Could we improve the accuracy of CT screening by adding a bl

[Guest LCSC Info]

Could we improve the accuracy of CT screening by adding a blood test to detect lung cancer?

March 30th, 2012 - by Dr. Jack West

http://blog.lungevity.org/2012/03/30/co ... ng-cancer/

Despite the fact that last year’s reporting of the National Lung Screening Trial (NLST) revealed a significant, 20% improvement in overall survival by doing annual CT screening for lung cancer in smokers from 55-75, this approach hasn’t yet emerged as a standard of care. I think there are many reasons for this, but a key element is that a vast number of scans would need to be done, at a very huge expense, to run such a comprehensive program. Even beyond the cost, a central challenge of CT screening is that it detects many nodules in lots of people that are actually not cancer, yet detecting one or more nodules typically leads a patient to experience a lot of anxiety, many more scans, and potentially an invasive procedure that can even include lung surgery to rule out cancer. In this setting, it’s easy to think that if we could refine our “hit rate” for identifying which patients are at greater or lesser risk for cancer, we could potentially use this new test to identify which patients should be CT screened at all, or if used with screening, which patients we’re more or less concerned about cancer in.

A few years ago, I wrote a post on the GRACE website that reviewed research with a test looking at the chemicals in the exhaled breath of patients in hopes of identifying a pattern that signals cancer and that could screen people easily and non-invasively; this research is ongoing. Another approach to assessing risk might be look for patterns in the bloodstream that signal higher or lower risk for lung cancer. Following this strategy is the EarlyCDT-Lung test, which screens a patient’s blood for specific auto-antibodies, a biomarker by product of the immune system, created by the body in reaction to a protein detected within. The Early CDT-Lung test looks for the presence of any of 6-7 that can potentially be seen in the setting of an immune response to a cancer.

There has been some research with this test, and I would say that the utility of this test and its promise are really in the eye of the beholder. For example, a review of the results in just over a thousand patients who had the commercially available test essentially confirmed the company’s early findings. Specifically, a positive test is seen in about 40% of the people actually found to have evidence of lung cancer, meaning that the Early CDT test misses 60% of cancers. At the same time, most of the tests that are positive don’t actually prove to have evidence of lung cancer: this means that your probability of actually having lung cancer on further workup in the face of a positive Early CDT test is only around 7%. People with a negative test were certainly likely to not have cancer, but the probability isn’t zero, but rather just under 2%.

These findings are arguably better than not having any further information, but my personal view is that they fall far short of solving the key issues. If the test is positive, it doesn’t mean that you have cancer, or even that cancer is likely — the probability is just 7%. And if the test is negative, it doesn’t mean you don’t have cancer. The risk is very low at 2%, but that’s not worlds apart from 7%. Overall, I can’t be especially clear about the meaning of a positive OR a negative result.

Importantly, there is further research being done to assess the potential utility of the test. A study is being developed to enroll about 1600 patients to receive simultaneous CT and blood screening, with a goal of learning whether the combination of screening approaches increases our ability to detect cancers more reliably while reducing unnecessary tests. In addition, Scotland’s National Health Service (NHS) is initiating a new study of 10,000 high risk smokers (at least 20 pack-years, or an average of a pack per day for 20 years), with half of the enrolled patients being randomized to undergo EarlyCDT-Lung testing followed by a low-dose chest CT only if the test is positive, those who test negative by Early CDT-Lung to not get CT screening; the other half of the enrolled patients will pursue “standard of care” management that will presumably not include CT screening in most cases.

In my estimation, the idea of limiting the proportion of patients being scanned is an understandable goal, since the NLST trial would lead to a lot of scans in a lot of patients to find a relatively small numbers of cancers, while many people will otherwise have scans that show nodules that don’t lead to a cancer diagnosis but do lead to a lot of additional testing and patient anxiety. But knowing that only 40% of people with an actual lung cancer test positive on the Early-CDT trial, I’m concerned that this approach being tested by NHS-Scotland will knowingly miss 60% of the lung cancers.

We’ll see how the research develops. I’m wary that the EarlyCDT-Lung test may not be able to deliver on the promise of the potential benefit, but I applaud the efforts to study the utility of the test thoroughly. I hope this is a springboard for further efforts to refine our ability to screen for lung cancer, and potentially other cancers, with a blood test.

So what do you think? Would you be reassured not to do CT screening based on a negative blood test, knowing that the risk is in the range of 2%, compared with 7% probability if your test returns as positive?