Mammographic density as a risk factor: to go out of a 30-year fog

Millions of healthy women with so-called dense breasts are identified in mammographic breast cancer screening programs. Mammographic density (MD) as a risk factor is resolutely another issue than the masking effect linked to potential undetected breast cancers in dense areas. MD as a risk factor remains hotly debated, as it intrinsically depends on so many radiological and technical parameters to be assessed and quantified, along with so many confounding factors and ill-studied density variations all over life (1). To go out of a 30-year fog, notably for radiologists, it is paramount to consider basic epidemiological steps.

First, a large body of literature supports the existence of a common organ dose-response relationship between anthropometric factor and cancer risk (2–4). The more morphologically exposed tissue, the higher the cancer risk. The breast is not an exception. Consequently, such a basic continuous risk relationship between the “fibro-glandular tissue” amount and the “risk of breast cancer” cannot justify itself as a significant risk factor.

Second, the major key of clarification is nested in the relative risk (RR) data analysis and the conditions to use a risk factor as a test to distinguish a significant risk category in the population (1). With the introduction of mammographic campaigns in the 1980s, this relationship has been addressed by establishing an improper comparison between two extreme density group quantifications: women with fatty breasts and women with very dense breasts resulting in some overestimated RR or odds ratio of about 4 or 5. These reported data were established only with classifications, which isolated extreme density categories and compared them. With a rigorous public health point of view, it is necessary to have a control group representative of the risk of the majority of the population. In such relevant conditions, RR dropped from 4–5 to 2 or non-significant RR for women with dense breasts (1). Based on this evidence, the over-estimation of MD as a risk factor has been widely validated (5).

To inform women about the trap linked to misleading use of epidemiological data is going to be necessary to stop anxious messages and to avoid the deleterious consequences in terms of false-positives (women with dense breasts but no significant increased breast cancer risk).