Screening for Occult Cancer after Unprovoked Venous Thromboembolism; Assessing the Current Literature and Future Directions

Abstract

While significant evidence has established an increased rate of thrombosis in patients with cancer, the risk of occult malignancy in the setting of an unprovoked thrombosis is less clear. Despite continued interest in developing an effective screening system for occult malignancy following unprovoked venous thromboembolism (VTE), discrepancies in the literature and guideline recommendations leave providers uncertain whether to screen or perform further diagnostics for this patient population. Evidence suggests that screening for malignancy can detect cancer sooner in patients with unprovoked VTE, but there is a lack of high-quality evidence demonstrating improvements in survival who receive early detection. In the following manuscript, we summarize VTE in relation to cancer epidemiology and pathophysiology. Our literature review summarizes the spectrum of testing strategies for occult malignancy following unprovoked VTE, including biomarker detection methods and various imaging approaches. We evaluate the benefit of additional diagnostic strategies, review current guidelines on the issue, and provide guidance to the reader on the best practice for investigating undiagnosed malignancy in patients with unprovoked VTE.

Introduction:

The relationship between venous thromboembolism (VTE) and malignancy was originally described by Trousseau in 1865¹. The hypercoagulable state in cancer patients is now known to be multifactorial in etiology; relative immobilization, pro-thrombotic side effects of some anti-cancer therapies (including chemotherapy, hormonal, adjuvant, anti-angiogenic therapies), surgical interventions, central venous catheter placement, and tumor-mediated procoagulant activity all play a role². Although the risk of thrombosis in patients with cancer is well established, the converse, namely the relative risk of occult malignancy in patients with seemingly unprovoked VTE, is less understood.

Given what is known about the risks of thrombosis in the setting of malignancy, as well as disparate management recommendations for VTE in patients with versus without malignancy, it is unsurprising that clinicians often embark on aggressive investigations for occult malignancy in patients with unprovoked VTE. However, determining the optimal approach for malignancy screening in this patient population has been challenging, complicated by frequent over-testing and unclear cancer-related mortality and morbidity benefit. Indeed, screening itself is not fully benign, with potential harms including needless radiation exposure, unnecessary procedures and complications thereof, and financial and psychological toxicity to name a few³. The prospect of utilizing VTE to identify occult malignancy is alluring especially with advancements in diagnostics, functional imaging with positron emission tomography (PET), direct endoscopic evaluation, and/or utilization of tumor-maker specific markers via peripheral blood testing. However, the true utilization of these advancements with unprovoked VTE are yet to be seen in clinical practice.

In this paper, we review the pathophysiology and epidemiology of occult malignancy following VTE, and summarize the published literature and guideline recommendations regarding screening methods that have been studied in the past. From this body of information, we provide practical guidance to clinicians facing this clinical scenario. Finally, we discuss future directions for research needed for cancer screening following unprovoked VTE.

The epidemiology of thrombosis and cancer:

Active cancer accounts for 20% of the overall incidence of VTE, and among ambulatory cancer patients, venous and arterial thromboembolism are the second most common non-cancer causes of death (following infection)^4,5. Cancer patients experience multiple high-risk periods for VTE, including at the time of diagnosis, during initial hospitalization, initiation of chemotherapy, and during periods of disease progression⁶. Risk factors for VTE in patients with cancer include older age, female sex, black race, immobility, underlying comorbidities, and previous personal history of VTE⁶. In one study, Black Americans had a higher incidence of cancer associated thrombosis (CAT) for all tumor types except myeloma, while Asians/Pacific Islanders had consistently lower incidence of CAT compared to non-Hispanic Whites, after adjusting for potential confounders⁷. In addition to the contributions from social determinants of health, these racial and ethnic differences may be related to variations in thrombotic predisposition by ancestry, though additional research and genomic analyses are needed⁷. Additionally, certain tumor types pose an elevated risk of thrombosis including lung, stomach, small intestine, colon, gallbladder, pancreas, soft tissue, ovary, and kidney/renal pelvis, melanoma, NHL, myeloma, and AML⁸. Another study encompassing total 15,348 patients with DVT and 11,304 with PE, found an increased risk of several cancer types after VTE diagnosis, with a standardized incidence ratio of 1.3 for malignancy. This increased malignancy risk was seen almost entirely during the first six months of follow-up after which the risk was determined to decline to a level slightly more than 1.0.⁹. While there is certainly variation in VTE incidence among different cancer types and patient populations it is still clear that they frequently occur together during all stages of cancer diagnosis and treatment.

