of such a test depends on existing options. For example, the development of a test for ovarian cancer, even one that is not highly accurate, could save many lives because there is no existing technique to detect early stage ovarian cancer. In contrast, a comparable test for breast cancer would be unlikely to save lives unless it is more sensitive or specific than mammography and gave localization information or was paired with mammography.
Recent headlines to the contrary, it will be many years—if ever—before blood tests replace mammograms. The most obvious reason is that a blood test would measure biomarkers (usually proteins) that have been released from cancerous tissue into the general blood circulation, which means they are highly diluted in the midst of a multitude of other proteins, and are a long way from their source. A blood test would have to be able to measure trace quantities of any biomarkers and, at best, a blood test would indicate that cancer was present somewhere in the body, but not where—unless the biomarkers were found only in breast tissue, which puts yet another restriction on possible tests. For example, a problem could arise if it was so sensitive that only a few cancer cells would result in a positive test. The cancer could not be physically located with current imaging technology within the breast and thus true positives could not be distinguished from false positives.
No existing blood test—for breast or any other cancer—rivals mammography as a screening method. Mammography has an acceptable sensitivity, and despite its modest specificity, it locates the tumor for definitive biopsy. Furthermore, mammograms provide richer data than would be possible from a low-dimensional biochemical assay that measures only one or a few substances; improvements described in Chapter 3 have the potential to increase the information available from mammography.
Many different biologically based approaches to detecting breast cancer are in development, but they face many of the same challenges if they are to become truly useful for improving outcomes for breast cancer patients. Certain themes recur throughout this chapter in the discussions of the different types of biologically based cancer detection technologies:
Biological methods may prove to be advantageous for screening high-risk populations, but are not likely to replace mammography.
Nonimaging biological techniques must be linked to imaging methods that can localize the cancer.
Statistical methods necessary for definitive analysis of large genomic and proteomic data sets are not yet defined or standardized.
Assays to detect cancer must account for the variability that exists among tumor types and among patients in order to be effective for widespread use.