The promise of personalized medicine has been around for many years, and growth in the field is rapidly gaining momentum. The concept is to use information from an individual's genomic, transcriptomic, and proteomic profiles to tailor a custom management plan for his or her disease based on an assessment of the disease's risks and aggressiveness (1). A recent revolution in personalized medicine has arisen after the introduction of “molecular profiling” approaches. These approaches, including high-throughput sequencing, microarray analysis, array comparative genomic hybridization, and mass spectrometry, provide an enormous amount of information by allowing screening of an individual's entire genome in a single experiment (2). The recent evidence has shown that the integration of molecular changes from multiple levels of analysis, including the genomic, the epigenetic, microRNA, mRNA, and the proteomic, can provide a much better understanding of the pathways that are affected in cancer. This evidence has led to a change in focus from the discovery of individual biomarkers to a search for “biological processes” that are altered in individual patients (3). After the initial period of optimism for an imminent revolution in medical practice whereby the molecular “fingerprint” for each cancer patient would replace the clinicopathologic parameters, it has now become clear that the transition from the research bench to bedside is more challenging than was previously expected. One important challenge is the ability to analyze and extract meaningful information from this overwhelming amount of data. Different strategies have been suggested to translate molecular-profiling data into the clinic. One approach is to use global analysis as a discovery tool to identify a limited number of potential biomarkers that could then be processed into a clinical assay; however, these assays usually have limited clinical usefulness, with their low sensitivities and specificities owing to the great heterogeneity in biomarkers among …