Unprovoked VTE and cancer:

Multiple studies have estimated that occult cancer, often advanced or metastatic disease, is identified in approximately 5% of unprovoked VTE cases^10,11. A study with 15,348 patients with unprovoked DVT and 11,304 with PE found an increased risk associated with several types of cancer with a standardized incidence ratio of 1.3⁹. In this study almost all VTE events were found to occur within six months of an eventual diagnosis of cancer⁹. Another study using data from the California Cancer Registry found that among 528 693 cancer cases, 596 (0.11%) were associated with a preceding diagnosis of unprovoked VTE within 1 year of the cancer diagnosis. The incidence of preceding VTE was increased over that expected only during the 4-month period immediately preceding the cancer diagnosis date. Given the timing and advanced stage of the unexpected cases, it is unlikely that earlier diagnosis of these cancers would have significantly improved long-term survival¹². Recurrent VTE appears to be an even more significant harbinger of malignancy; a cohort study reported that 17% of patients with recurrent VTE also had a newly diagnosed cancer within 2 years of follow-up, as compared with only 4.5% of patients with no VTE recurrence (odds ratio 4.3; 95% confidence interval [CI] 1.2–15.3, P = 0.024)¹³. Overall, the literature suggests to clinicians a window of opportunity exists for cancer screening and diagnosis in the setting of an unprovoked VTE. While these observational studies certainly suggest a strong association of unprovoked VTE and malignancy, randomized control trial and systematic review data as outlined in Table 1 paint a less promising picture.

Table 1.

Table 1 displays and summarizes selected studies that have investigated the results on unprovoked VTE and different levels of screening

