Abstract
Purpose
It is clinically important to refine strategies to manage women with dense breasts. We aim to identify changes in density patterns related to breast cancer risk.
Methods
This case-control study was sampled from the Joanne Knight Breast Health Cohort of 10,481 women cancer free at entry and followed from November 3, 2008 to October 31, 2020. Routine screening mammograms were performed every 1 to 2 years, providing a measure of breast density. 289 pathology-confirmed breast cancer cases were identified and approximately 2 controls were sampled for each case based on age at entry and year of enrollment, resulting in 658 controls. Screening mammograms were used to measure volumetric changes in breast density over time. We analyzed 8,710 craniocaudal view mammograms using longitudinal latent class to cluster women and estimate density change over time.
Results
The latent class and density trajectory were observed separately for dense and non-dense groups of women. The data defines 2 classes among both groups of women: those with decreasing density and those with increasing density over time. The odds ratio for cancer comparing decrease vs. increase is 4.85 (95% CI 1.66, 14.19) for women with dense breasts (BI-RADS C, D) and 6.17 (95% CI 2.89, 13.17) for women with non-dense breasts (BI-RADS A, B).
Conclusion and Relevance:
Latent class analysis confirms that change in density over time is important for risk classification for women with dense or non-dense breasts. Thus, a one-time measure of density is insufficient to define the level of breast cancer risk.