Mammographic density in relation to tumor biomarkers, molecular subtypes, and mode of detection in breast cancer

Mammographic density is an established risk factor for breast cancer; however, the relation to tumor pathological parameters including the androgen receptor and molecular subtypes has not been extensively studied. In the Malmo Diet and Cancer Study, 733 invasive breast cancers were diagnosed from 1991 to 2007. Mammographic density was defined qualitatively. Tumor biomarker information including estrogen receptor (ER), progesterone receptor, androgen receptor (AR), human epidermal growth factor 2 (HER2), and Ki67 was collected. Surrogate molecular subtypes were defined as luminal A, luminal B, HER2 positive and triple-negative breast cancer (TNBC). Among the 632 tumors with mammographic and pathological information, 352 tumors were screening-detected and 280 clinically detected. Higher mammographic density was associated with ER-negative tumors [ORadj 1.93 (1.04-3.59)] and TNBC [ORadj 2.44 (1.01-5.89), luminal A reference], in clinically detected breast cancer. Similarly, higher mammographic density was associated with AR-negative tumors [ORadj 1.77 (0.80-3.93)] in clinically detected breast cancer, though the evidence for this association was weak. In clinically detected breast cancer, but not in screening-detected, higher mammographic density was associated with ER-negative tumors including TNBC. This study highlights the need for taking mode of detection into consideration when addressing mammographic density and tumor biomarkers.