18043 Background: JMTO LC00–03, a randomized trial of VGD versus PC in patients with advanced NSCLC, showed differences between the two treatment groups in response rate (25 vs 37.1%) and regimens’ specific toxicities including taxane-related toxicities such as numbness (0 vs 16.2%) and neuropathy (0.5 vs 9.6%), but not in overall survival (MST, 13.6 vs 14.1 mos) and progression-free survival (MST, 5.5 vs 5.8 mos). (Kawahara et al, # 7013, ASCO, 2006). Methods: Patients with advanced NSCLC were randomly assigned to VGD or PC. The patients were assessed with Functional Assessment of Cancer Therapy- FACT-L and FACT-Taxane score in baseline, and at 6-, 12-, 18-weeks after the treatment. The longitudinal analysis was used to compare mean changes of the QOL score over the two treatment groups. Results: Sixty-eight patients from the trial (VGD: 34, PC: 34) who submitted both baseline questionnaire and at least one questionnaire over the course of the treatment were subjects of the study. The table presents the estimated changes in mean scores in treatment arms over the time period. The longitudinal analysis showed significant difference in FACT-Taxane (p<0.001) in the treatment over time, but no significant difference in the FACT-L score (p=0.261). The analysis assuming non-random missing mechanism resulted in slightly larger differences in the mean change. Conclusions: There was no statistically significant difference in FACT-L between the two groups, but VGD group was numerically better than PC group at any point from baseline. The significant difference in FACT-Taxane score favoring VGD would be due to the difference in frequency of neuropathy with docetaxel than with paclitaxel. [Table: see text] [Table: see text]