Longitudinal Study of Mammographic Density Measures That Predict Breast Cancer Risk

Area of Science:

• Epidemiology
• Biomarkers
• Radiology

Background:

• Mammographic measures including dense area (DA), percent dense area (PDA), and nondense area (NDA) are established predictors of breast cancer risk.
• These associations are adjusted for age and body mass index (BMI).

Purpose of the Study:

• To assess the longitudinal tracking of breast cancer risk-predicting mammographic measures over time.
• To estimate the extent to which DA, PDA, and NDA remain consistent within individuals over extended periods.

Main Methods:

• Utilized longitudinal mammogram data from 970 women (2,320 mammograms) from the Melbourne Collaborative Cohort Study and Australian Breast Cancer Family Registry.
• Measured DA, PDA, and NDA using Cumulus software, normalized with the Box-Cox method.
• Employed nonlinear mixed-effects modeling of Gompertz curves to estimate correlations in normalized measures over varying time intervals.

Main Results:

• Mean normalized DA and PDA showed age-related changes, decreasing after the early 40s. Mean normalized NDA increased nonlinearly with age.
• High within-woman correlations were observed for normalized DA, PDA, and NDA over time intervals up to 10 years, even after adjusting for age and BMI.
• Correlation estimates for normalized DA were 0.91-0.94 for mammograms 2-10 years apart.

Conclusions:

• Mammographic measures predictive of breast cancer risk exhibit strong correlations over time.
• These findings support the potential for early identification of women at increased risk, enabling timely screening and prevention strategies.