Study authors, year	Design	Sample size	Limited screening	Extensive screening	Follow up period	Detection outcomes	Mortality outcomes	Conclusion
Carrier et al., SOME trial, 201511	Randomized control trial	862	Typical cervical, breast, prostate cancer testing and simple blood tests with chest radiography	Limited screening + CT abdomen and pelvis	1 year	No statistically significant difference in the rates of cancer detection	No difference in cancer-related mortality during the follow up period	Routine screening with CT abdomen and pelvis did not provide a clinically significant benefit
Prandoni et al., 201628	Randomized control trial	195	Appropriate screening determined by physician clinical judgment and patient preference	CT chest, abdomen, pelvis in combination with fecal occult blood testing	2 year	No statistically significant difference in rates of cancer detection	No difference in cancer-related mortality during the follow up period	CT + FOBT did not provide a clinically significant benefit over more limited cancer screening for detecting occult cancer in patients with unprovoked VTE
Robin et al., MVTEP, 201629	Randomized control trial	394	Limited screening (physical examination, usual laboratory tests, and basic radiographs)	Limited screening + 18F-fluorodeoxyglucose (18F-FDG)	2 year	No statistically significant difference in rates of cancer detection	Risk of subsequent cancer diagnosis was lower in patients who had negative initial screening with the 18F-FDG PET/CT than in patients who had negative initial limited screening	18F-FDG PET/CT might be useful in a selected population of patients deemed higher risk of cancer
Piccioli et al., 200430	Randomized control trial	201	No further testing	Ultrasound prostate, CT abdomen and pelvis, EGD, colonoscopy, tumor markers, mammography, sputum cytology	2 year	Malignancies identified in extensive screening group were at an earlier stage and the mean delay to diagnosis for cancer was reduced from 11.6 months to 1.0 month	No difference in cancer-related mortality during the follow up period	Although early detection of occult cancers may be associated with earlier diagnosis, it is still uncertain whether this improves the prognosis in patients as reflected by the insignificant change in mortality
Robertson and Agarwal, 201540	Systematic review (RCT and quasi-RCT included)	2 studies	Limited	Extensive	-	Extensive screening for malignancy for unprovoked VTE patients led to an earlier stage diagnosis of cancer (approximately 10 months)	No difference in cancer-related mortality during the follow up period	There was insufficient evidence to conclude if additional testing was effective for reducing cancer and VTE related mortality and morbidity. Additional testing also had no effect on the number of cancer-related deaths and interestingly did not identify more people with cancer.
Carrier et al., 200841	Systematic review	50 studies + 1 abstract	History, physical examination, and usual laboratory tests	Limited screening + imaging techniques or tumor marker measurement	-	The period prevalence of previously undiagnosed cancer in patients with unprovoked VTE was 6.1% at baseline and 10.0% from baseline to 12 months. Extensive screening statistically significantly increased the proportion of previously undiagnosed cancer detected in patients with unprovoked VTE.	No determination of mortality benefit	Previously undiagnosed cancer is frequent in patients with unprovoked VTE, stating that many undiagnosed cancers are missed by the limited screening method. They suggested an extensive cancer screening strategy detects more malignancy than does a limited screening strategy and support extensive screening use in unprovoked VTE patients.
Hildyard et al., 201631	Retrospective chart review	62	History and examination, basic blood tests, dipstick urinalysis and chest radiograph	Limited screening + CT CAP and mammogram if female	1.5 year	Researchers were only able to identify one incidental cancer (early urothelial carcinoma of the bladder). However, 28 of these scans identified other incidental abnormalities, which necessitated extensive follow-up investigations.	No determination of mortality benefit	CT screening for cancer in patients who do not have signs or symptoms of cancer is not of overall benefit to patients or a good use of hospital resources, they would advise against extensive screening.
Mwirigi et al., 201532	Retrospective chart review	239	Full history and clinical examination, full blood count, renal and liver function, chest X-ray and urinalysis. Prostatic Specific Antigen (PSA) was measured in all 35 men, whilst 15 of the 23 women had a mammogram	Limited screening + Abdominal CT or ultrasound	2 year	None of the additional diagnostics helped identify a malignancy.	No determination of mortality benefit	Based on the results of the abdominal imaging, patients required further diagnostics including breast biopsy, colonoscopy, and MRI. Risk of further testing from additional imaging among this patient population without clear benefit for cancer diagnosis.

Scoring Systems for VTE Prognostication:

The RIETE cancer score was developed to assess cancer risk in patients with an unprovoked VTE¹⁴. The RIETE score retrospectively determines the following factors as higher risk for occult malignancy: male gender, age >70, chronic lung disease, anemia, elevated platelet count, prior VTE, and recent surgery. In one analysis, 11.8% of patients with a “high” score were ultimately diagnosed with cancer, compared to 3.6% of patients assigned a “low” score¹⁴. However, another study found that a high RIETE score was not significantly associated with cancer diagnosis, a finding ascribed to poor predictive ability in women¹⁵. Improved clinical tools are needed to identify patients at high-risk for occult malignancy who may benefit from more aggressive screening for cancer.

Pathophysiology of cancer associated thrombosis:

The typical coagulation pathway initiates when insult to tissue endothelium occurs, but otherwise the cascade remains quiescent. However, tumor cells can manipulate this tightly regulated process towards coagulation and eventually predispose to thrombosis. Tumor cells are sophisticated and have several mechanisms such as production of procoagulant factors, release of proinflammatory and proangiogenic cytokines that mediate endothelial dysfunction, or directly interact with blood cells by producing adhesion molecules that predispose to thrombosis¹⁶. Tissue factor (TF), a protein that works within the extrinsic coagulation pathway to facilitate factor C activation, fibrin synthesis, and platelet activation, has been implicated in both tumor progression and in VTE formation in patients with malignancy. Studies suggest that TF along with other components of the coagulation cascade are expressed at higher levels in tumor cells suggesting an environment perpetuated by cancer cells towards thrombosis¹⁶. Cancer cells can also produce plasminogen activator inhibitor-1, which inhibits the inherent fibrinolytic system¹⁷, or produce cancer procoagulant (CP) that directly activates factor X in the coagulation cascade¹⁸. At the cellular level the multiple biochemical processes that tumor cells initiate predispose toward an overall hypercoagulable state among cancer patients.

