OBJECTIVES:  Patients with metastatic non-small cell lung cancer (NSCLC) whose tumors harbor exon 19 deletions or exon 21 (L858R) substitution mutations in the epidermal growth factor receptor (EGFR) gene are eligible for first line treatment with erlotinib or afatinib, tyrosine kinase inhibitors (TKIs) approved for use in the U.S.  The relative clinical and economic impacts between these two agents remain unclear.  METHODS:   A Markov model was developed in which patients could transition through three health states: Pre progression, progression, and death.  Transition probabilities were derived from key clinical trials, peer-reviewed literature, and U.S. life tables. We assumed a progression free survival (PFS) hazard ratio (HR) of 1 in the base case, and a scenario analysis was performed using data from an indirect treatment comparison in which the PFS HR for erlotinib vs. afatinib was 0.44.  Costs included those related to drug utilization, drug administration, and adverse events. We calculated the discounted (3%) incremental life expectancy, quality-adjusted life expectancy, and costs from a U.S. payer perspective.  Probabilistic sensitivity analysis (PSA) was conducted to assess parameter uncertainty.  RESULTS: Initiating treatment with erlotinib rather than afatinib resulted in comparable QALYs and modest decreased costs (-$895) in the base case. PSA indicated that at a threshold of $100,000/QALY, there was approximately a 60% probability that erlotinib is cost-effective.  In the scenario analysis (PFS HR: 0.44) we found that erlotinib treatment resulted in $35K greater costs and 0.274 additional QALYs gained, resulting in an ICER of $128K/QALY.  CONCLUSIONS: Our analysis indicates that first-line treatment with erlotinib vs. afatinib may lead to similar quality adjusted life years gained but slightly lower costs. Depending on assumptions about the comparative effectiveness, erlotinib may be in the range of being cost effective compared to afatinib, even with significantly longer PFS and treatment duration.