[Skip to Navigation]
Sign In
Views 154
Citations 0
Comment & Response
February 2016

Differing Perspectives on Breast Cancer Chemoprevention—Reply

Author Affiliations
  • 1Breast Cancer Research, Women’s College Research Institute, Toronto, Ontario, Canada
JAMA Oncol. 2016;2(2):277. doi:10.1001/jamaoncol.2015.4395

In Reply Drs Cuzick and Jordan are both leaders in the field, and I am grateful to each of them and their co-authors for commenting on the important question of tamoxifen chemoprevention. Specifically, does tamoxifen prevent death from breast cancer, and if it does not, does it have any role to play in the preventive setting?

The statistical arguments made by Cuzick and colleagues may be convincing to other statisticians. Just because the trials were not designed to measure the beneficial effect of tamoxifen use on mortality does not mean that we should assume that there is any. It is the mandate of the authors to show that tamoxifen therapy prevents deaths and until the authors do so, I will hold off on recommending it. When the trials were designed in the 1980s, it was assumed that a reduction in incidence would precede a proportionate reduction in mortality. However, due to better understanding of overdiagnosis and heterogeneity in outcome, we no longer assume this to be the case and any conclusions regarding mortality should be based on mortality data and not on assumptions about mortality. Cuzick and colleagues state correctly that to demonstrate a 50% reduction in mortality with adequate power, 216 deaths would be needed and to date only 92 deaths have been reported. It is fair to say that to show that tamoxifen works would require many more than 92 deaths. However, I think it could be shown not to work with far fewer—suppose (for the sake of argument) that 82 of the first 92 deaths were in the tamoxifen arm. Cuzick et al may be correct in saying that I lack an understanding of the facts about prevention. Nevertheless, I try to muddle through counseling sessions and explain complicated concepts of risk to patients who are then expected to make a decision. I could recommend that they can begin tamoxifen therapy today and stick with it for 5 years but they should not expect to reduce their chance of dying of breast cancer in the next 20 years. I anticipate that (in the words of Dr Jordan) “adherence will dwindle.” I wish we had more to offer; to say that preventive strategies [for breast cancer] are “in their infancy” is a euphemism for the stubborn fact that (other than preventive mastectomy) there are no effective preventive strategies for breast cancer. Jordan makes the interesting point that better SERMs are now “on the way” that will target multiple diseases, including breast cancer. What is particularly interesting about this prediction is that it was also made in 20011 and anticipated our current preoccupation with personalized medicine.

Add or change institution