Mammogram data unlikely to change screening policy: overview analysis concludes that mammograms don't increase life expectancy. (New Analysis Called Flawed).

Mammography screening policies in the United States are unlikely to change despite a new overview analysis that concluded that mammography has not proved to cut overall mortality risk.

Even if the new analysis were correct and mammography did not increase the overall life expectancy of women, other study results indicate that mammography is able to cut the risk of death from breast cancer, several experts told this newspaper.

"The data are solid that screening mammography saves lives," said Robert Smith, Ph.D., director of cancer screening for the American Cancer Society in Atlanta.

"If you combine the results of all studies, you probably need to screen about 800 women aged 50 or older every year over 10 years to cut the rate of breast cancer deaths by 1," said Dr. Barnett S. Kramer, a senior scientist at the National Cancer Institute in Bethesda, Md.

The new analysis was a revision of a controversial report published last year by a pair of researchers at the Nordic Cochrane Centre in Copenhagen. Both last year's report and the new revision carefully reviewed seven published, randomized trials that assessed the efficacy of screening mammography. The more recent version provided a fuller analysis of the data.

Both analyses found flaws in several of these trials. The overviews focused on the proven ability of mammography to reduce total mortality rate because this is the only "reliable" measure of benefit, wrote Dr. Ole Olsen and Dr. Peter C. Gotzche (Lancet 358[9290]:1340-42, 2001). Studies that focused on breast cancer deaths missed tallying fatal cases that were misclassified or that were triggered by cancer treatment such as radiotherapy, they wrote in their latest article, which was published in late October.

An accompanying editorial written by Dr. Richard Horton, editor of the Lancet, praised the "rigorous and well-developed method" for conducting systematic literature reviews that is used by researchers in the Cochrane Collaboration. Based on the new review, "there is no reliable evidence from large, randomized trials to ...