NCI Reaffirms Guidelines for Mammography Screening

BETHESDA, Maryland—The scientific conflict about whether mammography saves lives has been reignited by the decision by an independent advisory committee to rescind its support for such screening and to emphasize the uncertainty of the evidence supporting it.

The decision once again puts PDQ (Physician Data Query) at odds with the recommendations of the National Cancer Institute. The NCI recently reaffirmed its recommendations.

PDQ is a database for health care providers and the public that provides evidence-based assessments of cancer treatments, screening techniques, and prevention approaches.

At a January 23rd meeting, the PDQ screening and prevention editorial board concluded that the evidence for mammography screening in women remains uncertain and does not warrant definite conclusions about its use. The group decided that changes needed to be made in PDQ’s website discussion of mammography, but did not formulate a new wording at its meeting.

Although the use of mammography to screen women between ages 40 and 49 has been most controversial, the board decided that some uncertainty exists for women of all ages.

The PDQ board acted in large part on the basis of two articles published in The Lancet by Ole Olsen, MSc, and Peter C. Gøtzsche, MD, of the Nordic Cochrane Center, Copenhagen, Denmark.

In January 2000, the two researchers reported their assessment of results from randomized trials that have been used to support screening mammography and concluded that the analysis failed to show that such screening actually saved lives. They also concluded that screening led to more aggressive treatment that increased the number of mastectomies by about 20% and the number of tumorectomies by about 30%. In their latest article, published in the October 20, 2001, issue of The Lancet, they reported that their new analysis of seven screening trials confirmed the 2000 report that "that there is no reliable evidence that [mammography] screening for breast cancer reduces mortality."

NCI Responds

PDQ is funded by the NCI, which hosts its website, but it is independent of the Institute. A series of editorial boards prepares the statements that appear on the P.D.Q. "Their decisions are not binding on NCI and they are not recommendations of the NCI," said Institute spokeswoman Caroline McNeil. "Their decisions do not have to be vetted by NCI."

NCI has recommended since 1997 that, starting at age 40, women should get a screening mammogram every 1 or 2 years. In a news release on January 31, the Institute announced that it would not change its current recommendations.

The NCI stated that it is "determined to continue to address the uncertainties in the analysis of complex information surrounding screening mammograms." One of its highest priorities, the statement said, "will be the monitoring and evaluation of new data on mammography from numerous studies now underway that will supplement the data from clinical trials."

Overall Mortality

In their Lancet article in October, Mr. Olsen and Dr. Gøtzsche ranked the seven screening trials as high quality (none), medium quality (Malmö and Canada), poor quality (Two-County, Stockholm, and Göteborg), and flawed (New York and Edinburgh). "We found that the results confirmed and strengthened our original conclusion," the two said.

The seven screening trials used breast cancer mortality as their major outcome measure. "This choice seems rational, since larger trials would be needed to show an effect on overall mortality," the two researchers wrote. "However, we showed that the assumption that a demonstrated effect on breast cancer mortality can be translated into a reduction in overall mortality rests on suppositions that are not correct. The only reliable mortality estimates are therefore those for overall mortality."

As a result they concluded, "Any hope or claim that screening mammography with more modern technologies than applied in these trials will reduce mortality without causing too much harm will have to be tested in large, well-conducted randomized trials with all-cause mortality as the primary outcome."

Researchers from Weill Medical College of Cornell University and McGill University, Montreal, undertook a review of fatalities from the Malmö, Sweden, study (Br Med J 297:943-948, 1988), which the Cochrane reviewers had found to be of acceptable quality.

They did not review the Canadian study, also considered valid by the Cochrane researchers, because screening in the Canadian study continued for only 3 to 4 years after study entry, compared with 10 to 11 years for the Malmö study. Furthermore, follow-up in the Canadian study "stopped at the point at which follow-up in the Malmö study started to show fewer breast cancer deaths among those screened," said Claudia Henschke, MD, of Weill-Cornell, and her colleagues in the February 2 issue of The Lancet (359:404-406, 2002).

The Cochrane investigators, ignoring these differences in screening duration and follow-up, pooled the overall results from the Malmö and Canadian studies to reach their conclusions.

In finding no benefit for screening in the Malmö study, the Cochrane researchers looked at overall deaths from breast cancer at 10 years (63 in the screening group vs 66 in the control group), without looking at when these deaths occurred.

The Cornell researchers, however, argued that the benefit of screening-associated early intervention for breast cancer in reducing fatalities becomes apparent only after a delay of several years.

They showed that during the first 5 years after study entry, breast cancer mortality rates for women 55 years of age or older at study entry were higher than those for controls. The rates were identical at 6 years, but, from year 7 onward, the rates were lower for the screened women. From years 8 to 11, the reduction in risk of death from breast cancer in these older women was 55%.