Spectrum of available cancer screening strategies following unprovoked VTE: Laboratory

The use of specific biomarkers for cancer screening via serum testing has brought attention to non-invasive cancer detection. Somatic tumor mutations such as STK11, KRAS, CTNNB1, KEAP1, CDKN2B, and MET have been associated with an increased risk of VTE in patients with solid tumors¹⁹. Cell-free DNA, which utilizes non-encapsulated DNA in the peripheral bloodstream, is a relatively new modality of peripheral-blood cancer screening. Through identification of somatic mutations in tumors or circulating tumor DNA (ctDNA) this approach has the potential to act as a liquid biopsy, although this approach is not without drawbacks²⁰. Thus far the approaches in epigenetics for early detection, treatment, and disease monitoring have been developed for colorectal, breast, glioblastoma, bladder, and lung cancers with varying degrees of success²¹. The use of ctDNA for screening thus far is limited by cancers with low presence in peripheral circulation, either due to the cancer type itself (gliomas), or due to an earlier stage with limited ctDNA peripheral shedding²⁰. Epigenetics and DNA methylation for colorectal and lung cancer detection have been in development with potential for widespread utility, but are still being refined^22,23(Powrózek et al., 2014). Regardless of drawbacks and limitations in peripheral blood testing for cancer, there is a clear demand and need among patients and clinicians to identify sensitive and specific biomarkers for early detection of cancer. Furthermore, the technology has not been utilized on the large-scale in patients with unprovoked VTE.

Thrombocytosis has been identified as an independent biomarker of malignancy. Tumor cells are thought to induce platelet formation through release of cytokine IL-6 which stimulates thrombopoietin production, megakaryocyte cell division in the bone marrow, and platelet formation²⁴. A nested case-control of 8,917,187 patients measuring elevated platelet count on CBC with matched controls, found that an elevated platelet count was associated with a cancer diagnosis within 10 years of the blood test. Another study found the association of thrombocytosis and a subsequent diagnosis of cancer was greatest for colon, lung, ovarian, and gastric cancers²⁴. Researchers from that study proposed platelet count could potentially be used as an independent biomarker for presence of undiagnosed malignancy. While it is quite simple to obtain a platelet count for patients, using the marker of thrombocytosis for malignancy must be clinically correlated in patients (especially in those patients who have thrombocytosis from reasons independent of cancer).

Tumor-derived tissue factor-positive membrane microparticles (MP-TF), have also been proposed as a biomarker for underlying malignancy. MP-TF have been shown to activate thrombosis in some patients²⁵ and have been found to be strongly associated with mortality in pancreatic cancer and other pancreaticobiliary cancers^26,27. However, in alternate studied MP-FP have been found not to have no association with future VTE risk in patients with pancreatic, gastric, colorectal or brain cancer²⁶. it’s clinical utility as a reliable prognostic biomarker remains unclear. Similar markers using cellular components or microparticles are continually being developed for cancer screening purposes.

Imaging:

Advancement in CT imaging technology have expanded their potential utility for cancer screening. Low-dose chest CT scans are widely used for lung cancer screening in higher risk individuals in the general population. While patients with unprovoked VTE may represent a higher risk population for occult malignancy, most available studies exploring this topic have not demonstrated clear benefit of more extensive screening.

The SOME trial was a multicenter, open label, randomized controlled trial (RCT) that compared a screening strategy for occult cancer using CT abdomen and pelvis to limited screening for patients with their first unprovoked venous thromboembolism¹¹. Limited screening included typical cervical, breast, prostate cancer testing and simple blood tests with chest radiography. Confirmed cancers that were missed by the initial screening strategy and detected by the end of a 1-year follow-up period were measured and compared to the advanced screening group. Results showed the prevalence of occult cancer was low among patients with their first unprovoked VTE, and that routine screening with CT abdomen and pelvis did not provide a clinically significant benefit. The SOME trial, demonstrated no difference in the rates of cancer detection, time to cancer detection, or in the overall mortality between the two groups. Similarly, no clinically significant benefit was found in a smaller randomized control trial using CT scans in combination with fecal occult blood testing for detecting occult cancer in patients with unprovoked VTE²⁸. Both these RCT displayed statistically insignificant benefit to additional CT screening in their patient cohorts, further supporting a minimalistic approach to screening for occult malignancy in unprovoked VTE patients.

