Background/objectiveNon-platinum-based chemotherapy is a potential alternative to platinum doublet therapy for advanced non-small cell lung cancer in selected patients. We determined the cost-effectiveness of gemcitabine/vinorelbine (GEMVIN), versus cisplatin/gemcitabine (PG) or cisplatin/vinorelbine (PV), from a government payer perspective. MethodsResults from a randomized trial of GEMVIN versus PG or PV demonstrated no significant difference in global quality of life (primary endpoint) or overall survival between regimens, but superior progression-free survival for platinum-based regimens. A cost analysis was conducted using direct medical costs of treatment, grade 3 or 4 toxicity management, and investigations for the mean number of cycles per study arm. Costs were calculated using Canadian dollars in 2005, and then in 2013 after drug patent expiry. ResultsIn 2005, GEMVIN was the most expensive regimen ($6868), and PV the least expensive ($4650), with an incremental cost of GEMVIN over PV of $2218. Diagnostic and administration costs did not differ significantly among regimens; GEMVIN had the lowest toxicity costs. The principal cost driver in 2005 was the cost of chemotherapy. In 2013, toxicity and administration costs emerged as major drivers; GEMVIN was less costly than PV and PG, (cost savings of $413 over PV). ConclusionDespite similar outcomes, GEMVIN was more expensive than PV or PG in 2005 because of higher chemotherapy costs. By 2013, after chemotherapy drug patent expiry, GEMVIN became the least costly regimen. Economic considerations in oncology change over time, and should be revisited in policy decisions based on cost.