Longitudinal Case-Control Study of Active and Passive Dense Mammographic Breast Tissue

Mammography is used as secondary prevention for breast cancer. Computer-aided detection and image-based short-term risk estimation were developed to improve the accuracy of mammography. However, most approaches inherently lack the ability to connect observations at the mammography level to observations of cancer onset and progression seen at a smaller scale, which can occur years before imageable cancer and lead to primary prevention. The Hurst exponent (H) can quantify mammographic tissue into regions of dense tissue undergoing active restructuring and regions that remain passive, with amounts of active and passive dense tissue that differ between cancer and controls at diagnosis. A longitudinal retrospective case-control study was conducted to test the hypothesis that differences can be detected before diagnosis and changes could signal developing cancer. Mammograms and reports were collected from 50 patients from Maine Medical Center in 2015 with at least a 5-year screening history. Age-matching patients within 2 years created a primary dataset, and within 5 years, a secondary dataset was created to test for sensitivity. The amount of passive (H [&ge;] 0.55) and active dense tissue (0.45 < H < 0.55) was calculated for each breast and was predicted by creating a linear mixed-effects model. Cancer status was a predictor for passive (p = 0.036) and active (p = 0.025) dense tissue using the primary dataset. However, when increasing the power, cancer status was a predictor for active dense tissue (p = 0.013), while breast status (p = 0.004), time (p = 0.009), and interaction (p = 0.038) were predictors for passive dense tissue. This suggests active dense tissue is a risk for cancer and passive dense tissue is an indication of developing cancer. Required Key MessagesO_LIMammographic dense breast tissue can be separated into regions of active and passive. C_LIO_LIThere is more active dense breast tissue in pathology-confirmed cancer cases than controls. C_LIO_LIIncreases in passive dense tissue in a breast could indicate a developing tumor. C_LI