Study questions value of mammograms, breast-cancer screening

A new study questions the value of mammograms for breast-cancer screening. It concludes a woman is more likely to be diagnosed with a small tumor that is not destined to grow than she is to have a true problem spotted early.

The work could shift the balance of whether screening's harms outweigh its benefits.

Screening is only worthwhile if it finds cancers that would kill and if treating them early improves survival versus treating when or if they ever cause symptoms.

Treatment has improved so much over the years that detecting cancer early has become less important.

Mammograms catch some deadly cancers and save lives. But they also find many early cancers not destined to grow or spread and become a health threat. There is no good way to tell which ones will, so many women get treatments they don't really need. It's a twin problem: overdiagnosis and overtreatment.

Whether to have a mammogram "is a close call, a value judgment," said study leader Dr. H. Gilbert Welch of Dartmouth Medical School.

Welch long has argued mammograms are overrated, and the study parallels work he published from the same data sources four years ago. This time, the authors include Dr. Barnett Kramer, a National Cancer Institute screening expert, although the conclusions are not an official position of the agency. The study was published Wednesday by the New England Journal of Medicine.

Researchers used decades of federal surveys on mammography and cancer-registry statistics to track how many cancers were found when small -- under 2 centimeters, or about three-fourths of an inch -- versus large, when they are presumably more life-threatening.

They estimated death rates according to the size of tumors for two periods -- 1975 through 1979, before mammograms were widely used, and a more recent period, 2000 through 2002.

In the earlier period, one-third of cancers found were small. In the later period, two-thirds were small. But the change was mostly because screening led to so many more cancers being detected overall, and the vast majority of them were small -- 162 more cases per 100,000 women, versus only 30 more cases of large tumors.

Assuming the true number of cases of cancer in the population was stable, this implies 132 cases per 100,000 women were overdiagnosed.

Next, they estimated how much of the drop in deaths since mammography started was because of early detection versus better ways to treat the disease. They concluded at least two-thirds of the drop was due to better treatments -- a trend other studies also have found.

Size matters, but it's not the whole story and isn't a proven sign of how aggressive a tumor is biologically, said Dr. Kathryn Evers, director of mammography at Fox Chase Cancer Center in Philadelphia.

Tumors have to be found before they can be treated, and so far, mammography is the best way to find ones that can't be felt, she said.

The study applies only to screening mammograms, not diagnostic ones done when a problem is suspected, and only to women at average risk.