e19056 Background: Erlotinib was approved in Dec 2007 in Japan, and incidence and pattern of ILD during its therapy for Japanese pts with NSCLC has not still been determined, although we had previously reported the frequency of ILD through the gefitinib treatment [PASCO2004, #7063]. In this study, we intended to elucidate this issue in pts receiving erlotinib therapy. Methods: We reviewed the clinical records of 159 pts who had initiated erlotinib therapy last year (cohort A), and of 330 pts receiving gefitinib between 2000 and 2003 (cohort B) for comparing the incidence and pattern of ILD during the both TKI treatments. Toxicity data during the first months after the initiation of TKIs were obtained. Results: The demographics of 489 pts were as follows; M:63%, Ad:75%, and PS 0–1:69%. None of pts in the cohort B received erlotinib therapy before the gefitinib treatment, whereas 66 of the 159 cohort A pts (42%) were given gefitinib before the erlotinib therapy. In 23% and 28% of the pts in the cohorts A and B, erlotinib and gefitinb treatments were discontinued within 1 month after the initiation of TKI therapy, respectively. Two pts (1.3%) developed ILD in the cohort A during the first month of erlotinib treatment, while 8 ILD-events (2.4%) were observed in the gefitinib therapy (cohort B) during the same treatment period. Both 2 pts who developed ILD during the erlotinib therapy had not had a history of prior gefitinib treatment. The toxicity grades of ILD were as follows: grades 1 and 2 in 1 each (cohort A) and grades 3, 4 and 5 in 1, 1 and 6 pts, respectively (cohort B). Statistically significant factors affecting the occurrence of ILD by multivariate analysis were presence of prior pulmonary fibrosis (OR=37.3, p<0.01) and poor PS (OR=6.4, p=0.02), but type of TKIs was not a significant risk factor for ILD. Conclusions: In this setting, the type of TKIs did not affect the incidence of ILD although its incidence after the initiation of erlotinib was somewhat low as compared with that during gefitinib therapy. In addition, the grade of ILD was less severe in the cohort A. These might be partly due to a patient selection based on the recent awareness of Japanese physicians regarding the risk factors for ILD events who learned it through the gefitinib treatment. No significant financial relationships to disclose.