OBJECTIVES: Even though chemotherapies of advanced non-small cell lung cancer (NSCLC) are improving patients’ outcomes, usually patients’ treatment preferences are underestimated by physicians. To identify those attributes related to chemotherapy treatment that is highly valued by patients with advanced NSCLC. METHODS: An observational, cross-sectional study was carried out. A four attributes treatment preferences questionnaire was elaborated in a pilot study: treatment mode of administration, time spent in hospital for administration, immediate toxicity, and level of symptoms control; distributed in 9 hypothetic scenarios. Second line NSCLC patients were asked to score by preferred scenario, through the conjoint analysis method. Socio-demographic, clinical and health-related quality of life (RSCL) characteristics were collected and linear regression models were performed. RESULTS: A total of 172 patients were enrolled. Mean (SD) age 61.9 (10.5) years. Seventy-five percent were men, 84.8% in NSCLC stage IV, 50% ECOG 1 20.3% of patients received cisplatine+gemcitabine in first line, 30.8% showed treatment’s adverse reactions and 52.9% were scheduled to erlotinib as second line. Psychological symptoms’ was the most affected RSCL dimension. The ICC of questionnaire’s scenarios was assessed (0.55-0.88).The most preferred attribute was symptoms’ control and the most preferred scenario was: oral administration, no time spent in hospital, low toxicity and total symptoms’ control (G scenario). Patients who started oral administration vs patients who started intravenous administration, gave more utility/importance to the attributes ‘treatment administration’ (0.30 vs -0.12) and ‘time spent in hospital’ (0.62 vs 0.45). Psychological and daily functioning affection, ECOG 1-2 and adverse reactions in first line were related to G scenario. CONCLUSIONS:The most valued attributes by advanced NSCLC patients are moderate-total symptoms’ control and oral administration. Patients with worst functional status and psychological affection preferred total symptoms’ control, oral administration, no time spent in hospital and low toxicity. Treatment preference’s questionnaire showed utility, feasibility and reliability for treatment decisions, allowing physicians to explore patients’ preferences before initiating treatment.