Background: WJTOG3405 has proven that chemotherapy-naïve patients with postoperative recurrent or stage IIIB/IV NSCLC harboring activating EGFRmutation have longer progression free survival (PFS) when treating with G than treating with CD. (9.2 months (mos.) for G vs. 6.3 mos. for CD, hazard ratio (HR) 0.489, 95% confidence interval (CI): 0.336-0.710). (Mitsudomi et al., Lancet Oncol. 2010) Although we reported updated overall survival results of this study after 34 mos. median follow-up period, the impact on overall survival (OS) was still unclear because of lack of survival events. (Mitsudomi et al, ASCO 2012). Methods: Overall survival (OS) was re-evaluated using updated data (data cutoff, 30 Sep, 2013, median follow-up, 59.1 mos.) for 172 patients. Results: One hundred twenty-seven events had occurred (73.8%). Median survival time (MST) for G arm was 34.8 mos. (95% CI: 26.0 – 39.5) which was not significantly different from 37.3 mos. (95% CI: 31.2 – 45.5) for CD arm (HR 1.252, 95% CI 0.883-1.775). Multivariate analysis using Cox proportional hazards model revealed that postoperative recurrence or IIIB/IV significantly affected OS among assessed covariates (treatment arm, smoking status, sex, age, postoperative recurrence or IIIB/IV, and mutation type). The overall survival of patients with postoperative recurrence was better than that of those with IIIB/IV stage disease (HR 0.459, 95% CI 0.312-0.673, p < 0.001). In the CD arm, 8 patients (9%) never received EGFR-TKI in their whole courses of therapy, whereas 31 patients (36%) never received chemotherapy in the G arm. MST for the former and the latter group were 13.5 months (95% CI: 6.5 -) and 44.5 months (95% CI: 21.3 - 51.8), respectively. Conclusions: This five-year follow-up OS analysis confirmed that G for advanced NSCLC with EGFR mutation offers survival benefit of 3 years. There was no difference in OS whether the initial treatment was G or CD, probably due to high cross over rate. Clinical trial information: 000000539.