Prevalence of right-to-left shunt in stroke patients with cancer

Abstract

Background and Objectives:

Cancer is associated with an increased risk of acute ischemic stroke (AIS) and venous thromboembolism. The role of a cardiac right-to-left shunt (RLS) as a surrogate parameter for paradoxical embolism in cancer-related strokes is uncertain. We sought to investigate the relationship between the presence of an RLS and cancer in AIS patients.

Methods:

We included consecutive AIS patients hospitalized at our tertiary stroke center between January 2015 and December 2020 with available RLS status as detected on transesophageal echocardiography (TEE). Active cancers were retrospectively identified and the association with RLS was assessed with multivariable logistic regression and inverse probability of treatment weighting to minimize the ascertainment bias of having a TEE obtained.

Results:

Of the 2236 AIS patients included, 103 (4.6%) had active cancer, of whom 24 (23%) were diagnosed with RLS. An RLS was present in 774 out of the 2133 AIS patients without active cancer (36%). After adjustment and weighting, the absence of RLS was associated with active cancer (adjusted odds ratio (aOR) 2.29; 95% confidence interval (CI), 1.14–4.58). When analysis was restricted to patients younger than 60 years of age or those with a high-risk RLS (Risk of Paradoxical Embolism Score ⩾ 6), there was no association between RLS and cancer (aOR, 3.07; 95% CI, 0.79–11.88 and aOR, 0.56; 95% CI, 0.10–3.10, respectively).

Conclusion:

RLS was diagnosed less frequently in AIS patients with cancer than in cancer-free patients, suggesting that arterial sources may play a larger role in cancer-related strokes than paradoxical venous embolization. Future studies are needed to validate these findings and evaluate potential therapeutic implications, such as the general indication, or lack thereof, for patent foramen ovale (PFO) closure in this patient population.

Keywords

Acute ischemic stroke right-to-left shunt patent foramen ovale cancer venous thromboembolism

Introduction

Cancer-related stroke is a growing area of research.^1,2 While patients with cancer face an increased risk for acute ischemic stroke (AIS) through mechanisms similar to those in stroke patients without cancer, cancer-associated mechanisms also apply.^1,2 Cancer-mediated hypercoagulability is thought to be an important driver of cancer-related stroke. Hypercoagulable stroke mechanisms can be arterial (cerebral intravascular coagulation, nonbacterial thrombotic endocarditis) or venous (paradoxical embolization of a venous thromboembolism (VTE) via a cardiac right-to-left shunt (RLS)).^2,3 In most cases, RLS is attributable to a patent foramen ovale (PFO), a fetal connection between the right and the left cardiac atrium.⁴ In approximately 25% of the general population, the foramen does not fully close in utero and persists throughout life.⁴ A PFO is observed in up to 40% of patients with cryptogenic stroke classified as embolic stroke of undetermined source (ESUS).^5,6 A less frequent cause of an RLS is an atrial septal defect (ASD), a more serious congenital defect which can be seen in people with or without PFO.⁷ Transesophageal echocardiography (TEE) and transcranial Doppler are considered to be the gold standards in RLS detection.⁸ Up to 20% of patients with cancer develop a VTE during the course of their disease.⁹ However, the role of RLS as a surrogate parameter for paradoxical embolism in patients with cancer, although often assumed, remains poorly investigated and, therefore, uncertain.^1,10–13 Earlier studies have been limited by their sample size, patient selection, or the diagnostic modality used for RLS detection.^12,13 This study investigated the association between the presence of RLS on TEE and the presence of cancer in AIS patients to explore the role of RLS in cancer-related stroke.

Methods

Study cohort

AIS patients hospitalized at our tertiary stroke center between 1 January 2015 and 31 December 2020 were evaluated for eligibility in this retrospective analysis of prospectively collected data from our institutional stroke registry (Figure 1). Patients were included if they met the following criteria: (1) imaging-proven ischemic stroke as detected on brain MRI (or CT if MRI was contraindicated) and (2) availability of TEE findings (either known from previous TEE studies or obtained from an examination during the index hospitalization) with documented RLS status. The study population was divided into two groups: patients with active cancer and patients without active cancer.

