THE premise is simple enough. Catch a deadly cancer early, before it has a chance to spread, and you can save lives. That鈥檚 why pap smears have been so successful, and why people submit to uncomfortable colonoscopies.
It鈥檚 also why mammography, which uses X-rays to detect lumps in women鈥檚 breasts long before they can be felt by hand, reduces a woman鈥檚 chance of dying from breast cancer. Or does it? After decades of widespread screening worldwide, that safe-looking assumption has been called into question. Everyone seems to agree that mammography can spot lumps earlier. But does this actually save lives?
Experts are split and the public is confused as yet again the statistical tools of science seem infuriatingly incapable of giving a clear-cut answer. Mammography is in danger of being assigned to 鈥渢he red wine department鈥�: may be good, may be bad, so let鈥檚 propose a toast and ignore the experts until they make up their minds.
Advertisement
It wasn鈥檛 always that way. Until about a year ago, there was a consensus that mammography did save lives, especially for women over 50. But in October 2001, two Danish scientists re-examined seven key epidemiological studies on mammography, and concluded that almost all were too flawed to offer clear evidence of any benefits. Worse, they said, because of mammography鈥檚 downsides, such as a high rate of false positives leading to unnecessary biopsies, it might even be harmful (快猫短视频, 22 June, p 34).
Not so, replied a second team. Its analysis of the same studies suggested it was the Danish report that was flawed: mammography did spare lives. That view was echoed by two further analyses of Swedish clinical trials.
Then, last month, more confusion. A report in The Annals of Internal Medicine of a Canadian trial of more than 50,000 women contended that for women in their 40s mammography was no better at saving lives than breast examination by a doctor. Yet in the same journal, another report concluded that mammography did reduce mortality rates in women aged 40 to 74. This came from the US Preventive Services Task Force and looked at eight international trials, including those analysed by the Danish team.
In spite of the contradictory studies, medical societies and health agencies continue to tell the public that the statistical aggregate of the evidence still points to a mortality benefit. Technically speaking, that may be true. Unfortunately, relying on this kind of narrow scientific reasoning has hardly cleared the air. It is forcing people to base personal decisions on concepts they don鈥檛 understand 鈥� statistical significance, odds ratios, lifetime risks. And it focuses the debate too heavily on science when there are other, less obvious reasons for continuing with mammography.
The first thing the public needs to know is that no scientific study is perfect, so it鈥檚 no surprise that mammography studies aren鈥檛. For example, the Swedish trials, which concluded that mammography works, compared mortality rates among women before and after screening had been introduced in different regions. Critics have said this is not ideal because the women did not come from identical populations, chosen at random. Other studies relied on mammography data collected by dozens of doctors 鈥� with varied skills. Can anything be concluded from them?
Those are valid concerns, yet they could be levelled just as easily at studies involving other diseases and interventions. So how come it鈥檚 so hard to reach a consensus for mammography? The answer lies in the magnitude of the mortality benefit. When all major trials are pooled, screening seems to reduce the risk of dying from breast cancer by about 20 per cent. This might seem a lot but is, by epidemiological standards, surprisingly modest. According to the US Preventive Services Task Force, it means that over 14 years, among 1224 women screened, one extra death can be prevented by mammography. That benefit, admits the task force, is 鈥渟mall enough that biases in the trials could erase it or create it鈥�.
If the 20 per cent figure is accurate, it would mean that detecting lumps earlier is not necessarily the key to a woman鈥檚 survival. Mammography could be spotting many tumours that would have remained localised anyway, and it may not be the right way to identify aggressive cases.
Then there is the problem that the 鈥渓atest鈥� epidemiological studies are based on data collected years (sometimes decades) ago. X-ray techniques have improved since then, as have breast cancer treatments. In fact, contrary to what most women fear, at least half of those diagnosed with invasive breast cancer will survive, regardless of screening.
The hard truth is that not one of the trials making headlines today can gauge the true benefits 鈥� or lack thereof 鈥� of mammography for women right now. Given a modest benefit based on outdated technologies, and the lack of knowledge of how recent developments might boost this benefit, authorities are erring on the side of caution and endorsing the technique for now. If only they would say as much in their official statements.