Longitudinal Change in Mammographic Density among ER-Positive Breast Cancer Patients Using Tamoxifen

Abstract

Tamoxifen-associated mammographic density (MD) reductions are linked to improved breast cancer survival. We evaluated MD at six time points to determine the timing of greatest reduction following tamoxifen initiation. We sampled 40 Kaiser Permanente Northwest estrogen receptor (ER)-positive breast cancer patients from a prior study of MD change, according to tamoxifen use duration and age at diagnosis: <4 years tamoxifen and ≤50 years (N = 6) or >50 years (N = 10) old; ≥4 years tamoxifen and ≤50 years (N = 13) or >50 years (N = 11) old. A single reader evaluated percent MD in the contralateral breast on baseline (pre-diagnosis) and five approximately yearly post-diagnostic (T1 to T5) mammograms. Mean MD change was calculated. Interactions with age (≤50 and >50 years), tamoxifen duration (<4 and ≥4 years), and baseline MD (tertiles) were tested in linear regression models. Overall, the largest MD decline occurred by T1 (mean 4.5%) with little additional decline by T5. Declines differed by tertile of baseline MD (Pinteraction < 0.01). In the highest tertile, the largest reduction occurred by T1 (mean 14.9%), with an additional reduction of 3.6% by T5. Changes were smaller in the middle and lowest baseline MD tertiles, with cumulative reductions of 3.0% and 0.4% from baseline to T5, respectively. There were no differences by age (Pinteraction = 0.36) or tamoxifen duration (Pinteraction = 0.42). Among ER-positive patients treated with tamoxifen and surviving ≥5 years, most of the MD reduction occurred within approximately 12 months of tamoxifen initiation, suggesting that MD measurement at a single time point following tamoxifen initiation can identify patients with substantial density declines.

Figure 1

Mammogram timing. One pre-diagnostic (baseline) and five post-diagnostic (T1-T5) mammograms of the contralateral breast were evaluated among 40 ER-positive breast cancer patients who were treated with tamoxifen. Mammograms were approximately one year apart, with the exception of the first post-diagnostic mammogram (T1) which was approximately 12 months after the initiation of tamoxifen and 18 months after the baseline mammogram.

Figure 2

Change in percent density over time is shown stratified by percent density at baseline. Lines represent smoothed mean values of percent density change at each time point (compared to the previous mammogram) and ‘x’, ‘o’, and ‘+’ markers represent values for individual patients. There was a notable reduction in percent density observed at T1 among women in the highest tertile of baseline density (solid red line, ‘x’), but not among women in the middle (dashed blue line, ‘o’) and lowest (dotted black line, ‘+’) tertiles of baseline density. For all subjects, relatively little change in density was observed at T2 to T5.