Definition of active cancer and occult cancer

Known cancer was considered active if it fulfilled the criteria from the Hemostasis and Malignancy Scientific and Standardization Committee of the International Society on Thrombosis and Hemostasis.¹⁴ Patients with cancer newly diagnosed within 1 year after the index hospitalization were classified as having “occult cancer.” Occult cancers were also considered active at the time of the index event in line with previous studies.^15,16 In accordance with previous research practice, patients with focal non-melanoma skin cancer and prior breast cancer with prophylactic hormone therapy at the time of index event were not considered to have active cancer.¹⁷

Indication for TEE

In our tertiary center, a TEE is usually performed in patients with suspected endocarditis, AIS of undetermined etiology (according to the Trial of ORG 10172 in Acute Stroke Treatment (TOAST) classification), multiple or recurrent cerebrovascular events, or a high likelihood that the AIS source is an RLS if present as defined by a Risk of Paradoxical Embolism (RoPE) Score ⩾ 6 points.¹⁸ RLS detection is part of every TEE examination. A TEE was not considered in patients with a known AIS etiology (e.g. cardioembolism or large artery atherosclerosis with ⩾50% luminal stenosis). In severely affected patients with refractory or terminal cancer, a TEE was performed only in case of approval by either the patient or relatives. Our institutional decision tree for TEE indications during the study time frame is provided in the Supplementary material (Supplementary Figure 1). In our center, other methods for RLS detection, such as transcranial Doppler, were not performed for that indication.

Standard protocol approvals, registrations, and patient consents

The study was approved by the local ethics committee (Project ID: 2022-01560; Kantonale Ethikkommission Bern). According to the ethics committee’s decision, no informed consent from individual patients was required for inclusion in the study. Study data can be made available on reasonable request to the corresponding authors, and after clearance by the local ethics committee. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines were followed in the reporting of this study.

Data collection

From our institutional stroke registry, we extracted the following variables: demographic and baseline patient characteristics, such as sex, age, pre-stroke functional independence (defined as a modified Rankin Scale (mRS) ⩽ 2), cardiovascular risk factors (including history of diabetes mellitus type II, hypertension, hyperlipidemia, smoking, stroke, and coronary artery disease), prior medications (antiplatelet or anticoagulant therapy), National Institutes of Health Stroke Scale (NIHSS) at admission, time from last known well to admission, and the site of the vessel occlusion. The presence of multi-territory infarction (involving at least two brain vascular territories) was determined from baseline neuroradiological reports and imaging. Stroke etiology was classified according to the TOAST and ESUS criteria, except that, for the purposes of this study, RLS-associated stroke was recorded as ESUS and not as cardioembolic stroke.^19,20

A neurologist (F.S.) blinded to patients’ cancer status abstracted the RLS status from the TEE reports and data on VTE from our institution’s electronic health record. RLS was defined as any PFO or an ASD resulting in blood flow through the interatrial septum, regardless of the shunted volume, demonstrated spontaneously or with the help of the Valsalva maneuver or the use of agitated saline contrast bubble or both. In the case of an ASD, the shunt can be predominantly left to right due to higher left atrial pressure. However, the right atrial pressure can transiently exceed the left atrial pressure, resulting in RLS that might cause paradoxical embolism.²¹ VTE was considered present if deep venous thrombosis or pulmonary embolism had been documented within 1 year before or after the index AIS. Documented VTEs were either symptomatic or asymptomatic and diagnosed as part of the AIS evaluation. Two neurologists (J.G. and M.B.) assessed the presence of known active and occult cancer and associated characteristics (histological type, localization, and cancer stage at the time of AIS^22,23) from our institution’s electronic health record. We also collected data on leukocyte count, hemoglobin, platelet count, international normalized ratio (INR), fibrinogen, D-dimer, and C-reactive protein (CRP).

