We found no difference in rate of bland thrombus formation
and/or embolization or survival in patients who did or did not receive
anticoagulation. However, we did find a significantly higher rate of
bleeding events in the anticoagulation group. Our study provides some
practical information and guidance for clinicians considering the use of
anticoagulation in these complicated patients both at risk for bleeding
and clotting. While we found no clear evidence that anticoagulation
administered in conjunction with cancer-directed therapies improves
outcomes in pediatric patients with tumor thrombus, further studies are
indicated to identify subgroups that may benefit from
thromboprophylaxis.
Conflict of Interest: Authors have no conflict of interest to
disclose.