Lung cancer has a very high recurrence rate and mortality, even in early stages of the disease. Current clinical staging techniques have limitations in terms of predicting which patients have an increased risk of recurrence, and they are not capable of sorting out who will benefit most from adjuvant therapy in term of survival advantage. The study of genomics has revolutionized how researchers are able to identify new molecular targets and improve patient care through the identification of 'genetic fingerprints or profiles' that might be able to predict responsiveness to therapy or prognosis. Techniques such as microarray-based gene-expression profiling have also allowed investigators to reveal different non-small-cell lung cancer subtypes that have been associated with different clinical outcomes, independently of histology and current clinical staging techniques. We review the current advances in gene-expression profiling and its potential role as a diagnostic and prognostic/predictive biomarker, and how this may translate into a 'personalized medicine' for lung cancer.