A CT scan of lungs.
Dr. William C. Black, professor of radiology at Dartmouth Medical School explains why early detection of lung cancer may not be beneficial:
Screening may be too early. That is, it a computerized tomography or CT scan might find cancers that might never cause a problem. Doctors call this "over-diagnosis."
CT screening detects lung cancers that are much smaller and more slowly growing than those that have been diagnosed in the past with chest X-rays or because the patient has symptoms. Some of these tumors would not cause any symptoms before the individual died from other causes, such as coronary artery disease or chronic obstructive lung disease, which are very common in smokers. For example, an untreated tumor three-eighths of an inch across which doubled in size every 500 days would not be expected to cause death for about 14 years.
Screening may be too late. The lung cancer may already have metastasized, or spread to other tissues. A tumor only 1 millimeter in diameter contains about 1 million cells. A 10-millimeter tumor contains about 1 billion cells.
The timing may be right, but the surgery might cause major complications or death. Any net benefit to those with lung cancer detected through screening —- about 1.5 percent in one recent study —- must be weighed against the harm from false positive tests in the vast majority of screened individuals who do not have lung cancer — the other 98.5 percent. In one recent national study, 12 percent of people who underwent screening had positive CT scans that led to negative lung biopsies —- showing no cancer.
Survival statistics can be misleading in cancer screening. They can contain what experts call statistical biases, such as:
1. Lead time bias. Cancers may be found earlier, but the patient doesn't live longer than he/she would have without the screening diagnosis.
2. Length bias. Screening tends to detect slowly growing tumors. Rapidly progressing tumors tend to show themselves by causing symptoms in between screening rounds or after the patient has stopped getting screening CT scans.
3. Overdiagnosis bias. Screening detects tumors that would not have become clinically significant before the patient dies of other causes.
4. Survival statistics may only pertain to the small minority of screened individuals who are diagnosed with the cancer through screening, not those diagnosed outside of screening programs.
What clinicians should tell their patients in light of the existing controversy on CT screening for lung cancer: If they currently smoke, stop smoking. This would be much more effective than CT screening could ever be at prolonging life expectancy and quality of life.
If they are former smokers, focus on reducing risk from coronary artery disease. Most former smokers are more likely to die from heart disease than from lung cancer.
In either case, if they are still interested in CT screening, they should get information on what is currently known. They can estimate their own risk of developing lung cancer using a risk calculator developed by Dr. Peter Bach of Memorial Sloan-Kettering Cancer Center. They can get information on the potential benefits and harms of CT screening from the National Cancer Institute. If they want to proceed with screening, they should go to a medical center that has experience with screening and is willing to provide counseling.
The National Lung Screening Trial is designed to resolve some of the questions about the usefulness of CT screening. It's a federally funded study of CT screening versus chest X-ray to detect lung cancer. It involves 53,000 volunteers. Results are expected in early 2010. — Richard Knox
Ninety million Americans are smokers or former smokers, and they're all at risk of lung cancer, the deadliest cancer of all.
For years, many have hoped that computerized X-ray machines called CT scanners could spot lung cancers earlier when they're more curable. Last fall, one large study concluded that CT screening for lung cancer saves lives.
Today, another study casts doubt on that claim.
Despite the growing debate, many health-care centers are offering CT screening for lung cancer.
Influenced by last fall's study, which appeared in the New England Journal of Medicine, Stamford Hospital in Connecticut recently decided to offer the service, even though patients have to pay $350 out of their own pockets for it. Most insurers don't cover CT screening for lung cancer unless it results in a positive diagnosis.
"For people at high risk who have smoked a number of packs of cigarettes a day for years, if we catch cancers early and treat them, then we'd be doing a huge service to our patient population," says Liz Manfredo of Stamford Hospital.
Dr. Peter Bach of Memorial Sloan-Kettering Cancer Center in New York is a skeptic. He led the study that appears in this week's Journal of the American Medical Association.
"We have done a terrific job in public health of telling people, 'Catch it early, it will save your life,'" Bach says. "And we did that to promote proven screening tests like mammography and the Pap smear."
But when it comes to lung cancer, Bach says, unfortunately, it doesn't work out that way.
Bach and his colleagues looked at the experience of three lung-cancer screening programs — two in this country, one in Italy. Out of 3,246 current or former smokers, lung CT scans found cancer in 144 people. That's three times the number expected from such a group.
That should be a good thing. It's what screening is supposed to do. But Bach says it didn't work out as intended.
"We saw 10 times as many surgeries being done because of the CT scanning," he reports. "And no one's life was saved. Not one case of advanced cancer was prevented."
There were 39 deaths from lung cancer among the three groups — the same number that Bach calculates would occur among a similar population that never got CT screening. Bach says it was a big disappointment.
"We all hoped that CT would be sensitive enough to pick up lung cancers before they spread, before they caused illness, before they killed people," Bach says. "But it looks like those cancers we don't seem to be able to catch with CT very often."
Unfortunately, the lung cancers destined to be deadly tend to arise very quickly – often in between scans.
On the other hand, many doctors are coming to understand that many people — smokers and non-smokers — harbor small lung tumors that look just like cancer on CT scans, and even under the microscope after they're removed. But they don't behave the way you'd expect lung cancer to behave.
Dr. William Black, professor of radiology at Dartmouth Medical School, says this is because many lung tumors are "indolent" – they grow slowly.
"We do know from autopsy studies that we can find a lot of small cancers, and many of these small cancers do not appear to have been related to the patient's death," Black says. "Now, if you were to find such a small cancer with CT screening and treat it, you might falsely assume that you've prevented a death — when, in fact, that death never would have occurred."
Perhaps the patient would have died from something else, such as heart disease. Or perhaps the lung tumor would never have progressed to the life-threatening, metastatic phase during the patient's lifetime.
The new JAMA study is under attack from those who think CT lung screening has already been proven to work. Dr. Claudia Henschke of Weill-Cornell Medical College is foremost among them.
Henschke criticizes the new study's method, especially its reliance on "expected" lung-cancer deaths from populations that are supposedly comparable to the screening populations. Henschke stands by her study from last fall, claiming that CT screening greatly lengthens lung cancer survival.
"There are more and more people that are convinced that screening works," Henschke says. "And certainly there are more people being screened, I'm sure, in the country."
But many in the field say we won't really know whether lung cancer screening works until results of a $220 million federal study are in. That won't be until early 2010.