By Sue Rochman
How do we know if screening tests save lives?
By Sue Rochman
The difficulty of ascertaining a test’s true benefits is arguably most evident in the controversy that continues over mammography screening. Eight randomized mammography trials were initiated in the 1960s and 1970s. If mammography had a huge benefit, these studies would have found it. Instead, we have disparate, conflicting findings that have left researchers debating the overall benefits of mammography for years.
Statistics indicate that mammography is doing what it was designed to do: find small tumors. But some of these tumors are so slow-growing they would have never gone on to cause harm—and we don’t know which ones. New tumor tests are being developed to help oncologists assess which women are in need of the most aggressive treatment, but the general consensus in the breast cancer field is that many women receive chemotherapy who probably don’t need it. Some might say it’s the price we need to pay for reducing breast cancer deaths. And that might be true if mammography had a dramatic impact, but that’s not the case: Mammography, which remains the best screening test for breast cancer, is only about one-third as effective in terms of reducing death rates as the Pap smear (though the total number of deaths from breast cancer is much higher than from cervical cancer).
There are other concerns, too. Mammography has led to a dramatic increase in diagnoses of ductal carcinoma in situ (DCIS), a precancer that can go on to become cancer, but doesn’t always. With widespread mammography screening, “DCIS went from a rare disease to a common one,” says Kramer. And because we don’t yet know how to assess which cases of DCIS are truly dangerous and which aren’t, there’s no watchful waiting. It’s always treated—with surgery, radiation and hormone therapy.
Right now, says Brawley, “few disagree that for women over 50, a well-done mammography lowers the death rate by 20 to 35 percent.” But not all women, he says, are getting the high-quality mammo-grams they need. For women between 40 and 50, assessing the benefit of mammography has been more difficult. That’s because most women in their 40s still have dense, premenopausal breast tissue, which appears white on mammography, as do breast cancers. And that can make looking for a cancer on an X-ray like looking for a white marble in a pile of fresh snow. The end result: not just a higher rate of missed diagnoses than in older age groups, but a large number of false positive results, too.
What’s bad about that? Isn’t it better to be safe than sorry? Maybe. But researchers have to consider the psychological impact of a woman thinking she may have cancer as well as the physical impact of the medical procedures that follow a positive result. “An interesting thing happens,” says Buys. “We’ll find an abnormality on a mammogram, biopsy it and find it benign. The woman will say, ‘I’m so grateful, thank you for saving my life.’ But we didn’t save your life. You had an unnecessary procedure that led to a diagnosis of something we didn’t need to know about.”
Lung Cancer: The Next Frontier
There is no doubt that the misplaced hope that surrounded chest X-ray screening in the 1950s and 1960s is at the back of many people’s minds when they think about CT screening for lung cancer. “It was a hard lesson to learn,” says Kramer, “and it shows that preliminary evidence from uncontrolled study designs can give you qualitatively incorrect answers.”
But there is something else at play. Many people in the lung cancer community believe lung cancer is being held to a different standard. “There are plenty of data that show that PSA screening” has resulted in overdiagnosis and “isn’t saving lives,” says Tracey Weigel, a lung cancer surgeon at the University of Wisconsin–Madison, who is studying CT screening. “But doctors haven’t stopped doing them, and almost every man gets them.”
“And there are tons of data that show there is no benefit for mammography screening for women under 50,” she adds, but no one would stop a woman from having it. When it comes to lung cancer, the perception is that “smokers give it to themselves,” she says, and “there is this nihilistic attitude that lung cancer is not curable, so why screen for it?”
Kramer disagrees. The problem, he says, isn’t the type of cancer but the type of research. A person who chooses to have a CT scan based on what we know right now is making “a rational choice” he says. But that doesn’t mean “it’s not crucial to know the right answer,” he continues. “The randomized studies we are doing are going to save lives or money or both. And I reserve judgment until those trial results are in.”
But will the two large controlled studies that are now under way provide the answers we need? Or will this be another mammography quagmire? The PLCO, which was started 13 years ago, is evaluating chest X-ray, and few people, if any, really expect the data to reveal anything not already known. More on target is the NCI-sponsored National Lung Screening Trial, which is comparing CT with chest X-ray in 50,000 current or former smokers. (Kramer is an investigator on this trial.)
Henschke says she remains surprised by the amount of opposition and resistance to her findings. “We had hoped that … the sheer logic that if you find many stage I tumors, you are saving lives, would lead [to CTs being] made more available to more people sooner.”
And although that logic is spurious to her critics, Henschke has begun to see her hope become a reality. Spurred by Henschke’s work, in November 2005, the University of Nebraska Medical Center initiated the first ever free statewide CT screening program for all residents who are at high risk for lung cancer due to smoking.
In the years to come, what will happen when researchers compare mortality rates from lung cancer in Nebraska with, say, its neighbor Kansas? Will we see the Canadian neuroblastoma statistics all over again? Or will we find that people in Nebraska have not merely better survival rates but the lower death rates that really matter? Depending on your perspective, the data in support of Nebraska’s decision to implement widespread CT screening may, or may not, be evident. But one thing is crystal clear: The test has begun.