In your recent editorial on breast cancer detection [1], you legitimately stress the importance of breast cancer screening for women over 50 years of age. Unfortunately, you do a disservice by suggesting that women 40 to 49 years old should no longer be offered breast cancer screening because there is no supporting "scientific" evidence. There are no data that absolutely prove efficacy for the screening of women 40 to 49 years old because there have been no properly designed or performed prospective studies. If the studies that have been done are analyzed the way they were designed, efficacy has been clearly shown for screening women 40 to 64 years old and 40 to 74 years old [2, 3]. Only retrospective analyses with unplanned stratification by age have raised any question, but none of the trials had sufficient numbers of women to have adequate statistical power, when stratified by age, to address these women separately.

The only clinical trial that was designed to prospectively evaluate screening women 40 to 49 years old was the National Breast Screening Study of Canada (NBSS). The poor quality of the mammography in the NBSS is not a trivial problem but rather critical to the outcome of the trial. There was no training for the technologists performing the mammograms or for the radiologists interpreting the mammograms; the centers were allowed to use available equipment as long as they maintained a low dose. The poor quality of the NBSS mammograms was pointed out to the investigators at the outset and repeatedly. Two successive advisors to the NBSS, Dr. Wende Logan-Young and Dr. Stephen Feig, resigned over the refusal to correct the problems. Having participated in an external review of the NBSS mammography, I can attest to this "fatal" flaw. The investigators admit that for much of the study, fewer than 50% of the mammograms were graded as even "acceptable" [4] and that the (untrained) radiologists should have detected 25% of the breast cancers on mammograms 1 to 5 years earlier than they were eventually diagnosed [5]. The investigators have still not determined how many cancers were not even imaged because of poor positioning.

The NBSS was so poorly done that the results do not even show a benefit for women 50 to 59 years of age where unequivocal benefit has been shown in other studies. Why should the results for women 40 to 49 years of age be considered valid?

The NBSS did not publish cancer size distribution data, probably because 50% of the "screen"-detected cancers in women age 40 to 49 were stage II by node status alone. Size criteria would bring this figure closer to 70% or worse, a poor reflection on the quality of the mammography.

Other major questions remain unexplained. Women were not randomized blindly but after having had a physical examination; they were not disqualified if a mass was found. There were significantly more women with advanced cancers (40% more based on node positivity alone) "randomized" to the screened group, raising questions about the randomization.

The editors cite other studies that "fail to show efficacy," but each of these trials had significant problems. In the Malmo trial, there were too few women in the under-50 age group for statistical validity; 35% of the control women had mammograms outside the trial, and at least 20% had their cancers detected by these mammograms. In the NBSS, 25% of the "controls" had mammography. The already insufficient statistical power is further diluted by "crossover" in these studies.

Contrary to the statement made in the editorial, the Swedish data reveal that, after 8 years, the mortality curves are diverging (with a 13% benefit thus far) and that a benefit is beginning to accrue for women age 40 to 49. These same results were seen in the Health Insurance Plan of New York (HIP) study.

Rather than rushing out with an editorial that may affect the lives of women all over the world, the available data should have been reviewed in greater depth, and "science" should not have been invoked where none exists. Instead of withdrawing support for screening, we should be working to provide the highest quality mammography and clinical breast examinations as efficiently and inexpensively as possible. We should not reproduce the mistakes of the past. Because the lead time for mammography for women age 40 to 49 is approximately 2 years, these women should be screened every year and not the consensus compromise of every 1 to 2 years.