Mammography Saves Lives—Or Does It?

As a consumer advocate who has followed the controversies over mammography since the technology was introduced in the1970s, I’m convinced that the current controversy has resulted from the premature promotion of mammography to younger women. Both the American Cancer Society and the National Cancer Institute had recommended mammography for women ages 35 and older—before any research supported the advice. Theoretically, the controversy over mammography recommendations ended in 2002 with a reaffirmation of its value to women over age 40 when the U.S. Preventive Services Task Force reviewed seven randomized, controlled trials. Two found no lifesaving benefit to mammograms for women at any age, but five showed a lower breast cancer death rate among all screened women, ages 40 to 69 years. Pool the results of all trials for women ages 50 and older, and the breast cancer mortality drops by 22% over 14 years in relation to the unscreened women. Narrow the group to women in their 40s, and the mortality-reduction benefit is 15%. One of the trials that showed no lifesaving benefit to mammography is the Canadian National Breast Screening Study, which included 50,000 women ages 40 to 49 years. The study’s seven-year results revealed 10 more deaths in the mammography-screened group than in the control group. This difference was dismissed as not statistically significant—that is, until the same slight increase in breast cancer deaths also showed up in the early years of five other trials. Each trial’s result may not be statistically significant, but when a result shows up so consistently, it becomes alarming. Michael Retsky, PhD, and colleagues have suggested that the deaths are due to the surgery that often follows an abnormal mammogram. They suspect that a surgical procedure, performed during a time of hormonal fluctuations within the premenopausal breast, could trigger the growth of invasive cancers in those with dormant micrometastases. Retsky and colleagues examined the recurrence patterns in premenopausal women treated in an era when surgery was the only option. They found that the slight increases in breast cancer deaths among these women parallel those in the early years of the mammography trials. They speculate that biopsy or surgery stimulates angiogenesis of dormant micrometastases. This is supported by the contention of leading breast cancer researcher Bernard Fisher, MD, that micrometastases have occurred in most breast cancers before clinical or radiologic presentation. This means that 5% to 6% of premenopausal women diagnosed as a result of screening are dying two years earlier than they would have if they had not had regular mammograms. Combine that with the findings of an analysis that disputes conventional medical wisdom: women have been told that mammography screening will lead to less radical treatment, but a 2001 review of all randomized controlled trials showed 20% more mastectomies among women given mammograms. In the Canadian trials there were 40 more cases of nonpalpable invasive breast cancer and 42 more cases of ductal carcinoma in situ in those screened with mammography than in those screened with a skilled clinical breast examination. Sounds good, doesn’t it? This demonstrates that the much-criticized mammography techniques in the Canadian trial were of high quality. But this ability to find more cancers in younger women without reducing the odds of dying suggests two major drawbacks. The technology leads to the unnecessary treatment of some cancers that might have remained latent, and it also detects invasive breast cancers so slow-growing that women will have long lives regardless of when tumors are found. In spite of these troubling findings, the media debate of 2002 ended with a simplified, incomplete message to younger women: mammography is still better than nothing. You may wonder why all the information has not been made known to the public. So do I.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.