OBJECTIVES: The purposes of this study were to investigate redistribution properties of EQ-5D-5L and to compare the validity and reliability of each EQ-5D version for Korean cancer patients. METHODS: Patients visiting one ambulatory cancer center self-administered the two EQ-5D versions and EORTC QLQ-C30 questionnaire. Redistribution properties are examined between EQ-5D-5L and EQ-5D-3L. Convergent validity of these two versions was evaluated by comparing EQ-VAS, ECOG performance status and EORTC QLQ C30 subscales. Discriminant ability was evaluated based on Shannon index and ceiling effect and test-retest reliability was evaluated using kappa statistics and intraclass correlation coefficient. RESULTS: Inconsistent rate was 3.5% between two versions. In convergent validity, EQ-5D-5L demonstrated similar or higher correlations with EQ-VAS, ECOG performance status and EORTC QLQ-C30 compared with EQ-5D-3L. Absolute informativity in EQ-5D-5L was improved but relative informative is similar to the EQ-5D-3L. Ceiling effect was decreased from 16.8% in the 3L version to 9.7% in the 5L version. Agreements by Kappa were fair to good in 4 dimensions except usual activities in both EQ-5D instruments. Intraclass correlation coefficient of EQ-5D-5Lindex was 0.77. CONCLUSIONS: The EQ-5D-5L version appear a valid and reliable quality of life instrument in cancer patients. EQ-5D-5L enhanced descriptive richness and diminished ceiling effect compared with EQ-5D-3L.