November is Lung Cancer Awareness month. In Atlanta this week you may have heard a dulcet-voiced doctor on the radio announcing that "anybody can develop lung cancer," and the alarming factoid that women who have never smoked are the fastest-growing segment of people getting the diagnosis. In California, a press release announced that anyone who has smoked for more than 10 years is a "patient at risk" for lung cancer, and another advised that "prior smokers over 50 years of age"would benefit from screening.
But these are not public-service announcements intended to raise awareness of the No. 1 cancer-killer. They are advertisements promoting computed tomography lung cancer screening from St. Joseph's Hospital in Atlanta, Wilshire Radiology Associates in Beverly Hills, Calif., and the Bonnie Addario Lung Cancer Foundation (which sponsors a CT screening program at Sequoia Hospital in Redwood City, Calif.). There are more examples, too—coming from hospital Web sites and physicians' groups. And they are all dangerously misleading.
On Nov. 4, the National Cancer Institute held a press conference announcing that a large randomized trial of CT screening had shown an astounding 20 percent reduction in the risk of death from lung cancer among participants. This is a huge step forward in lung-cancer prevention, no doubt. But the Institute said little about who should be screened, or the risks that are involved.
They should have. Cancer screening is fundamentally inefficient: Hundreds, and sometimes thousands, of people must be screened to help just one or two. Each person who undergoes the test may suffer consequences from it, even though most will never get any counterbalancing benefit. This is why the recent study—called the National Lung Screening Trial—focused on a narrow, "high-risk" subgroup of the adult population who ranged from 55 to 74 years old and had smoked at least a pack a day for 30 years. If they had quit, it was within the past 15 years.
These are strict criteria, and they ensured that the patients had a meaningful chance of developing lung cancer during the course of the decadelong study—and thus an opportunity to benefit from being screened. But people who are younger or have smoked less than the test subjects are at much lower risk for the disease (although no one is entirely safe from it).
If you have a lower risk of lung cancer, there's less of a chance that screening will help you, as you can't prevent something that wasn't going to happen. But that doesn't mean there's less of a chance you'll be harmed by the procedure. Taking a CT scan of the chest can uncover something that looks abnormal but ends up being nothing. Along the way there are more scans, biopsies and, sometimes, unnecessary surgeries. In the NCI study one in four people had these false positives. A prior study from the University of Pittsburgh pegged the rate at around two out of five, and in that study one in 100 subjects had parts of their lung removed for no good reason.
CT screening can also uncover small lung cancers that you'd be better off not knowing about, because they would be unlikely to progress and make you sick. This problem of overdiagnosis is familiar in prostate cancer—many cancers found by the PSA test are not dangerous. A study from 2007 showed that CT screening may uncover one overdiagnosed cancer for each real cancer it turns up. The NCI has not yet released the equivalent numbers from its new trial, but the data we do have suggest that the ratio is about the same—about one in 70 patients is told he has lung cancer when the condition might end up being harmless.
Leaving aside false positives and overdiagnoses, screening is not a panacea. Not even close. The NCI study showed that regular scans prevented one in five lung cancer deaths, which means that four out of five sneaked through. It's amazing to save even one in five people, but it also means that the number who benefit from screening is a lot smaller than the number who test positive. All told, a death from lung cancer was prevented in one out of every 300 people in the study.
The potential harms listed here are more than offset by this one important benefit. But that calculation changes if we stop focusing on those patients at very high risk of disease. For everyone else, the risks may outweigh the gains, and the overall cost in dollars would be prohibitive.
Insurance doesn't cover the test, and the NCI says that a scan costs about $300. But when I called about 50 places around the country, I heard numbers more like $1,800, with Sequoia Hospital in Redwood City topping the charts at $4,000. (Update, Nov. 29, 2010: The Bonnie Addario Foundation reports that a research study on lung cancer screening at Sequoia Hospital charges high-risk patients $399 out of pocket for the procedure.) To put this in perspective , there are about 7.5 million people in the United States at high risk of lung cancer like those the NCI studied. The price tag for screening all of them once would be at least $2 billion, and more like $13 billion if the prices I found are more accurate.
But the new round of CT scan advertisements released following the Nov. 4 press conference target a much larger population for screening. The Bonnie Addario Foundation's criteria suggest that some 46 million people should be screened at a cost of between $14 billion and $84 billion (about the budget of the U.S. Department of Education). Beverly Hills Radiologists? It says 77 million people and $23 billion to $138 billion (similar to the numbers for the Department of Labor). And if you go with St. Joseph's in Atlanta, which encourages screening for nonsmoking women and all men with any smoking exposure whatsoever, that's 166 million people—about half the U.S. population, at a cost of around $50 billion to $300 billion (half the Medicare program).
All of this would be for the first scan. Then there are the follow-up tests associated with false positives. If we stick to screening the high-risk people it will probably be worth it. Researchers at the Massachusetts General Hospital in Boston reported that it would cost around $150,000 for each quality-adjusted life year saved if we focused on screening those particular people. That's much more costly than, say, mammography, but it's in the neighborhood of many other cancer treatments. Still, these researchers noted that more could be gained at lower cost by getting people to stop smoking. Those who quit their habits reduce their risk of getting lung cancer by about 50 percent, which is more than double the benefit of being screened, without any of the radiation, false positives, overdiagnoses, and unnecessary surgeries.
Some day CT screening will save lives—hopefully a lot of them. It will harm some people, too. We can stay ahead in this tradeoff if we are circumspect about whom we screen, and if we don't believe every radio ad we hear.