Increasing evidence demonstrates that target based-agents are active only in molecularly selected populations of patients. Therefore, the identification of predictive biomarkers has become mandatory to improve the clinical development of these novel drugs. Mutations of the epidermal growth factor receptor (EGFR) or rearrangements of the ALK gene in non-small-cell lung cancer, and BRAF mutations in melanoma are clear examples of driver mutations and predictive biomarkers of response to treatment with specific inhibitors. Predictive biomarkers might also identify subgroups of patients that are not likely to respond to specific drugs, as shown for KRAS mutations and anti-EGFR monoclonal antibodies in colorectal carcinoma. The discovery of novel driver molecular alterations and the availability of drugs capable to selectively block such oncogenic mechanisms is leading to a rapid increase in the number of putative biomarkers that need to be assessed in each single patient. In this respect, two different approaches are being developed to introduce a comprehensive molecular characterization in clinical practice: high throughput genotyping platforms, which allow the detection of recognized genetic aberrations in clinical samples, and next generation sequencing that can provide information on all the different types of cancer-causing alterations. The introduction of these techniques in clinical practice will increase the possibility to identify molecular targets in each individual patient, and will also allow to follow the molecular evolution of the disease during the treatment. By using these approaches the development of personalized medicine for cancer patients will finally become possible. J. Cell. Biochem. © 2012 Wiley Periodicals, Inc.