Validation Study: Response-Predictive Gene Expression Profiling of Glioma Progenitor Cells In Vitro

摘要: Background In a previous publication we introduced novel approach to identify genes that hold predictive information about treatment outcome. A linear regression model was fitted by using the least angle algorithm (LARS) with expression profiles of construction set 18 glioma progenitor cells enhanced for brain tumor initiating (BTIC) before and after in vitro tyrosine kinase inhibitor Sunitinib. Profiles from treated allowed predicting therapy-induced impairment proliferation vitro. Prediction performance validated leave one out cross validation. Methods this study, used an additional validation serum-free short-term cell cultures test properties signature independent cohort. We assessed rates together transcriptome-wide Sunitinib each individual culture, following methods publication. Results We confirmed treatment-induced changes our set, but failed predict inhibition. Neither re-calculation combined dataset all 36 BTIC nor separation samples into TCGA subclasses did generate prediction. Conclusion Although gene published exhibited good prediction accuracy validation, were not able validate data set. Reasons could be mean, moderate numbers samples, or too low differences response inhibition At stage based on presented results, conclude does warrant further developmental steps towards clinical application.