Should We Screen All Smokers for Lung Cancer?

November 15, 2006


Commentary by Deborah Shapiro, M.D.  Attending Physician, Dept. Veterans Affairs

The International Early Lung Cancer Action Program Investigators. Survival of patients with stage I lung cancer detected on CT screening. N Engl J Med 2006;355:1763-1771

Over the last few weeks there has been major press in the lay media about this lead article in the NEJM from 10/26/06.  This study was designed to diagnose lung cancer at an earlier stage and therefore to decrease mortality rates. The authors had two study groups, one in New York and one in Japan comprising almost 60,000 people who were considered high risk for lung cancer. They underwent annual chest CT. Of these people, 484 cancers were diagnosed and of those 412 had stage I. The authors go on to report that screening high risk patients could prevent 80% of the deaths from lung cancer.

There are several issues with this paper. As with many screening articles there are issues of lead time bias and overdiagnosis bias.  (For a review of all the different screening biases click here).  There is no discussion of complication rates among people who were not diagnosed with cancer. There are over 5000 people in this study who had suspicious nodules. We can only assume that all of the people had some sort of invasive diagnostic procedure to determine that they did not have cancer. It’s difficult to understand why this data was not included and makes one suspicious that this was intentionally left out.

The only other screening program that we can liken this to is breast cancer. The majority of women who go for breast biopsies do not have cancer. But, a fine needle aspiration of breast tissue does not carry the morbidity and mortality associated with any of the procedures required to diagnose a lung nodule. I believe the authors have not dealt with the important issue of “number needed to harm” that can come from a screening catscan of the lungs.  Add to that the psychological morbidity of knowing you have a nodule that may be nothing or may turn out to be a cancer and you start to realize that screening is not always a benign procedure.  The definitive answer to this will more likely come from the randomized controlled trials-PLCO trial and the NLST trial both underway with preliminary results due soon.