Therapy for all non-small cell lung cancers has been non-selective for many decades. This was a "blind" process due to lack of solid scientific data in terms of tumor biology and molecular medicine focused in this disease. In the last decade, the study of lung cancer genome, molecular analysis, proteomics, and many others have resulted in an avalanche of knowledge which has allowed us to tailor specific treatments based on patients' tumor fingerprints ("phenotype") at the molecular level. Efforts are being conducted to see which available treatments may induce major responses that may translate into survival advantage. As an example, several predictive biomarkers in nonsmall cell lung cancer have been recently elucidated such as DNA repair genes (e.g., ERCC1, RRM1), gene expression profiling, certain gene mutations (e.g. EGFR, K-ras), and others. Recently, the histology of the lung tumors has become a clinical feature driving therapeutic decision. Nowadays, we consider that histology in lung cancer matters and we must secure enough tumor tissue, which can yield not only an accurate pathologic diagnosis but also for further biomarker tests. More research is needed in this arena to define the best treatment selection and therapeutic algorithm for our patients. Many variables besides patient-related factors play a crucial role in the therapeutic decision-making.