Many people in the lung cancer world consider the National Lung Screening Trial (NLST) that demonstrated a 20% improvement in survival from CT-screening higher risk people for lung cancer as a major advance in the field, befitting coverage in Dr. Weiss's summary of lung cancer highlights from 2011, but this hasn't yet led to wholesale adoption of the practice.  Why not? Some people say that it's just one study and that we need more evidence, but it's hard not to believe that implications for health care resource utilization (i.e., cost and practical implementation) aren't a big part of the challenge.  Annual CT scans will/would add a lot of cost when applied for the millions of people just in the US who would be appropriate candidates by the definition of the study (age 55-75 and with a 30 pack-year smoking history).  In addition, we know that CT screening detects a lot of nodules that require plenty of follow-up scans and cause significant anxiety but ultimately prove to not be cancer.  What if we could add another factor that could raise or lower our suspicion and potentially enable us to modify the frequency of scanning and/or our enthusiasm for escalating the workup?

I've covered one such approach in a prior post describing a test analyzing chemical compounds in the exhaled breath of patients, and this research is ongoing.  But another strategy is to analyze the blood of patients in search of preliminary evidence of cancer that might be detectable at the time of or even before the earliest imaging findings suggestive of cancer.  The EarlyCDT-Lung test is designed to do this by looking for immuno-biomarkers ("auto-antibodies" created by the body in reaction to a protein detected within), any of a panel of 6-7 that could signify early cancer if one or more of these is elevated.   

That's the idea at least, and there's some evidence that it can be helpful.  It's really a question of how helpful that I think is debatable.  The evidence from approximately 1000 patients who underwent the test, which is commercially available, largely corroborated what their earlier testing suggested: a positive test would be present in about 40% of patients who were found to have lung cancer (so of course this means it misses 60% of them), and your chance of actually having cancer by imaging if the test is positive is only about 7%.  In contrast, if you had a negative test, your chance of actually having cancer was less than 2%.  Not zero, but low.

In my mind, these results, while arguably better than not having any information to go with the imaging results, don't qualify as the answer to my prayers.  I don't know that I'm very assured knowing that if the test comes back positive, there's only a 7% chance of actually having cancer, or a negative result, knowing that 60% of actual lung cancers are in patients with a negative test and that the risk of actually having a cancer is still in the 2% range -- not way, way lower than the 7% in the positive group.  To me, this doesn't mean that I can be completely confident about the meaning of a positive result OR a negative result.

Further research with this approach is ongoing. In the US will be developed with approximately 1600 patients who will undergo CT screening and blood screening simultaneuously, to see if this significantly increases our ability to detect cancers more reliably while reducing unnecessary tests.  And the National Health Service (NHS) in Scotland is about to initiate a new trial of 10,000 high risk smokers (at least 20 pack-years, or an average of a pack per day for 20 years), half randomized to EarlyCDT-Lung testing and then screening low-dose chest CT if the test is positive, while those randomized to the no EarlyCDT-Lung testing arm will be followed with "standard of care" (presumably no CT screening).   Overall, I think the idea of refining the idea of who to screening and being able to modify our level of suspicion based on imaging is a very good goal, given that the NLST trial would have a lot of people getting a lot of scans, most of which won't show cancer even when they show nodules that will lead to a lot of further imaging and anxiety.  However, I'm concerned that if only 40% of people with an actual lung cancer are detected by the Early-CDT trial, we're going to defeat the purpose of screening by missing the other 60%.   

I'll be very interested in seeing how further research with these approaches actually improves our screening process and whether we can develop a screening process that spares a lot of people scans and biopsies while capturing the vast majority of the cancers.  I'm not sure that the EarlyCDT-Lung test in its present iteration has that potential, but I do think that it's good that the trials are being done that could lead to better ways to complement imaging with non-invasive tests.

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?