Likewise, Robin et al, conducted the MVTEP trial in 2016 that compared a strategy of limited screening with more extensive 18F-fluorodeoxyglucose (¹⁸F-FDG) PET/CT. While screening with ¹⁸F-FDG PET/CT was not associated with a higher rate of cancer diagnosis. The risk of subsequent cancer diagnosis was, interestingly, lower in patients who had negative initial screening with the ¹⁸F-FDG PET/CT than in patients who had negative initial limited screening. This study sheds light on whether ¹⁸F-FDG PET/CT might be useful in a selected population of patients deemed higher risk of cancer²⁹.

Benefit of enhanced testing for occult malignancy in unprovoked VTE:

An RCT by Piccioli et al. compared extensive testing (ultrasound, CT) versus tests at physicians’ discretion in a 2-year follow up period. The outcome measured cancer-related mortality and characteristics of diagnosed cancer (rates of different types of cancers early vs late-stage). Overall, malignancies were identified in the extensive screening group at an earlier stage and the mean delay to diagnosis was reduced from 11.6 months to 1.0 month. There was, however, no difference in cancer-related mortality during the follow up period. This study demonstrated that although early detection of occult cancers may be associated with earlier diagnosis, it is still uncertain whether this improves the prognosis in patients as reflected by the insignificant change in mortality³⁰. Despite the potential for detection of earlier stage malignancy, no study has demonstrated a survival advantage.

Furthermore, additional screening carries risks including cost, exposure to radiation, procedures, and potentially needless patient stress and anxiety. A UK-based retrospective single-center evaluated 62 CT scans obtained in the setting of unprovoked VTE. With the imaging results, researchers were only able to identify one incidental cancer (early urothelial carcinoma of the bladder)³¹. However, 28 of these scans identified other incidental abnormalities, which necessitated extensive follow-up investigations³¹. Another retrospective review of all ambulatory patients diagnosed with VTE in a London hospital, followed 239 patients after unprovoked VTE over a 24-month period. These patients were offered abdominal CT or ultrasound. Based on the results of the abdominal imaging, patients required further diagnostics including breast biopsy, colonoscopy, and MRI. None of the additional diagnostics helped identify a malignancy. From their patient cohort 2 patients with abnormal chest x-rays and long smoking history were found to have lung cancer prior to abdominal imaging³². These studies highlight the risk of further testing from additional imaging among this patient population without clear benefit for cancer diagnosis.

Most studies have tried to evaluate the utility of CT, PET-CT, endoscopy, and ultrasound into screening practices with unprovoked VTE. Some studies have seen an earlier diagnosis of malignancy with advanced early imaging, however, this was also paired with increased false positive testing that can cause psychological stress and further actions such as biopsy or interval imaging. The degree to which these technologies can influence cancer screening on the large scale remains undetermined, with most studies suggesting that extensive screening among patients with unprovoked VTE is typically costly and overly invasive without a clear mortality or morbidity benefit. Limited screening strategies typically include a history and physical examination, basic blood work, chest radiograph, and age-specific and gender-specific cancer screening. Extensive screening strategies add other tests including CT or PET–CT of the chest, abdomen, and pelvis, ultrasound of the abdomen and pelvis, endoscopy, mammography, and/or serum tumor markers. Multiple studies that have found that extensive cancer screening in patients with unprovoked VTE may result in earlier detection of occult malignancy but generally does not result in mortality benefit as seen in Table 1 below.

Society Guidelines:

Guidelines regarding testing for occult malignancy in patients with unprovoked VTE have evolved over time in response to the progression of data and evidence, as summarized below.

