Non-small cell lung cancer (NSCLC) creates a large economic and disease burden worldwide. In an era of evidence-based medicine and increasing cost pressures, it is important to understand the relative clinical and economic impact of the many drug treatment strategies available for NSCLC. A systematic review of the peer-reviewed literature for pharmacoeconomic evaluations in the primary treatment of NSCLC published over the past decade (1 June 1997 to 1 June 2007) was conducted using the PubMed, EMBASE, BIOSIS Previews, Harvard Review of Economic Analyses, National Institute for Health and Clinical Excellence and Canadian Agency for Drugs and Technologies in Health databases. A total of 19 studies met the inclusion/exclusion criteria. Of these studies, 58% were cost-effectiveness studies, 37% were cost-minimization studies and 5% were cost-utility studies. Most were from the EU (63%), were from the payer perspective (89%), were in advanced (stage IIIB/IV) NSCLC (84%) and were funded by drug manufacturers (68%). Drug treatments generally were found to be cost effective compared with best supportive care. In addition, cisplatin alone or in combination appeared to provide better value than carboplatin alone or in combination. We did not identify any studies of recently approved therapeutics (e.g. erlotinib or bevacizumab). The quality of studies varied but the majority did not meet recommended guidelines for economic evaluations, with only 43% using direct comparisons, 5% of studies being cost-utility studies and 26% using either statistical analysis of patient-level data or probabilistic sensitivity analyses. In conclusion, there are a multitude of studies examining drug treatment for NSCLC; however, few of these utilized methodological approaches consistent with recommended guidelines. Despite these limitations, it appears that drug therapy compared with no treatment provides reasonable value for money, but carrying out more detailed comparisons of various agents is challenging. Given the absence of studies on newer therapeutics and the lack of cost-utility studies, additional studies are warranted.