Fig. 3From: Widely targeted quantitative lipidomics and prognostic model reveal plasma lipid predictors for nasopharyngeal carcinomaVariable importance in the proposed model based on lipid and clinical biomarkers. The feature importance score in predicting DMFS (A), PFS (B) and OS (C) was output in the DeepSurv model representing the decrease in the concordance index when the feature’s information was destroyed. The color of the bar represents the variable category. Abbreviations: DMFS, distant metastasis-free survival; PFS, progression-free survival; OS, overall survival; N, node stage; T, tumor stage; Hex2Cer, dihexosylceramide; CAR, acylcarnitine; EBV, Epstein-Barr virus; Cer, ceramides; PC, phosphatidylcholines; FFA, free fatty acid; DKLCA, diketolithocholic acidBack to article page