Fig. 6From: Integrative analysis of plasma metabolomics and proteomics reveals the metabolic landscape of breast cancerSpecific metabolic signature-based diagnostic biomarkers for BC. A ROC curves of the prediction efficacy for the metabolites-based predictors in training cohort using SVM (AUC = 1). B, C The validation for the performance of the prediction model in testing cohort using SVM, BC vs. HC (AUC = 0.794) (B), benign vs. HC (AUC = 0.879) (C). D ROC curves of the prediction efficacy of the CA15-3 (AUC = 0.722) and CEA (AUC = 0.757)Back to article page