Assessment of mammographic density changes on plain film mammograms in postmenopausal women on hormone replacement therapy.

ABSTRACT

BACKGROUND & OBJECTIVES: Mammographic screening is an effective tool for the early detection of breast cancer. Hormone replacement therapy (HRT) has been shown to increase mammographic density and thus may hinder early detection of small tumours. We undertook this study to determine and compare the frequency and degree of change in mammographic density in postmenopausal women in HRT using two different methods: the classical Wolfe classification and a new semiquantitative method, which we named as the comparison wheel. METHODS: This study included 285 women, 206 under hormone treatment, and 79 control subjects. All women underwent baseline mammographic study before the beginning of treatment. Mean interval of the follow up mammograms was 16 months. The methods were compared in evaluating the effects of three types of hormone therapies on mammographic density. RESULTS: The frequency of change was only significant in the combined hormone replacement group when Wolfe classification was used. However, the frequency of increase in density (estrogen group 21%, combined therapy group 42%, tibolone group 28%) was markedly higher when the comparison wheel was used. The inter-rater Kappa value was calculated as 0.977 for the first and 0.957 for the second readings of the two radiologists for the comparison wheel, and 0.973 and 0.968 for the Wolfe classification. The intra-rater Kappa values were determined as 0.972 and 0.957 for the first and and 0.963 and 0.926 for the second radiologist for comparison wheel and Wolfe classification respectively. INTERPRETATION & CONCLUSION: Our findings indicate that the estimated increase of mammographic density depends on the selected hormone regimen, as well as the method of evaluation. The comparison wheel is a semiquantitative method of evaluating changes of mammographic density and is sensitive and reproducible with high inter- and intra-rater Kappa values. This method can be used as an alternative for comparison of digital mammographic applications in the future.