Breast density and breast cancer-specific survival by detection mode

Background: Breast density is known to affect breast cancer risk and screening sensitivity, but it may also be associated with breast cancer survival. The interpretation of results from previous studies on breast density and survival is complicated by the association between detection mode and survival. Here, we studied the effect of breast density on breast cancer-specific survival for different detection modes (screen-detected, interval <= 24 or > 24 months, non-participant). Methods: Data from the Nijmegen (Dutch) breast cancer screening programme were used. Women diagnosed with invasive breast cancer between 1975 and 2011 were included. Breast density was assessed visually, based on a dichotomized Wolfe scale: 'fatty breasts' (<= 25%) and 'dense breasts' (> 25%). Cox proportional hazard regression was used to obtain hazard ratios (HR). Results: We identified 2742 eligible women, with a breast pattern available for 2233 women. A diagnosis of interval cancer (HR 2.06, 95% CI 1.62-2.61) led to a significantly increased risk of breast cancer death compared with screen-detected cancer. No significant cause-specific survival difference between women with dense and fatty breasts was observed (HR 0.94, 95% CI 0.77-1.15). The hazard was only higher for women with dense breasts among interval cancers <= 24 m (HR 1.07, 95% CI 0.74-1.56). The hazard appeared to be lower for women with dense breasts than for women with fatty breasts among screen-detected (HR 0.77, 95% CI 0.53-1.11) and interval cancers > 24 m (HR 0.80, 95% CI 0.53-1.20). None of the effects were statistically significant. Conclusions: Detection mode is strongly associated with breast cancer death. No clear association is apparent between breast density and breast cancer death, regardless of detection mode.