CT scans may break the impasse over lung cancer screening, although many questions remain.
Screening people for lung cancer ought to work. Lung cancer is common, relative to other cancers, so you're not looking for a needle in a haystack. If it's caught at a very early stage, it often can be treated effectively with surgery and even cured. Current and former smokers are easy to identify, so focusing on a high-risk group is pretty straightforward. And the technological means to spot suspicious growths and lesions in the lungs has existed for decades.
The chest x-ray came first. As early as the 1950s, researchers were investigating whether chest x-rays and inspection of sputum, the mucus produced by a hard cough, would find early lung cancer. Using CT scans for lung cancer screening became feasible in the mid-1980s, when the scans got fast enough to create clear images during the few seconds most people can hold their breath.
But lung cancer screening has been at an impasse for almost as long as it has been a possibility. Screening tests are supposed to find diseases before they cause symptoms, but that is just the means to an end. The goal is for fewer people to die from the disease. Several large, well-designed trials of chest x-rays have found that while they do a good job of identifying early lung cancers, that hasn't translated into fewer deaths.
And the jury has been out on CT scan screening, partly because of the size and design of studies that have been done so far. In the absence of a verdict, there's been debate. Skeptics have argued that while CT scans would certainly detect lung cancer at an earlier stage than chest x-rays (and no screening at all), it would still be too late for treatment to be very effective. Supporters said, sure, some incurable lung cancers will be found, but so will many curable ones, and that large, well-designed studies will show that when people are screened with CT scans, fewer of them will die from lung cancer.
An important study
One such large, well-designed study has now been done. In November 2010, the National Cancer Institute (NCI) announced preliminary results from the largest randomized trial of lung cancer screening ever conducted. The study, called the National Lung Screening Trial (NLST), included over 53,000 current and former heavy smokers, ages 55 to 74, who had no signs or symptoms of lung cancer when they agreed to be in the study. A heavy smoker was defined as someone with a smoking history of at least 30 "pack years." Thirty pack years means someone has smoked, on average, a pack (20 cigarettes) a day for 30 years or the equivalent (for example, two packs a day for 15 years is also 30 pack years).
Half of the study volunteers had their lungs screened with a CT scan three times: once at enrollment, a year later, and two years later. The others were screened with chest x-rays at the same intervals. The tests were conducted at over 30 sites across the country.
More than twice as many lung cancers (649 vs. 279) were discovered in the people screened with CT scans as with chest x-rays, which isn't at all surprising. Results from previous studies suggest such a ratio, and CT scans give a much more detailed and complete picture of the lungs than chest x-rays, making it easier to spot very small cancers that don't show up on a chest x-ray.
But early detection wasn't the point of this trial. It was set up to find out whether the screening method made a difference in the number of deaths from lung cancer. So researchers followed the two groups for several years, keeping track of who died — from lung cancer and from other causes. The trial was designed to detect a 20% difference between the two groups with 90% certainty, and in October 2010, that threshold was reached, so the outside board monitoring the trial voted unanimously to stop the study early. (Dr. Robert Mayer, a member of the Health Letter's editorial board, was a member of the monitoring board.)
During the follow-up period of about five years, 354 of the people in the study screened with CT scans (roughly 1.4%) died of lung cancer. Over the same period, 442 of those screened with chest x-rays (roughly 1.7%) died of the disease. There were also 7% fewer deaths from all causes in the CT scan group compared with the chest x-ray group.
The sheer size of the NLST makes these findings important. In randomized trials, large numbers put results on a firmer statistical footing. This was also the first large study of CT screening for lung cancer to include a control group, which is the best way of making reliable calculations of risk reduction relative to other screening tests.
The fact that the NLST had lung cancer deaths as its main outcome is also crucial. Cancer screening trials that use other outcomes, such as how long people survive after a diagnosis, can be misleading. Showing improvements in survival after diagnosis, without showing whether the number of deaths has been affected, may mean a screening test is just moving up the time at which a diagnosis is made, instead of leading to early treatment that truly lengthens life.
Time to get screened?
So, should heavy smokers, present or past, get a CT scan of their lungs? It's too soon to make a recommendation. The full results of the trial haven't even been reported yet. The NCI has said publication in a peer-reviewed journal is scheduled for some time in 2011. The full results should have breakdowns by gender, race, smoking status (current vs. former), and other details that may be revealing about who will benefit most from screening with CT scans.
But it will take months, if not longer, for groups like the American Cancer Society and the U.S. Preventive Services Task Force to mull over the results and issue guidelines. Medicare and other insurers may wait to make coverage decisions until those groups weigh in.
Many difficult issues need to be resolved. Do only heavy smokers get screened? How about people with shorter smoking histories of, say, 15 or 20 pack years? Or people with other risk factors, such as exposure to asbestos? Screening in the NLST was limited to three annual scans, but should people be screened later on even if the first screens didn't find anything? The NLST radiologists were highly trained. Will the CT scans be interpreted correctly outside of a carefully done, government-funded study?
If screening with CT scans does become the norm, people will have to be warned about the very real possibility of false positives: a scan that finds an abnormality — and there are lots of them in the lungs, especially in smokers — that turns out not to be cancer. Almost one in every four people who were screened with CT scans in the NLST had a false positive, compared with about one in every 14 in the x-ray group. A positive finding may just lead to another CT scan, but the ultimate determination of whether cancer is present may involve a biopsy.
Another concern is added radiation exposure. According to the NCI, the dose from a screening CT scan of the lungs is 20% of the dose from a diagnostic scan (1.4 millisieverts vs. 7) and less than the dose from a two-view mammogram (2.5 to 3.5 millisieverts). But if suspicious findings lead to follow-up scans, the amount of radiation exposure could start to add up.
Cancer screening tests certainly have a lot of appeal. If a cancer is found, we're relieved it was caught early. If the test comes back negative, we're happy to be in the clear. Even if the test turns out to be falsely positive, producing unnecessary worry, more radiation, and even invasive procedures like biopsies, we may still consider that a price worth paying.
The psychological comforts should not blind us to some of the problems with cancer screening. Results from the NLST suggest that about 300 heavy smokers will need to be screened with CT scans, at a cost of perhaps several hundred dollars each, to prevent just one death from lung cancer over a five-year period. That ratio actually compares pretty well to screening mammography, especially when used for detecting breast cancer in younger women. But it's still a reminder that screening is inherently an effort that affects many to benefit a few. (Of course, we don't know who those few are until the larger group is screened.) From a public health perspective, it may be that the dollars spent on CT scan screening of heavy smokers could be better spent on helping people to quit — or preventing them from smoking in the first place.
Overdiagnosis is possibly another problem. Screening tests (PSA screening for prostate cancer is a prime example) can lead to detection of slow-growing cancers that would never have caused illness or might even have regressed. Furthermore, competing illnesses may dwarf any effects that an indolent cancer is having on someone.
Lung cancers don't seem to linger in an indolent state, so screening with CT scans isn't expected to cause a big overdiagnosis problem, but it isn't out of the question, either. Longer-term follow-up of the people in the NLST may shed some light on the issue.