Predictive factors of gefitinib antitumor activity in East Asian advanced non-small cell lung cancer patients

Background: Gender, smoking history, adenocarcinoma histology, performance status, and East Asian ethnicity were predictive factors of gefitinib response in previous analysis. However, these factors tend to be correlated with each other; it is not clear whether gender, smoking history, and adenocarcinoma histology were all independent predictors for response in East Asian populations. Methods: Tumor response, survival and predictive factors of gefitinib response of advanced non-small cell lung cancer patients treated between May of 2002 and November of 2004 were collected retrospectively from three medical centers in Taiwan. Univariate and multivariate logistic regression models were used to test potential predictive factors associated with response to gefitinib. Overall survivals between groups with different predictive factors were compared by log-rank tests. Multivariate analyses were performed to identify factors that independently predict for survival. Results: A total of 428 patients were analyzed. The median follow-up duration for living patients was 19.5 months (range, 10.2-39.9). Objective tumor response was observed in 114 patients (26.6%, 95% confidence interval [CI]: 22.4%-30.8%) and disease stabilization in 129 patients (30.2%). Response rate was statistically significant higher in adenocarcinoma, good performance status, and chemonaive patients in multivariate analysis. The median survival was 7.4 months (95% CI: 5.8-9.0) and 1-year survival was 34.3% (95% CI: 29.0%-38.0%). Significant independent predictive factors associated with longer survival in multivariate analysis were good performance status (p < 0.001) and responsiveness to gefitinib (p < 0.001). In 286 chemotherapy-treated patients, the response rate was 22.7%. Median and 1-year survival was 7.9 months and 36.7%, respectively. Good performance status was predictive of tumor response (p < 0.001) and better survival (p < 0.001) in multivariate analysis. Response to gefitinib was predictive of better survival (p < 0.001). Conclusions: Gender and smoking status were not, but good performance status (PS), no previous chemotherapy, and adenocarcinoma histology were independent predictive factors in multivariate analysis for gefitinib response in Taiwanese advanced non-small cell lung cancer population. In patients previously treated with chemotherapy, only good PS was an independent predictor for tumor response in multivariate analysis.