Is response rate or stable disease rate a surrogate endpoint for survival in the treatment of advanced non-small cell lung cancer using epidermal growth factor receptor tyrosine kinase inhibitors?

14600 Background: Clinical trials have been intensively conducted on advanced non-small cell lung cancer (NSCLC) using epidermal growth factor tyrosine kinase inhibitors (EGFR-TKIs). In these trials, molecular biomarkers have been often examined with response rate (RR) as a surrogate endpoint instead of survival time. However a relationship between RR and survival time has not been determined in patients treated with EGFR-TKIs. The purpose of this study is to clarify the significance of RR and stable disease rate (SDR) as a surrogate endpoint for survival in the treatment of advanced NSCLC with EGFR-TKIs, through a systematic review of publications. Methods: Prospective trials of EGFR-TKIs monotherapy in NSCLC were extracted from MEDLINE, EMBASE, and the grand meetings in 2007 of the American Society of Clinical Oncology, European Cancer Conference, and World Conference on Lung Cancer. Correlation between clinical response and survival was examined in each study using SAS Version 8.2. Because of the presence of interaction for RR and SDR, two covariate analysis was applied. Multiple regression analysis was also performed on a per study basis to adjust imbalance of such patient characteristics in each study as gender, performance status (PS), histologic subtypes of cancer, number of prior chemotherapies, ethnicity, and types of EGFR- TKIs. Results: We identified 17 phase II trials, 4 phase III trials and 7 prospective trials with a total of 6163 patients and 30 treatment arms including 22 arms for gefitinib group and 8 arms for erlotinib group. In the two covariate analysis, RR strongly correlated with MST (P<0.0001), while SDR did not. In the multiple regression setting, only RR and PS significantly contributed to MST prolongation (P=0.003 and 0.04, respectively), and MST increased 0.21 month with each 1% increase in RR. RR had also significant correlation with median time to progression (MTTP, P=0.002), and SDR did (P=0.04) in the multiple regression setting. Conclusions: We found a significant relationship between RR and MST in clinical trials with EGFR-TKIs. SDR was a surrogate endpoint just for MTTP in this study. No significant financial relationships to disclose.