COMMENTS AND OPINIONS
Spontaneous Regression of Breast CancerT he recent article published by Zahl et al 1 brought up some very interesting discrepancies in incidence rates of invasive breast cancer between screened and unscreened groups of women in the Norwegian Breast Cancer Screening Program. At the end of the study, there was a much lower rate of invasive breast cancer diagnosed in the unscreened group, which the authors contribute to spontaneous regression. According to their theory, if all invasive tumors that develop are persistent, then at the end of the study period, the screened group and the control group should have a similar incidence, and since this is not observed, then some tumors must spontaneously regress.The authors test a lot of alternative theories to their conclusions and seem to prove that their theory still holds. However, there is a potentially fatal mistake that the authors make when refuting the alternative hypothesis that hormone therapy (HT) differences between the 2 groups can account for the observed difference in invasive breast cancer incidence. The authors assume that HT accounts for a 24% increase in the diagnosis of invasive breast cancer, and they base this assumption on the conclusions from the Women's Health Initiative study, which was conducted on a group of US women. 2 Based on this assumption, as the authors point out in their study, a 24% increase cannot account for the observed difference seen in their study. Unfortunately, previous research has shown that Norwegian women have substantially higher rates of invasive breast cancer due to HT than US women. A study published by Hofvind et al, 3 which analyzed essentially the same screened group of women as did Zahl et al, 1 found that attributable risk due to HT in the population of Norwegian women was 58%. In this scenario, if we use the most extreme example used by Zahl et al 1 in which there was no HT use in the control group and 46 000 women in the screened group with HT use (58%ϫ46 000/119 000=22.4%), then the observed difference based on HT alone would account for a 22% difference in cumulative incidence. Therefore, according to this most extreme example, the entire difference seen in the cumulative incidence of invasive breast cancer between the screened group and the control group could be due to HT alone and have nothing to do with spontaneous regression.