Fig. 4

Validation of the DRG score model and establishment of a disulfidptosis predictive nomogram for HCC. A–C Distribution of patient survival status A and Kaplan–Meier OS curves B and time-dependent ROC curves C for patients in the high- and low-DRG score groups in the ICGC − LIRI-JP validation cohort. The group with high DRG scores had a worse survival rates (P = 0.00095), and the AUC values of the DRG scores for 1 year and 3 year OS were 0.674 and 0.741, respectively. D Multivariate Cox regression showed that high stage (P = 0.003, HR = 1.39) and DRG score (P < 0.001, HR = 3.69) were independent risk factors for worse OS. E The establishment of a nomogram to predict the 1-, 3-, and 5 year overall survival probability for HCC patients in the TCGA cohort. F The ROC curves demonstrated that the nomogram model's AUC values for one-year OS, 3 year OS, and 5 year OS were 0.806, 0.768, and 0.730, respectively. G and H Compared with stage or DRG score alone, the nomogram model showed a higher AUC value for 1 year G and three-year H OS, which suggested that the model was effective. I The DCA curve suggested that the nomogram model performs better in clinical practice than stage or DRG score alone. P values are shown as * P < 0.05; ** P < 0.01; *** P < 0.001; ns not significant. DRGs Disulfidptosis-related genes, ROC receiver operating characteristic, ICGC international cancer genome consortium, HCC hepatocellular carcinoma, AUC area under curve, DCA decision curve analysis