The advent of high-throughput technologies in the 1990s generated great anticipation that patterns of gene expression or of other molecular characteristics of tissues or tumors would allow refinements in diagnosis, prognosis, and treatment selection to revolutionize the treatment of cancer. A voluminous literature was generated. Companies were formed. New statistical methods were developed to analyze the thousands of data elements that were analyzed in scores of samples. A large number of important discoveries have come from the ability to perform wholesale analyses of gene expression, gene methylation, and DNA alterations. However, application of high-throughput technologies to clinical practice and to decision-making that affects patient care has been slow to evolve from these findings. Thus early anticipation that macromolecular profiles of cancers would replace conventional diagnostics and provide prognostic insight has largely been unfulfilled. There are some important instances where the management of some cancers has incorporated information that was originally gleaned from high-throughput analyses. This chapter will summarize different approaches to high-throughput analysis and enumerate the instances where results from these approaches have impacted patient care.

The importance of biomarkers

Cancer treatment is characterized by the application of morbid and toxic therapies to all patients with a particular stage of a cancer to benefit a subset of those patients. We have long sought to discriminate between those patients who will benefit from a therapy and those who will not. In the age of molecular oncology, clinical discriminators have been sought among biomarkers. A “biomarker” is “a characteristic that is objectively measured and evaluated as an indicator of normal biological processes, pathogenic processes, or pharmacologic responses to a[n]…intervention” (1). For example, the serum cholesterol level is a biomarker that indicates risk for cardiac disease and indicates the use of cholesterol-lowering drug therapy. Similarly, biomarkers in cancer treatment include expression of estrogen receptor (ER) to indicate the need for hormonal therapy of breast cancer. Other examples of useful biomarkers in cancer therapy are α-fetoprotein and β-human chorionic gonadotrophin, indicators of active and recurrent germ-cell tumor. In clinical oncology there are a limited number of individual biomarkers that are useful for diagnostic, prognostic, or predictive purposes.