biostudies-otherElloumi F54https://www.ebi.ac.uk/biostudies/studies/S-ECPF-GEOD-22384EurocanPlatformResults: Normal tissue contamination caused misclassification of tumors in all predictors, but different breast cancer predictors showed different susceptibility to normal tissue bias. Sensitivity and negative predictive value (NPV) of the PAM50 assay was improved by accounting for normal tissue. Conclusions: Normal tissue sampled concurrently with tumor tissue is an important source of bias in genomic predictors. Adjustments for normal tissue contamination could improve the application of breast cancer genomic predictors in both research and in clinical settings. Reference x breast tumor samples.biostudies-othertranscription profiling by arrayEurocanPlatformUnknown4BMC medical genomicsHomo sapiensElloumi FLi YHu ZTroester MAParker JSGulley MLAmos KDfalseSystematic Bias in Genomic Classification Due to Contaminating Normal Tissue in Breast Tumor SamplesResults: Normal tissue contamination caused misclassification of tumors in all predictors, but different breast cancer predictors showed different susceptibility to normal tissue bias. Sensitivity and negative predictive value (NPV) of the PAM50 assay was improved by accounting for normal tissue. Conclusions: Normal tissue sampled concurrently with tumor tissue is an important source of bias in genomic predictors. Adjustments for normal tissue contamination could improve the application of breast cancer genomic predictors in both research and in clinical settings. Reference x breast tumor samples.2016-04-14T13:44:09Z20112016-04-14T13:44:09Z2016-04-14T13:44:09ZS-ECPF-GEOD-22384GSE22384E-GEOD-22384EFO_0000635EFO_0000305E-GEOD-22384