The accelerated expansion of the knowledge of genetic and molecular basics of cancer, together with the recent development of molecular biology techniques, have had a significant impact in the field of oncology, among other medical disciplines. So, over the last few years, we are crossing from an empiricism-based model to an evidence-based model in which drugs are adapted depending of the molecular alterations which result crucial for tumor development (both for carcinogenesis and acquisition of an aggressive phenotype leading to tumor invasion and resistance to therapy). The molecular alterations /variations offer the possibility of being detected and used as biomarkers in clinical practice. Biomarkers may have multiple applications in the field of oncology, from determining the risk to suffer the disease to prediction of response to therapy, including diagnosis, prognosis and disease monitoring, with the final aim of performing a more personalized medicine and achieving greater efficacy for the therapies selected, diminishing each therapy's own adverse events. Considering the importance biomarkers may get to have in clinical decision making, it is basic that their development is performed under straight evaluation and validation rules. In this article we review the various types of biomarkers and the basic methodological principles for their development, validation and subsequent clinical application.