Outcome variables and statistical analysis

Baseline characteristics were reported with median and interquartile range (IQR) for continuous variables and frequency (percentage) for categorical variables. Differences between included and excluded patients were assessed with Fisher’s exact test for categorical variables and the Wilcoxon rank-sum test for continuous variables. The same analyses were performed in included patients with and without active cancer.

Our first set of analyses used multivariable logistic regression models to determine potential associations between active cancer, RLS, and the following covariates (male sex, age at admission, prior antiplatelet drugs, prior anticoagulant drugs, VTE, CRP, D-dimer, hemoglobin, multi-territory infarction on baseline imaging, and ESUS). These covariates were selected based on previously published studies and pathophysiological considerations.^2,16,24 Interaction analyses were performed to identify conditions that could have influenced the association between active cancer and RLS (i.e. VTE status × absence of RLS interaction term with active cancer as dependent variable). Logarithmic transformation was applied to continuous variables with a skewed distribution. Patients who received intravenous thrombolysis before blood could be drawn for laboratory analysis (i.e. outside the hospital) were excluded from analyses that included blood biomarkers.

Our second set of analyses calculated propensity scores and used the inverse probability of treatment weighting (IPTW) method to minimize the confounding effects of TEE indication.²⁵ IPTW was applied with the use of stabilized weights to adjust for the covariates listed above to minimize the imbalance between the groups in the propensity scores.²⁶ Adjusted odds ratios (aORs) were reported with their corresponding 95% confidence intervals (95% CI).

We performed subgroup analyses based on age (< 80 and < 60 years of age), the presence of occult cancer only (occult cancer versus cancer-free patients after exclusion of patients with known active cancer), and among patients whose stroke would be more likely to be caused by PFO, if present (RoPE Score ⩾ 6) versus all others.

Cross-tabulations using the chi-square test were set up to assess, in patients with active cancer, the association between RLS and VTE. Distribution of cancer stage in cancer patients with and without diagnosis of RLS was assessed using median, IQR, and the Wilcoxon rank-sum test. Further analyses of the predictive value of RLS status and performance of predictive models are reported in the supplementary material (Supplementary Method and Results I).

No imputation was applied to compensate for missing data. An alpha error of less than 0.05 was considered statistically significant. Analyses were performed with Stata 16 (StataCorp LLC).

Results

Of the 5012 patients with AIS treated at our tertiary stroke center from January 2015 through December 2020, 2236 patients (44.6%) with available TEE and assessable RLS status were included in this study (Figure 1). Active cancer was identified in 103 of these patients (4.6%), of whom 32 (1.4% of the total study population) had occult cancer. The detailed distribution of histological type and location of cancer is shown in Supplementary Figure 2. A comparison between included and excluded patients is provided in Supplementary Table 1. Compared to excluded patients, included patients were on average younger (median age (IQR) 68 (58–76) versus 78 (68–85) years) and had a lower prevalence of active cancer (7.3% versus 4.6%).

The baseline characteristics of included patients with and without active cancer are shown in Supplementary Table 2. Compared to patients without active cancer, patients with active cancer were significantly more often treated with anticoagulation before AIS, had higher NIHSS scores at admission, lower RoPE scores (median (IQR): 4 (3–5) in patients with active cancer versus 5 (4–6) in cancer-free patients), and higher frequency of VTE diagnoses (14.6% versus 2.5%). When subcategorized, deep venous thrombosis (7.8% versus 1.9%) and pulmonary embolism (6.8% versus 0.6%) were both more common in the active cancer group.

