Read the full EBM Focus and earn CME credit.
Reference - N Engl J Med 2015 Aug 20;373(8):697 (level 2 [mid-level] evidence)
Unprovoked venous thromboembolism (VTE), a VTE occurring in the absence of triggering events or risk factors, has been associated with the presence of occult malignancies. The prevalence of undiagnosed cancer may be as high as 10% at 1 year after VTE and missed diagnoses at the time of VTE was found to be about 4% in a systematic review (Ann Intern Med 2008 Sep 2;149(5):323). The Institute for Clinical Systems Improvement (ICSI) recommends evaluating patients with unprovoked VTE with a basic cancer work-up consisting of a history and physical exam, complete blood count, erythrocyte sedimentation rate, renal and liver function tests, urinalysis, and chest x-ray (ICSI 2013 Jan PDF). The addition of imaging tests (such as CT or ultrasound) and molecular biomarkers have been proposed to reduce the rate of missed cancer diagnoses at the time of VTE. However, in previous studies that compared limited screening to extensive screening procedures, there were no significant differences in overall cancer detection or in cancer-related mortality with extensive testing (J Thromb Haemost 2004 Jun;2(6):884, J Thromb Haemost 2011 Jan;9(1):79). To look at the specific role CT might play in cancer detection, a recent randomized trial compared limited cancer screening plus pelvic and abdominal CT vs. limited cancer screening alone in 852 adults (mean age 53 years, 67% male) with their first unprovoked symptomatic VTE.
Limited cancer screening included a complete history and physical exam, basic blood tests (complete blood count, serum electrolyte level, creatinine level, and liver-function testing), chest x-ray, and sex-specific cancer screening if not performed in the prior year (breast, cervical, and prostate). The pelvic and abdominal CT screening included a virtual colonoscopy and gastroscopy, biphasic enhanced liver CT, parenchymal pancreatography, and uniphasic enhanced distended bladder CT. Ninety-five percent of patients completed the 1-year follow-up. The rates of cancer detection at screening were 3.3% with limited screening plus CT vs. 2.3% with limited screening alone (not significant). At the 1-year follow-up, 4.5% of patients who had limited cancer screening plus CT vs. 3.2% of patients who had limited screening alone were diagnosed with cancer resulting in rates of missed diagnosis at screening of 1.18% vs. 0.93%, respectively (not significant). These results correspond to 5 of 19 cancers (26%) missed with the more extensive protocol and 4 of 14 cancers (29%) missed with limited screening alone. There were also no significant differences in the mean time to cancer diagnosis, recurrent VTEs, overall mortality (1.4% vs. 1.2%, p = 1), or cancer-related mortality (1.4% vs. 0.9%, p = 0.75).
The results of this randomized trial suggest that after the initial cancer screening following VTE, the rate of cancers detected during the first year is not affected by the use of extensive CT scanning of the abdomen and pelvis. The key issue is whether there is a slight increase in the rate of cancers detected at the time of initial work-up with extensive screening. While not statistically significant, the 1% higher detection rate at time of initial diagnosis is consistent with previous systematic reviews, randomized trials, and studies that have shown an approximate 1%-2% additional yield for cancer detection when other forms of extensive screening are used (J Thromb Haemost 2004 Jun;2(6):884, J Thromb Haemost 2011 Jan;9(1):79, Ann Intern Med 2008 Sep2;149(5):323, Cochrane Database Syst Rev 2015 Mar 6;3:CD010837). Significant limitations to the interpretation of these results are the low event rates and the wide confidence intervals, which could not rule out a small benefit associated with CT imaging with respect to cancer detection. Regardless, whether a small increase in cancer detection rates translates in to a meaningful clinical outcome remains unproven, and the use of multiphasic CT of the abdomen and pelvis carries a risk for causing a subsequent radiation-induced malignancy that has been estimated at 1 in 460-500 for a 40 year old patient (Arch Intern Med 2009 Dec 14;169(22):2078). For now, the use of limited screening, along with routine screening for cancer based on age- and sex-based recommendations for the general population seems to remain appropriate for most patients.
For more information, see the Venous thromboembolism - testing upon diagnosis topic in DynaMed Plus. DynaMed users click here.