Digitized assessment of mammographic breast density – Effects of continuous combined hormone therapy, tibolone and black cohosh compared to placebo

Abstract

Objectives

To determine the effects of continuous combined hormone therapy, tibolone, black cohosh, and placebo on digitized mammographic breast density in postmenopausal women.

Study design

A prospective, double-blind, placebo-controlled study of 154 postmenopausal women randomized to estradiol 2 mg/norethisterone acetate 1 mg (E2/NETA), tibolone 2.5 mg or placebo and a prospective, open, uncontrolled drug safety study, of which 65 postmenopausal women were treated with black cohosh. Mammograms, at baseline and after six months of treatment, were previously classified according to visual quantification scales.

Main outcome measures

Reanalysis of assessable mammograms by digitized quantification of breast density.

Results

Treatment groups were comparable at baseline. During treatment, both E2/NETA and tibolone significantly increased breast density (mean increase 14.3%, p < 0.001 and 2.3%, p < 0.001, respectively), while black cohosh and placebo did not. Twenty-four out of the 43 women on E2/NETA had an increase in density exceeding 10% and 6 women had an increase of 30% or more. In the tibolone group, only one woman had an increase in density of more than 10%. The difference in increase in breast density between E2/NETA on the one hand and tibolone, black cohosh and placebo on the other was highly significant (p < 0.0001).

Conclusions

Digitized mammographic breast density is a highly sensitive method confirming significant increase in density by standard E2/NETA treatment and to a lesser extent by tibolone, whereas black cohosh does not influence mammographic breast density during six months treatment. Digitized assessment also yields data on individual variation and small increases left undetectable by visual classification.

Introduction

Mammographic breast density has been defined as a strong and independent risk factor and it has even been suggested as an intermediate phenotype for breast cancer [1], [2], [3], [4]. Density in individual women seems to reflect the net influence of endogenous and exogenous hormonal stimuli and their background genetics on the breast during the lifespan [5], [6]. During conventional hormone therapy (HT), an increase in mammographic breast density has repeatedly been demonstrated to occur in a significant proportion of postmenopausal women [7], [8], [9]. It also seems that the degree of change during HT follows our current perception of breast cancer risk from different treatment principles. In agreement with data from the Women's Health Initiative (WHI) study, most combined estrogen/progestogen treatments will increase density whereas estrogen alone has very little effect on this risk factor [10].

Many different methods have been used to classify and quantify mammographic breast density. Much of the older information on the effects of HT has been based on visual classifications according to different scales, e.g., Wolfe, BI-RADS and percentage scales [8], [11], [12]. All these traditional visual classifications represent a rather crude measurement where an increase in density of about 20–25% could be required for an upgrading of one class. Apart from these limitations, visual assessment may also be subject to bias and inconsistency. Since mammographic breast density should be regarded a continuous variable, digitized methods allow a more sensitive measurement and are able to quantify more discrete changes than traditional visual classification. Currently, there is a rapid development of computer assisted techniques for individual assessment of density and the monitoring of change following interventions [13]. Using visual classifications, we previously reported an increase in density to occur in about 50% of women treated with 2 mg estradiol (E2)/1 mg norethisterone acetate (NETA), but in only 6% of women treated with tibolone [8]. During treatment with black cohosh and placebo none of the women showed any increase [14]. However, in these studies the degree of change in individual women was not quantified.

Therefore, we carried out a reanalysis of the previous material using a highly sensitive digitized technique in order to quantify the individual response and to see if we could confirm the differences between treatments.