Association of active cancer with RLS status

An RLS was detected in 35.7% of study patients (n = 798/2236). This comprised 23.3% (n = 24/103) of patients with active cancer and 36.3% (n = 774/2133) of patients without cancer (p = 0.008). Among patients with active cancer (n = 103), there was no association of comorbid VTE with the diagnosis of RLS (chi-square test: p = 0.323, Supplementary Table 3). Specifically, an RLS was diagnosed in 2 of the 15 patients with VTE (13%; 95% CI, 3%–41%) versus 22 of 88 patients without VTE (25%; 95% CI, 17%–35%). Information about cancer stage at the time of AIS was available in 93 patients. There was no difference in the cancer stage distribution in cancer patients with versus without the diagnosis of RLS (median (IQR) 3 (2–4) versus 3 (2–4), p = 0.82). In the first set of adjusted analyses using multivariable logistic regression, active cancer was associated with the absence of RLS (aOR, 2.62; 95% CI, 1.28–5.38) (Figure 2). The presence of VTE did not influence the association between active cancer and the absence of RLS (p for interaction = 0.115, Figure 2). The association between active cancer and absence of RLS remained significant in patients younger than 80 years of age (n = 1913; aOR, 3.50; 95% CI, 1.52–8.07; Supplementary Figure 3) and patients in the occult cancer group (aOR, 5.40; 95% CI, 1.18–24.74; Supplementary Figure 4). However, among patients younger than 60 years of age (n = 646), the absence of RLS was no longer significantly associated with active cancer (aOR, 1.41; 95% CI, 0.17–11.53; Supplementary Figure 5). Similarly, when restricted to included patients with a RoPE Score ⩾ 6 (n = 218/1415), no association existed between active cancer and the absence of an RLS (aOR, 0.41; 95% CI, 0.10–1.58, p = 0.193).

In the second set of adjusted analyses using IPTW, the stabilized weights appeared normally distributed around one, with a few weights above two, which could have substantially impacted the results (Supplementary Figure 6). Using IPTW did not change the association between active cancer and the absence of RLS for the overall cohort (aOR, 2.29; 95% CI, 1.14–4.58), for patients younger than 80 years of age (aOR, 3.03; 95% CI, 1.39–6.62), and for patients with occult cancer at the time of AIS (aOR, 5.28; 95% CI, 1.12–24.88). The association between active cancer and RLS remained nonsignificant among patients younger than 60 years of age (aOR, 3.07; 95% CI, 0.79–11.88) and those with a RoPE score of ⩾ 6 (aOR, 0.56; 95% CI, 0.10–3.10).

Discussion

Of the 2236 patients with AIS who underwent TEE at a tertiary care stroke center, we found that an RLS was less prevalent in patients with cancer than in patients without cancer. This inverse association was also present among patients aged below 80 years and among those assumed to have occult cancer at the time of AIS. Conversely, when restricted to patient subgroups considered to be at high risk for paradoxical embolization, namely those younger than 60 years with a high RoPE score, or a known VTE, there was no significant association between the presence or absence of an RLS and active cancer.

An RLS, predominantly from a PFO, is observed in approximately 25% of the general population and may be present in up to 40% of ESUS patients.^4–6 Based on the results of multiple randomized trials, the American Heart Association/American Stroke Association 2021 guideline update states that PFO closure for secondary stroke prevention might be beneficial in selected patients younger than 60 years with a RoPE score ⩾ 6.²⁷ Meanwhile, the role of PFO in cancer-associated AIS remains controversial. As VTE risk is increased approximately fivefold in patients with cancer, it is reasonable to assume that the presence of RLS may increase the risk of paradoxical embolization and subsequent AIS in some patients.^2,3,24 However, epidemiological data to support this supposition are scarce. To our knowledge, only one study by Iguchi et al (n = 184 total AIS patients; n = 11 with AIS and cancer) has investigated the association between RLS and active cancer in AIS patients.¹² In this analysis, an RLS was more frequent in patients with active cancer than in patients without cancer (55% versus 15%, p = 0.001), suggesting that paradoxical embolism may be an important cause of stroke in patients with cancer. However, this study was limited by its small sample size, enrichment of patients with advanced stage cancers, and reliance on transthoracic echocardiography (TTE), instead of TEE, to detect RLS.^8,12

In our study, which had a sample size more than 10-fold larger than that studied by Iguchi et al, an RLS was less prevalent in AIS patients with cancer than in those without cancer. We did not find a difference in cancer stage distribution in patients with versus without a diagnosis of RLS.

Furthermore, even among patients with known VTE, there was no positive association between the presence of an RLS and active cancer status. These data suggest that arterial stroke mechanisms through both distal embolism and in situ thrombosis predominate in cancer-related AIS. At the same time, paradoxical embolization from venous thrombi through an RLS seems less frequent.

