OBJECTIVES: To analyze the cost-effectiveness of treating all chronic-phase chronic myeloid leukemia (CML) with imatinib initially compared to physician-choice between imatinib or the second-generation tyrosine kinase inhibitors (TKIs) dasatinib or nilotinib. Imatinib will lose patent exclusivity between 2015-2016 and its price is expected to drop 60-90% within one year throughout Europe.   METHODS: A Markov model simulating “step-therapy” compared to “physician-choice” in treating CML in 2015 through 5 years. The model assumes a European societal perspective. In both approaches, if initial treatment fails, patients are switched to a second-generation TKI. Patients are assumed to switch if they fail to meet efficacy endpoints: complete cytogenetic response (CCyR) or major molecular response (MMR). The model assumes stabilized prices of second-generation TKIs, but discounts the price of imatinib: 100% for first 6-months; 60-80% for second 6-months; and 10-30% thereafter. For each drug, tolerance, efficacy and the probabilities of treatment choice, switching and failure were drawn from published clinical trials. Quality-adjusted life years (QALYs) were based on U.K. preference weights (Szabo et al. 2010). According to Hoyle et al. (2011), direct medical costs per patient were: £20,244 for imatinib; and ~£30,000 for dasatinib and nilotinib. Additional costs included patient monitoring and allogeneic transplantation. Costs and QALYs were discounted at 3% (British Pounds Sterling (£); 2013). Sensitivity analyses tested parameters for impact on results at a willingness-to-pay of £50,000/QALY.   RESULTS: Step-therapy costs less and offers clinically-equivalent utility (£62,388; 2.864 QALYs) compared to physician-choice (£71,268; 2.879 QALYs), at an ICER of £592,000/QALY. The results are robust to changes based on univariate analyses of each parameter. Multivariate probabilistic sensitivity analyses found step-therapy cost-effective in 99.9% of 10,000 Monte Carlo simulations.   CONCLUSIONS: When imatinib loses patient protection between 2015-2016 throughout Europe, it will be the cost-effective initial treatment strategy for CML compared to second-generation TKIs.