Original Research

Role of computed tomography in detecting incidental venous thromboembolism in cancer patients

Fatumah Ssewaali, Morontshe M. Ramantsi, Paul Ruff, Yusuf Mayet

Received: 05 Mar. 2025; Accepted: 13 June 2025; Published: 22 Aug. 2025

Copyright: © 2025. The Author(s). Licensee: AOSIS.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

Background: Cancer is one of the major risk factors for venous thromboembolism (VTE). The prevalence of incidental VTE among cancer patients in South Africa is unknown.

Aim: To determine the presence and location of VTE, find out if VTE is more common in certain cancer types, and to determine the role of human immunodeficiency virus (HIV) in the incidental finding of VTE among cancer patients.

Setting: This study was conducted at the Radiology Department at the Charlotte Maxeke Johannesburg Academic Hospital located in Johannesburg, South Africa.

Methods: There were 214 staging computed tomography (CT) scans of eligible adult cancer patients from January 2018 to June 2018. These were identified using a picture archiving and communication system (PACS) and analysed retrospectively by three consultant radiologists. Univariate analysis and multivariable logistic regression models were used to investigate associations between VTE and HIV, age, gender, ethnicity and common cancers.

Results: The mean age was 53 years (standard deviation [s.d.] 13.98) with the age range being 18–86 years, and 64% of these patients were female. Incidental VTE was 14.9% (95% confidence interval [CI] 10.7% – 20.4%). Pulmonary embolism (PE) accounted for 11.2%, while abdominal deep vein thrombosis (ADVT) accounted for 3.7%. Moreover, HIV-positive cancer patients were three times more likely to have VTE compared to HIV-negative cancer patients, with adjusted odds’ ratio 3.2 (1–10), p = 0.04. There was no association found between cancer type and VTE.

Conclusion: This study revealed cancer and HIV infection as risk factors for patients developing VTE. Pulmonary embolism was more common than ADVT. Actively search for VTE in cancer patients when reviewing staging CT scans.

Contribution: This is the first research in South Africa to determine the prevalence of incidental VTE among cancer patients.

Keywords: thrombosis and embolism; pulmonary embolism; deep vein thrombosis; cancer patients.

Introduction

Cancer is a major global health problem. Cancer may predispose patients to venous thromboembolism (VTE).^1,2,3,4 Around 4% – 20% of patients with cancer will develop VTE at some stage. The risk is at its highest in the acute period following diagnosis, mostly due to curative interventions like chemotherapy.¹ A two-way relationship exists between cancer and VTE. A higher cancer burden points to higher risk of VTE, while at the same time a high occurrence of VTE indicates that the cancer is aggressive and points to a poor prognosis for the patient.⁵

The risk of developing VTE is higher by twofold in admitted cancer patients relative to ambulatory counterparts due to immobility. Mortality is also higher among patients with cancer and VTE relative to those without, with some showing symptoms, some asymptomatic and others were diagnosed on autopsy.⁶ A study by Van der Hulle et al.,⁷ from the Netherlands found that mortality depended on the stage and type of cancer and location of pulmonary embolism (PE), with metastatic cancer, advanced stage of cancer and centrally located incidental PE, faring worst. They established that the ‘weighted pooled 6-month mortality’ was 10% higher among cancer patients who did not get treatment with low molecular weight heparin (LMWH) in comparison to those who received LMWH.^1,7

Studies by Goldstein and Wu from South Africa⁸ and Sule AA et al.,⁹ from Singapore suggest that human immunodeficiency virus (HIV) infection is being recognised as a risk factor on its own for patients developing VTE, despite prior controversy about whether an association between HIV and VTE exists.^8,9 The prevalence of HIV in South Africa in 2019 was 13.5% of the population of 58.78 million (i.e. 7.97 million living with HIV).¹⁰

Although modern multi-detector computed tomography (MDCT), unlike computed tomography pulmonary angiography (CTPA), is not the gold standard for diagnosing PE, it is possible to detect VTE on these scans due to its high spatial and temporal resolution along with the amount of contrast used.¹¹ There is still not enough information about incidental VTE among cancer patients in the South African population.

