Original Article
Identification of predictive biomarkers for ZD-6474 in lung cancer
Abstract
Background: Lung cancer is the leading cause of cancer-related death globally. Even though numerous research efforts have been devoted to improving the treatment of patients with lung cancer, the overall five-year survival rate is still below 20%. The major challenge in improving this survival rate is the highly heterogeneous cancer genome. Recently, targeted therapy, and especially drugs that target EGFR, has been shown to be a promising therapeutic method against lung cancer. Challenges arise, however, in trying to classify patients as responders or non-responders to the drugs. Patients who receive no benefits from the treatment may still suffer from its adverse effects. One way to address this issue is to identify genomic biomarkers that predict the responses of patients to drugs before treatment.
Methods: This study demonstrated the identification of predictive biomarkers for responses to drugs by analyzing the gene expression profiles of lung cancer cell lines. For each cell line, microarray data were normalized using the quantile algorithm. A linear regression model was applied to select probes that are associated with drug efficacy, and a prediction model was developed using the support vector machine algorithm.
Results: ZD-6474 has the strongest inhibitory effects of the three drugs that target EGFR, and a total of 24 probes displayed significant associations with the efficacy of ZD-6474 (P<2×10−5). The prediction model demonstrated approximately 80% accuracy in the leave-one out cross-validation test of 89 lung cancer cell lines.
Conclusions: In conclusion, gene expression profiles can serve as potential predictive biomarkers for predicting patients’ responses to drugs so the treatment plans for different individuals may be improved by considering their genetic statuses.
Methods: This study demonstrated the identification of predictive biomarkers for responses to drugs by analyzing the gene expression profiles of lung cancer cell lines. For each cell line, microarray data were normalized using the quantile algorithm. A linear regression model was applied to select probes that are associated with drug efficacy, and a prediction model was developed using the support vector machine algorithm.
Results: ZD-6474 has the strongest inhibitory effects of the three drugs that target EGFR, and a total of 24 probes displayed significant associations with the efficacy of ZD-6474 (P<2×10−5). The prediction model demonstrated approximately 80% accuracy in the leave-one out cross-validation test of 89 lung cancer cell lines.
Conclusions: In conclusion, gene expression profiles can serve as potential predictive biomarkers for predicting patients’ responses to drugs so the treatment plans for different individuals may be improved by considering their genetic statuses.
