Limited cancer screening seems appropriate after unprovoked PE

Cancers can cause pulmonary embolism, and an unprovoked PE may signal an undiscovered cancer lurking in the body. In older studies, as many as 1 in 10 patients with unprovoked PE were diagnosed with cancer within a year. (The current terminology is that a PE is provoked if associated with a known cancer or another provoking factor, but considered unprovoked if cancer has not yet been discovered.)

Whether (and to what lengths) doctors should go looking for cancer in patients with unprovoked pulmonary embolism is unknown. A new randomized trial in the New England Journal of Medicine provides the first satisfying answers to the question to date.

At multiple Canadian centers, Marc Carrier et al randomized 854 patients with a first unprovoked pulmonary embolism to either undergo limited cancer screening (breast exam or mammography, Pap smear, PSA or prostate exam, CBC, blood chemistry with liver function tests, chest X-ray) or to also undergo comprehensive CT of the abdomen and pelvis with multiple contrast/imaging protocols (attuned to the colon, stomach, liver, pancreas, and bladder).

(Colorectal cancer screening was not included as part of the limited-screening protocol, although many patients in both groups had had it. Nor was lung cancer screening with low-dose chest CT routinely done. Both lung and colorectal cancer screening, among other tests, are recommended by U.S. health authorities for patients when they accumulate sufficient risk factors.)

After a year, only 4% of patients with unprovoked pulmonary embolism had been diagnosed with cancer.

Both screening methods detected and missed about the same number of cancers. Limited screening missed 4 of 14 cancers or 29%, compared to more extensive screening which missed 5 of 19 cancers or 26%, a nonsignificant difference with p=1.0.

Overall mortality was similar (1.4% with limited screening vs 1.2% with additional CT imaging) as was cancer-related mortality (1.4% vs 0.9%) and recurrent DVT/PE (~3%).

Screening was not compared to no screening, so it's impossible to say whether screening itself was beneficial. It's likewise impossible to draw conclusions about the discovered cancers, their risk, any benefit of catching them early, or their causative contribution to the PEs: staging was not described, and significant numbers of lymphomas, prostate, and cervical cancers (which are often indolent) were among those discovered. (To varying degrees, all cancer screening produces lead-time bias and overdiagnosis, both of which tend to exaggerate the benefits of cancer screening).

About a quarter of cancers that develop in the year after an unprovoked pulmonary embolism, in other words, are probably undetectable despite the extent of the workup. Assuming all people with unprovoked PE get at least limited screening described here, the risk is probably about 1% for an occult cancer to become manifest despite an adequate search. It's unclear from this study what risk any of the cancers posed, or whether detecting some of them early was helpful.

Clinical Takeaway: Limited cancer screening seems appropriate after an unprovoked pulmonary embolism. Although screening for breast, cervical, and prostate cancer were included here, screening for lung and colorectal cancer for eligible patients according to age and published guidelines also seems advisable. Additional CT imaging of the abdomen and pelvis does not seem to detect more cancers or save lives.

Read more: Marc Carrier et al.  Screening for Occult Cancer in Unprovoked Venous Thromboembolism. N Engl J Med 2015; 373:697-704.