for biopsy following a suspicious mammogram, if the bioassay reduced the number of false positives (increased specificity) without sacrificing sensitivity. Reliance on such an assay in cases of questionable mammographic results (for example, BI-RADS® 3-4) would be prudent only if the positive predictive value of the assay is extremely high, particularly when the result is to forgo biopsy. This is especially critical in the United States, where biopsy is the current standard of care for virtually every suspicious lesion. Bioassays may also be performed on the sample of suspicious cells obtained at biopsy in order to inform treatment decisions, but such tests will not supplant pathological examination of the biopsied tissue sample for the primary diagnosis.
As long as biomarkers continue to have low sensitivity and specificity, their use in primary diagnosis will be limited, and histological examination of biopsied tissue will remain the gold standard. But even then, biomarkers are likely to be useful as adjunct to other procedures, including:
Differential diagnosis or prognosis, such as distinguishing among types of ductal carcinoma in situ;
Assistance in the choice of therapy and evaluation of its outcome; or
Monitoring patients with ongoing disease before or after therapy.
Results of preliminary studies suggest that pre-operative serum levels of CA 15-3 are as good, if not better, predictors of patient outcome than traditional measures such as tumor size and nodal status.18 Tissue levels of estrogen and progesterone receptors and the erbB2 receptor, as determined by immunohistochemical analysis, are considered in the selection of therapy.18,42 CA 15-3, approved in 1997 by the Food and Drug Administration (FDA) for the detection of recurrent breast cancer, may also prove useful in monitoring response to therapy for metastatic breast cancer.18
The path to biomarker-based assays for breast cancer, and particularly for the early detection of the disease, is far from smooth. The considerable challenge of identifying highly sensitive and specific screens that rival the effectiveness of mammography is made more difficult by biological heterogeneity among humans, as well as among cancers, and even among different cell populations within a single tumor.55 A successful bioassay for breast cancer will need to overcome variability associated with cancers of different histologic types, expression patterns within histologic types, additional (noncancerous) patient conditions, and intrinsic human biochemistry. For now, as noted by Kenneth Pritzker, “our conceptual framework of