Etiology and early detection of breast cancer: Biomarkers, lifestyle and mammographic density

Abstract

In the Netherlands, one in seven women will develop breast cancer at some point during her lifetime. Although breast cancer mortality has declined, still 23% of the breast cancer patients dies from the disease within 10 years after initial diagnosis. This emphasizes the importance of research on breast cancer prevention and screening. This thesis aims to find clues for primary (part I) and secondary (part II) prevention of breast cancer. To design primary prevention strategies, knowledge on the etiology is required. In part I, studies on the association between estrogen metabolites, lifestyle factors and (subtype-defined) breast cancer risk are described. We observe that high urinary estrogen and estrogen metabolite concentrations are associated with higher postmenopausal breast cancer risk. Besides on the exploration of the underlying pathway, this thesis zooms in on the effects of weight changes later in life and the consumption of vegetables and fruits on subtype-defined breast cancer risk. We observe that high vegetable intake is associated with lower overall breast cancer risk. The inverse association is most apparent for hormone receptor-negative breast cancer. Furthermore, we find evidence that a high weight gain in middle adulthood is associated with slightly higher breast cancer risk. The association seems to be more pronounced for breast cancer diagnosed before age 50. In part II, we focus on the early detection of breast cancer within the scope of the national population-based screening program. In the Netherlands, women between the ages of 50 and 75 years are screened for breast cancer every two years using mammography. In women with high mammographic density (i.e. high amount of fibroglandular tissue and low amount of fat tissue in the breast) the sensitivity of mammography is reduced and these women have a higher breast cancer risk. This dual effect makes mammographic density a promising key determinant in the evaluation and personalization of the national screening program. In this thesis, we investigate whether regional variation in mammographic density is present within the population-based screening program. We find evidence that women living in urban areas have a higher mammographic density compared to women living in nonurbanized areas. As we observe regional variation in mammographic density, we conclude that mammographic density is an important parameter to incorporate into the evaluation of screening performance. Furthermore, this thesis describes the rationale and design of the DENSE trial. The DENSE trial is a multicenter parallel-group randomized trial that investigates the (cost-)effectiveness of screening with mammography and MRI, compared to mammography alone in Dutch breast cancer screening participants with extremely dense breasts. In conclusion, the results of part I underline that nutrition, weight and hormones play a role in the etiology of breast cancer. It is important to take tumor subtype into account when studying (the etiology of) breast cancer. Based on part II, we advise to incorporate mammographic density as a parameter for monitoring the quality of the national screening program. The question whether screening participants with extremely dense breasts benefit from additional screening with MRI will be answered in the coming years.