Lung cancer is the leading cause of cancer-related death in both developed and developing countries with unacceptably high mortality even for early-stage cancers. Present guidelines do not recommend adjuvant chemotherapy for stage I non-small cell lung carcinoma and prompts a search for a prognostic marker that would separate stage I patients into 2 groups, those who would benefit and those who would not benefit from adjuvant treatment. Studies during the last decade showed that the gene expression profiling can be the biomarker being sought. Many gene expression profiling have been found and reported from the analysis of surgical specimens of resected lung cancers during the last decade, and many of them had been shown to have an excellent predictive accuracy. These profiles used in the studies had not only different gene combinations but also different number of genes and methods of identification. Researchers have used microarray assays, RT-quantitative polymerase chain reaction, and more recently microRNA-based techniques to achieve this goal. Unfortunately, most of the profiles were not sufficiently validated and/or were not used in prospective phase III studies. This review focuses on major studies in the field, future prospects, as well as the lessons learned so far. It is shown that the gene expression profiles have a good chance of being implemented in future everyday practice.