Can breast cancer simply disappear?

A new study in the Archives of Internal Medicine has been getting a lot of attention for its extraordinary claim that breast cancer can spontaneously disappear without any treatment. If it’s true, it raises a host of important questions: Should some women with breast cancer forgo toxic treatments? How does cancer simply disappear? Should we stop aggressive breast cancer screening?

The operative word, of course, is “if.” The case of the disappearing breast cancers described in the scientific paper is powerful but entirely circumstantial. The authors did not see even a single cancer disappear. They hypothesized that some breast cancers disappeared because they did not find as many cancers as they expected.

The paper is The Natural History of Invasive Breast Cancers Detected by Screening Mammography written by Zahl, Maehlan and Welch. The authors recruited over 200,000 Norwegian women ages 50-64 and divided them into two groups. The tested group had mammograms every other year for 6 years (a total of 3 mammograms). The control group had only one mammogram at the end of the 6 year period. The authors were very careful to make sure that the two groups were as similar as possible.

Over the course of the 6 year interval, 1909 cases of breast cancer were diagnosed per 100,000 women who had screening in every other year. In contrast, only 1564 cases of breast cancer were diagnosed per 100,000 women who had only one mammogram at the end of the 6 year interval. The authors concluded that over 300 cases of breast cancer were “missing” from the control group and they believe that those cancers arose and spontaneously disappeared.

Is that what really happened? First, we must consider the fact that the 1564 cases diagnosed in the control group does not include all cancers in the group. Screening mammography is not perfect and some cancers will be missed that might be detected on a subsequent scan.

Second, the use of estrogen replacement therapy in the study group was substantially higher than in the control group. Since hormone therapy has been associated with increases in breast cancer risk, that may account for part of the observed difference.

Third, the incidence of breast cancer in Norway increased over the life of the study. The women in the study group (1996-2002) had a known higher incidence of breast cancer than the women in the control group (1992-1998).

Even taking these factors into account, there is almost certainly a real difference in the cumulative number of breast cancer cases between the study group and the control group. In other words, routine mammography may over diagnose breast cancer by finding cancers that would disappear on their own if not treated.

That’s not a surprising finding. The body has many mechanisms for controlling the growth of runaway cells that could lead to cancer. The appearance of an actual cancer represents a failure of these mechanisms. It is possible, particularly in the case of small cancers detectable only by mammography, that there is still one last fail safe mechanism that could even then attack the tumor and destroy it. This study may be demonstrating that phenomenon.

There are some serious limitations to this study. The study only looks at incidence of cancer. It does not look at outcome and life expectancy. If it turns out that the women in the study group have a much lower incidence of death from breast cancer, because they are treated early and aggressively, it will justify the apparent over diagnosis of breast cancer. That is a very real possibility, because the data show that although the cumulative incidence of breast cancer in the control group was lower, many cases of breast cancer in the control group were not diagnosed until the mammogram at the end of the 6 years.

Finally, and most importantly, there is no way to tell the difference on mammography, or by any other technique, between the cancers that will disappear and the ones that will go on and kill the woman. Without a practical way to separate those who need to be treated from those who do not, the finding is intriguing and worthy of further investigation, but cannot guide us in determining the best way to screen for breast cancer and the best way to treat it.