OBJECTIVES: Since the launch of imatinib in Taiwan, survival of chronic myeloid leukemia (CML) has significantly improved with the cost of numerously increasing drug expenditure. Neither health-related quality-of-life (QoL) of CML patients nor long-term cost-effectiveness of imatinib, however, has been investigated in Taiwan. This study aimed to explore the feasibility of preference-based utility measures on CML patients with imatinib treatment. METHODS: This cross-sectional survey was conducted at a medical center in southern Taiwan from June 2011 to January 2012 after approved by the Institutional Review Board. CML outpatients receiving imatinib were invited to participate the interviews; their QoL were measured by EuroQol group 5-dimension (EQ-5D) questionnaire, 100-mm visual analogue scale (EQ-5D VAS) and time-trade-off (TTO) method, and then transformed into utility value. Japanese preference weight was used to transform EQ-5D index into utility. Patients' demography, socioeconomic characteristics and progression history were collected to adjust utility values by multiple regression. RESULTS: Of the 42 (mean age: 50.0±16.0 years, male: 54.8%) participants, 36 (85.7%) have been regularly followed up over 1.5 years, and 13 (31.0%) had history of progressing to accelerated or blast phase. In EQ-5D survey, most participants choose “no problem” in the domains of mobility, self-care, and usual activities, and 10 (23.8%) and 7 (16.7%) participants choose “some problem” in pain/discomfort and anxiety/depression, respectively. Mean utility values and response rates for EQ-5D, EQ-5D VAS, and TTO were 0.80±0.09 (100%), 0.78±0.13 (97.6%), and 0.80±0.18 (92.9%) and there was no significant difference in utility between three measures. No association between participant' characteristics and utility was found, except for utility derived from EQ-5D VAS was significantly associated with education level. CONCLUSIONS: Preference-based utility measured are applicable to CML patients receiving routine treatments. Since no Taiwanese preference weight for transferring EQ-5D assessment, further large-scale studies are needed, however, to validate and cross-validate the measurements.