To confirm this hypothesis, an analysis of thrombus composition in cancer-related stroke patients in the presence and absence of RLS should be carried out in future studies. Arterial thrombi are believed to generally contain higher fibrin and platelet fractions than venous thrombi, which tend to be more red blood cell (RBC)-rich.^28,29 Härtl et al. recently demonstrated in cryptogenic stroke patients with large vessel occlusion treated with mechanical thrombectomy that the RBC proportion was higher in patients with a PFO than in those without.¹¹ Their findings histologically support the concept of PFO having a causative role in cryptogenic ischemic stroke through paradoxical embolism of venous thrombi. Meanwhile, retrieved thrombi from cancer-related strokes have previously been shown to be fibrin- and platelet-rich.^30–32 However, the presence of a PFO was not taken into account in these studies. Therefore, it remains uncertain from such histopathological studies whether the presence of PFO in cancer-related stroke impacted the composition of retrieved thrombi from large vessel occlusion and contributed to stroke development.³³

The findings from our study do not support the potential clinical benefit of PFO closure in patients with cancer and AIS, as discussed by Potugari et al. in their case report.¹³ Given the inverse association we found between active cancer and the presence of RLS, one could argue that the focus of secondary stroke prevention in patients with active cancer and AIS, and concomitant PFOs should be antithrombotic therapy and not PFO closure.^14,34,35 Due to its medium sensitivity, low specificity and low positive predictive value (PPV) value, the absence of RLS alone is not sufficient for the prediction of underlying cancer in AIS patients (see Supplementary Method and Results I). As the difference between multivariable predictive models including and excluding RLS status was not significant, it is not suitable to characterize the absence of RLS as a candidate biomarker for cancer-related strokes (see Supplementary Method and Results I).

Limitations

This study has several limitations. First, due to its retrospective cohort design, ascertainment bias of having a TEE obtained, is possible. Patients with cancer who harbored an RLS may have been less likely to undergo TEE because they were viewed as sicker or less likely to undergo subsequent PFO closure if an RLS was detected. This concern is somewhat mitigated by our institution’s systematic decision tree for determining who should undergo TEE and our use of IPTW analyses to statistically account for potential differences in indication. Second, a systematic screening for VTE in AIS patients was not performed at our institution. The number of VTEs in this study may be underestimated, as only symptomatic VTEs or asymptomatic VTEs discovered during routine evaluation for the index AIS have been reported. Third, the frequency of patients with occult cancer is likely underestimated as we were unable to capture cancer diagnoses made at other centers after AIS. Moreover, the last clinical follow-up date at our center is unavailable in the database. It made it impossible to determine the proportion of included patients with a completed follow-up time (of at least 1 year) regarding the identification of occult cancer at the time of AIS. Fourth, due to the small number of patients considered to have occult cancer, the association analyses for this subgroup were imprecise. Fifth, our study was conducted at a tertiary care referral center in Europe where the patient population was predominantly White and therefore, our findings may not be generalizable to other settings.

Conclusion

This study raises doubts concerning the assumed role of paradoxical embolism as a frequent cause of ischemic stroke in cancer patients because of the demonstrated negative correlation between RLS and cancer. Future research to determine alternative causes of stroke, such as arterial thromboembolism, in this patient population is warranted and could inform clinical practice and the potential utility or lack thereof for PFO closure.

