Session III : Breast Breast imaging for risk assessment , cancer preven - tion , and early detection

Epidemiologic models such as the Gail model are validated for populations but have suboptimal discriminatory accuracy, making it difficult to advise women regarding the need or urgency for prevention drugs or prophylactic surgery. A number of reversible risk biomarkers have been established for use in risk assessment [1]. The two most powerful are mammographic breast density and evidence of precancerous changes in diagnostic or nondiagnostic tissue samples. Women with > 75% breast density have a five-fold increase in risk (for at least the next five years) relative to those with essentially no breast density [2]. Women with histologic or cytologic evidence of atypia have a fivefold increase in cancer risk over the next 5–10 years compared with those with no evidence of atypia [3,4]. Women with atypical hyperplasia tend to have higher mammographic breast densities (and vice versa), but the two biomarkers may have independent predictive significance. Both mammographic breast density assessed by breast imaging, reporting, and data system (BIRADS) and cytomorphologic atypia (from samples obtained by random periareolar fine-needle aspiration of high-risk women),when added to the Gail model, improve discriminatory accuracy as assessed by the concordance statistic [5,6]. With the use of BIRADS breast density categorization, however, this improvement is very modest. Especially for young high-risk women, an inexpensive imaging technique is needed which would avoid the use of radiation, accurately quantitate the portion of the breast encompassed by stroma and epithelium, and if possible, permit detection of early breast cancers. Breast spectroscopy is attractive in this regard. Breast-specific reversible risk biomarkers such as mammographic density, breast cytomorphology, and proliferation may serve both to help select appropriate candidates for prevention and monitor response to standard prevention therapy, or as part of a clinical trial to evaluate a new agent [7]. Tamoxifen, a standard prevention drug, substantially reduces mammographic density in women younger than 45 but not in women older than 55 [8]. It is not clear at present whether any drug likely to be effective in primary prevention substantially reduces density in postmenopausal women. Mammographic density may be most useful as a response biomarker in preand perimenopausal women. At present, it is unclear whether any drug, including selective estrogen receptor modulators or aromatase inhibitors, substantially alters tissue morphology after 6–12 months of treatment, the interval commonly used in Phase II prevention trials. This is in part due to the difficulty in quantitating the amount or volume of benign breast disease. A woman is considered to have atypical hyperplasia if 1% or 70% of epithelial cells captured in a tissue sampling procedure are atypical. Proliferation in areas of hyperplasia has been identified as a risk factor in a cross-sectional study [9] and is currently being used as a response biomarker in Phase II trials [7]. Tamoxifen is known to reduce proliferation in benign breast disease and breast cancer. Further, in cancer treatment trials, reduction in proliferation predicts clinical response (reviewed in [7]). If spectral imaging could be performed inexpensively, it would be ideal for assessing a response to an intervention, as no radiation is involved. Mammography and sonography are currently the two primary modalities for early detection. Screening breast MRI is more sensitive than mammography but, given the expense of imaging, is generally reserved for women from hereditary breast cancer families.