Can breast cancer regress?

Routine mammography is the gold standard for breast cancer screening in the United States and most of the Western world. When an invasive tumor is detected, treatment generally follows. Because of this, the natural history of invasive breast cancer picked up on mammography screenings is not clear. A study from Norway suggests some breast cancers detected by routine screening might spontaneously regress if left untreated.

“We first became interested in this topic following the Cochrane-reviews of Peter Gøtzsche and co-workers suggesting that overdiagnosis in mammography screening programs is a serious problem that may lead to overtreatment and other harms,” said Jan Mæhlen, MD, from the department of pathology at Ullevå University Hospital in Oslo. “When we discovered a way to analyze the Norwegian data, the study was not difficult to perform. Our time and effort have mostly been spent on convincing reviewers that the study design is sound.”

How the Study Was Conducted

Professor Mæhlen and his group compared cumulative breast cancer incidence in age-matched cohorts of women in four Norwegian counties with biennial mammography screening programs. The screened group consisted of women who were invited for 3 rounds of mammography between 1996 and 2000. The control group included women who would have been invited for screening between 1992 and 1997 had such a program existed at that time. Women in the control group were offered a single prevalence screening. Counts of invasive breast cancers were obtained from the Norwegian Cancer Registry.

The 4-year cumulative incidence of breast cancer was significantly higher in the screened group than in controls (1268 vs 810 per 100,000 women). Even after the controls underwent prevalence screening in years 5 and 6, the cumulative incidence of invasive breast cancer remained 22% higher in the screened group (1909 vs 1564 per 100,000 women). Higher incidence was seen in screened women across every age bracket.

“Our main finding is that there were 22% more cancers during 6 years among women offered 3 biennial screenings than in women offered a single screening at the end of the 6-year period,” he said. “This suggests that the extra cancers among screened women are pseudocancers that spontaneously regress and disappear in un-screened women.”

There is currently an ongoing study of similar data sets from Sweden and Norway. If confirmed, Professor Mæhlen said it is likely that they will conclude that screening leads to harmful overdiagnosis instead of early diagnosis. What that means from a clinical perspective remains to be seen.

“Something like ‘watchful waiting’ would be an option if we knew which tumors are bound to regress,” he said. “At the moment we have no method to tell which cancers progress and which cancers regress. Research is needed to find histological, molecular or genetic markers for the pseudo-cancer (if such markers exist) and to rule out the possibilities that cancer regression is due to an attack by the immune system or other defense systems of the patient.”

Questioning the Results

“One of the things the researchers don’t talk about that could potentially effect their results is the quality of mammography and the techniques used,” said Ellen Shaw de Paredes, MD, communications chair for the Breast Cancer Commission of the American College of Radiology and director of the Shaw de Paredes Institute for Women’s Imaging in Richmond, Virginia. “They don’t discuss this, yet they are making many assumptions about screening mammography. In addition, there is not much information given about any changes in the way they performed their studies over this time period.”

Dr. Paredes also expressed concern about the possibility of contamination between the control and study groups that may have skewed results. “The screening group was composed of those who were invited to be screened, not necessarily those who actually were screened,” she said. “The control group [comprised] women not invited to screenings because there was no program in place. Apparently some of the women in the control had mammograms and some who were invited to be screened either never were or had them less frequently than offered.”

Dr. Paredes also expressed concern that investigators attributed the the lesser incidences of cancers seen in the control group to cancer regression when other explanations were possible. “There are certainly cancers missed by radiologists,” she noted. “If you have 3 radiologists looking at 3 mammograms in the screening group, there is a much larger opportunity for detection than 1 radiologist looking at a single film. I don’t think from this study, one can conclude that 22% of cancers might be regressing.”

Ethical Concerns About Watching and Waiting

Even if the results are confirmed and some breast cancers do regress, legal and ethical concerns render the clinical significance of the finding moot. Since there is currently no method to determine those that will regress and those that will grow, the only reasonable practice is to remove all lesions. Dr. Paredes said there is no equivalent of “watchful waiting” in breast cancer and she sees no way ethically or legally to complete these types of studies. “I don’t see any way clinically that we can presume to identify which women do not need to have their cancers removed,” said Dr. Parades.

“We have learned so much over the years about the value of mammography in reducing breast cancer deaths. To step back and suggest that mammography is finding breast cancers that would have regressed any way is dangerous,” she concluded.

Zahl P-H, et al. The natural history of invasive breast cancers detected by screening mammography. Arch Intern Med. 2008;168:2311-2316.