Supplemental Material

sj-docx-1-wso-10.1177_17474930241260589 – Supplemental material for Prevalence of right-to-left shunt in stroke patients with cancer

Supplemental material, sj-docx-1-wso-10.1177_17474930241260589 for Prevalence of right-to-left shunt in stroke patients with cancer by Fabienne Steinauer, Philipp Bücke, Eric Buffle, Mattia Branca, Jayan Göcmen, Babak B Navi, Ava L Liberman, Anna Boronylo, Leander Clenin, Martina Goeldlin, Julian Lippert, Bastian Volbers, Thomas R Meinel, David Seiffge, Adnan Mujanovic, Johannes Kaesmacher, Urs Fischer, Marcel Arnold, Thomas Pabst, Martin D Berger, Simon Jung and Morin Beyeler in International Journal of Stroke

Footnotes

Author contributions

F.S. contributed to data acquisition, interpretation of data, and writing of the publication. P.B. contributed to the conception and design, interpretation of data, writing of the publication, and supervision. E.B. contributed to the conception and design, interpretation of data, and critical revision of the article for important intellectual content. Ma.B. contributed to analysis and interpretation of data and critical revision of the article for important intellectual content. J.C. contributed to data acquisition and critical revision of the article for important intellectual content. S.J. contributed to the conception and design, critical revision of the publication for important intellectual content, and provided supervision. Mo.B. contributed to the conception and design, analysis and interpretation of data, critical revision of the publication for important intellectual content, and provided supervision. All other authors contributed to the critical revision of the article for important intellectual content.

Declaration of conflicting interests

The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: Mo.B. reports research support from the Department of Neurology, Inselspital, Bern University Hospital and the University of Bern, Switzerland.

None of the other authors report any conflicts of interest in relation with this study.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by grants provided by the “Kurt und Senta Herrmann-Stiftung” (grant no. WNK-228), the Department of Neurology, Inselspital, Bern University Hospital, and University of Bern (the Research Grant Diversity and Excellence in Clinical Research 2022) and the University of Bern (UniBE Doc. Mobility grant).

ORCID iDs

Philipp Bücke

Mattia Branca

Babak B Navi

Martina Goeldlin

Bastian Volbers

David Seiffge

Adnan Mujanovic

Morin Beyeler

Supplemental material

Supplemental material for this article is available online.

References

Bang

Chung

Lee

, et al. Cancer-related stroke: an emerging subtype of ischemic stroke with unique pathomechanisms. J Stroke 2020; 22: 1–10.

Navi

Kasner

Elkind

MSV

et al. Cancer and embolic stroke of undetermined source. Stroke 2021: 1121–1130.

Navi

Reiner

Kamel

, et al. Association between incident cancer and subsequent stroke. Ann Neurol 2015; 77: 291–300.

Bang

Lee

Ryoo

et al. Patent foramen ovale and stroke–current status. J Stroke 2015; 17: 229–237.

Kraywinkel

Jauss

Diener

, et al. Patent foramen ovale, atrial septum aneurysm, and stroke. An examination of the status of recent evidence. Nervenarzt 2005; 76: 935–942.

Vermeiren

Haven

Dikranian

, et al. Patent foramen ovale and cryptogenic stroke. JBR-BTR 2009; 92: 265.

Webb

GM.

Atrial septal defect in the adult. Circulation 2006; 114: 1645–1653.

Mojadidi

Mahmoud

Elgendy

, et al. Transesophageal echocardiography for the detection of patent foramen ovale. J Am Soc Echocardiogr 2017; 30: 933–934.

Blom

Doggen

CJM

Osanto

et al. Malignancies, prothrombotic mutations, and the risk of venous thrombosis. JAMA 2005; 293: 715–722.

10.

Windecker

Stortecky

Meier

Paradoxical embolism. J Am Coll Cardiol 2014; 64: 403–415.

11.

Härtl

Berndt

Poppert

, et al. Histology of cerebral clots in cryptogenic stroke varies according to the presence of a patent foramen ovale. Int J Mol Sci 2022; 23: 9474.

12.

Iguchi

Kimura

Kobayashi

et al. Ischaemic stroke with malignancy may often be caused by paradoxical embolism. J Neurol Neurosurg Psychiatry 2006; 77: 1336–1339.

13.

Potugari

Priyanka

Komanapalli

, et al. Ovarian cancer presenting as cryptogenic stroke from patent foramen ovale. Clin Med Res 2019; 17: 97–101.

14.

Khorana

Noble

Lee

AYY

, et al. Role of direct oral anticoagulants in the treatment of cancer-associated venous thromboembolism: guidance from the SSC of the ISTH. J Thromb Haemost 2018; 16: 1891–1894.