National Institute of Health (NHS)

The UK National Institute for Health and Care Excellence (NICE) published guidelines for malignancy screening in unprovoked VTE in 2012 and more recently in 2020³³. The 2012 guidelines suggested extensive screening individuals >40 years of age who presented with an unprovoked VTE. They recommended further additional imaging, such as CT of the abdomen/pelvis or ultrasound, along with other testing such as, mammography, sputum cytology, tumor markers, fecal occult bloods and colonoscopy. At the time it was argued that the benefit of these interventions in terms of earlier diagnosis would outweigh harms, although this recommendation did receive notable criticism for encouraging over-testing in an asymptomatic patient population³⁴. However, in 2020, based on evidence from the Robertson and Agarwal 2017 Cochrane review, NICE updated their guidelines to only recommend routine screening (including a history, physical exam, and basic laboratory testing). The updated guidelines did not recommend further investigations for cancer to people with unprovoked DVT or PE in the absence of other clinically-relevant signs or symptoms.

International Society on Thrombosis and Haemostasis (ISTH)

In October 2017 the Scientific and Standardization Committee (SSC) of the International Society on Thrombosis and Haemostasis (ISTH) published guidelines on unprovoked VTE and cancer screening. They recommended limited cancer screening in first unprovoked VTE, including a thorough medical history and physical examination, laboratory evaluation (complete blood count, calcium, urinalysis, and liver function tests), and chest X-ray in addition to recommended age- and gender-specific cancer screening (cervix, prostate, lung, colon)³⁵. This recommendation has not been updated since, but has been widely accepted as standard of care for most US-based clinicians.

Future studies:

Further studies in these areas include an upcoming French-Canadian clinical trial, Screening for Occult Malignancy in Patients with Unprovoked Venous Thromboembolism (MVTEP2/S The MVTEP2 Trial)³⁶. This trial seeks to determine if among higher risk patients (≥ 50 year-old) with a first unprovoked VTE, a cancer screening strategy including a FDG-PET/CT decreases the number of missed occult cancers detected over a 1-year follow-up period as compared with a limited screening alone (NCT04304651).

Conclusion:

Few other diseases are screened for as extensively as cancer among the general population. The risk of missing a potentially early-stage diagnosis of cancer by not screening has very real implications to patients and clinicians alike. The event of an unprovoked VTE in a seemingly cancer-free patient muddies the water further. In the clinical scenario of unprovoked VTE, clinicians act quickly to rule out genetic and situational causes for VTE (surgery, immobilization). They are then left in the precarious position to determine appropriate screening for an occult malignancy given the well-known connection between cancer and VTE.

Our paper explored the epidemiology and pathophysiology of VTE and cancer, displaying that a well-known relationship does exist between the two. The article also delved into two camps of screening in patients with unprovoked VTE, extensive versus a limited approach. Our review of the literature display that extensive screening for cancer using PET/CT, endoscopy, and additional laboratory testing do not provide patients with unprovoked VTE any survival advantage even if they are diagnosed with a cancer. A few studies have shown an earlier time to diagnosis among the unprovoked VTE population. This could certainly still be an appealing advantage to clinicians and patients alike and should be a point for clinicians to highlight in their discussions with patients. More work is needed to develop an appropriate serum testing for cancer screening in this patient population. Few RCT have implemented screening for cancer with biomarkers. Developments for screening for lung, breast, and prostate cancer screening in the general population are increasingly moving toward epigenetics and serum biomarkers but are still far away from being useful for this patient population^37,38,39.

The identification and diagnostics that commence after an unprovoked VTE diagnosis can be extensive, costly, financially and psychologically draining. In accordance with current guidelines proposed by the ISTH, we recommend against extensive screening of asymptomatic patients with unprovoked VTE beyond age and gender-specific cancer screening, unless otherwise guided by history and physical examination. The proposed recommendation will save patients and clinicians time, money, and undue stress and will provides a more universal approach to managing patients with unprovoked VTE.

Novelty Statement:

Our article offers a review of multiple published studies from randomized control trial to retrospective analyses in one manuscript. This provides readers a body of evidence to reference when deciding to screen for cancer in patients with unprovoked venous thromboembolism. The central finding of our work demonstrates that additional screening with advanced imaging and lab testing do not improve mortality or outcomes among this patient population. The clinical relevance of this work will impact clinicians as they navigate clinical shared decision-making with their patients who have unprovoked venous thromboembolism.

Data availability statement:

The investigators did not generate any novel data for the manuscript.