Therefore, the aim of the study was to determine the prevalence of VTE among adult cancer patients at the Charlotte Maxeke Johannesburg Academic Hospital. Our primary objective was to determine the prevalence of VTE among cancer patients, while secondary objectives were to document the anatomical location of thrombi and the effect of HIV on the susceptibility of patients developing VTE.

Methods

This was a retrospective descriptive, cross-sectional study carried out in the Radiology Department at the Charlotte Maxeke Johannesburg Academic Hospital located in Johannesburg, South Africa. Age, gender, HIV status and cancer type data were extracted from clinical information on the radiology reports. Information on ethnicity and, in some instances, HIV status, was obtained from review of patients’ records.

In this study, three consultant radiologists, each with at least 2 years’ experience, re-read the anonymised computed tomography (CT) scans. The readers were blinded to the patients’ clinical details and each other’s imaging findings.

The study population consisted of adult patients (18 years and older) with histologically proven cancer who came in for initial staging CT of the chest, abdomen and pelvis. This was extended to head and neck, which is included in lymphoma patients. Consecutive sampling was used to select CT scans over a 6-month period, from 01 January 2018 to 30 June 2018, at Charlotte Maxeke Johannesburg Academic Hospital.

All adult patients with any type of cancer who came in for a baseline staging CT scan before cancer treatment were included. Those receiving treatment for PE or deep vein thrombosis (DVT) at that time were excluded. This information was obtained from the documented patients’ history on the radiology request form from the picture archiving and communication system. Patients’ records were only reviewed to obtain the HIV status and ethnicity.

The data were checked for inconsistencies in gender for sex-specific cancers and for missing data.

Statistical analysis

All statistical analyses were conducted in Stata version 15. A p-value < 0.05 was considered significant.

Descriptive statistics

Continuous variables such as age were described in terms of mean and standard deviation (s.d.). Categorical variables such as gender, ethnicity, HIV status, cancer type, and presence and location of VTE were described using frequencies and proportions. A histogram was used to describe age distribution.

Inferential statistics

Univariate analysis and multivariable logistic regression models were used to investigate associations between VTE and HIV infection, age, gender, ethnicity and common cancers. Variables with p < 0.3 in univariate analysis were included in the final model. Cohen’s Kappa statistic was used to assess interrater agreement.

Ethical considerations

Ethical clearance to conduct this study was obtained from the University of the Witwatersrand, Human Research Ethics Committee (Medical) (reference number: M200257).

Results

Of the 222 CT scans collected, eight CT scans were excluded due to inadequate images. The mean age was 53 years (s.d. 13.98) with a range of 18–86 years, and 64% of these patients were female and 69% were black. Most of the patients had breast cancer followed by colorectal cancer (Table 1). The presence of VTE was higher among black patients at 11.68%, while non-black patients accounted for 3.27%. This, however, was not statistically significant (Table 1). Locations of VTE varied among study participants. Six of the 32 cases had VTE in the sub-segmental arteries (Table 2). Results of the multi-variable analysis indicate that a cancer patient with HIV was 3 times more likely to have VTE compared to an HIV-negative cancer patient. This is demonstrated by the adjusted OR 3.2, 95% CI (1–10) and p-value 0.04 (Table 3).

There was no association between VTE and age, gender, ethnicity, cancers of the breast, colorectal, lung, cervix and lymphoma. It was found that HIV infection was the only factor with statistically significant association with VTE. Moreover, HIV-positive cancer patients were three times more likely to get VTE compared to cancer patients without HIV. Venous thromboembolism was commonest among patients with breast cancer (adjusted odds ratio 0.5 [0.1–1.5], p = 0.2).