15.

Rioux

Touma

Nehme

et al. Frequency and predictors of occult cancer in ischemic stroke: a systematic review and meta-analysis. Int J Stroke 2021; 16: 12–19.

16.

Beyeler

Birner

Branca

, et al. Development of a score for prediction of occult malignancy in stroke patients (occult-5 score). J Stroke Cerebrovasc Dis 2022; 31: 106609–106609.

17.

Beyeler

Belachew

Kielkopf

, et al. Absence of susceptibility vessel sign in patients with malignancy-related acute ischemic stroke treated with mechanical thrombectomy. Front Neurol 2022; 13: 930635.

18.

Kent

Saver

Ruthazer

, et al. Risk of paradoxical embolism (RoPE)-estimated attributable fraction correlates with the benefit of patent foramen ovale closure: an analysis of 3 trials. Stroke 2020; 51: 3119–3123.

19.

Hart

Sharma

Mundl

, et al. Rivaroxaban for secondary stroke prevention in patients with embolic strokes of undetermined source: design of the NAVIGATE ESUS randomized trial. Eur Stroke J 2016; 1: 146–154.

20.

Nouh

Hussain

Mehta

, et al. Embolic strokes of unknown source and cryptogenic stroke: implications in clinical practice. Front Neurol 2016; 7: 37–16.

21.

Pristipino

Sievert

D’Ascenzo

, et al. European position paper on the management of patients with patent foramen ovale. General approach and left circulation thromboembolism. Eur Heart J 2019; 40: 3182–3195.

22.

Rosen

Sapra

TNM Classification. Treasure Island, FL: Statpearls Publishing, 2024.

23.

Cheson

Fisher

Barrington

, et al. Recommendations for initial evaluation, staging, and response assessment of hodgkin and non-hodgkin lymphoma: the lugano classification. J Clin Oncol 2014; 32: 3059–3067.

24.

Xiong

Current status of treatment of cancer-associated venous thromboembolism. Thromb J 2021; 19: 21.

25.

Shiba

Kawahara

Using propensity scores for causal inference: pitfalls and tips. J Epidemiol 2021; 31: 457–463.

26.

Chesnaye

Stel

Tripepi

, et al. An introduction to inverse probability of treatment weighting in observational research. Clin Kidney J 2022; 15: 14–20.

27.

Kleindorfer

Towfighi

Chaturvedi

, et al. 2021 guideline for the prevention of stroke in patients with stroke and transient ischemic attack: a guideline from the American Heart Association/American Stroke Association. Stroke 2021; 52: e364–e467.

28.

Chernysh

Nagaswami

Kosolapova

, et al. The distinctive structure and composition of arterial and venous thrombi and pulmonary emboli. Sci Rep 2020; 10: 1–12.

29.

Weisel

Litvinov

RI.

Visualizing thrombosis to improve thrombus resolution. Res Pract Thromb Haemost 2021; 5: 38–50.

30.

Chen

Lin

, et al. Fibrin and platelet-rich composition in retrieved thrombi hallmarks stroke with active cancer. Stroke 2020; 51: 3723–3727.

31.

Park

Kim

, et al. Histological features of intracranial thrombi in stroke patients with cancer. Ann Neurol 2019; 86: 143–149.

32.

Kataoka

Sonoda

Takahashi

, et al. Histopathological analysis of retrieved thrombi from patients with acute ischemic stroke with malignant tumors. J Neurointerv Surg 2022; 14: 017195.

33.

Sun

Kwak

et al. Different thrombus histology in a cancer patient with deep vein thrombosis and recurrent strokes. J Stroke 2022; 24: 300–302.

34.

Jang

Lee

, et al. Comparison of enoxaparin and warfarin for secondary prevention of cancer-associated stroke. J Oncol 2015; 2015: 502089.

35.

Gladstone

Lindsay

Douketis

, et al. Canadian stroke best practice recommendations: secondary prevention of stroke update 2020. Can J Neurol Sci 2022; 49: 315–337.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

1.05 MB