There were 32 patients with VTE, 14.9% (95% confidence interval [CI] 10.7–20.4). Among those, 12 had VTE in multiple locations (n = 12, 6%). Pulmonary embolism accounted for 11% (n = 24), while abdominal deep vein thrombosis (ADVT) accounted for 4% (n = 8) (Table 2). The commonest locations for PE were the segmental (n = 13, 6.1%) and sub-segmental (n = 11, 5.1%) arteries, and for ADVT, the most common locations were portal vein (n = 5, 2.3%) and the left external iliac vein (n = 3, 1.4%) (Table 2).

This study revealed that the presence of VTE was higher among black patients at 11.7%, while non-black patients accounted for 3.2%, with adjusted OR 0.4 (0.1–1.4) and p = 0.2 (Table 1 and Table 3).

Looking at inter-reader agreement, reader 1 found no VTE at 69%, VTE present at 25%, and could not comment on 6%. Reader 2 found no VTE at 65%, VTE present at 25%, could not comment on 10%, while reader 3 found no VTE at 62%, VTE present at 13%, and could not comment on 24%. Cohen’s kappa among readers 1, 2 and 3 was 0.24 (95% CI: 0.15– 0.31). This indicates fair agreement among the 3 readers of the presence of VTE in the study participants. Examples of cancer associated VTE are shown in Figure 1, Figure 2 and Figure 3.

FIGURE 1: Filling defect in the right lobar artery in a patient with metastatic adenocarcinoma (carcinoma of head of pancreas) (green arrow).

FIGURE 2: Filling defect in the left lobar artery in a patient with diffuse large B-cell lymphoma (green arrow).

FIGURE 3: Filling defect in the portal vein in a patient with metastatic breast cancer (green arrow).

Discussion

We have reported on the presence or absence of VTE, patient demographics and different types of malignancy among cancer patients in this study. This is the first report on incidental VTE among cancer patients at a tertiary-quaternary hospital in South Africa. The results revealed that incidental VTE is common among cancer patients with a prevalence of 14.9%. This is much higher than the range of prevalence (1.47% – 3.03%) documented in several other studies.

In a study by Escalante et al. from the United States (US) (2017), involving 1090 patients, of which 546 (50.09%) were females, the prevalence of incidental VTE was 1.8%. Similarly, Shinagare et al. from the US (2011) studied 13 783 patients and found prevalence of VTE to be 1.47%.^1,12 Additionally, Di Nisio et al. from the Netherlands (2018), involving a study on 330 patients, of which 134 (40.6%) were females, found a prevalence of 3.03%.¹³

A review by Abdol Razak et al. from Australia⁶ found that black patients were more likely to have VTE (5.1%) compared to Hispanic and white patients (4.0%). In our study population, 69% of the patients were black; however, the p-value for ethnicity was 0.214 (> 0.05), showing no significance.

Similarly, in a study by Addo-Tabiri et al., from the US where 33.32% were black and 54.92% white, proved higher incidence of VTE in black patients, whereby the incidence of VTE was n = 97 (1.8%) among black patients and n = 57 (0.6%) among white patients.¹⁴

The commonest locations for PE were segmental and subsegmental arteries. These results were similar to those of a study by Escalante et al., where subsegmental PE were (n = 6, 43%),¹ likewise an estimated incidence of segmental PE of 1.51% by Di Nisio M et al.¹³ Abdominal DVT was predominantly in the portal vein and left external iliac vein.

This study revealed that VTE was more common among female patients (n = 21, 65.6%) which could also be due to the higher number of overall female patients compared to the males. This finding was similar to that documented by Escalante et al. (n = 14, 70%),¹ in addition to Abdol Razak et al.,^1,6 and Fernandes et al.⁴ Various possible reasons include pelvic disease, prior oral contraceptive use and raised body mass index.

Among the cancers, the prevalence of VTE differed widely by cancer type, with the highest prevalence reported in breast cancer and lymphoma (n = 6, 2.8%) – however, with p-value of 0.2 – followed by lung cancer (n = 5, 2.3%), with p-value of 0.4, and prostate cancer (n = 1, 0.5%), with p-value of 0.5.

Patients with lymphoma and breast cancer both had the highest VTE at 2.80% each, in agreement with the study by Hohaus et al., from Italy, where lymphoma was mentioned as having the highest risk for VTE.¹⁵

There was an association found between HIV and VTE, similar to the findings by Nkoke et al., from Cameroon,¹⁶ where HIV was found to be a risk factor for developing VTE as well as a study by Vululi et al., where they found that HIV patients on antiretroviral therapy (ART) were more at risk of getting DVT.¹⁷ There was no association found between age and VTE when logistic regression was used and age was categorised. This is at odds with Abdol Razak et al., who stated that with increase in age came increase in the risk of VTE in cancer patients undergoing surgery when contrasted with those < 60 years (OR = 2.6, 95% CI: 1.2–5.7).⁶

The mean age was 53 years (range 18–86 years) (Figure 4), which did not differ significantly from that documented in a prospective study by Escalante et al.,¹ where mean age was 58 years, with s.d. of 13.98 and age range being 18–87 years.

FIGURE 4: Age distribution of cancer patients: Age followed a normal distribution (range 18–86 years; mean 53.12 [standard deviation = 13.98] years).

A retrospective study by Ageno et al. from Italy and prospective studies by Vululi et al. from Uganda and Shinagare et al. looked at either PE or DVT, and this study, although retrospective, enabled us to look at both PE and DVT without being restricted to only PE or DVT.^12,17,18

Limitations

This study was limited by the use of retrospective analysis. Realising that request forms for imaging generally have limited or scanty clinical data, collaborating with clinicians in the oncology department would have increased the robustness of the dataset; for instance, obtaining from them information about treatment for VTE at the time rather than relying on the records. Moreover, the impact of risk factors such as immobility, recent surgery or trauma, medical conditions like heart failure or chronic lung disease, medical treatment such as hormone replacement therapy, and other factors like contraception, obesity as well as smoking were not explored.

Recommendations

We recommend that a large multi-centre prospective study be conducted that will enable investigators to obtain more generalisable results. The correlation between lymphoma and incidental VTE is another aspect to explore. In this study – albeit with small numbers – lymphoma was shown to have high numbers of VTE. In view of this, a study looking at VTE in lymphoma patients could validate the results. We recommend that those reporting staging CT scan go to great lengths to look for VTE whenever reporting staging CT scans, due to morbidity and mortality associated with VTE in cancer patients.

Conclusion

This study revealed cancer and HIV infection as risk factors for patients developing VTE. Pulmonary embolism was more common than ADVT. The presence of VTE as shown in our study should prompt active search for thrombi in cancer patients, just like with any other patients due to the morbidity and mortality associated with VTE.

Acknowledgements

The authors extend their thanks to Dr Neels van der Merwe, Dr Hoda Eshraghi and Dr Nadegei Ntumba who helped to re-read the scans of all the patients. This article is partially based on the author’s thesis entitled ‘Role of computed tomography in detecting incidental venous thromboembolism in cancer patients’ towards the Master’s degree in Radiology in the Department of Diagnostic Radiology, University of the Witwatersrand, South Africa, December 2024, with supervisors Dr Morontshe M. Ramantsi, Prof. Paul Ruff and Dr Yusuf Mayet.

The authors declare that they have no financial or personal relationships that may have inappropriately influenced them in writing this article.

The authors, F.S., M.M.R., P.R. and Y.M., contributed to conceptualisation, drafting and reviewing of the article. M.M.R., P.R. and Y.M. supervised the research.

The authors received no financial support for the research, authorship and/or publication of this article.

The data that support the findings of this study are available from the corresponding author, F.S. upon reasonable request.

The views and opinions expressed in this article are those of the authors and are the product of professional research. The article does not necessarily reflect the official policy or position of any affiliated institution, funder, agency or that of the publisher. The authors are responsible for this article’s results, findings and content.

References